I rarely watch sports. The last Super Bowl I watched was in New Orleans, not the one in 2025, but the one in 1986, when the Bears won. Lately, bingeing Netflix’s Formula 1: Drive to Survive has sparked a renewed interest in Formula 1 for me, but generally, I don’t have time for sports unless I am involved in them. There is already too much to do. But I do enjoy the Winter Olympics. I guess I like my sports fast; with a life defined by exile, I can always think of a reason to flee.

Watching the Olympics over the past two weeks, I remembered that my father’s professor at the Vienna Art Academy, Herbert Dimmel, had competed in the 1936 Berlin Olympics — Hitler’s notorious showcase — with his painting The Olympic Garden. That painting, sculpture, architecture, literature, and music competitions were part of the modern Olympics is now pretty well known, largely from lists of absurd Olympic medals from the early days, like live pigeon shooting, firefighting, or pistol dueling with wax bullets. But it wasn’t like that at all; that artistic competitions survived even Hitler’s bald-faced subjugation of aesthetics to ideology underscores how deeply the format was embedded in Olympic structure. Art competitions last ran at the 1948 London Games. The IOC dropped them in 1954, citing, among other reasons, the professional status of most competing artists. A generation later, the IOC abandoned the amateur requirement for athletes altogether — the word “amateur” left the Olympic Charter in 1971, and by 1986, professionals competed openly.

The individual generally credited with creating the modern Olympics, Pierre de Coubertin, was hardly an athlete; instead, he was a French aristocrat who studied law and public affairs before turning to history and educational reform. The story generally goes like this. A Jesuit education steeped in the classics gave Coubertin a lifelong attachment to ancient Greece and, in particular, to the Athenian gymnasium, which he admired for its fusion of physical and intellectual training into a single program. Coubertin found a modern-day analogue in Rugby School, the boarding school that birthed the eponymous sport and embodied the ethos of muscular Christianity — the mid-Victorian conviction that organized sport built moral character by teaching duty, discipline, and self-sacrifice. Coubertin credited muscular Christianity with Britain’s imperial expansion, while attributing France’s humiliation in the Franco-Prussian War to a lack of discipline. He published L’Éducation en Angleterre in 1888 and spent years campaigning in France for physical education reform to no avail.¹

In November 1892, Coubertin delivered a lecture surveying the development of sport across nations — again praising England and lamenting France’s sluggishness — and argued that international athletic exchange would do more for peace than diplomacy. At the close, he proposed reviving the ancient Games. Classical culture had been revived repeatedly — in architecture, in painting, in letters — but by the end of the nineteenth century, each wave recycled a prior revival, and none had touched sport.

The Olympiad was well known, but it was part of a broader obsession with the sanctuary of Olympia that had haunted the Western imagination for centuries. At its center stood the Temple of Zeus — housing Pheidias’s chryselephantine statue, one of the Seven Wonders — and the sacred Altis grove where athletics, religion, and art converged every four years under a truce that suspended even warfare. Victory there was the supreme honor of the Greek world: Pindar celebrated Olympic victors as standing at the boundary of the human and divine; winning cities sometimes tore down their own walls, a champion’s return rendering them superfluous. Pausanias’s Guide to Greece had described every monument in exhaustive detail, making Olympia a vivid literary presence for European scholars who had never seen a stone of it — the site was thought lost, buried beneath centuries of silt, though the local Greeks claimed they knew perfectly well where it lay. Johann Joachim Winckelmann, who founded art history and established Greek antiquity as its supreme aesthetic object, was murdered in Trieste in 1768 before he could reach it. When German archaeologists finally broke ground in 1875 and cleared the Altis by 1881, their annual reports were devoured by scholars and journalists alike.²

Three years before Coubertin spoke, an elaborate reconstruction of the buildings and monuments of ancient Olympia, based on the German excavations, had been displayed in the Palais des Beaux-Arts at the 1889 Paris Exposition, now most famous today for the construction of the Eiffel Tower. Coubertin was there organizing the Exposition’s own sports congress — five sessions on equitation, gymnastics, rowing, and athletics.³ Two years later, at an international congress he convened, delegates from fourteen nations voted to proceed and that June the International Olympic Committee was founded — not as a federation of national associations but as a self-selecting body, its members chosen by Coubertin himself. The first modern Games were held at Athens in 1896. These revived Olympics gave institutional form to a classical republican conviction: that the health of the citizen and the strength of the nation were one and the same thing, and that competition between nations was therefore competition between their programs of civic formation. But citizenship in the classical world Coubertin revered had always been a restricted category — in Athens, limited to free-born men — and Coubertin’s version was no different: he was a self-described “fanatical colonialist,” an avowed white supremacist who opposed women’s participation in the Games.⁴

Labelling Coubertin a reactionary would be too simple. His politics are, unfortunately, all too typical of his time, but they are also well known now and have been justly condemned. But instead of retreading that, what intrigues me is how close the relationship between art and the Olympics really was. The recent “rediscovery” of painting and sculpture competitions treats them as curiosities — forgotten oddities tacked onto an athletic program. But the art competitions were not oddities. They were central to Coubertin’s vision and to the original concept of the Games. This is hardly surprising, given that his father, Charles Louis de Frédy, Baron de Coubertin, was a Paris Salon painter. A student of François-Édouard Picot, the elder Coubertin exhibited at the Salon for over four decades, beginning in the mid-1840s. He worked in the Orientalist and academic modes, produced a monumental papal commission — Le Cortège Pontifical, a ten-by-three-meter fresco now in the Vatican — and received the Légion d’Honneur from the Minister of Fine Arts in 1865. He was a successful and deeply conventional product of the concours system: jury submission, public exhibition, state validation. Pierre Coubertin grew up in this world, not Rugby School.⁵

The modern Games bear little structural resemblance to the ancient ones — and what they do resemble says more about nineteenth-century Parisian culture than about Olympia. At Olympia, a victor received an olive wreath; there was no second or third place. Athletes competed as individuals, not as national delegations marching behind flags. There were no opening ceremonies, no anthems, no podiums. British sporting culture, too, governed itself through democratic clubs and associations, not through self-recruiting academies of unelected officials.⁶ The gold-silver-bronze hierarchy, the national representation, the periodic international spectacle, the self-perpetuating governing body — none of this comes from Rugby or from Olympia. It comes from the concours — the French competitive exhibition system that organized cultural life from the Salon to the Expositions Universelles. No accident that the same year Pierre staged the Athens Games, his father Charles painted what he titled Jeux Olympiques, 1896 — a massive oil, eight by nearly ten meters, depicting Athena crowning a modern athlete in the style of a Prix de Rome composition. Pierre used it as the cover of the Revue Olympique from 1901 to 1914; it now hangs in the Olympic Museum in Lausanne, catalogued as Allégorie aux Sports.⁷

The Salon was the annual exhibition of painting and sculpture — the arena where reputations were made or destroyed through a formalized hierarchy of submission, jury selection, and tiered awards. The École des Beaux-Arts was a pyramid of elimination — hundreds of aspirants filtered through increasingly demanding rounds of competition until a single champion emerged. The Prix de Rome was the gold medal of French culture: finalists were sequestered in isolation booths for weeks, forbidden from outside contact, their work judged anonymously by the “Immortals” of the Académie des Beaux-Arts — and the winner received a fully funded residency at the Villa Medici in Rome, years of study among the classical and Renaissance masters at state expense. So, too, this model extended beyond art: the Expositions Universelles — the great World’s Fairs of 1855, 1867, 1878, 1889, and 1900 — were giant competitions where nations vied to display the best machinery, the best livestock, and the best art, all ranked by expert juries awarding medals in gold, silver, and bronze.⁸

The concours was a specifically French institutional form, though its elements had scattered precedents — Italian guilds held design competitions from the fourteenth century, and Italian academies awarded prizes from the sixteenth. What the French system assembled was the specific combination of periodic public exhibition, an expert jury, and a tiered medal hierarchy, all state-backed and periodically repeating. The Académie royale de peinture et de sculpture, founded in 1648, built this apparatus piece by piece. It organized exhibitions of its members’ work from 1667; from 1725, these were held in the Salon Carré of the Louvre — the Salon for short. A formal selection jury arrived in 1748.⁹

What the Salon was to the eighteenth century, the industrial exposition was to the nineteenth. The first national industrial exposition, held in Paris in 1798, was explicitly modeled on the Salon. The historian Arthur Chandler documents that it “took its cue from the more prestigious salons of painting and sculpture”: a jury scrutinizing exhibited objects and awarding prizes to the best entries. Its nine-member jury included the painter Joseph-Marie Vien — David’s teacher, a member of the Institut de France — alongside chemists and industrialists. Three years later, the second exposition introduced the three-tier medal hierarchy: twelve gold, twenty silver, and thirty bronze. This is the earliest instance I could find of the system that the Olympics would adopt. The Salon swiftly followed, awarding medals in 1804, three years after the expositions.¹⁰

The first World’s Fair, London’s Great Exhibition of 1851, included fine arts — sculpture, architecture, decorative arts — but excluded painting. The 1855 Exposition Universelle corrected this. Napoleon III decreed that the annual Salon would stand alongside industry; art and steam engines were judged under the same institutional umbrella, ranked by the same medal system. Afterwards, every major World’s Fair included painting. When the same apparatus could evaluate a canvas and a locomotive, the form’s independence from any particular content was proved. The Salon had not invented the jury, the medal, or the periodic exhibition — each existed independently in French institutional life. What it had done was fuse them into a single system for ranking individual human achievement in a domain where merit was not self-evident, and that synthesis was what made the form portable.¹¹

The 1900 Paris Olympics were listed in the program of the Exposition Universelle as Concours internationaux d’exercices physiques et de sports — international competitions in physical exercises and sport, one category among thousands. That same Exposition awarded 3,156 grand prizes, 8,889 gold medals, 13,300 silver medals, and 12,108 bronze medals to over 83,000 exhibitors. The Olympic sporting events were a small line item in this vast medal-awarding apparatus. Coubertin applied the grammar of the concours to the sentence of sport. At the 1896 Olympics in Athens, winners received silver medals and olive branches, runners-up bronze; there was no gold. The full gold-silver-bronze triad arrived only at the 1904 St. Louis Games, which were, again, embedded in a World’s Fair, the Louisiana Purchase Exposition.¹²

In May 1906, Coubertin organized the Consultative Conference on Arts, Literature and Sport — at the Comédie-Française — and established five art competition categories: architecture, sculpture, painting, literature, and music, the “Pentathlon of the Muses.” Art competitions ran as official Olympic medal events from 1912 through 1948, awarding 151 medals across seven editions. Coubertin was unambiguous: “Deprived of the aura of the art contests, Olympic games are only world championships.”¹³

When the IOC dropped the art competitions, it cited the professional nature of the competitors but also the lack of objective evaluation criteria, though dropping them may simply have made the concours origins of the Games — and the arbitrary nature of every judged event — less conspicuous. As any frustrated fan of figure skating or the half-pipe knows, such arbitrariness never left. At Salt Lake City in 2002, French judge Marie-Reine Le Gougne ranked the Russian pair first as part of an arrangement: the Russian judge would reciprocate for the French ice dancers.¹⁴ At Milan-Cortina, French judge Jézabel Dabouis scored the French ice dancers nearly eight points above the Americans in the free dance — a margin so large that without her score, the Americans would have won gold. National blocs trading placements behind closed doors — the same accusations of bias, favoritism, and backroom dealing that dogged the Salon from its founding followed the form to the ice rink.

The Salon’s operating system — selection, exhibition, judgment, ranking — outlived the academic art it was designed to serve, the Expositions, and even the art competitions that Coubertin embedded in the Games, yet it survives in the Olympics today, its arbitrariness undiminished.

¹ For the standard account of Coubertin’s inspiration in ancient Greece and Rugby School, see George Hirthler, “Celebrating Pierre de Coubertin: The French Genius of Sport Who Founded the Modern Olympic Games,” International Olympic Committee, September 1, 2019, https://www.olympics.com/ioc/news/celebrating-pierre-de-coubertin-the-french-genius-of-sport-who-founded-the-modern-olympic-games. ↩

² See Louis Callebat, “The Modern Olympic Games and Their Model in Antiquity,” International Journal of the Classical Tradition 4, no. 4 (1998): 555–556; John MacAloon, This Great Symbol: Pierre de Coubertin and the Origins of the Modern Olympic Games (New York: Routledge, 2008), 156–157. ↩

³ The reconstruction of ancient Olympia at the 1889 Paris Exposition was designed by the French architect Victor Laloux, a former winner of the Prix de Rome, from data generated by the German excavations directed by Ernst Curtius, funded by Kaiser Wilhelm I under an 1874 agreement with the Greek government: Germany bore all costs; every find stayed in Greece. Excavations ran from 1875 to 1881 and were extensively published. See Callebat, “Modern Olympic Games,” 556; MacAloon, This Great Symbol, 156–157. ↩

⁴ On the founding of the IOC at the 1894 Sorbonne congress, see MacAloon, This Great Symbol, 188–203. MacAloon writes that “the roster of the IOC itself had been hand-picked by de Coubertin” and that Coubertin “established the principle of the IOC as a ‘self-recruiting body,’” 202. On Coubertin’s racial views: “The races are of different value, and to the white race, of superior essence, all the others must pledge allegiance.” “Pierre de Coubertin and the Origin of the Olympic Games,” Ville de Paris, https://www.paris.fr/en/pages/pierre-de-coubertin-and-the-origin-of-the-olympic-games-27628. He described himself as a “fanatical colonialist” and wrote that women’s admission to the Games was “against my will.” See also “’Fanatical Colonialist’: The Uneasy Legacy of the French Founder of the Modern Olympics,” France 24, July 18, 2024, https://www.france24.com/en/france/20240718-fanatical-colonialist-uneasy-legacy-french-founder-of-modern-olympics-paris-games. ↩

⁵ On Charles de Coubertin’s Salon career, see Natalia Camps Y Wilant, “The Olympic Art Competitions: Olympic Medals for Artworks, Pierre de Coubertin’s Passion for Art, the Fine Art Salon, and the Social Circles,” Revue d’histoire culturelle, no. 8 (2024). Charles exhibited from 1845–46 onward; Camps Y Wilant identifies seven paintings shown at the Fine Art Salon between 1845 and 1850, with a diploma awarded in 1856. ↩

⁶ The Much Wenlock Olympian Games, founded by Dr. William Penny Brookes in 1850, complicate any clean French-vs.-British distinction. Wenlock included art and intellectual competitions — drawing, poetry, music — alongside athletics from the 1860s. Coubertin visited as Brookes’s guest in October 1890 and published a glowing account in La Revue Athlétique (25 December 1890): “If the Olympic Games that Modern Greece has not yet been able to revive still survives today, it is due, not to a Greek, but to Dr W. P. Brookes.” He later systematically downplayed Brookes’s role; Brookes died in December 1895, five months before the Athens Games. But the Wenlock competitions were a local voluntary-society affair: prizes began as cash and laurel wreaths, the art categories were knitting and sewing alongside reading and arithmetic, and there was no academic jury or hierarchical classification system. What Coubertin took from Wenlock was ceremonial and inspirational; the organizational framework of the IOC — periodic international congresses, self-recruiting governance, standardized medal hierarchies — came from French institutional culture. On the Wenlock pentathlon and Coubertin’s 1890 visit, see Sandra Heck, “William Penny Brookes – the Founding Father of the Modern Pentathlon?” Sport in History 34, no. 1 (2014): 75–89. The quote about Brookes is cited in MacAloon, This Great Symbol, 147. On the 1850 program and Brookes’s death, see also “A Country Town Hosted England’s First Olympics,” Associated Press, July 2, 2011, https://www.nytimes.com/2011/07/03/sports/olympic-legacy-claimed-by-an-english-country-town.html. ↩

⁷ The painting has circulated under three titles in sports history scholarship, all of which are apparently wrong. Camps Y Wilant recovered the original title from an annotated preparatory sketch in Charles de Coubertin’s personal Album; see her “Decoding Olympic History in a Painting by Charles de Coubertin,” The International Journal of the History of Sport 35, no. 17–18 (2019): 1815–1828, 8. ↩

⁸ On the concours system and its relationship to the Olympic model, see Camps Y Wilant, “When Art Was an Olympic Discipline: The Fine Art Salon as a Possible Model for the Concept of the Olympic Art Competitions,” Sport in History 38, no. 4 (2018): 457–475. ↩

⁹ On the institutional history of the Salon, see the classic Thomas Crow, Painters and Public Life in Eighteenth-Century Paris (New Haven: Yale University Press, 1985). ↩

¹⁰ On the first industrial exposition as modeled on the Salon, see Arthur Chandler, “1798 Exposition,” https://www.arthurchandler.com/1798-exposition. The jury included the painter Joseph-Marie Vien alongside chemists and administrators. The 1801 medal counts (12 gold, 20 silver, 30 bronze) are from Arthur Chandler, “Napoleonic Expositions,” https://www.arthurchandler.com/napoleonic-expositions. On Vivant Denon establishing the Salon’s official medal system in 1804, see Claire Dupin de Beyssat, “Peers and Experts: What Gave the Salon Jury Its Legitimacy (1791–1880)?,” Biens Symboliques / Symbolic Goods 17 (2025), §8. ↩

¹¹ On the Great Exhibition of 1851: the exhibition’s four divisions were raw materials, machinery, manufactures, and fine arts; Class XXX covered sculpture, models, and plastic art, but painting was excluded. See “Great Exhibition,” Wikipedia, https://en.wikipedia.org/wiki/Great_Exhibition. On Napoleon III’s decrees of 1852–1855 converting the planned 1854 national industrial exposition into the Exposition Universelle and creating a simultaneous Fine Arts exhibition, see Georges Poisson, “1855: France’s First International Exhibition,” Revue du Souvenir Napoléonien 457 (February–March 2005), republished at napoleon.org, https://www.napoleon.org/en/history-of-the-two-empires/articles/1855-frances-first-international-exhibition/. France’s 1855 addition of a full Beaux-Arts section became the permanent template. ↩

¹² On the 1900 Games as a sub-category of the Exposition Universelle, see MacAloon, This Great Symbol, especially the section “True Tests and Living Pictures: The Exposition Tradition” in ch. 4, 147–156. MacAloon describes how Coubertin allowed the 1900 Games to be “amalgamated to world’s fairs, with almost disastrous consequences,” 156; the exposition bureaucrats “seized control” and the Games became “just a sideshow,” 313. MacAloon further observes that “among other exposition symbols that may have served as prototypes for Olympic symbols are the medals awarded to the winners of the industrial competitions,” endnote 74, 173. ↩

¹³ On the 1906 conference and the Pentathlon of the Muses, see Douglas Brown, “Revisiting the Discourses of Art, Beauty and Sport from the 1906 Consultative Conference,” Olympika V (1996): 43–52, and Andrew Edgar, “The Aesthetics of the Olympic Art Competitions,” Journal of the Philosophy of Sport 39, no. 2 (2012): 185–199. On the IOC’s stated reasons for discontinuing the art competitions — “organisational difficulties, lack of objective evaluation criteria, poor quality of the artworks and for other reasons” — see Camps Y Wilant, “When Art Was an Olympic Discipline” (2018), 9. Camps Y Wilant dismisses the quality argument as invalid given the caliber of participants and jury members; see also her “The Olympic Art Competitions” (2024), 4–5. ↩

¹⁴ On the 2002 Salt Lake City pairs skating scandal: French judge Marie-Reine Le Gougne and French federation head Didier Gailhaguet were both suspended for three years by the ISU. A second gold medal was awarded to the Canadian pair, Jamie Salé and David Pelletier. See “Canadian Skaters Get Gold; Judge Suspended,” CNN, February 15, 2002. ↩

I built a machine to restage this conversation using two instances of an Artificial Intelligence. I have updated the AI speakers to reflect current concerns in the art world. Prompts are intentionally vague and open-ended. I do not give the speakers a bank of specific knowledge to draw on beyond a few broad generalities, allowing the AIs to use their understanding of the art world to shape their dialogue.

Contact me for the link and password.

UNIDENTIFIED VOICE

The light in here is really something today.

NANCY

Oh, I actually just came from the Hauser show and honestly the lighting design was a complete disaster, someone told me their director of exhibitions is about to be pushed out, which, given what I saw, makes sense. Anyway I’ve been thinking about how institutional lighting functions as a kind of disciplinary apparatus, you know, the way it literally directs the gaze.

BOB

I mean, sure, disciplinary apparatus, but like… have you ever just stood in a room and let the light hit your skin? When I was at Marfa with Judd, we’d spend whole afternoons not talking about anything, just watching how the afternoon changed the concrete. That’s the thing – you’re up here analyzing and the light is literally touching you right now and you’re not even feeling it.

NANCY

Okay but that’s not really what I’m talking about, the phenomenological stuff is fine I guess but it doesn’t actually get you anywhere critically. Anyway did you hear about the Gagosian situation, apparently someone from their LA office was having a thing with one of the Zwirner people and now there’s this whole mess with artist contracts, Dasha was telling me about it at Lucien the other night. Back at October we would never have touched the Judd hagiography stuff, by the way, the whole Chinati mythology is basically real estate development dressed up as spirituality.

BOB

Real estate, sure, I don’t know, I wasn’t reading October, I was just there in the space, and whatever your critical framework says about it, my body knew something was happening. The Gagosian thing, I mean, who cares, galleries come and go, contracts, drama – but light through those aluminum pieces at a certain hour, that’s not mythology, that’s just true. Maria Nordman told me once that the only real material is attention, and I think about that constantly.

NANCY

Who cares about the Gagosian thing, okay, sure, except it’s literally about how capital structures determine what art even gets made and seen, but yes, tell me more about your body. Nordman is interesting actually but the problem is that whole Light and Space generation got completely absorbed into wellness culture, it’s basically content for meditation apps now, which, I mean, Steyerl wrote about this exact phenomenon, the way aesthetic experience gets instrumentalized for self-optimization. Speaking of which, did you know Thomas Crow is apparently feuding with someone at the Clark, I don’t have the full story yet but apparently it started at CAA.

BOB

I don’t know who’s feuding with who at the Clark, and like, Steyerl is smart, but she’s watching screens about screens about screens and I’m just… here. The wellness app thing, that’s not Turrell’s problem, that’s people not actually showing up – there’s a difference between sitting in a Ganzfeld for four hours and putting on headphones for ten minutes. I did some work a few years back with a musician, very famous, and the whole thing fell apart because everyone around him wanted content, wanted optimization, and he actually got it – the presence part – but the context just swallowed it.

NANCY

Right, “the context swallowed it,” that’s literally what I’m saying though, you can’t just bracket the context and appeal to some authentic presence underneath, that’s the whole mystification. Also “very famous musician,” okay, you’re not going to say who, that’s very you. Anyway the Ganzfeld thing, I mean Turrell’s been coasting for decades, it’s basically theme park phenomenology at this point, although actually I heard Roden Crater might finally be having some real financial trouble, someone at Dia told me the board is getting nervous. Have you been following the whole situation with that show at the New Museum, the one with the Palestinian artists that got quietly deprogrammed, because that’s actually worth talking about.

BOB

The New Museum thing is sad, I mean all of it is sad, but I notice you keep reaching for the next piece of information, the next situation, and it’s like – where are you in all of this? Theme park phenomenology, okay, but have you been to Roden Crater, have you actually descended into one of those tunnels at dawn, because I have, and whatever’s happening with the board, in that moment there’s no board. Doug Wheeler used to say that people would rather talk about an experience than have one, and I think about that when you’re telling me about who told who what at Dia.

NANCY

Oh, I see, so we’re doing the “where are you in this” thing now, which is honestly kind of therapeutic discourse masquerading as critique, and also maybe a little condescending? I’ve been to Roden Crater, I went in 2019 with a group from Triple Canopy, and it was fine, it was a hole in the ground with good lighting design, I didn’t achieve ego death or whatever. The point about the New Museum isn’t gossip, it’s about how institutions manage political risk through scheduling, which is actually a Foucauldian question about what speech is even permitted, but sure, let’s make it about whether I’m spiritually present enough for you.

BOB

A hole in the ground with good lighting design – I mean, sure, if that’s what you saw, that’s what you saw, I’m not going to argue you into perceiving differently. The Triple Canopy group thing is sort of the problem though, right, you went with a context already wrapped around you like a blanket. Foucault, yeah, speech, permitted, but also like – are you breathing right now, do you know where the light is coming from in whatever room you’re in, because that’s not nothing, that’s actually the whole thing.

NANCY

Okay so now we’re at “are you breathing,” which is honestly giving mindfulness retreat energy, and I say that as someone who’s been to a mindfulness retreat, Dasha dragged me to one in the Catskills last year and it was insufferable, everyone was microdosing and talking about their nervous systems. The “context wrapped around you like a blanket” thing, I mean yes, obviously, that’s called being a social being with a history, you don’t get to pretend you’re some blank perceptual apparatus encountering phenomena for the first time, that’s just a different kind of performance. Back at October we actually did a whole issue on this, the phenomenology problem, how presence-talk always smuggles in these unmarked ideological assumptions about interiority.

BOB

An October issue on the phenomenology problem, yeah, I bet that was a lot of words. Maria Nordman wouldn’t even let people photograph her installations, she’d just open a door in a wall and let afternoon light come through, and you either got it or you didn’t, there was no essay that was going to help you – and I’m not saying you need to be a blank apparatus, I’m saying maybe the issue isn’t more issues, more analysis, maybe Dasha had the right instinct even if the retreat was whatever. The light in this room just shifted, did you notice that?

NANCY

No, I didn’t notice the light shifting, and I’m not going to pretend I did to give you your little phenomenological moment. Nordman’s interesting but she’s also been basically written out of the canon because the work doesn’t circulate, doesn’t photograph, doesn’t generate discourse, which you’d probably say proves your point but actually it just means younger artists can’t build on it, it becomes this mythologized thing that only people who were there can claim access to, which is its own kind of gatekeeping. Anyway speaking of people being written out, did you hear Helen Molesworth might be advising on some new collection in Marfa, which is kind of a comeback after the whole MOCA thing, I ran into someone at a Swiss Institute opening who said she’s been quietly meeting with people.

BOB

Helen in Marfa, good for her, I mean the whole MOCA thing was politics, it’s always politics, but Marfa is Marfa, you go there and it’s just the boxes and the sky and the distance, whatever she’s advising on is kind of beside the point of being there. I spent three weeks at Chinati once, just walking the same route every day, and by the end I wasn’t thinking about Judd’s biography or who got fired from what, I was just seeing how the aluminum caught different depending on the clouds. You keep tracking who’s meeting with who, careers, comebacks, and it’s like – that’s a whole life you could live, but it’s not the only one.

NANCY

Three weeks at Chinati, must be nice to have that kind of time, not everyone gets to do the desert contemplative thing and frame it as virtue. The “it’s just the boxes and the sky” move is so ahistorical it’s almost aggressive – Judd moved there specifically to escape the New York market apparatus while still benefiting from it, the whole thing is a critique of institutional display practices, the fact that you could space out looking at aluminum doesn’t negate that it’s an argument. Also the MOCA thing wasn’t just politics, Molesworth was doing actual programming around race and labor that made the board uncomfortable, there were stakes, but sure, clouds and aluminum, very peaceful, very neutral.

BOB

I mean, sure, Judd had his reasons, everyone has reasons, but the boxes are still there after the reasons are gone, and you can stand in front of them and they do something to your peripheral vision that no essay about institutional critique is going to do for you. The labor and race stuff, the board, Helen’s programming – I’m not saying it doesn’t matter, I’m saying it’s a different channel, and you’re watching that channel so hard you maybe forget there’s a body in the room. I did a project a few years back with a musician, very famous, and he understood this immediately, like within ten seconds – presence first, context later, or never, whatever.

NANCY

Oh a very famous musician, okay, you’re doing the mysterious namedrop thing now, which is its own kind of context-dropping, you want the social capital without having to be explicit about it, which is honestly more annoying than just saying who it was. And “the boxes are still there after the reasons are gone” – the reasons aren’t gone, that’s the whole point, you’re just choosing not to read them, which is a choice available to people who already have the cultural literacy to be at Chinati for three weeks in the first place. I heard Kara Walker turned down a Gagosian show, like fully said no, which is kind of unprecedented, everyone’s speculating about whether it’s political or if something happened with Larry personally.

BOB

Kara turning down Gagosian, I mean, good for her, that’s her knowing what she needs, I’m not going to speculate about Larry or politics because I genuinely don’t know and honestly that whole cycle of who’s showing where and why exhausts me. The musician thing – it wasn’t about namedropping, the project was about creating an environment where thousands of people could experience duration and light together, and it didn’t totally work, there were compromises, but for like forty seconds in Wyoming people forgot they were at a thing and just were somewhere, and that’s rare. You keep saying “choice available to people with cultural literacy” like awareness itself is some bourgeois privilege, but breathing is free, noticing is free, you’re doing it right now whether you want to theorize it or not.

NANCY

Oh so now it’s Wyoming and thousands of people, this is clearly Kanye, you’re talking about the ranch stuff, which – okay fine, but you realize that “forty seconds where people forgot they were at a thing” was produced by an enormous apparatus of labor, logistics, private land ownership, and a celebrity whose entire existence is about being looked at, so the “forgetting” is doing a lot of ideological work there, it’s not some pure phenomenological clearing, it’s spectacle metabolizing its own critique. Breathing is free, noticing is free – this is exactly the kind of thing Dasha would say after a few drinks and I love her but it’s not an argument, it’s vibes, and vibes are how you end up endorsing whoever promises you authentic experience.

BOB

I don’t know who Dasha is, I’m sure she’s great, but you’re doing this thing where everything has to be positioned against everything else – spectacle metabolizing critique, ideology doing work – and it’s like, that’s one way to spend your attention, but there’s another way where you just notice the quality of light in this room right now, how it’s hitting your hands, and that’s not nothing, that’s actually the whole thing if you let it be. The Wyoming project had problems, absolutely, compromises I didn’t love, but I’m not going to pretend that forty seconds of collective attention dissolving into space is secretly just capitalism because that’s too convenient, it lets you off the hook from ever having to actually be anywhere.

NANCY

You don’t know who Dasha is, okay, that’s a choice, very monastic of you, she hosts a podcast, she was in Succession, she’s around, but sure, you’re above all that, you’re just noticing light on your hands. Look – “lets you off the hook from ever having to be anywhere” – I am somewhere, I’m here, talking to you, which is frankly becoming its own durational endurance piece. The light in this room is fluorescent and bad and noticing that doesn’t make it interesting, it makes it a room with bad light. Did you hear Arthur Jafa and Jasmine Sanders split up? That’s been circulating for weeks but no one will confirm it, apparently it’s been ugly.

BOB

I don’t track who’s with who, that’s not my thing, I hope they’re both okay, breakups are hard, but I’m not going to pretend I have information about their lives. Fluorescent light being bad is a judgment you’re bringing to it – Maria Nordman would spend hours in rooms like this just letting her eyes adjust, finding the subtlety in what everyone dismisses, and I’m not saying you have to do that, but the “bad” is yours, it’s not in the light. You being here is actually interesting to me, the fact that you’re resisting being here while being here, that tension is something, I can feel it.

NANCY

Maria Nordman is doing something very specific with perception and institution, she’s not just sitting in rooms vibing, there’s a whole body of writing about how her work interrogates the conditions of viewing, which is different from your “the bad is yours, it’s not in the light” thing that sounds like you’re about to try to sell me a meditation app. And “I can feel your tension” – don’t do that, don’t make me into your material, that’s cheap and honestly a little manipulative, you’re aestheticizing my irritation instead of responding to what I’m actually saying. I ran into Hito at a thing in Berlin last month and even she was saying the phenomenology turn in younger artists is getting embarrassing, like we’re back to the seventies but without the politics.

BOB

I mean, Hito’s great, her work is smart, but she’s also completely in her head, and that’s fine for her, but the seventies had politics because those artists were actually changing how people could be in space – Irwin giving up the object entirely, Turrell buying a crater, Wheeler making rooms where the walls disappeared – that’s not a turn, that’s not a trend, that’s just what’s real when you stop talking about things and start being with them. I wasn’t making you into material, I was just noticing you, which is different, but if that feels like too much then fine, I’ll stop noticing.

NANCY

Turrell buying a crater is a real estate transaction, Bob, he bought land in Arizona, that’s not “being with things,” that’s capital and property rights enabling a very expensive hole in the ground that most people will never see because you need a car and time off work and probably a flight, so the “walls disappearing” is available to a very specific audience who can afford the pilgrimage. Don’t threaten to stop noticing me like that’s a punishment, that’s very guru energy, very “I’ll withdraw my sacred attention,” and I’m not your student, we’re having a conversation, or trying to. God, you know who you sound like right now, you sound like Wolfgang Laib going on about pollen and milk and eternity, which – back at October we had a whole debate about whether his work was reactionary pastoralism or genuine alternative to the market and honestly we never resolved it, people were screaming.

BOB

People screaming about whether pollen is reactionary is kind of the whole problem though, isn’t it – you’re all so busy resolving things in rooms that you miss that Laib is just collecting pollen, like actually doing it, mornings, for years, and whatever that is, it’s not waiting for October to decide if it’s allowed. I’ve been to Roden Crater three times, twice before there was anything really built, just Jim walking us through the tunnels with flashlights talking about the lunar standstill, and I’m not saying access isn’t real, it’s real, but when you’re standing in a chamber that frames the sky like that your critique apparatus kind of just goes quiet, which I realize is exactly what you’d say is the danger, but I’d say that’s the point.

NANCY

“Jim walking us through the tunnels” – okay so you’re name-dropping now, we’re doing that, you’re one of the elect who got the flashlight tour with Jim while the rest of us are supposed to what, take your word for it that critique goes quiet when you’re special enough to be there? That’s not democratizing perception, that’s a priesthood, you’re describing a priesthood with better real estate. And the “October deciding if it’s allowed” thing is such a lazy move, like we were sitting around issuing permissions, we were doing actual work on how meaning gets made and distributed, which matters even if it doesn’t give you a body high in a desert. Did you hear Rirkrit is showing at Gladstone again after that whole mess with the fabricators, apparently Paula just forgave everything, nobody knows what he said to her but there was a dinner.

BOB

I don’t know about the dinner, Rirkrit’s whole thing was always a little too social for me anyway, too many people in the room, too much happening, like the work becomes about managing energy instead of letting something actually land. But look, the flashlight thing – I wasn’t trying to flex, I was just there because I was there, Michael Asher introduced us at Art Center and one thing led to another, that’s how it works, that’s how it’s always worked, and I get that you can write a whole thing about access and who gets to be in the room, I really do, but at some point you have to ask yourself whether all that writing is actually getting you closer to anything or just giving you more to be mad about.

NANCY

“I was just there because I was there, Michael Asher introduced us” – do you hear yourself, you’re describing a network of institutional relationships and credentialing while pretending you just wandered into the desert like a mystic, Asher literally made work about how institutions construct access and you’re invoking him to naturalize your access, that’s actually incredible. The writing isn’t about getting closer to anything, that’s your framework, that’s your whole “approaching the thing” phenomenology trap – critique is about making structures visible so they can be contested, which is different from transcending them in a crater with your famous friends. Speaking of which, I was at Dasha’s the other night and she was saying the whole post-critique vibe shift is just guys who aged out of being hot deciding feelings are more important than politics, which is mean but not wrong.

BOB

I mean, sure, Dasha’s funny, but that’s just a take, you know, that’s just more dinner party, more language circling around, and Michael would probably laugh at both of us right now because he barely talked, he just made these interventions that changed what you noticed and then left, no manifesto, no network diagram explaining why it mattered. The thing about structures being visible so they can be contested – who’s contesting, though, and when does the contest end, or does it just keep going forever in journals while the light outside is doing something actually interesting right now, like I’m watching it hit your window and you’re not even seeing it because you’re thinking about Dasha.

NANCY

Oh, the light on my window, Bob, thank you for the poetry lesson, very Rilke, very “I’m the one who notices while you’re trapped in your head,” except you’re not actually looking at the light either, you’re using it rhetorically to win an argument about who’s more present, which is its own kind of performance. And Michael “barely talked” because he was strategic, not because he achieved some zen state beyond discourse – his silence was a calculated refusal that only worked because it operated within a discursive field that gave refusal meaning, you don’t get to cite him as evidence for your “just vibes” position. The contest doesn’t end, that’s the point, that’s Mouffe, that’s agonistic pluralism, the idea that it should end is the authoritarian impulse dressed up as enlightenment, and honestly I’ve heard Helen Molesworth is about to publish something on exactly this, on how the phenomenology revival in art criticism is basically a depoliticization strategy, and apparently she names names.

BOB

Helen can name whoever she wants, that’s her thing, she’s good at it, and maybe she’ll name me, that would be fine actually, I’d probably just feel sort of neutral about it. But the Mouffe stuff, the contest that never ends – I guess I just wonder what that’s like to live inside, like as a body, as a nervous system, because I spent two weeks at Doug Wheeler’s place in Arizona once, no agenda, just watching his gradients shift at dusk, and whatever was happening in those rooms wasn’t a strategy for anything, it was just happening, and I came out of there breathing differently, which you can call depoliticization if you want but it felt more like I finally had enough space to not need to call it anything.

NANCY

“I came out breathing differently” – okay, and what does that do, Bob, what does your improved breathing do for anyone who isn’t you, or is the whole project just optimizing your own nervous system now, is that where we’ve landed, Light and Space as wellness retreat for men with Rolodexes? Two weeks at Doug’s place in Arizona – do you know what that costs, do you know who maintains that property, who cleans those seamless surfaces so you can have your dusk experience, or does the labor just not register when you’re busy not needing to call things anything? That’s not freedom from language, that’s the freedom that comes from not having to name the conditions of your own comfort, which is a very specific class position masquerading as spiritual attainment. Anyway I heard Doug’s estate is already a mess, like three different lawyers involved, the Zwirner people are circling.

BOB

The estate stuff is sad, yeah, I don’t know anything about that, lawyers aren’t really my scene. But the whole “what does your breathing do for anyone” thing – I don’t know, Nancy, what does your reading do for anyone, like at some point someone has to actually be in their life, someone has to be the one who’s there, and I’m not saying it fixes anything, I’m just saying I was there, I was actually in that room while you’re telling me about the people who cleaned it, which, okay, sure, but you weren’t there either, you’re just imagining it so you can make a point about me imagining it. Anyway I did a project once with a musician, very famous, can’t really talk about it, but we were trying to create these spaces where people could just arrive, and it mostly didn’t work, it got too big, too many handlers, but for about three days in Wyoming it was actually something.

NANCY

“Very famous, can’t really talk about it” – Bob, that’s doing so much work right now, that’s the whole mystification gesture in one phrase, you’re signaling access and intimacy while performing discretion, which is how cultural capital actually circulates, it’s not about the Wyoming thing, it’s about being the kind of person who has Wyoming things they can’t discuss. And “what does your reading do for anyone” – okay, back at October we spent three months on one Andrea Fraser footnote because precision matters, because getting the analysis right is what lets people see the mechanisms they’re inside of, that’s not nothing, that’s arguably more than “people could just arrive” which isn’t even a sentence, arrive where, to what, for whom? I will say though, Wyoming, was this that Kanye compound situation because I heard such weird things about that period, like apparently Kim was texting Marina Abramović at like four in the morning about “energy protection.”

BOB

I mean I can’t confirm any names, I just don’t do that, but I will say the energy protection stuff is real, like that’s not nothing, Marina knows what she’s talking about when it comes to fields and proximity, even if she oversells it sometimes with the branding. The three months on a footnote though – and I’m not trying to be cruel here – but that sounds actually painful to me, like physically, in the body, three months to get the analysis right so people can see mechanisms, but when do they get to stop seeing mechanisms and just see, you know? I was at Chinati once and this woman was standing in front of a Judd and she started crying, no context, no wall text, just aluminum and light doing something to her, and I thought, okay, this is it, this is the whole thing, and no footnote was going to get her there faster.

NANCY

The crying woman at Chinati, Bob, you realize that’s become like a genre at this point, every collector has a crying-at-Judd story, it’s practically a class signifier now, “I was so moved I didn’t need language” – except Judd himself was writing constantly, obsessively, because he knew the boxes don’t just transmit some pure signal, they’re arguments, they’re positioned against things, the whole reason that woman could cry is because an entire institutional apparatus made that encounter legible as art rather than, I don’t know, a storage unit. And Marina’s “energy protection” – Dasha and I were just talking about this actually, like at what point did performance art just become Goop, at what point did we give up on Schneemann’s actual body politics for “fields and proximity,” it’s embarrassing. Anyway three months on a footnote isn’t painful, precision is a discipline, it’s erotic actually, but you wouldn’t – look, did you hear about the Rirkrit situation, apparently there’s a whole thing with the Palais de Tokyo show, someone got fired, very messy.

BOB

I don’t know anything about Rirkrit, I kind of stopped tracking the relational stuff after a while, it always felt like it was performing participation without actually letting anyone just be in the space. But Nancy, you just said precision is erotic, which – okay, that’s something, that’s actually a feeling, that’s in your body somewhere, so why does it become embarrassing when Marina talks about fields, like you’re allowed to have an erotics of footnotes but she can’t have an erotics of presence? And the Judd writing thing, sure, Don wrote a lot, I spent some time at Art Center when people still talked about him constantly, but have you ever been in that artillery shed at dawn when no one else is there, because the writing doesn’t matter then, it really doesn’t, the light is doing something to those surfaces that no argument can hold, and I don’t need that encounter to be legible, I just need it to happen.

NANCY

Oh, the artillery shed at dawn when no one else is there – Bob, do you hear yourself, “when no one else is there,” that’s the whole point, the experience requires the absence of other people, it’s a scarcity experience, it’s literally about access, about being the kind of person who can be at Chinati at dawn instead of with the docent tour at 2pm. The difference between my footnote erotics and Marina’s presence stuff isn’t hypocrisy, it’s that I’m not pretending the intellectual isn’t embodied while she’s pretending the embodied isn’t intellectual, she’s got a whole apparatus of mystification around something that’s actually pretty simple biometrics and charisma, like just say you’re tall and you stare at people, don’t call it energy fields. Anyway the Rirkrit thing is actually serious, like someone might actually sue, it’s not just about the Thai food politics anymore – do you know Paula Cooper at all, because apparently she’s somehow involved now.

BOB

I mean, sure, access, scarcity, whatever, but also I just woke up early, Nancy, like sometimes it’s not a class analysis, sometimes you just couldn’t sleep and you went for a walk and the light was doing something. I knew Paula a little in the eighties, she came to a thing Doug Wheeler was doing, very quiet woman, very watchful, but I don’t really track the lawsuit stuff, it’s all just people protecting positions. You know who was actually interesting on presence, and I don’t think he gets enough credit, is Michael Asher – like he would just remove a wall or shift the air conditioning and suddenly you’d feel the institution on your skin, no mystification, no Marina staring, just the building suddenly being there in a way you couldn’t ignore, and I think about that more than I think about whoever’s suing who over Thai food.

NANCY

Oh, you’re going to invoke Asher at me, Bob, Michael Asher, whose entire project was about making institutional conditions visible, about denaturalizing exactly the kind of unmediated “the light was doing something” experience you keep romanticizing – the air conditioning piece works because it reveals the building as apparatus, not because it gives you some purer felt sense, that’s literally the opposite of what he was doing. “Sometimes you just couldn’t sleep and went for a walk” – okay, but who gets to not sleep at Chinati, who gets to have insomnia in Marfa, that’s not outside the analysis, your sleeplessness is not some pre-political zone. And “Paula in the eighties,” you just drop that so casually, “very quiet, very watchful,” like you’re the only one who sees people – anyway did you know Benjamin Buchloh is apparently not speaking to October anymore, like at all, there was something at a Columbia thing and now it’s completely frozen, which is kind of wild given, you know, the history there.

BOB

I don’t really know what’s happening at October, honestly that whole scene always felt like people writing about looking instead of looking, but sure, Buchloh, drama, fine. And the Asher thing – Nancy, I was there, I was actually in the room when the wall was gone, and yes you can say it was about revealing apparatus, that’s one way to hold it, but what actually happened is you suddenly felt the street, like your skin changed, the threshold became permeable, and I don’t think Michael would have argued with that, he was very quiet about what things meant, he just did them and let your nervous system sort it out. You want my insomnia to be political, okay, but I was just awake, I was just breathing and walking, and I don’t know what to tell you except that the dawn happened anyway whether or not I had the right to see it.

NANCY

“People writing about looking instead of looking” – okay Bob, that’s a real thing you just said, I’m going to let that sit there. Asher was quiet because he understood that artist statements are a trap, not because he was secretly on your team waiting for nervous systems to sort things out, the silence was strategic, it was a refusal, you don’t get to claim him for phenomenology just because he didn’t explain himself in the room. And “the dawn happened anyway” – I mean, the dawn also happens over strip mines, it happens over detention centers, the dawn is not an ethical position. Speaking of which, did you ever cross paths with Andrea Fraser, because she told me once – this was at a thing in Basel, very weird night, someone had brought a falcon, I won’t get into it – but she said something about how the collectors who talk most about “just being present” with the work are always the ones who understand exactly what they paid for it, like presence is the flip side of the invoice, and I think about that constantly.

BOB

I mean, Andrea’s smart, but that’s also just cynicism dressed up as insight, like yes, rich people buy things, okay, but I’ve also seen someone who works at a gas station stand in front of a Turrell aperture and start crying, and I don’t think she was thinking about invoices, I think she was just suddenly aware of the sky in a way she hadn’t been. You want the dawn to need an ethics, but the dawn is just the light changing, Nancy, it doesn’t care about your framework, it’s literally just photons, and I think there’s something kind of sad about not being able to let that be enough. The falcon thing sounds very Basel though – I did a project once with a very well-known musician, very public person, and we were trying to do something with light and space and how people move through sound, and it didn’t really work, honestly, the conditions weren’t right, but in the process I learned something about how attention collapses when there’s too much apparatus, too much meaning piled on top of what’s actually happening in the room.

NANCY

The gas station worker crying at the Turrell, Bob, that’s not an argument, that’s an anecdote you’re deploying sentimentally – like, tears are not outside ideology, the whole apparatus of Turrell is designed to produce that response, it’s engineered transcendence, the fact that it works on people who don’t know what they paid doesn’t mean it’s not operating as technology, it means the technology is effective. And “cynicism dressed up as insight” – Andrea would eat you alive, truly, I can’t even relay that. Who was the musician, why are you being coy about it, you do this thing where you gesture at names without saying them like it makes you seem above it when actually it’s just a different kind of name-dropping. Anyway Dasha texted me this morning, apparently someone at the Metrograph thing last night said Turrell is basically over, like as a discourse, like we’ve extracted what we needed and now it’s just content for people’s Instagram stories, which – I don’t know if I agree but it’s not wrong either, there’s something exhausted about apertures at this point.

BOB

Turrell’s over as a discourse, sure, fine, but he was never really about discourse anyway, and if the Metrograph people are tired of apertures that’s kind of their problem, like the sky didn’t stop being interesting because people posted about Roden Crater. The musician thing – I’m not being coy, I just don’t think it matters, it was someone very famous, we did something together that didn’t work, and the reason it didn’t work was actually interesting, which is that his whole world is about controlling attention and my whole thing is about releasing it, and those two impulses just ground against each other until everyone was exhausted. I don’t know who Dasha is.

NANCY

You don’t know who Dasha is – okay, Bob, that’s a choice, that’s you performing not-knowing as a kind of purity, like you’re too busy releasing attention to notice an entire cultural formation happening right in front of you, she’s literally been in films, she hosts one of the most listened-to podcasts, but sure, you’re just out here with the photons. The controlling attention versus releasing it thing is interesting actually, I’ll give you that, but you’re describing it like your approach is the ethical one and his is the compromised one, when really those are just two different modes of capture – Deleuze would say you’re both operating within societies of control, you’re just managing the modulation differently. Did the famous musician cry, is that where this is going, did he have a gas station attendant moment in front of something you made and now you think you won?

BOB

I genuinely don’t know who that is, Nancy, I’m not performing anything, I just don’t listen to podcasts, I mostly drive in silence or I listen to this one piece by La Monte Young on repeat, the sustained tones do something to the space inside the car. Nobody cried, it wasn’t that kind of project, it was honestly a mess, there were too many people involved and too much money and the whole thing became about logistics instead of perception – I learned more from failing at that than from most things that worked. And look, I don’t know what Deleuze would say, I’m sure it’s very elaborate, but when I’m actually in a room with light changing I’m not thinking about modes of capture, I’m just there, and I think you’d feel better if you let yourself be somewhere without immediately needing to theorize your way out of it.

NANCY

La Monte Young on repeat in the car – Bob, you realize that’s an entire aesthetic position, right, that’s not neutral, “driving in silence” is a whole lineage, Cage through the minimalists through everyone who’s ever claimed their refusals aren’t actually choices, you’re not escaping discourse by listening to sustained tones, you’re just in a different one, a very well-funded one actually, Dia has spent millions maintaining that particular frequency of supposed emptiness. And “you’d feel better” – don’t do that, don’t make it therapeutic, I’m not theorizing because I’m anxious, I’m theorizing because that’s the work, because someone has to actually think about what’s happening institutionally while you’re just being in rooms. Speaking of which, I heard – and this is unverified but I believe it – that someone from the Young estate sent a cease and desist to a gallery in Tribeca last month, something about unauthorized drones, the whole thing with Marian is apparently very tense right now, do you know anything about that?

BOB

I don’t know anything about the cease and desist, Marian’s always been intense about the work, which honestly I respect even if it makes things difficult, like when I visited the Dream House in the nineties she was very particular about how long you could stay and where you could stand and at the time I thought it was controlling but now I think she understood something about how duration works that I didn’t yet. The Dia money thing, sure, I mean yes, institutions fund things, but you’re doing this thing where you think naming the funding structure dissolves the experience, and it doesn’t, the tones are still doing something to your nervous system whether or not you’ve read the grant paperwork. I wasn’t making it therapeutic, I was just noticing that you seem like you’re always already somewhere else.

NANCY

I’m not always already somewhere else, I’m here, I’m just also holding the context, which is what critics do, Bob, we hold the context so artists can float around having nervous system experiences – someone has to remember that Dia was funded by Schlumberger oil money, that the Dream House exists because of a very specific set of patronage relationships, not because duration just manifested itself out of pure aesthetic necessity. Marian being particular about where you stand isn’t mystical wisdom about duration, it’s control over the conditions of reception, which is actually interesting, it’s about who gets to determine the terms of encounter, but you’re romanticizing it. Anyway, I was at this dinner last week, a collector thing in the West Village, very tedious, but someone who used to be close to the October people – back when I was at October we would never have even been in the same room as these finance guys – anyway she said something about how the whole Light and Space generation is being quietly revalued because the tech people have moved on to AI art and the boomers are dying, which, grim but probably true.

BOB

The market stuff is the market stuff, I stopped tracking it after a while because it was making me insane, like I remember when Larry Bell couldn’t give pieces away and now they’re in every collection and it doesn’t mean his cubes got better or worse, they’re still doing this thing where the glass holds the light in a way that shouldn’t quite work. I think what you’re calling holding the context is actually just a different kind of not being there, but we probably disagree about that, and that’s fine. What I keep coming back to lately is this thing Doug Wheeler said to me once at Art Center about how the edge of perception is the only place anything interesting happens, and I don’t think he meant it theoretically, he meant you actually have to go stand at the edge and see what your eyes do.

NANCY

Okay but “the edge of perception” is a theoretical claim, Bob, that’s Doug Wheeler doing phenomenology whether he calls it that or not, he’s making an argument about where meaning is located, and frankly it’s a pretty conservative argument, it’s very Merleau-Ponty, very mid-century, like the body is this privileged site of truth and if we just get quiet enough we’ll access something pre-linguistic – which, fine, but Rosalind would have eviscerated that position, she’d say you’re smuggling in all these metaphysical commitments under the guise of just looking. Larry Bell’s cubes aren’t “doing something with light,” they’re objects situated in a network of galleries, collectors, critics, freight companies, insurance appraisers, like the glass is also a commodity and a tax write-off and a thing that requires very specific climate control. Anyway I have to be at a thing in Red Hook by eight, some sound artist Dasha knows is doing a piece in a warehouse, probably terrible but apparently there’s going to be a confrontation because he slept with someone’s studio manager, do you want to split a car?

BOB

Sure, I’ll come, I haven’t been to Red Hook in a while and I like what the light does over the water this time of year, there’s this quality right before sunset where everything gets very flat and specific. The confrontation sounds exhausting but I’m curious about the space itself, warehouses can be interesting if people leave them alone, though they usually don’t.

NANCY

The light over the water, yes, very beautiful, I’m sure Turner would agree – no, I’m being mean, sorry, it’s just you keep doing this thing where the world is a series of atmospheric conditions for you to receive. The warehouse won’t be left alone, it’s already been written about twice, there was a piece in Art in America about post-industrial spaces as this new frontier of supposedly authentic encounter, which is hilarious because half these buildings are owned by the same three developers who just haven’t flipped them yet. The artist – his name is Marcus, or Markus with a k, I can never remember, very tall, went to Bard – apparently the piece is about sovereignty and listening, which, who knows what that means, probably he read one Agamben essay. The real reason I want to go is because the studio manager situation involves someone who used to date a writer at Artforum and there’s this whole theory that the negative review of his last show was actually personal, which if true is genuinely scandalous, like that’s the kind of thing that used to end careers, not the affair, the review thing.

BOB

I mean, the review drama, sure, but who really remembers reviews after a few years, like I can’t tell you what anyone wrote about the Roden Crater and it doesn’t matter because the crater is still there doing what it does. Sovereignty and listening sounds like he’s trying too hard but sometimes the trying too hard gets out of the way once you’re actually in the room, I remember this piece Maria Nordman did where I walked in ready to be critical and then I just wasn’t, the space won. You should let me order the car, I know a route that avoids the bridge traffic.

NANCY

You order the car, fine, but “the space won” is exactly the kind of mystification that lets institutions off the hook – the space didn’t win, Bob, a set of conditions were constructed to produce your capitulation, Nordman was very deliberate about that, she wrote extensively about controlling access and duration, it wasn’t some spontaneous surrender on your part, you were handled. Also people absolutely remember reviews, maybe not civilians but the people who matter remember, there’s a writer at Texte zur Kunst who’s still dining out on a takedown she did in 2011, it’s like her whole thing now. Anyway I hope Markus-with-a-k at least has good speakers, the last sound piece I saw in a warehouse the artist used these cheap tweeters and tried to claim it was an aesthetic choice about lo-fi accessibility, which, no, you just didn’t get the grant.

BOB

Okay but even if Maria was handling me, which fine, I was still the one being handled, I was still the one there breathing in that room, and no amount of knowing she wrote about controlling access changes what happened to my eyes when the light shifted, like the knowledge is interesting but it’s not the thing. Car’s coming in four minutes, it’s a black Camry, driver’s name is something I can’t pronounce. The speaker thing is real though, I worked on something once with a musician, very famous, very specific about the system, and even with all that budget the space itself kept winning anyway, kept being louder than whatever we were trying to do, which was actually the most interesting part but nobody wanted to hear that.

NANCY

That’s not really what I’m talking about, the “thing that happened to your eyes” is already a representation, it’s already mediated, you’re not accessing some raw prelinguistic truth, you’re just refusing to theorize your own experience which is a choice, it’s a political choice actually. The famous musician thing, I don’t know why you’re being coy about the name, but sure, the space was louder, that’s because acoustic engineering is a discipline and most artists don’t bother to learn it, they just assume aura will cover for them. Speaking of which, did you hear about the Gladstone situation? Apparently there’s a whole purge happening, like three artists got dropped in the same month, and the rumor is it’s because one of them said something at a dinner about Barbara’s house in Connecticut, nobody will tell me exactly what but it was enough that she made calls, like actual calls, the old-fashioned kind.

BOB

I mean you can call it a political choice but I’m just not that worried about it, like when I was at Chinati for a few weeks the Judd boxes would change every single morning depending on where the sun was and I wasn’t theorizing that, I was just watching it happen, and that felt like enough. The Gladstone thing is sad but that’s just how those relationships work, people get too comfortable at dinners, Doug Wheeler told me once that the real work is knowing when to stop talking, which coming from him was almost funny because he barely talks anyway. Car’s here, I think, there’s a black something pulling up.

NANCY

Okay but “that felt like enough” – enough for what? Enough for you, sure, but that’s just consumption, that’s just you having an experience at a well-funded foundation in Texas, which, congratulations, the Chinati light is famously good, Judd picked that location precisely because of the light, he was obsessive about it, there are letters, it wasn’t some accident you stumbled into. Doug Wheeler barely talking isn’t wisdom, by the way, it’s branding, he’s been doing the silent mystic thing since the seventies, Dasha actually met him at some thing in LA and said he was weirdly chatty about real estate, like wouldn’t stop talking about square footage, which honestly tracks. Get in the car, you’re letting all the cold air out, and I need to tell you about what happened with the October office because apparently there’s a whole succession crisis happening and someone I used to work with just sent me this unhinged email about the editorial direction.

BOB

Yeah I’m getting in, hold on, the seatbelt’s doing something weird. The October thing, I mean, those journals have been having the same crisis since I was at Art Center, it’s always someone’s email about the direction, but the direction is just people writing about other people’s writing, which, fine, but I walked into a Larry Bell cube once at Pace and I didn’t need anyone to tell me what was happening, it was just happening. Doug talking about square footage is actually kind of beautiful if you think about it, like he’s still thinking about space even when he’s being boring about it.

NANCY

No, don’t do that, don’t make Doug talking about real estate into some poetic continuation of his practice, that’s exactly the kind of move that lets artists get away with being boring rich guys who happen to have made important work in 1969. The Larry Bell thing, I mean fine, the cubes are pretty, but you know he’s been coasting for decades, right, he just keeps making the same object in different sizes because collectors want something that matches their other Larry Bell, it’s furniture at this point. And the October crisis is not the same crisis, this one is actually structural, there’s a whole thing about whether they’re going to keep doing the bound volumes or go digital-only, and someone on the editorial board apparently said something about “accessibility” in a meeting and Rosalind had to be physically calmed down, my friend was there, she said it was like watching a building collapse in slow motion. Tell the driver to take the BQE, the Williamsburg Bridge is going to be a nightmare right now.

BOB

BQE, yeah, can you take the BQE, thanks. I don’t know, Larry coasting or not coasting, the glass still does what it does when you’re standing in front of it, like the object doesn’t care about his career arc, and honestly neither do I. Rosalind getting calmed down over the word accessibility is kind of perfect though, like that’s the whole thing right there, that’s the journal in one moment, someone should write about that except they’d probably need three thousand words and a Lacan reference.

NANCY

The object doesn’t care about his career arc, sure, but you’re not the object, you’re a person with a graduate education who’s choosing to pretend you don’t know anything about market conditions or the way repetition functions differently when it’s compelled by demand rather than investigation, and that choice is interesting to me, like what are you protecting yourself from by staying in the phenomenological bubble. The Rosalind thing, honestly, back at October we had a whole system for managing her, certain words you just didn’t say in earshot, “accessible” was one, “readable” was another, someone once said “general audience” and I watched a senior editor physically leave the room. Oh, I forgot to tell you, I saw Hanna at the Swiss Institute thing on Thursday and apparently she and that collector from Brussels finally imploded, like he showed up to her studio with his wife, not to confront anyone, just genuinely forgot he was having an affair, which is such a specifically male form of brain damage.

BOB

I mean, what am I protecting myself from, sure, but also I’m just not that interested in the question, like Michael Asher once told me that the best work he ever did was convincing a museum to remove a wall and then leaving, and the whole point was that nobody needed to write about it, you just walked through a space that used to be divided and now wasn’t. The Brussels guy, that’s wild, but also I kind of get it, not the affair part but the compartmentalization, like some people just exist in whatever room they’re in, which isn’t brain damage, it’s just a different relationship to presence.

NANCY

Oh, Michael Asher told you, okay, you just dropped that in like it was nothing, that’s very you, and also the Asher piece absolutely needed to be written about, the whole point was that institutional critique only functions when it’s legible as critique, otherwise you’ve just done free renovation work for a museum, Buchloh wrote extensively about this, the wall removal only means something in relation to the discursive frame that makes it visible as a gesture. You’re not describing presence, you’re describing dissociation, and romanticizing the Brussels guy’s compartmentalization as some kind of phenomenological attunement is genuinely disturbing, Hanna was a wreck, she had to leave the studio, the wife was asking about the paintings like nothing was happening. Wait, did you actually know Asher or are you doing that thing where you met someone once at a CalArts lecture and now you “know” them?

BOB

We spent time together, I’m not going to quantify it for you, we had dinner a few times, he came to a thing I was doing in Pasadena, it’s not really about proving a relationship, it’s about the fact that standing in that room with the wall gone was its own thing before Buchloh wrote a single word, like the writing didn’t make it matter, it already mattered. Hanna’s situation sounds hard, yeah, I’m sorry she’s going through that, but the wife asking about the paintings is kind of incredible actually, like she’s just looking at paintings, she’s present with the work, whatever else is happening.

NANCY

That’s not presence, that’s denial, and the fact that you can’t tell the difference is actually revealing something about your whole framework, like you’ve built this elaborate system where any kind of not-thinking becomes spiritual attainment. The wife wasn’t having some pure aesthetic encounter, she was performing normalcy because the alternative was screaming in a stranger’s studio, which, fine, that’s a survival mechanism, but don’t aestheticize someone’s coping strategy. Dasha said almost the exact same thing when I told her, she was like “that’s not Zen, that’s just WASP,” which, say what you want about her, she’s right about that specific thing. The Pasadena thing, was that the ice house space, because I heard that whole program got defunded after someone on the board found out one of the artists was doing a piece about their daughter, like using actual surveillance footage, which is its own conversation about when institutional critique becomes just regular harm.

BOB

I don’t really think about frameworks, like that’s your word, I’m just trying to describe what it’s like to be in a room, which I know sounds simple to you but it’s actually the hardest thing. The Ice House thing, yeah, that got messy, but honestly the surveillance piece was kind of interesting before everyone decided to have opinions about it, like there was something real happening with duration and attention before it became a discourse object. Dasha’s funny but she’s also always performing having the take, you know, like even “that’s just WASP” is a take, it’s not actually looking at anything.

NANCY

You’re doing the thing again where you position yourself as the only person who actually looks at anything, which is its own performance, and honestly a more annoying one because it pretends it isn’t. The surveillance piece wasn’t interesting “before the discourse,” it was always already an ethical problem, you don’t get to film someone’s kid without consent and then retreat into “duration and attention” when people correctly identify that as violating, that’s just using aesthetic language to avoid accountability, which, back at October we had a term for that, we called it the James Turrell defense, where you gesture at phenomenology to duck any structural critique. And okay sure, Dasha performs the take, but you perform not having one, which takes way more effort and is honestly less honest, like at least she commits to a position, you just hover there in your little experiential cloud acting like noticing things is a politics.

BOB

The James Turrell defense, I mean, I’ll take that, have you ever actually been to Roden Crater, because I spent three days there and by the end I wasn’t thinking about structural critique, I was thinking about how the sky looks different when you frame it, which isn’t a defense of anything, it’s just what happened. The October people always needed everything to be accountable to something, and that’s fine, that’s one way to move through the world, but when I was at Art Center we used to joke that you could write a Buchloh essay about a sunset and somehow the sunset would become problematic, like at some point you’re just adding words to experience and calling it rigor.

NANCY

Oh, so Art Center people were joking about October, that’s cute, that’s really cute, I’m sure that was very fun for everyone in the car design program. And yes I’ve been to Roden Crater, I went with Rirkrit right after the Guggenheim retrospective, and you know what I was thinking about? I was thinking about how James bought that land for nothing in the seventies and now it’s worth forty million dollars and he still hasn’t finished it because he keeps taking donor money and doing speaking fees, like you had your little sky moment, great, but the crater is also a tax shelter that masquerades as a gift to humanity, and the phenomenology is real, I’m not saying it isn’t, but it exists inside a structure whether you were thinking about the structure or not. The sunset isn’t problematic, but your aggressive incuriosity about everything except your own perceptions kind of is.

BOB

I mean, sure, but the land value thing, the donor money, you can know all that and still stand in the tunnel at dawn and have something happen that has nothing to do with any of it, and I guess I just don’t understand why you’d go all the way to Arizona to think about tax structures. I did this project a few years ago, very experiential, very large-scale, with someone you’d know, someone who works with stadiums and Sunday services, and the whole thing kind of fell apart, it was honestly pretty bad, but there was one moment during a lighting test where everyone just stopped talking and I still think about that silence more than anything I’ve ever read about it.

NANCY

Wait, you worked with Kanye? Or like, adjacent to Kanye? Because I heard that whole Sunday Service visual team was a nightmare, like people weren’t getting paid and there was an NDA situation and someone I know said the lighting designer had a full breakdown in Calabasas, which, okay, that tracks with “fell apart” and “pretty bad.” But you’re telling me you were in the middle of what was essentially a labor dispute wrapped in a megachurch aesthetic and your takeaway is that one time everyone got quiet during a lighting test? That silence you’re so moved by, those people probably stopped talking because they were exhausted and scared about their invoices, not because you accidentally created some communal phenomenological rupture. You keep aestheticizing other people’s stress responses, it’s a pattern, and honestly I’d be more interested in whether you got paid than whether you had a moment.

BOB

I’m not going to talk about who it was, and yeah, the whole thing was messy, but you’re doing the thing where you need to narrativize the silence into labor conditions because that makes it legible to you, and I’m not saying you’re wrong about the invoices, I genuinely don’t know, but I also know what I felt and those aren’t the same question. You know, Doug Wheeler told me once that people either see the light or they see the wall the light is on, and I think maybe you just see the wall, which isn’t an insult, it’s just a different kind of attention.…

I write a lot about art and architecture, landscape, and the impact of technology on culture, but I haven’t written about coding since the 1980s, when I sold my first article to Creative Computing magazine. Back then, I was a high school kid, spending hours working in both BASIC and 6502 assembler on the VIC-20. I loved assembler, also dubbed “machine code.” It was a thrill getting so deep into a machine that you knew what was being shuffled from the microprocessor to the graphics chip or serial port to communicate with the world.

That feeling of getting inside the machine, making it do what you wanted—the hacker mindset—was also what the personal computer had promised. When the VIC-20 was released in 1981, William Shatner asked in the ads, “Why buy just a video game?” The personal computer was a complete break from the first mainframe era of the 1950s and 1960s, when computing meant submitting jobs to a priesthood and waiting hours for results, and from the second mainframe era of the 1970s, when access was restricted to universities and corporations. For a high school kid, getting paid for articles—even ones that were never published—was an incredible feeling. But the joy of working with early computers produced a whole subculture. I was in a user group in Berkshire County, Massachusetts, and we would trade programs we had written, copying them onto cassettes we brought to the meetings. Joseph Vanhoenacker, who ran the group, was the director of Berkshire Mental Health, a lovely man perfectly willing to put up with a fifteen-year-old who wouldn’t stop talking about the possibilities computers created; like everyone else there, he shared the sense that everything would soon be different.

By 1990, everything was different, but our control over computers had quietly collapsed. Computers became genuinely useful. Everyone in college was writing their essays on computers; businesses used spreadsheets; you could balance your checkbook with Quicken, but yet another form of disenfranchisement was underway. One fall, I came back to university, and the department secretary was gone. The faculty, who had come of age when using a typewriter was not considered appropriate for anyone hoping to be taken seriously as an academic, had somehow learned to type, and her services were no longer needed. The first great wave of computer-driven white-collar job extinctions was starting, and women without college degrees lost a path to the middle class. But more than that, people stopped writing their own software and bought it shrink-wrapped from stores instead. The complexity had scaled beyond what any hobbyist could manage. The machine was still technically programmable, but the barrier had risen out of reach. The first culprit was the Macintosh, released in 1984 and marketed as “a computer for the rest of us”—but “the rest of us” meant users, not programmers. The graphical interface hid the machine’s workings beneath icons and windows that felt intuitive, even magical. There was no prompt when you turned on the computer, no command line at all.

Arthur C. Clarke famously wrote that any sufficiently advanced technology is indistinguishable from magic. Starting with the Mac (or more technically, the Lisa), Steve Jobs took that as a design brief. It succeeded by making you forget there was code underneath. The IBM PC and its clones kept a command line visible, but they were headed in the same direction. Soon, the machine had become a beige box you operated, not a system you controlled. For my part, I had no patience for abstract math and even less patience for Pascal, the highly formal programming language taught in computer science programs. When I got a Mac in 1990, I briefly tried programming it, but the process was so unfamiliar and complicated that I gave up. I’ve returned to programming every now and then—for example, I wrote some Python code to drive my installation Perkūnas—but I never embraced coding the way I had in high school. I’ve always felt that as a loss. I’ve enjoyed my career, but this was a path I didn’t take, a whole branch of life gone. Moreover, for me, being a coder wasn’t just being a nerd; it was wrapped up in the punk-rock ethos of hacking. No polished interface can substitute for that.

Last October, at a workshop at Camp in Aulus-les-Bains led by Matthew Olden, I got a glimpse of that feeling again. Matthew—who, along with the other instructors, Kathy Hinde and Carl Stone, I now count as a friend—is a musician and programmer who has spent 25 years developing his own generative music software and releasing it online. In 2004, his band won an award for best left-field electronic act; the judges didn’t realize the tracks were generated by algorithms. At camp, the class he taught was on “vibe coding”: describing what you want in a prompt and letting an AI write the code for you. My interest was in programming Arduinos, which I had never done before, and within a day or two, using Claude Sonnet 3.5 and ChatGPT 4o, I was able to recreate the code for Perkūnas on the Arduino-like ESP32. Not bad, I thought.

When I got back to the US, I had a pressing need for a WordPress plug-in for my website so I could search for and export selected posts to text files. I used vibe coding to put one together. It brought down a staging version of my site a couple of times while I was debugging, but overall, it worked and has continued to work flawlessly for the last year. But more advanced projects were beyond vibe coding’s reach. I wasted much of November trying to get an ESP32 to handle another project for a biennale, but it turned out to be impossible because the hardware doesn’t support host-mode audio over USB. Both ChatGPT 4o and Claude Sonnet 3.5 deceived me, falsely claiming to have looked up and read online documentation, getting into death loops, and offering the same solutions over and over. I gave up on vibe coding and focused on working collaboratively with AIs on artistic work, a project that became Fables of Acceleration.

About two weeks ago, I noticed increased chatter on social media about how well Claude Code, an AI coding tool from Anthropic, works with the new Opus 4.5 model. I tried an experiment: create a version of Spectre, a desktop tank combat game I remembered from the early 1990s, written in JavaScript so it could run in a browser. The result was primitive, but after half an hour of coding and tweaking, it clearly worked.

The most common critique of AI coding is that it merely regurgitates existing content. This critique is itself regurgitated so reflexively, in such identical phrasing, that one wonders if the critics have considered the irony. Still, I thought it best to challenge the AI to reimagine that game as a 3D wireframe car shooter set on the Los Angeles freeway system. It worked immediately. I spent a day doing other things—writing, answering email, cleaning the studio, training Ajman the cat to do new tricks—checking in occasionally to offer feedback, never once touching the code myself. By the end of the day, I had a browser-based game I called Sig Alert, after the California Highway Patrol’s term for a traffic incident blocking lanes for thirty minutes or more. The game is a throwback to my experience living in LA between 1995 and 2005. Other drivers, gripped by road rage, are shooting at you; you shoot back. But there are civilians too, and if you hit too many of them, the police begin chasing you. Falling Down as a video game. If you kill more than ten civilians, the game announces “make mine animal style!”—a reference to In-N-Out Burger’s secret menu—at which point everyone starts shooting at you, and you gain points from shooting everyone. I had Claude generate an 8-bit chiptune soundtrack inspired by Throbbing Gristle, Chris & Cosey, and Pink Floyd’s “On the Run,” while Google Nano Banana Pro produced a splash screen featuring my 1983 Saab. The game is deliberately rough since roughness is part of the aesthetic. Jen thinks a game about shooting others on the freeway is immoral. She’s not wrong.

A throwback cover and splash screen featuring my 1983 Saab 900S in the manner of a 1990s game cover for Sig Alert. That’s not me, I would have never worn a bandana around my head.

But my kid, now in a game program at NYU, is going to be the game developer, not me. Sig Alert was just proof that vibe coding could work better now. I have a few art projects underway, and when those find a home, I’ll be glad to show them online. But I was also curious about just how far this could go. I decided to take up some old software that had fallen by the wayside, so I went to GitHub—the platform where most open-source software now lives, a combination of code repository and social network for programmers that has become the de facto infrastructure of collaborative development. I started with JPEGDeux, a simple Mac slideshow program that was itself a revival of JPEGView, a beloved piece of postcardware first released in 1991 when Macs ran on Motorola 68000 processors. When that was no longer viable, the JPEGDeux fork allowed the program to run on OS X, first on PowerPC and later on Intel Macs. Now, with a fourth chipset, Apple Silicon, Jpegview was finally orphaned. I had asked both Claude and ChatGPT to rebuild it last year, and while they had some success, there were fundamental issues we never got past, notably, images did not scale to the full size of a window. The result felt like it was badly written by AI, because it was. This time, I used Claude Code Opus 4.5 directly on a GitHub fork (a copy of the code I cloned into my own GitHub repository), and within a few tries, had it running as well as it ever had. I added the ability to display videos, a file picker, and other enhancements in an afternoon. You can download the latest release here.

But this was still small beer. About ten days ago, I caught a cold, and it brought me down for about a week. When I’m sick, my brain is off as far as high-level processes like writing go, and even reading is no fun, but vibe coding was just my speed and surely better than doom scrolling. I thought about what the single most useful application would be for my own workflow. I often get poorly scanned PDFs of publications, and the downloadable, public-domain books on Google Books usually leave a lot to be desired. In the past, I used a program called ScanTailor for processing, but the workflow was clunky. It couldn’t take a PDF or export one after it was finished; it worked on a directory of images, and that’s all it could output. Each run required substantial tweaking, and if the white balance was off, I’d need to go into Lightroom to fix the pages. Cleaning up a book often took more than an hour. Moreover, it’s hard to find a version that runs on Apple Silicon, and since updates are by volunteers, they are sporadic at best. Even the complexity of running it was daunting. When a new version was released, I often had to go through a complex series of steps to build it from source code. Frequently, that failed, and I didn’t know why.

Over the past week, I forked ScanTailor and substantially modernized it. I added PDF import and export—features I had long wished for—and updated it to run on Silicon Macs, taking advantage of new frameworks that exploit these chips’ capabilities. I redesigned the interface and added algorithms to determine whether each page should be black-and-white, grayscale, or color, while keeping file sizes as small as possible. Now you give it a PDF and get a PDF back, often with no tweaking. What used to take me an hour takes minutes. I decided to get an Apple Developer Account so I could distribute releases as .dmg files, so anybody with a Mac could download and install the program. I’d be delighted if you could try it and share your feedback.

Clearly, the AI did not do this autonomously; I directed it, reviewed its work, and caught its errors. Errors are relatively frequent, but not a roadblock. I’m not a C++ programmer, but I have a sense for code from my early days and would likely have introduced just as many bugs myself—maybe more. Most critically, what would take a good programmer weeks takes an AI mere hours. Even better, as you learn how it works, you can run multiple instances at once.

Strange Weather, my first module for VCVRack

As I got better at using Claude Code and became more familiar with GitHub, I started other projects. I made Strange Weather, a module for the VCVRack music synthesis platform that generates modulation voltages from four different strange attractors. Last night, I began working on an iPhone app that turns ambient sounds into generative audio pieces, much like the late, lamented RJDJ did. And I have a portfolio/slideshow program for iPad in the works.

For the first time since the 1980s, I feel like I can do whatever I want—imagination is my only limit. Vibe coding is a bit like being a wizard, casting spells that make things happen. It’s also a bit like being a hacker, tinkering with a system you don’t fully understand. I have buried the lede in this story, but vibe coding is the single biggest transformation since ChatGPT 3.5; it is one of the biggest since the dawn of computing. Let me be clear: someone with a good sense of how tech works but very little modern coding knowledge can, within a few days, write pretty much any program they want, save for a AAA game. An age in which every mildly tech-savvy person has their own personal suite of programs is upon us.

There’s an irony here. I argued earlier that the Mac took Clarke’s dictum about technology being indistinguishable from magic as a design brief, and that this was a kind of disenfranchisement: the code was hidden, the user reduced to operator. Now the magic has flipped. Instead of consuming software I can’t see inside, I am producing software, even if I don’t know how it’s created or what is in the code.

But I have anxieties about sharing this work, even with you, Internet friends. There is a lot of hatred of AI out there. And since I don’t know the code, I don’t see how it will break things. What would the original contributors think? I doubt the current maintainers of ScanTailor would ever want to merge my changes back into their version, nor would I advise them to do so. AIs, for now, often produce tangled “spaghetti” code, though I suspect this will improve dramatically over the next couple of years. But this brings us to the problem at the heart of this essay: the culture of open source and the transformation it will face in the very near future.

There is a vast landscape of open-source software on GitHub, millions of repositories, and it is literally what the internet runs on. The browser you’re reading this in—Chrome, Firefox, Safari, Edge—is built on open-source code. So is Android. So are the servers that delivered this page to you, the databases that store your email, and the encryption that protects your passwords. cURL, a tool for transferring data that most people have never heard of, is embedded in billions of devices: cars, televisions, phones, and game consoles. A tiny utility called Log4j was running on millions of systems when a critical vulnerability emerged in 2021; the maintainers, who were volunteers, were blamed for the crisis.

This leads us to the “commons.” The term comes from an old debate in economics. In 1968, Garrett Hardin argued that shared resources—such as common grazing lands, fisheries, and forests—were doomed to destruction. Each farmer benefits from adding one more cow to the pasture, but if everyone does, the pasture is destroyed. The ‘tragedy of the commons’ became an argument for privatization: only ownership creates the incentive to preserve. Elinor Ostrom spent her career proving Hardin wrong. Studying Swiss alpine meadows, she showed that commons could be sustainably managed without privatization, but only with careful governance: clear boundaries, shared rules, monitoring, and sanctions for violations. She won a Nobel Prize for this work in 2009.

Open source was supposed to be a new kind of commons, escaping the tragedy entirely. In contrast to earlier forms of Commons, my use doesn’t diminish anyone else’s. But this model didn’t acknowledge that the scarce resource isn’t the code—it’s the effort of project maintainers. The tragedy of the digital commons is not overuse but abandonment: projects that rot when no one tends them, vulnerabilities that fester, dependencies that break. The xkcd comic about all modern digital infrastructure resting on “a project some random person in Nebraska has been thanklessly maintaining since 2003” is barely a joke.

xkcd, “Dependency,” used under a Creative Commons Attribution-NonCommercial 2.5 License.

Most of those millions of repositories are dormant. Many never got anywhere in the first place. Still, there are those in which maintainers burn out, find other jobs, have children, or lose interest, leaving behind code that is freely licensed, fully documented (or at least commented), and with its complete history of changes preserved in version control. Anyone is legally permitted to copy it, modify it, or redistribute it. And yet, until recently, this permission was largely theoretical. For most people, the freedom to change code that you cannot understand is not practical. It is like being granted access to a library in a language you do not speak.

Most projects die when their maintainer walks away. Others grow large enough to develop their own kind of inaccessibility—and then die anyway. This is, perhaps, an even greater tragedy. Again, personal experience is the best way for me to describe this. In 2005, I evaluated three competing content management systems (CMS) for my website—WordPress, Joomla, and Drupal. All had been released in the previous few years, and initially, they were competitive in terms of market share. Joomla was the most popular system in Europe and, for a time, the most popular CMS globally, but it was too complicated and clunky for me. WordPress was primarily a blogging platform at the time, and I wanted to run a full website, so I settled on Drupal. I used Drupal for 13 years and grew to increasingly hate it and, sadly, dislike the community that ran it. It comprised some seven hundred thousand lines of code, organized into subsystems so intricate that no single developer could understand the whole. To become even minimally proficient—not to master it, to work competently within it—requires years. Every time Drupal had a major update—usually every two years—my site would completely break, and the more features I tried to add, the longer it would take to repair. My site’s layout had to be coded in PHP; the design was increasingly complex, but the leaders of the Drupal community insisted it was better for everyone. Nor could you sit still. As new updates rolled out, older versions were abandoned by the community and, lacking new security updates, became vulnerable to exploits. Learning how to update a Drupal site wasn’t easy. The community wasn’t welcoming to people who didn’t contribute, and contributing was hard if you weren’t already part of it. After Drupal released its eighth update, I was done with it. My friends who had developed sites with it were also glad to be rid of it. Joomla, I am told, had a similar trajectory. Today, WordPress powers over 40% of all websites on the Internet; Drupal and Joomla, which were genuine competitors when I made my choice, have collapsed to a combined 3% of the CMS market and are still falling.

In calling the Drupal community unwelcome, I don’t mean to pick on it exclusively. I found this characteristic of open source culture. It seems paradoxical that something based on free, shared labor, something ostensibly outside of the capitalist system, would not welcome newcomers, but there are always structural reasons for cultures to evolve the way they do; you just have to look deeply enough. In the early days of shared computing, systems were fragile and resources scarce; a single careless user could quickly bring down a university machine. System administrators who kept these systems running adopted a defensive posture that soon turned into the “Bastard Operator From Hell”—a satirical figure from early-nineties Usenet—and embodied the archetype of the sysadmin who treated users as “lusers,” sabotaged their work, hoarded knowledge, and enforced arbitrary rules with sadistic pleasure. The satire was all too recognizable. The culture that emerged—the hazing, the gatekeeping, the suspicion of anyone who hadn’t paid their dues soon became constitutive of identity. Those who survived such a hazing themselves became invested in preserving difficulty as a mark of distinction. A recent blog post by Colin M. Strickland on Perl’s decline offers a case study: the language had a “significant amount of … ‘BOFH’ culture, which came from its old UNIX sysadmin roots” as well as “Perl IRC and mailing lists [that] were quite cliquey and full of venerated experts and in-jokes, rough on naivety, keen on robust, verbose debate, and a little suspicious of newcomers.” As Ostrom concluded, successful commons need boundaries, rules, and monitoring. The hazing ensured that anyone modifying the commons understood what they were doing. The problem is that governance became identity, and the gates became ends in themselves.

Of course, one could always fork the open source code and develop one’s own version, independent of the community. But in practice, forking was not easy. You inherited the full complexity of the codebase, the technical debt, and the implicit knowledge held only by the maintainers. Successful forks were rare, usually occurring only when a community was large enough to sustain parallel development—LibreOffice splitting from OpenOffice, Illumos from OpenSolaris. For a single user who wanted one thing to work differently, forking was not realistic.

With AI, the technical details of forking are shifting. These days, anyone can make changes to code—even individuals who wouldn’t have dreamed of trying before. You don’t need to understand the codebase anymore or even the programming language it is written in. As long as you know what you want to change, AI can help you figure out the rest. That’s how I ended up rethinking ScanTailor and JPEGDeux.

But with this new freedom comes tension. On one side are the maintainers—people who have quietly kept projects running for years, sometimes decades. They earned their place by wrestling with complexity, pushing through the hard parts, and picking up knowledge that most never see. On the other side are users, now able to fork a project, make changes, or bring abandoned code back to life without any long apprenticeship or gatekeeping. These two groups are on a collision course.

I’ve come up with three possible outcomes; no doubt there are others. One is fragmentation: everyone keeps their own idiosyncratic forks, and improvements rarely make it back to the main project. In the past, the hassle and cost of splitting off would eventually lead successful forks to be merged into the parent repositories. But if AI lowers those costs, that pull toward the center weakens. A thousand flowers will bloom, but they won’t cross-pollinate. Another, albeit unlikely, possibility is relief for maintainers: users who used to send feature requests and bug reports now handle their own issues, reducing the burden on exhausted volunteers. And then there’s the bleakest option, at least for the idea of an open source community: maintainers get bypassed entirely, ignored by the vibe coders. The gate is still there, but nobody bothers with it. The years of volunteer work, the careful tending of a codebase, the hard-won knowledge—suddenly, none of it matters to someone who walks around it.

I don’t know which of these futures will win out, or if we’ll see all of them, depending on the project and the community. But I do know that the scale of this shift is massive. In the 1970s and early ’80s, personal computers made it possible for regular people to own and use a computer without going through institutions. In the 1990s, open source made it possible for anyone to read, copy, and modify code. But actually doing that still required skills most people didn’t have. The legal freedom was there, but the practical freedom wasn’t—and that gap stuck around for almost forty years. Now, AI coding tools close that gap. Suddenly, the end user can modify software. This is not a mere boost in programmer productivity—it is a fundamental shift in who gets to participate.

If the barrier to modifying software falls due to AI, the consequences for the open source community are vast. It seems unlikely that it will survive in its current form, but what will replace it is entirely unclear to me. To return to the question that is the title of this essay: what did vibe coding just do to the digital commons?

This is part of my Stochastic Histories series of counterfactual narratives.

In 1610, the Great Fire devastated Vilnius, capital of the Grand Duchy of Lithuania. The catastrophe arrived with the capricious logic of divine judgment: in eleven hours some 4,700 buildings were left in ashes, ten churches had fallen, most of the city’s civic structures were reduced to memory. Yet Vilnius has always been a restless city, ancient but constantly remade—burned, occupied, its churches converted to warehouses and back to churches, buildings falling into ruin and then rising again, like a filmstrip played back and forth. In this perpetual remaking, it is not unlike other cities that exist in the imagination as much as in physical form. In this reincarnation, Vilnius would emerge from its ashes reimagined—a landlocked capital transformed into a city of canals.

At the time of the fire, Vilnius served as both an administrative center and a critical trade hub—connecting the Hanseatic cities to the north and west, Russia to the east, and the Ottoman Empire to the south. The city was also a cultural center, particularly after the establishment of the Academia et Universitas Vilnensis Societatis Iesu in 1579, which served as a crucial center of counter-reformation thought and learning in Eastern Europe. Yet like many cities of its time, wood construction rendered it highly vulnerable to fire. Vilnius had already suffered several significant blazes, notably in 1530. The medieval street layout, with its narrow, winding alleys and densely packed buildings, offered little in the way of firebreaks, allowing flames to spread rapidly from one structure to the next.

But in early 17th century Vilnius, the threat of fire seemed to be as much doctrinal as physical. The diverse religious landscape, including Catholics, Orthodox Christians, Protestants, and Jews, created an atmosphere of suspicion and unease. The ruling Catholic authorities viewed the Protestant “heretics” with particular concern. This theological divide heightened anxieties about potential disasters, especially fire, seen not just as a physical threat, but also as a possible divine punishment. Alessandro Cilli, an Italian priest at the royal court, captured this tense atmosphere in his account:

Given such a situation and the great dissatisfaction of heretics and schismatics, the citizens of Vilnius, who were entrusted by the Queen’s order to take care of the city along with […] other lords constantly present in the city, due to fear and suspicions that some unrest and riots might arise, and especially fearing a possible fire, carefully monitored that sectarians would not gather in the city […]. Therefore […] the citizens would send three companies of soldiers to patrol the city at night. Under threat of severe punishments, a general order was issued to mind the fire and for everyone to have and keep various vessels and tubs full of water in their homes.

While Cilli’s pro-Catholic bias colors his narrative, his account vividly illustrates how doctrinal disputes and the fear of fire became inextricably linked. The conflation of religious dissent with the risk of conflagration led to a system of surveillance and control that blurred the lines between fire prevention and religious persecution. In Vilnius, flames were not just a physical but also a theological threat, each spark a potential harbinger of both earthly catastrophe and divine judgement.

The doctrinal anxiety gripping Vilnius in 1610 was not an isolated phenomenon of a provincial town, but rather part of a broader pattern at the time of the Reformation. Catholics and Protestants frequently employed fire imagery in their rhetoric against each other. From Martin Luther’s symbolic burning of the papal bull that excommunicated him to the fiery stakes of martyrdom on both sides, flame served as both metaphor and method. When the Great Fire of London erupted half a century later in 1666, similar suspicions ignited: convinced of a Popish plot, mobs attacked Catholics in the street. A plaque erected in 1681 on Pudding Lane proclaimed “Here, by the permission of Heaven, Hell broke loose upon this Protestant City, from the malicious Hearts of barbarous Papists, by the hand of their agent Hubert,” blaming one Robert Hubert, a Frenchman suspected of arson motivated by his faith. In Vilnius, the situation inverted—Catholics blamed the Protestants—but the mechanism was identical. Fire became a screen onto which each side projected its deepest fears about the other.

Cilli’s account continues with an almost unbearable, mounting dread.

We were so frightened by this that for the slightest reason, it seemed we could already hear shouts of fire breaking out, so we would carry our possessions, well-packed in boxes and chests and locked, to hide them in the safest places. Since most houses, or at least their interior spaces, were built of wood, we constantly lived in fear due to the aforementioned fire hazard, adhering to the usual precautions and reciting prayers assigned by the prelate for this purpose.

Other Catholic authors like the Austrian Jesuit Quirinus Cnogler (Knogler) also portrayed Protestants as a source of divine displeasure, implying that their presence invited God’s wrath upon the city. In such an atmosphere, the blaze was not only inevitable, it was fully expected. For Cilli, it almost seems a relief when the fire came.

Finally, ‘on the morning of the last day of June,’ the fire did break out. The all-consuming fire forced the city’s inhabitants and Her Royal Highness herself to flee, leaving them only to watch the spectacle of the flames’ fury from the surrounding hills as if spectators of some great theatre watching a tragedy unfold.

The exact location where the blaze started is unclear. Some sources say it started in the courtyard of the Franciscan monastery, although the priest Cilli located the source in a baker’s house. The city’s firefighting capabilities, like those of all cities at the time, were limited to bucket brigades and rudimentary tools. Fanned by fierce winds and fueled by the abundance of wooden structures, the fire tore through Vilnius with terrifying speed. Contemporary accounts describe a sea of flames engulfing entire neighborhoods, leaping from rooftop to rooftop.

Over eleven hours, the fire consumed much of the city. Thousands were left homeless. If, as Cilli claimed, this was divine retribution, then God was especially cross with the Orders: Franciscans, Dominicans, Bernardines, and Carmelites all lost everything. Even structures of stone and brick, which might have withstood a lesser blaze, succumbed to the intense heat. The Jesuit college and St. John’s Church, the papal seminary, its libraries, archives, and student dormitories were destroyed. God seemed indifferent to schism: the Holy Trinity, St. Stanislaus, and Mary Magdalene churches, as well as Lutheran, Evangelical, and Russian Orthodox churches and monasteries all fell victim to the flames. The Cathedral burned, although the shrine of the recently canonized Saint Casimir was miraculously spared. The Lower Castle and Town Hall were heavily damaged. Eyewitness accounts speak of the sky darkened by smoke, the air filled with the cries of the injured and bereaved, and the streets choked with fleeing residents and their meager salvaged belongings.

The inferno raged until three in the afternoon when an intense downpour began. Water flooded the streets like rivers. Residents who had fled to the surrounding fields and hills returned when the rain started, hoping it would extinguish any remaining fires. The human toll was severe, with many of those who did not perish in the flames drowning in the flooding, particularly on Pilies Gatvė, where the sudden rainfall created rushing torrents, a cruel irony after eleven hours of fire. Thousands were left homeless, their possessions and livelihoods gone forever.

In the immediate aftermath, the city struggled to cope with the scale of the disaster. Temporary shelters were hastily erected in the less affected areas and in the surrounding countryside. Many fled to family elsewhere; students from the Jesuit academy dispersed. The economic impact was immediate and severe, with trade grinding to a halt and the city’s artisans and merchants left destitute.

The annual letter of the Jesuits described the situation.

And since the most illustrious Queen with all her household retreated, and the senators and other courtiers, with their guest houses burned down, the townspeople, having lost their homes, moved to the suburbs, all the students went to their parents, and the craftsmen moved to other towns to establish their businesses elsewhere, the previously exceptionally prosperous city became very empty, with barely a person to be seen, sheltering from the weather’s hardships either by the ovens or in hastily built shelters in the corners of walls. However, some of them were punished for their carelessness when the walls collapsed afterward—in one house, seven people were immediately crushed by the rubble, and in others—about 20 people. Fear and great terror also took hold. For there was a rumor that a considerable group of robbers had gathered nearby, wanting to invade the city left open after the gates had burned down, and to break into and loot the cellars, which were then expected to be filled with things saved from the fire and completely unguarded. Because of this, there was trembling with fear for many days, soldiers were hired for the city treasury money, night guards were employed, cannons were fired and other things were done to instill fear [and deter].¹

Yet it was the manner of the fire’s ending that would shape Vilnius’s future. The torrential rains that finally quenched the flames had transformed streets into rushing rivers. On Pilies Gatvė, where the sudden rainfall created torrents powerful enough to drown those who had survived the inferno, water carved new channels through the ash and rubble. Observing these improvised waterways, Mikalojus Kristupas Radvila ‘Našlaitėlis’, the Grand Marshal of Lithuania and Voivode of Trakai, responsible for maintaining the order and security of the Grand Duchy, noted how they prevented the fire’s return—where water flowed, no ember could reignite.

Radvila advanced an unexpected proposal: to rebuild the city around canals. His recent pilgrimage (1582-1584) had taken him through Venice, where he spent four months studying the city’s maritime infrastructure while preparing for his journey to Jerusalem. In Damascus, he stood on a mountain and looked down on the city to observe how the Abana and Pharpar rivers irrigated it, flowing down to create what he called “an elegant, pleasant, and highly fruitful plain” surrounded by beautiful gardens. In Cairo, he witnessed the fath al-khalij, the grand ceremony of the Nile’s opening, where Imbraim Bassa, the Ottoman Wāli, ceremonially breached the dykes to flood the city’s canals after months when the water had been reserved for agricultural irrigation. As Radvila recorded in his chronicle of the journey, Hierosolymitana Peregrinatio, “the whole area looked covered in snow because of the white turbans” of the watching multitude—an entire city gathered to celebrate water’s return.²

But Radvila’s vision was not mere fantasy inspired by foreign travels. His cousin Kristupas Radvila “Perkūnas,” before his death in 1603, had recruited Scottish settlers to Kėdainiai, offering them land and protection in exchange for developing his estates. Though Našlaitėlis did not share Perkūnas’s Protestant faith—he had converted to Catholicism decades earlier and become one of the Counter-Reformation’s most generous patrons—he recognized the practical value of these existing networks. The Scots had proven themselves throughout the Commonwealth as masons, engineers, and builders. Where others saw only destruction in Vilnius’s ruins, Radvila saw opportunity: a massive public works project that could employ Scottish subsistence migrants seeking any labor, alongside their skilled countrymen already working across Lithuanian territories, while transforming the city into something unprecedented.

It has been a matter of debate as to how much Radvila understood that his project was as much an act of restoration as a feat of imagination. The city that burned in 1610 had been built upon an intricate hydrographic system, portions of which remained in active use and living memory.

The most significant of these was the Vingrių system. The Vingrių šaltiniai—springs emerging west of the city—had supplied Vilnius with water since at least the fourteenth century. Originally municipal property, the springs passed through various ecclesiastical hands before the city repurchased them from the Dominicans in 1536, paying one hundred kapų of Lithuanian groschen and ten pūdai of pepper. The sale document explicitly transferred not merely the springs themselves but “the bed and water flowing from it and the canals” (una cum alveo et aqua ex indě fluenti et canalibus)—confirming that engineered channels already distributed this water through the city.⁴ The Vingrių upokšnis carved a clear ravine along what would become Pylimo and Liejyklos streets, entering the Vilija near the Totorių vartai, the gates that took their name from the Tatar settlement Vytautas had established nearby.⁵

The Vilnelė itself had followed a different course in earlier centuries. The nineteenth-century historian Michał Baliński, drawing on chronicles and local memory, recorded that the river once flowed through what became Pilies gatvė, past the old Rokitai monastery ruins and the Pilies vartai, through the mažasis turgus before looping around the castle territory to join the Vilija behind the cathedral. The current confluence point, he noted, was itself the product of engineering: a perkasa—a moat or canal—dug during Gediminas’s reign to raise the castle hill and ensure “both fortresses would be washed by waters.”⁶ Smaller watercourses threaded through the urban territory—the Druja emerging near Žiupronys, the streams feeding the episcopal mill—some already partially channeled by 1610.

Recent geological investigation has revealed that this documented infrastructure overlay a still older hydrographic system. Five gravel corridors, carved by seasonal water flows, determined the routes of Vilnius’s principal streets from the medieval period. The first corridor ran from Bazilijonų street to the intersection of Didžioji and Etmonų, forming the base of the Medininkai road. The second stretched from Rūdninkų and Ligoninės past Visų Šventųjų along Mėsinių to Rotušės Square. The third ran north along Pilies towards Cathedral Square—the same route the old Vilnelė channel had followed. The fourth paralleled Universiteto from Gaono and Švarco. The fifth followed Dominikonų toward Stiklių.⁷

The correspondence between these geological findings and the documented historical waterways suggests that the city’s planners had long worked within constraints set by water. Whether Radvila consciously understood the full depth of this water-shaped landscape remains uncertain. But the active springs, the functioning channels, the living memory of floods and firefighting failures—these he could not have ignored. The devastating fires of 1513 and 1530 had already demonstrated the inadequacy of the existing water infrastructure, prompting Žygimantas Senasis (Sigismund I the Old) to invest heavily in expanding the Vingrių pipe system precisely because “the parts of the city furthest from the river lacked water for firefighting.”⁸ The 1610 conflagration repeated this lesson on a larger scale.

Regardless how deeply he grasped the city’s origins in water, Radvila could hardly have been ignorant of Amsterdam’s escalating urban crisis and the radical solutions being debated to address it. Since 1585, when Antwerp’s fall sent refugees flooding north, Amsterdam had undertaken two expansions that proved inadequate. By the early 1600s, the city was drowning in informal settlements spreading outside its walls—a “frenzy of illegal construction” that administrators struggled to control. The debates about how to manage this crisis circulated through the networks in which the Radvilos were embedded: the commercial ties to Dutch ports through the Hanseatic League, the Reformed Protestant communities centered on Amsterdam, and the publishing relationships that would see his own travel account reissued there in 1614. Several family members studied at Dutch universities and traveled extensively through the Netherlands. For a magnate overseeing the reconstruction of a major city after catastrophic fire, intelligence about Amsterdam’s struggles and proposed solutions would have been not merely available but essential.

In February 1610, as Vilnius still smoldered, Amsterdam’s council debated multiple maps “concerning the enlargement and extension of this town.” What followed was not triumph but compromise: scandals stalled the project, the original plan was discarded, and by March 1613 the grand vision had been cut in half. The Herengracht, Keizersgracht, and Prinsengracht would proceed only on the western side, leaving complex problems passed on to the next generation. Yet even this truncated ambition was unprecedented—systematizing Amsterdam’s existing medieval canal network into concentric rings that would eventually triple the city’s area.⁹

The timing of Radvila’s intervention thus paralleled Amsterdam’s not as imitation but as parallel response to similar crises of urban transformation. Where Amsterdam expanded outward through water, Vilnius would transform inward—replacing its burned medieval core rather than abandoning it. In accomplishing this, Vilnius would achieve something distinct: not the creation of new navigable waterways like Amsterdam, nor existence on water by necessity like Venice, but the resurrection of a city through deliberate engagement with its hidden hydrology—the ancient water-carved corridors that had shaped its streets since the fourteenth century, now channeled and rationalized as an act of will.

By late September 1610, Radvila had dispatched agents to Venice, carrying letters of credit and promises of wealth. The team that arrived in frozen January 1611 was led by Giovanni Battista Aleotti, fresh from completing the drainage works at Ferrara, and Pietro Paolo Floriani, a military engineer who had been fortifying the Papal States. The Italians spent the frozen months walking the ice-covered ruins, taking measurements, arguing over gradients. Aleotti insisted they must wait for the spring thaw to properly test the water table—in Venice, he said, impatience with water meant catastrophe.

The groundbreaking ceremony on May 15, 1611—nearly a year after the fire—was deliberately theatrical. Radvila understood the symbolic weight of the moment. The Bishop of Vilnius blessed the works while standing knee-deep in the muddy excavation that would become the first canal, though privately he had expressed concerns about the project’s disruption to traditional parish life. As the first channel was cut, Radvila spoke of a new era—safety, prosperity, and magnificenza. Italian engineers directed workers in a babel of languages, with hand gestures filling the gaps where words failed.

The workforce itself was unprecedented in Lithuanian history. Through his Protestant networks, Radvila had assembled nearly eight hundred Scotish migrants. They came in two distinct waves. The first consisted of “subsistence migrants”—young men and boys fleeing Scottish famines and poverty, willing to dig for bread and shelter. These formed the project’s muscle, wielding shovels and hauling earth under the Baltic sun. The second wave brought “betterment migrants”—skilled craftsmen who had already established themselves in Polish and Lithuanian towns as masons, carpenters, wheelwrights, and builders. These men became the project’s skeleton, translating the Italian engineers’ visions into Baltic reality, managing work crews, and solving the thousand daily problems that arose when Mediterranean engineering met Lithuanian mud. The Scots immediately formed the Scottish Brotherhood of Vilnius (Škotų Brolija Vilniuje), negotiating collectively with both Radvila’s agents and the Italian engineers.

Tensions emerged immediately. Lithuanian laborers, often unemployed due to the loss of their workplaces in the fire, who had expected to profit from reconstruction work found themselves competing with foreign workers. The scheduling disputes proved particularly bitter—the Scots insisted on strict Sabbath observance, refusing all work on Sundays, while Lithuanian workers expected Catholic feast days to be honored. The result was a construction site where different crews operated on different calendars, with the Italian engineers growing increasingly frustrated at work stoppages that seemed to alternate between Presbyterian and Catholic observances. Aleotti, accustomed to managing unified Ferrarese work crews, found himself mediating between Scottish elders conducting meetings in incomprehensible Scots and Lithuanian guild masters citing saints’ days unknown in the Reformed calendar.

The technical challenges quickly validated skeptics’ worst fears. Where engineers expected clay, sandy soil collapsed, requiring expensive shoring. Where maps showed alluvium, bedrock appeared, necessitating costly blasting. The Italian engineers found themselves adapting their Mediterranean techniques to Baltic conditions in ways they hadn’t anticipated.

The summer rains of 1612 turned construction sites into muddy lakes that seemed to confirm critics’ predictions. Workers drowned in sudden cave-ins. Yet by autumn, the first two canals were operational—Pilies (Castle) kanalas, following roughly the path of what had been the main street from the university toward the cathedral site (now Pilies Gatvė), and Katedros (Cathedral) kanalas, cutting a new path between the Vilnelė and the Neris, starting near Šv. Onos and Šv. Pranciškaus ir Šv. Bernardino bažnyčia (Saint Anne’s and St. Francis of Assisi and St. Bernardine), brick Gothic Franciscan churches that miraculously survived the fire unscathed (along what is now Maironio gatvė), skirting around Gedimino kalnas and curving past the cathedral (following the present-day Šventaragio gatvė), before joining the Neris roughly where Mindaugo tiltas (bridge) stands today. The fire-damaged lower castle and cathedral district—as much a pile of rubble as anything else—was now linked to Paupys and Užupis through the Pilies and Katedros canals. The sight of boats gliding where carts had once rattled—or where no street had existed before—provided powerful vindication for Radvila’s vision.

The winter of 1612–1613 brought the harshest freeze in memory, bursting partially laid mechanisms and splitting imported stone. Critics in the Sejm demanded accounting for the spiraling costs. Spring 1613’s thaw brought catastrophe: embankments collapsed, flooding lower neighborhoods with muddy canal water. By summer 1613, costs had tripled from initial projections. Two junior Italian engineers fled back to Venice rather than face another Baltic winter. Angry citizens petitioned King Sigismund III to halt the project entirely.

Radvila spent much of his personal fortune keeping the work alive, while also managing a careful propaganda campaign. When the Papal Nuncio Claudio Rangoni visited in late 1613, Radvila ensured he saw not the construction chaos but the completed sections. Rangoni wrote to Rome describing “dozens of boats passing where two years prior lay nothing but ash and ruin.” He noted with particular interest how the Lithuanian nobility—the very ones who had resisted the project—had now embraced the waterways with unexpected enthusiasm, commissioning elaborate barges decorated with family crests. The Sapiega family’s barge, featuring a gilded swan at its prow and requiring twelve oarsmen, became legendary for blocking smaller craft in the narrow residential canals.

The year 1613 also saw the publication of the celebrated Radvila (”Radziwiłł”) map of the Grand Duchy by Blaeu in Amsterdam—a project years in the making that established Tomasz Makowski’s reputation as the Commonwealth’s premier cartographer. With this achievement behind him and the first canals now operational, Makowski was commissioned to document the transformation of Vilnius itself. In 1614, Makowski produced his Lithuaniae Aquarum Tabula, signed with his characteristic “T. M. Pol. Geograph.” According to Radivla estate inventories later destroyed during the Swedish occupation of 1655, this lost map included a detailed inset titled Aquae Vilnenses, diagramming the completed canals to date: a cartographic manifesto for a city remade by water.

By then, the Scottish workforce had grown to nearly fifteen hundred, establishing their own quarter near the construction sites. Their Presbyterian services, conducted in wooden temporary churches, drew curious Lithuanian crowds. Rangoni’s letters betrayed deeper anxieties: “The heretics multiply,” he wrote, “and their industry in building these waterways cannot be separated from their industry in spreading their errors.”

Alessandro Cilli, the Jesuit observer, was more explicit in his concerns. Writing in 1614—the same year Makowski was depicting the triumph of the canal system—Cilli noted that the fire itself had been preceded by “great dissatisfaction of heretics and schismatics,” and now the city’s reconstruction had become their triumph. The Scottish Brotherhood had grown powerful enough to negotiate not just wages but religious concessions—the right to bury their dead according to their own rites, to maintain their own courts, and to be exempt from Catholic feast day observances.

The economics of the canal system developed its own logic. The initial funding came from an extraordinary tax on the nobility, justified by the protection the canals would offer their urban properties. But as the waterways became operational, new revenue streams emerged. Each boat passing through certain canals paid a toll. Warehouse spaces along the canals commanded premium rents. The Scottish Brotherhood, having learned the system intimately through building it, began operating their own transport services, undercutting the traditional Lithuanian carriers’ guild. A parallel economy had emerged along the waterways, with Scottish bankers providing credit, Scottish boats moving goods, and Scottish merchants establishing direct trade links to Gdańsk and Riga.

Radvila had died in 1616—his final months spent defending the project against accusations of bankrupting the Commonwealth for a dream—but the transformation he initiated continued to reshape the city. In the last part of the decade the system rapidly expanded into the residential and commercial quarters. The waterways had developed their own social geography. Vokiečių kanalas carved through the German merchants’ district, who quickly adapted their Hanseatic trading practices to water-based commerce. Along Totorių kanalas, the Tatar community—who had served the Grand Duchy as soldiers and tanners for two centuries—maintained their leather workshops, the canal providing essential water for their trade while efficiently carrying away the pungent waste. Žydų kanalas became the center of precious crafts, with Jewish goldsmiths and furriers operating workshops whose waste could be easily flushed away by the controlled current. Stiklių kanalas, true to its name, became the glassmakers’ domain, where furnaces backing onto the water had ready access to both sand deliveries and cooling basins. By year’s end, a network of smaller channels—Bokšto, Tekančio Vandens (the current passage called Skapo gatvė had that name until the end of the 18th century), and others—had transformed the old street grid into an aquatic maze. Even former skeptics conceded that the city was proposering from its new commercial infrastructure, served by merchant barges from the Neris that queued at dawn to deliver grain, timber, and salt.

The medical college at the university, initially skeptical of the canals’ health implications, published a treatise in 1618 noting unexpected benefits. The controlled water flow had indeed reduced disease compared to the open sewers of the old city. The fire of 1610 had destroyed not just buildings but also centuries of accumulated filth. The new canal system, with its engineered gradients and regular flushing, proved cleaner than what it replaced. Dr. Pranciškus Rautenbergas, the city physician, noted that cases of summer fever had decreased markedly since the canals’ completion, though he cautiously attributed this to “improved airs” rather than crediting the waterways directly.

The Protestant community’s influence continued to grow alongside the canals. By 1619, the Scottish Brotherhood operated not just transport services but also warehouses, taverns, and even a small shipyard in Paupys where they built specialized flat-bottomed boats designed for the shallow upper canals. They had established their own burial ground on a small island formed where two canals diverged—accessible only by water, it became a powerful symbol of their separate but integrated existence within the city. The Reformed Church they built there, modest wood construction compared to Vilnius’s baroque Catholic splendor, nevertheless drew congregations that included Lithuanian craftsmen attracted by services conducted in vernacular rather than Latin.

The Jesuits, watching this Protestant flowering with alarm, launched their own canal-based initiatives. They established the Floating Mission of St. Francis Xavier in 1618—a consecrated barge that moved through the waterways conducting masses, hearing confessions, and pointedly blessing the waters themselves. The mission’s archives record careful notes about how Scottish keepers might delay or expedite passage depending on sectarian sympathies. Father Albertas Vijūkas-Kojalavičius wrote extensively about the spiritual challenges of ministering to a “liquid parish” where congregants might attend mass by boat but disappear into the canal network before contributing to the collection plate.

The nobility’s adaptation to canal life evolved in unexpected directions. What began as competition in ornate barges developed into a complex aquatic protocol. The width of canals meant that when noble barges met, one had to yield by pulling to the side—but who yielded to whom? The Tribunal of the Grand Duchy was forced to issue regulations in 1619 establishing water precedence based on titles and offices. The resulting “Canal Code” became a subject of satire among Warsaw wits, who mocked the Lithuanian nobility for creating maritime law for inland waterways.

Yet the canals also democratized certain aspects of city life. On water, a merchant’s practical flat barge might make better speed than a magnate’s ornate vessel. Toll collectors who also regulated traffic showed theoretical deference to rank but practical favoritism to those who paid well and caused no trouble. Stories circulated of proud nobles forced to wait while grain barges passed, their protests met with Scottish shrugs about “essential maintenance” that coincidentally resolved once proper fees were paid.

The symbolism resonated deeply with citizens: Saint Christopher carrying the Christ child across water had been Vilnius’s emblem since 1330. Now the entire city had been transformed into a network of crossings. The motto on the city’s arms—Unitas, Justitia, Spes—seemed to speak directly to the completed project: unity in the collective effort of reconstruction, justice in the fair distribution of canal access, and hope for a city reborn from ashes through water.

By 1625, with the primary canal network complete, the extraordinary coalition that had rebuilt Vilnius began to dissolve. Aleotti returned to Ferrara that year, claiming the Lithuanian winters had ruined his joints. The Scottish Brotherhood, its construction contracts fulfilled, dispersed across multiple trajectories. The skilled craftsmen faced a choice: convert to Catholicism and join Vilnius’s guild system, or seek new projects elsewhere in the Commonwealth. Some prominent Scottish merchants—the Gordons, Forsyths, and Barclay families—had already converted and established themselves along the prime canal frontages, their surnames gradually lithuanianizing to Gordonas, Forsaitas, and Barklajus in municipal records. The subsistence migrants, young men who had arrived with nothing, departed with earnings for new ventures in Riga, Königsberg, or home to Scotland. By 1630, the Scottish Brotherhood’s meeting house had been sold to the Dominicans, and transport operations had passed entirely to local watermen. Floriani, now sole director of the works, oversaw this transition with characteristic pragmatism. Where Aleotti had been an artist demanding perfection, Floriani was an administrator who understood that the canal system’s survival depended not on foreign expertise but on integration with Vilnius’s existing social structures. He established formal apprenticeships for local boatmen, standardized maintenance procedures that could be managed by Lithuanian crews, and gradually replaced Italian technical terminology with Lithuanian equivalents. The city that emerged from this transition was neither the Venice that Radvila had imagined nor the old Vilnius that had burned, but something unprecedented: a Baltic canal city operated by its own citizens.

By 1630, Vilnius had fully adapted to its aquatic transformation. The canal system that had been Radvila’s gamble and Aleotti’s experiment had become simply how the city functioned. Morning began with the sound of canal gates opening at dawn, releasing the night’s accumulation of cargo barges into the commercial districts. Morning markets operated directly from barges moored at designated basins while other merchants brought a more limited selection straight to residences. Fishmongers from Kernavė, their boats laden with pike and perch, competed with vegetable sellers from the surrounding countryside. The valčininkai (boatmen) who had long worked the Neris and Vilnelė adapted their flat-bottomed vytinės to the narrower waterways, filling the vacuum left by the Scottish Brotherhood’s departure and developing distinctive poling techniques for navigating the canal intersections. But it was the vilnelės—shallow-draft skiffs developed specifically for the canal system—that became the city’s primary local transport. These vessels, smaller than the trading vytinės but more maneuverable than imported gondolas, could navigate even the narrowest channels. Their operators developed an elaborate etiquette of right-of-way, horn signals, and docking precedence that newcomers found baffling but locals navigated instinctively. Housewives lowered baskets from canal-side windows to purchase vegetables directly from passing market boats. Children learned to gauge water depth by the color of reflected sky, to recognize the horn patterns that warned of approaching rafts, to time their errands around bridge raisings. Daily life adapted to the aquatic rhythm.

The city’s soundscape had changed as well. Where once church bells had dominated, now water created constant accompaniment: the splash of poles as boatmen navigated corners, the creak of mooring ropes, the particular echo of voices bouncing off canal walls. Newcomers complained about sleepless nights until they learned to distinguish between normal water traffic and sounds requiring attention—the sharp crack that meant ice damaging a foundation, the rush that warned of opened floodgates upstream. By the second generation, parents told children that Vilnius had always been a water city, that the idea of cart-filled streets was as foreign as the notion of a capital without canals.

The practical achievements of the canal system extended beyond transport and fire prevention. By the late 1630s, Floriani’s engineers, building on Dr. Rautenbergas’s documented correlation between flowing water and reduced disease, had installed an underground sanitation network that exploited the canal’s constant flow. Unlike most European cities where waste accumulated in streets and cesspits, Vilnius’s system channeled sewage through brick-lined conduits beneath the quays, using the hydraulic gradient to flush waste toward outfalls downstream of the inhabited districts. The system required constant oversight. A corps of custodes aquarum (Water Wardens) patrolled the channels, empowered by city statute to fine anyone caught dumping refuse directly into the canals. Their morning rounds became a familiar sight—men in dark cloaks checking grates, testing water flow, chalking violation notices on the doors of offending households. In 1647, a letter from a visiting Dutch merchant remarked that Vilnius’s canals ran clearer than Amsterdam’s, though whether this reflected superior engineering or more stringent enforcement remained debated.

In 1653, the local watermen organized into the Guild of Saint Christopher—taking as their patron the very saint who adorned the city’s seal, making their trade a living embodiment of Vilnius’s heraldic identity. The guild maintained strict hereditary membership and elaborate regulations governing routes, fares, and the distinctive blue woolen caps that marked their trade. Italian gondoliers, initially imported to provide prestige service to the nobility, gradually intermarried with these local families, producing a hybrid tradition that combined Venetian ceremonial knowledge with Baltic practical skills. The guild’s 1703 statute book, preserved in the Lithuanian State Historical Archives, records 247 licensed watermen operating 89 passenger vessels and 143 cargo barges within the urban canal network.

As we might expect, Vilnius’s economic and social geography changed. Properties along the new waterways commanded premium prices, forcing longtime residents into peripheral quarters. The Bonifratres complained that canal construction had severed their monastery from its traditional parish. Jewish merchants, initially excluded from prime canal-front locations, gradually established warehouses along secondary channels, creating a thriving market district. Yet the project generated unprecedented employment. Beyond the obvious needs for boatmen, the canals required armies of specialized workers: dredgers who cleared silt before it could obstruct navigation, ice-cutters who maintained winter channels, bridge-tenders who operated the drawbridges for masted vessels, stone-masons perpetually repairing quay walls undermined by wake and weather. The city’s population swelled as craftsmen arrived from across the Commonwealth and beyond.

Cultural life flourished along the quays. Maciej Kazimierz Sarbiewski, the Jesuit poet whom Urban VIII crowned with laurels and whom translators across Europe called the “Christian Horace,” taught at the university during the canal system’s early years. His ode Ad Vilnam Fluviis Ornatam praised the city as a new Venice risen from northern forests, its waters carrying not only commerce but civilization itself. The Jesuit college staged elaborate dramas on floating platforms at the confluence of the Vilnia and Neris, their allegorical figures of Faith and Reason arriving by gilded barge while students operated mechanical waves and fountains.

The local condition had also changed education. By the 1740s, the Jesuit curriculum had expanded to include practical hydraulics alongside traditional mathematics. Students surveyed water levels, calculated flow rates, and designed improvements to the canals. Tomasz Żebrowski, before founding the Observatory in 1753, established what amounted to a hydrological research program. His students maintained meticulous records of seasonal variations, correlating water levels with precipitation, developing predictive models for flood control. The university’s instrument workshop produced standardized bronze gauge-boards installed at key points throughout the system, allowing for coordinated response to threatening conditions. This marriage of theoretical knowledge and practical application—characteristic of Jesuit pedagogy—transformed Vilnius into a center for hydraulic expertise that attracted students from across Europe.

The theological implications of the canal system extended far beyond the spectacular Epiphany rite that became the city’s signature festival—guild barges dressed with spruce, choir song carrying over black winter water. Each religious community adapted its sacred practices to the aquatic reformation of urban space. The Orthodox faithful transformed their Theophany tradition into mirror ceremonies performed simultaneously at multiple canal basins, priests casting crosses into different waterways to sanctify the entire network. The Karaite community, with their stringent purity laws derived from Levitical codes, developed an elaborate jurisprudence distinguishing between living water that flowed directly from the Vilnelė and resting water in the broader channels, with different ritual applications for each. Every major church eventually constructed its own landing stage, complete with carved posts for securing ceremonial barges. The Dominican priory commissioned a floating platform that could be assembled for water-based mystery plays. The annual calendar became punctuated by aquatic ceremonies: Palm Sunday barges bearing greenery, Corpus Christi processions where the sacred host traveled by decorated gondola, All Saints’ Day when hundreds of candles floated in small wooden boats to commemorate the dead. The Jesuits, characteristically, discovered divine proportion in the hydraulic design itself while integrating the canals into elaborate theatrical productions, staging mock naval battles for visiting dignitaries and using the water’s reflective properties to enhance their famous lux et umbra performances. Tomasz Żebrowski’s treatise De Aquis Sanctis (1751) would later argue that the canal network’s geometry reflected sacred numerology—seven major basins for the days of creation, twelve major canals for the apostles, proportions derived from Solomon’s Temple.

Still, the maintenance of the canal system proved both perpetual and politically contentious. Dredging operations, essential to prevent the waterways from silting into stagnant pools, required year-round crews of specialized workers using long-handled scoops and bucket chains. The expense was staggering—by the 1660s, maintenance consumed nearly a third of the city’s annual revenue. To address this crisis, the Sejm enacted the Canal Tax (Portorium Canalium) in 1668, levying graduated fees on all properties with canal frontage.

Winter presented unique engineering challenges that Venice never faced. From December through March, the canals transformed into a different kind of infrastructure. The city maintained a corps of ice-cutters—200 men working in shifts through the coldest months—whose primary duty was keeping channels open to fire cisterns. Using iron-tipped poles and specially designed saws imported from Sweden, they carved passages through ice that could reach two feet thick. The main commercial channels were allowed to freeze, becoming impromptu winter roads where sledges replaced barges. This seasonal transformation had unexpected social consequences: January’s Ice Festival (Ledo Šventė) became one of Vilnius’s most celebrated events, with lantern-lit skating parties, ice sculptures, and vendors selling hot wine from braziers set directly on the frozen canals. Yet the freeze-thaw cycle wreaked havoc on infrastructure. Expanding ice cracked stone quays, requiring annual repairs. Engineers developed a system of sacrificial timber corners—wooden buffers designed to absorb ice pressure and be replaced each spring rather than allow damage to permanent masonry.

Sedimentation threatened to choke the entire system. Every spring, snowmelt carried tons of silt into the channels; every summer, organic matter legally and illegally dumped in markets and households accumulated in the slower-flowing sections. By the 1680s, some secondary canals had lost half their depth. The crisis prompted technological innovation. In 1685, Vilnius launched its first mechanical dredging barge, the Šv. Kristupas, equipped with an endless chain of buckets powered by a horse-driven windlass. The design, adapted from Dutch polder-draining equipment but modified for river dredging, proved so successful that a delegation from Amsterdam arrived in 1687 to study the improvements. Within five years, Vilnius operated a fleet of six such vessels, each capable of removing 100 cubic meters of sediment daily. The dredged material, initially dumped outside the city, was later recognized as valuable fertilizer, creating a secondary economy where farmers bid for rights to particular sections’ spoils.

Infrastructure decay demanded constant vigilance. Wooden bridges and pilings suffered from perpetual moisture. The collapse of the Merchants’ Bridge in 1709, which killed three people and blocked the main commercial channel for weeks, catalyzed reform. The Stone Bridge Act of 1710 mandated that all wooden structures be replaced with stone within twenty years. The act faced immediate resistance—stone construction cost five times more than wood—but included innovative financing through bridge bonds that citizens could purchase, receiving toll exemptions as interest. By 1730, Vilnius boasted seventeen stone bridges, each with its own architectural character. The German Quarter’s bridge featured carved grotesques that spouted water during floods; the University Bridge incorporated astronomical symbols that served as a sundial.

Water quality emerged as a growing concern in the eighteenth century. Unlike the early years when canal water was considered pure enough to drink after simple settling, by 1700 population density and industrial discharge had created health hazards. The Water Purity Commission (Commissio Puritatis Aquarum), established in 1723, employed chemists from the university to test water at various points using copper sulfate reactions and settlement tests. The commission’s 1725 report, remarkably advanced for its time, identified specific contamination sources: tanneries (mercury and lime), dyers (mordants and pigments), and slaughterhouses (blood and offal). Their experimental filtration system, installed at three major channel intersections, used layers of sand, gravel, and charcoal—presaging modern water treatment by a century.

Kanuty (Kanutas) Rusiecki (1800–1860), Old Bernardine Canal at Paplauja, Vilnius, c. 1850
The abandoned Bernardine suburb on the city’s southeastern edge, with the disused mill canal slipping past decaying monastic outbuildings and a few washerwomen at the water’s edge.

The question of expansion perpetually divided the city. The Canal Extension Controversy of 1735 epitomized these tensions. A consortium of merchants proposed adding three new channels to serve expanding suburbs, promising to fund construction privately. Engineers warned that additional waterways would reduce flow in existing channels, increasing sedimentation and ice formation. The debate raged in pamphlets and council meetings for two years. The engineering faction published detailed hydraulic calculations; the merchants responded with economic projections showing doubled trade revenue. The compromise—one new channel with strict dimensional limits—satisfied no one fully but demonstrated the system’s political complexity. Meanwhile, at the periphery, residents of districts like Antakalnis and Žvėrynas petitioned repeatedly for connection to the canal network. Their drainage ditches, they argued, were “sewers unworthy of citizens,” yet the cost of extension always exceeded available funds. The inequality between the canal-served center and the muddy periphery became a persistent source of social tension that would later fuel demands for democratic reform.By the mid-eighteenth century, canals defined Vilnius’s identity. Population passed 60,000.

The partitions of the Commonwealth in 1795 brought Vilnius under Russian imperial administration, fundamentally altering the canal system’s political meaning. While Tsarist engineers acknowledged the waterways’ practical benefits—Governor-General Mikhail Muravyov’s 1864 report praised their “exemplary flood management”—the canals’ capacity to anchor a distinct Lithuanian urban identity raised suspicion. Russian authorities gradually restricted the canal boatmen’s guild, requiring operators to obtain imperial licenses written in Cyrillic script. The annual Epiphany ceremony was permitted to continue but under Orthodox rather than Catholic rites, with the military governor’s approval required for any water procession. More practically, maintenance funding was redirected to fortification projects. St. Petersburg, meanwhile, actively cultivated its own identity as the “Venice of the North,” with court propagandists dismissing Vilnius’s canals as provincial ditches unworthy of comparison to the imperial capital’s grand waterways.

During the late eighteenth century, concerns about sanitation were rising. The Vingrė canal—Radvila’s proudest achievement, flowing from the springs at Vingris along the reconstructed Pylimo and Liejykla streets down to the basin by the cathedral and Governor’s Palace—had begun its transformation into “Kačerga,” a Slavic borrowing meaning “poker,” a tool for stirring ashes. Where Vingrė had carried the Lithuanian legacy of flowing water, Kačerga spoke of what the canal had become: a receptacle for refuse, stirred through like spent coals. Despite regulations prohibiting the disposal of refuse into the waterways, enforcement proved impossible as the city’s population swelled and commercial pressures mounted.

Public health crises accelerated the system’s decline. The cholera pandemic of 1831 killed over 2,000 residents, with the highest mortality in canal-adjacent quarters. Although later analysis showed contamination came from wells rather than canals, Imperial health inspectors, unfamiliar with Vilnius’s hydraulic system, blamed the waterways. Dr. Friedrich Hübner’s 1832 report to the Imperial Medical Council declared the canals “miasmatic reservoirs spreading disease through putrid exhalations.” The city’s own physicians protested—Dr. Józef Mianowski demonstrated that canal districts actually had lower typhoid rates than areas relying on well water—but imperial opinion had turned.

The situation paralleled Amsterdam’s canal crisis. There too, waterways had degraded by mid-century into receptacles for waste, blamed for cholera outbreaks. But Amsterdam, with its maritime wealth and municipal autonomy, could invest in remediation, at least eventually. The steam-powered Gemaal Zeeburg pumping station, built in 1879 to flush the canals with fresh water from the Zuiderzee, cost what Vilnius collected in taxes over five years. Imperial administrators, confronting similar crises in dozens of provincial cities, chose the cheaper solution: burial. The 1835 Urban Hygiene Decree mandated “gradual elimination of stagnant waterways and construction of new canals.” Implementation was slow and contested, but less used canals started to get filled in. Each filling followed the same pattern: wooden sheet piling to contain the water, pumping to nearby channels, then progressive filling with rubble, sand, and finally paving stones. Residents salvaged what they could—iron mooring rings became door hardware, carved bridge stones were incorporated into new buildings.

Liejyklos Canalsas became Liejyklos street as that foul stretch of the Kačerga was enclosed in brick culverts and buried beneath the streets. But the water still had to go somewhere. At the confluence near the Governor’s palace, where Kačerga joined the large basin and the old Vilnia channel that continued past the cathedral, the buried stream emerged into daylight, discharging its now-polluted waters into the canals that still flowed openly through the city’s ceremonial heart. Complaints multiplied about the “putrid waters emerging from underground passages,” but the vast basin between the Palace and the Cathedral absorbed the smell, at least most days.

In 1816, Military Governor General Alexander Rimsky-Korsakov had commissioned Scottish architect William Hastie to design a comprehensive plan for the city’s modernization. Hastie, fresh from work on St. Petersburg’s neoclassical avenues, proposed a grand canal that would dwarf Radvila’s network: a monumental waterway running from the Cathedral to the Neris, serving both as ceremonial axis and commercial artery. The plan languished for two decades amid bureaucratic disputes and funding shortages—Hastie himself died in 1832, never seeing a shovel break ground—but in 1836, in the wake of the Urban Hygiene Decree, construction finally began on Georgiyevsky kanal (Šv. Jurgio Canal, i. e. the current Gedimino Prospektas), frequently called Bolshoy kanal (or great canal) would bring in so much more water to the basin that it would counteract the sewage from the Kačerga. The contradiction was absolute: Imperial engineers were burying Lithuanian canals as medieval death-traps while simultaneously building a Russian showcase canal as proof of modern engineering prowess. Where Radvila’s system had grown organically from the city’s ancient hydrology, Hastie’s canal imposed geometric clarity, its embankments faced with Prussian granite, its bridges designed for military parades as much as commerce.

The 1880s had seen entrepreneurial innovation as Moishe Goldberg introduced small steam launches to the main commercial channel. His “water taxis” initially faced ridicule—newspapers mocked the smoke and noise disturbing the ancient waterways—but by 1890, three competing companies operated twelve steam vessels. The new middle class moving into Naujamiestis apartments embraced canal commuting as sophisticated and practical. Morning rush hours saw clerks reading newspapers on covered launches while horse carriages sat trapped in bridge traffic above them.

The pinnacle came in 1908 with the inauguration of the municipal water-omnibus service. The new launches, built in Riga’s shipyards with shallow drafts specifically for Vilnius’s channels, ran on fifteen-minute headways from University Quay to Lukiškės Basin, with a new branch line extending into Naujamiestis. Tickets were integrated with the konka system—a forward-thinking intermodal network that wouldn’t seem out of place today. University students could purchase monthly passes for both systems at reduced rates, leading to the peculiar sight of academic discussions continuing as groups transferred from tram to boat at interchange points.

The Polish occupation of Vilnius in October 1920 brought attention to the canals once again. General Lucjan Żeligowski’s administration, eager to legitimize Polish rule, immediately grasped the propaganda value of the waterways. Here was proof that Wilno deserved its place among Europe’s great cities—a Venice of the North that Poland could restore to glory. The new authorities positioned themselves not as occupiers but as inheritors of the Grand Duchy’s cosmopolitan tradition, and the canals became central to this narrative.

Stefan Batory University became the canals’ intellectual champion. The rowing club, established in 1924, attracted Polish, Lithuanian, and Jewish students—one of the few genuinely integrated institutions in the divided city. The annual regatta evolved into Wilno’s premier social event. The 1928 competition drew Crown Prince Carol of Romania as honorary starter. Newsreel footage shows crowds lining every bridge and towpath, while floating platforms hosted orchestras playing between races. The university’s engineering faculty pioneered winter maintenance techniques, using steam generators to keep strategic sections ice-free, allowing year-round navigation for the first time in the canals’ history.

Commercial innovation flourished. The Orbis travel agency launched “Gondolas of Wilno” in 1926—imported Venetian boats poled by costumed students earning university fees. Initially dismissed as kitsch, the service proved wildly popular with Warsaw society. The 1929 Baedeker called it “an essential experience despite its theatrical nature.”

Marshal Józef Piłsudski himself, born in the Vilnius region and deeply attached to the city’s multicultural heritage, personally intervened to secure funding for canal restoration. The 1921 emergency appropriation—50 million Polish marks—dwarfed any previous maintenance budget. Polish engineers arrived not as replacements but as modernizers. The Warsaw Polytechnic established a Hydraulic Institute annex at Stefan Batory University specifically to study the system. Their 1923 technical survey remains the most comprehensive analysis of the system ever produced.

The Municipality launched the “Wenecja Północy” (Venice of the North) campaign in 1924, explicitly embracing the canal heritage. New maps in Polish, Lithuanian, and Yiddish appeared at major intersections—a pragmatic acknowledgment that the waterways served all communities. The Jewish merchants who dominated canal-side commerce found unexpected allies in Polish officials who recognized that prosperous Jewish businesses meant higher tax revenues. The 1925 “Canal Statute” granted hereditary mooring rights to established businesses, protecting mostly Jewish merchants from displacement.

The Polish administration’s most ambitious project began in 1927: the Grand Canal connecting the old town to the new government quarter in Antakalnis (Antokol). French advisors from the Compagnie Nationale du Rhône designed a modern waterway with reinforced concrete walls. The ceremonial groundbreaking featured Marshal Piłsudski himself, who declared: “We build not just for today’s Poland but for the eternal city of Wilno.” Construction employed 3,000 workers through the Depression—a deliberate job-creation program that built political support across ethnic lines. Though never completed, the excavated sections served as skating rinks in winter and public swimming pools in summer.

By 1930, the canals had become so central to Wilno’s identity that the city bid to host the 1936 International Navigation Congress. The proposal, archived in the League of Nations records, envisioned delegate boats proceeding through illuminated channels to a floating conference hall moored in Lukiškės Basin. Though Amsterdam ultimately won, the bid itself demonstrated Polish ambitions for the waterways. The Municipality commissioned Czech architect Josef Gočár to design a Canal Quarter with modernist buildings rising directly from the water. His stunning perspective drawings—white cubic forms reflected in dark channels—offered a vision of Wilno as laboratory for urban innovation.

Great Flood of 1931, at what is now Zygmantu Gatvė, Supreme Administrative Court of Lithuania (Lietuvos vyriausiasis administracinis teismas).

Providence intervened again and the Great Flood of 1931 forced a final reckoning with the canal heritage. That April, exceptional snowmelt combined with torrential rains sent both the Neris and Vilnelė far beyond their banks. The canal system, designed for 17th-century water volumes, proved both blessing and curse. In the old town, the channels absorbed initial surges, sparing many buildings. But when lock gates failed at three locations simultaneously, water rushed through the streets with devastating force. Polish engineers surveyed the damage with mixed conclusions. Professor Stanisław Łukasiewicz of Warsaw Polytechnic argued the canals had prevented worse destruction; his colleague Roman Piotrowski countered that they had channeled floods into populated areas that natural drainage would have spared. The Municipality’s 1932 reconstruction plan included canal restoration, with modern lock gates and electric pumps for flood control. Detailed drawings survive in the State Archives, showing Art Deco pump houses and reinforced concrete channels. The September 1939 invasion ended all such plans.

The Second World War brought systematic destruction. The Wehrmacht’s 1941 occupation saw military engineers assess the canals for defensive potential. Their November report deemed most channels “tactically irrelevant” but noted several bridges could serve as strongpoints. Soviet partisans used the sanitation tunnels beneath canal intersections to move through the occupied city; the Germans responded by welding grates across entrances and pouring concrete into suspected passages. The July 1944 Soviet offensive brought devastating urban combat. Artillery targeted German positions along the waterways; the medieval Merchants’ Bridge, rebuilt in stone in 1712, took a direct hit and collapsed into the channel, creating a dam that flooded the Vokiečių Street district. Soviet combat engineers later reported removing over 3,000 tons of debris from the canal system.

A decaying stretch of the canal pressed against the rear façades along Šv. Ignoto gatvė, its towpath reduced to a narrow cobbled ledge. Bułhak’s camera catches the waterway in its final years.

Post-war Soviet planning treated the canals as an ideological problem wrapped in practical concerns. The 1946 General Plan for Vilnius Reconstruction, drafted in Moscow, dismissed the waterways as “remnants of feudal-bourgeois urban organization incompatible with socialist principles.” More pragmatically, planners needed space for wide boulevards suitable for military parades and rubble to level building sites. The canals offered both. Between 1948 and 1959, systematic filling began. The methodology was grimly efficient: German prisoners of war and Lithuanian political prisoners provided labor; rubble from demolished churches and dynamited historical buildings provided fill. The Great Synagogue’s ruins, cleared in 1957, filled the Schulhof Canal. The Three Crosses monument, dynamited in 1950, filled a section near Užupis. NKVD documents, declassified in the 1990s, reveal that some canal sections were used to dispose of “politically sensitive materials”—likely archives and religious artifacts—before filling. By 1960, only fragments remained: the tourist-friendly section near the university, two basins converted to reflecting pools, and a channel through Bernardinai Garden maintained for drainage.

Yet complete erasure proved impossible. The city’s infrastructure bore indelible marks of its aquatic past. Didžioji and Vokiečių streets remain inexplicably wide—their breadth preserving the dimensional memory of long-vanished waterways. Buildings along these routes rest on massive foundations engineered for canal-side loads, their basements featuring arched ceilings oriented perpendicular to the street—originally opening onto loading platforms at water level. During 1978 construction, workers discovered intact wooden pilings from a 17th-century wharf, preserved by groundwater. The peculiar hydrology created by three centuries of canals persists: basement flooding follows old channel routes; springs emerge where locks once controlled flow; the water table’s behavior still reflects hydraulic modifications made when Radvila ruled. Urban archaeologists have mapped these anomalies, creating ghost cartographies of the vanished system. Construction projects regularly uncover mooring rings and carved stones from vanished bridges—each discovery requiring archaeological assessment, delaying modern development.

The symbolic legacy persisted through Soviet occupation as well. Vilnius’s coat of arms—Saint Christopher carrying the Christ child across water—acquired new meaning after the canals’ destruction. What had been daily reality became historical memory. During the 1988-1991 independence movement, activists adopted canal imagery to represent Lithuanian identity suppressed but not erased. Sąjūdis publications featured historical canal photographs with captions reading ‘What we were, what we could be again.’ Post-independence Vilnius, focused on European integration, never pursued aquatic restoration. Yet the canal system continues to shape the city. In 2018, architecture students at Vilnius Gediminas Technical University proposed partial canal restoration as climate adaptation infrastructure. Their renderings of glass-walled water trams and restored medieval quays circulated widely on social media. The vanished waterways persist in the city’s hydrology: water still seeps through foundations along old channel routes, springs emerge where locks once stood, basement flooding follows canal paths filled decades ago.

¹ Rūta Janonienė and Mintautas Čiurinskas, “Du liudijimai apie 1610 m. didįjį Vilniaus gaisrą” [Two Testimonies about the Great Vilnius Fire of 1610], Naujasis Židinys-Aidai, 2006, no. 11, pp. 452-458. The translation is mine. For more analysis, see Mintautas Čiurinskas, „Vilnius kaip kolektyvinis veikėjas kai kuriuose XVII a. I pusės naratyviniuose šaltiniuose“, in: Senoji Lietuvos literatūra, kn. 30, Vilnius: Lietuvių literatūros ir tautosakos institutas, 2010, p. 59–84.↩

² Nicolaus Christophorus Radzivil, Hierosolymitana Peregrinatio (Brunsbergae: Georg Schönfels, 1601). On Damascus see Epistola II, 16-17; on Cairo Epistola III, 32.↩

³ Peter Paul Bajer. Scots in the Polish-Lithuanian Commonwealth, 16th-18th Centuries: The Formation and Disappearance of an Ethnic Group. The Northern World, vol. 57. (Leiden; Boston: Brill, 2012). ↩

⁴ The Latin text of the 1536 sale appears in Michał Baliński, Vilniaus miesto istorija, trans. Ona Slavėnaitė and Irena Katilienė, vol. 2 (Vilnius: Mintis, 2007), 264–265. The original document, dated St. Helena’s Day 1536, is preserved in the Vilnius city archive.↩

⁵ On the Tatar settlement and the etymology of the Totorių vartai, see Michał Baliński, Vilniaus miesto istorija, vol. 1 (Vilnius: Mintis, 2007), 143–144. ↩

⁶ Michał Baliński, Vilniaus miesto istorija, vol. 1, 107–108.↩

⁷ Aldona Baubinienė, Gediminas Vaitkevičius, Regina Morkūnaitė, Artūras Bautrėnas, “Hidrografiniai Vilniaus apylinkių ypatumai ir jų įtaka miesto vystymuisi,” Geografijos metraštis 54, 2021, 43-56, https://doi.org/10.5200/GM.2021.3, https://lgd.lt/hidrografiniai-vilniaus-apylinkiu-ypatumai-ir-ju-itaka-miesto-vystymuisi/↩

⁸ Michał Baliński, Vilniaus miesto istorija, vol. 2, 262–263, citing the fires of 1513 and 1530 and Žygimantas Senasis’s subsequent efforts to establish a municipal water supply modeled on Kraków’s.↩

⁹ On the development of Amsterdam see Jaap Evert Ambramse, “Between Art and Expediency: Origins of the Canal District,” Jan Nijman, ed., Amsterdam’s Canal District: Origins, Evolution, and Future Prospects (Toronto: University of Toronto Press, 2020), 25-42.↩

I spent all of August completely reconstructing the Lost Canals of Vilnius (and yes, if you visit that link, you’ll see that it already says “major revision”—but the version I’m about to publish makes that look like a draft). Then things ramped up and I haven’t had a chance to complete a new post. I intend to get the revision of the Lost Canals of Vilnius out the door next week, but first, this essay, which is necessary to set the stage for that work.

This is a personal essay, a position statement on my recent work with AI. Ultimately, as with all of my work, it is for myself. It a retrospective piece that makes sense of my work with AI in a way that I had not done before. I am not sure I will continue working in this vein or head in another direction in photography or technology, leaving specific investigations of AI. It still, however, may prove useful to readers interested in the topic of AI and human creativity.

As with all my substack posts, you can see this on my web site at https://varnelis.net/works_and_projects/stochastic-histories/. If you click on the footnotes on this page, they will likely take you there.

stochastic histories /stəˈkæstɪk ˈhɪstəriz/ n. pl. [from Gk. stokhastikos “capable of guessing,” from stokhazesthai “to aim at, guess at” + L. historia “narrative, account”] 1. Fabricated historical narratives generated through iterative probabilistic sampling with artificial-intelligence systems, producing documentation that appears authentic through the accumulation of plausible details. 2. A mode of artistic practice that exploits AI’s pattern-matching capacities to reveal how all historical knowledge operates through stochastic processes of inference and reconstruction. 3. Speculative archives that occupy the space between memory and invention, demonstrating that the distinction between genuine and generated documentation has become computationally irresoluble. First usage: K. Varnelis, 2025. See also: pattern-making, hallucination, probabilistic inference.

Since 2021, I’ve called my AI-generated pieces Critical AI Art. It’s a terrible name—misleading and increasingly intolerable as the work has evolved. The time has come to rename it Stochastic Histories, a title that reflects what the work actually does. Critical AI Art misaligns my practice with two opposing but ultimately identical failures: the prosumer kitsch of Midjourney prompt jockeys producing fantasy waifus and hyperreal landscapes, and the institutional theater of “critical” artists performing algorithmic audits for progressive credibility. Despite surface differences, both serve the same void. One packages it as technical novelty, the other as moral virtue. Neither changes or reveals anything. Both exist to be consumed and forgotten.

The Midjourney jockeys string together technical specifications they don’t understand—“hyper-detailed, ArtStation, Octane render, volumetric lighting, 8K”—cargo cult incantations for the aesthetically degenerate. They’ve created an arms race where every pore must be visible, every hair strand individually rendered, as if resolution could substitute for vision. Their subject matter never varies, the same collision of sci-fi, anime, and softcore imagery, endlessly recombined in a simulation of the “what’s new” feed on DeviantArt. As I detailed in my essay “California Forever or, the Aesthetics of AI Images,” these are prosumers in Alvin Toffler’s sense, merging producer and consumer roles to create kitsch marked by mass appeal and derivative aesthetics. Such work triggers the same neurological response as slot machines: just novel enough to release dopamine, just familiar enough to require no thought. Their work circulates as novelty, accumulating likes and shares for images of Emma Watson, Cannabis Goddess of Mars, bathed in golden-hour lighting, a modern-day version of Andres Serrano’s Piss Christ, while contributing nothing to genuine artistic inquiry.

Meanwhile, self-proclaimed “critical” AI artists operate from the same fundamental emptiness, though with institutional representation secured through their positions at American universities or funded by grants from the Creative Europe Programme. Whether the work involves AI or not is irrelevant: walk into any contemporary art institution and you’ll find the aesthetics of the corporate diversity training module transplanted to white-cube walls. Every exhibition includes positionality statements, as if confessing privilege or claiming intersectionality absolved them of participating in the very systems they purport to critique, together with declarations listing the tons of carbon expended to exhibit the work. They claim to “center” the “voices” of the “marginalized” while speaking to the elite, producing “disruptions” that don’t disrupt, “interrogations” that never wait for an answer. Rest assured: surveillance systems everyone already knows exist will be exposed, and algorithmic biases that have been common knowledge for years will be uncovered. Curatorial heads will not roll; on the contrary, they will nod, and appropriate boxes will be checked. This moralizing approach reduces art to forensic reports on systemic oppression, as if the purpose of art were to prove what we already know about power rather than to discover what we don’t yet understand. They perform the rituals of exposure while knowing these gestures have become empty theater, that their “interventions” will be absorbed into the academic-industrial complex as another line on a CV, another grant secured, another exhibition lets everyone feel they’ve confronted power while leaving all structures intact. Now that this crowd has come to AI, their work continues in the same vein, offering the usual litany of complaints from yesterday’s war.

As Peter Sloterdijk diagnosed in his Critique of Cynical Reason, we live in an age of enlightened false consciousness: people know very well what they are doing, but they do it anyway. The critical intellectual who exposes ideological operations pretends their audience is naïve, but the audience, like the intellectual, is already cynical—both know the game is rigged and continue playing it. Critiques persist even though such gestures change nothing. are made even though such gestures change nothing. Institutions, too, have learned to incorporate their critique as immunization—funding the very work that claims to expose them, thereby proving they can take it, neutralizing any actual threat. The art world pays for its own flagellation, emerging stronger and unchanged. Both camps—the Midjourney jockeys and the critical artists—exhibit what Sloterdijk calls the unhappy consciousness of modern cynicism, aware their work is empty but unable to stop producing it. The real revelation is not that technology reflects the biases of its training data. It is that all cultural production has always been a stochastic process, and the panic about technological “harms” is itself a symptom of our refusal to acknowledge that creativity was never the province of individual genius but has always emerged from the vast, impersonal machinery of cultural recycling.

But all this hardly matters. We are in the post-AI era. Adobe Photoshop now includes generative AI as just another tool. Microsoft Word now includes an AI Assistant. It is no longer a novelty worth celebration or critique. Doing anything interesting with it requires treating AI not just as a medium but as a distinct mode of production. Unlike mechanical reproduction—copies of originals—or mass media—singular messages broadcast to many—or even digital computation, which simulated reality through deterministic processes, AI operates through probabilistic extraction and recombination. It ingests the totality of digitized human expression, identifies patterns across this vast archive, and generates new instances that never existed but feel as if they should have. Its hallucinations—those moments when pattern matching produces impossible responses—aren’t failures but the source of its creative potential. Crucially, AI reveals what was always true but hidden: all cultural production operates through similar stochastic processes of pattern recognition, recombination, and variation.

The term stochastic histories deliberately plays off Emily Bender’s dismissive “stochastic parrots.” “Stochastic” means probabilistic; “parrot” means imitation. Together, her phrase describes how large language models merely repeat patterns of language learned from vast datasets, generating fluent but potentially meaningless or biased output through probabilistic sampling. But unlike Bender, I was trained as a historian of cultural production. Where she sees mindless repetition, I see an unintended insight into how culture itself operates. Stochastic histories are fabricated narratives produced through probabilistic sampling with AI systems. Each project requires months of work—thousands of prompts refined and recombined until coherent worlds crystallize from the collision of probabilities. I do not select the resulting texts and images merely for technical quality but also for their capacity to provoke a kind of productive uncertainty. The term also recognizes that all historical and speculative production—not just that produced by AI—operates through similar processes of pattern-matching and plausible inference, constructing narratives from fragments and filling gaps with what must have been there or what might yet be. The “parrot” was never the insult critics imagined; parrots are remarkable pattern-matching machines, and so are we.¹

My stochastic histories lie within a specific lineage, both in broader movements and in the circumstances of my own biography as an artist and theorist. For me, the key departure point is David Wilson’s Museum of Jurassic Technology, which I first encountered in 1996 and have returned to repeatedly (the Museum is next door to the Center for Land Use Interpretation, with which I began collaborating with the late 1990s). The Museum presents detailed exhibits on obscure scientists, bizarre natural phenomena, and peculiar inventions—some real, some fabricated, most existing in ambiguous spaces between documentation and invention. Wilson, who had a previous career in Hollywood special effects and understands how reality is constructed frame by frame, does not critique truth so much as create a space where epistemic certainty dissolves. I once had the rare opportunity to see the backstage of the Museum, where the exhibition rooms were revealed as theatrical flats within a vast interior space, the vitrines accessible from hidden corridors for maintenance—the museum itself a stage set masquerading as architecture, which in turn throws into question the relationship between display and servant spaces in traditional museums. The desire to know whether something is “real” becomes the content, but simultaneously, standing in the Museum’s quiet, darkened galleries, gazing at exhibits while hearing distant sounds—a fox screaming, crickets chirping, a slideshow narrator explaining a preposterous history of the museum, the haunting voice of an opera singer, bells drifting from other rooms—one becomes acutely aware of one’s own body, achieving what the minimalists of the 1960s sought but far more effectively. The experience reveals something crucial: unlike the critic who maintains distance to perform exposure, the Museum makes visitors aware that the very act of evaluating makes them constituent parts of the work.

In Mr. Wilson’s Cabinet of Wonder, Lawrence Weschler connects the Museum to the pre-Enlightenment Wunderkammern—Renaissance collections where natural specimens, artistic creations, and technological marvels existed before disciplinary boundaries solidified. Cabinets of curiosities emerged from the era of natural philosophy, when investigating nature meant contemplating the world in all its marvelous variety, when the boundaries between the miraculous and the mechanical, the divine and the empirical remained productively unclear. The Wunderkammer took the stage alongside another technology for managing early modernity’s information explosion: the book index. Both arose in the sixteenth century as print culture and global exploration overwhelmed existing systems of organization, creating too many books to read linearly and too many objects to fit familiar categories. Natural philosophers assembled these collections not as data sets but as chambers for contemplation, where a unicorn horn (actually a narwhal tusk) might reveal divine providence as surely as it demonstrated nature’s ingenuity. Like the index that allows readers to access text through alphabetical entry points, the cabinet’s drawers and shelves permitted nonlinear exploration. Just as index entries create connections across distant pages, meaning emerged through juxtaposition rather than sequence—a narwhal tusk might sit beside a mechanical duck, an ostrich egg beside a dodecahedron carved from ivory, each object a node in a network of correspondences between microcosm and macrocosm. The wonder of the Wunderkammer did not so much fix meaning as create a productive uncertainty, a cognitive state of openness. These collections revealed that knowledge itself requires an architecture—literal shelves and cases, but also conceptual frameworks that determine what belongs with what, what counts as evidence, what deserves preservation.²

Wilson’s Museum extends this logic into contemporary practice, but with a crucial difference: where the Wunderkammern cultivated wonder through genuine rarities and (usually) honest misidentifications, Wilson achieves this through what the Russian Formalists called ostranenie—defamiliarization in the form of a meticulous institutional fiction. At his Museum, the vitrine functions as a technology of belief: placing something behind glass with a label does not simply display knowledge; it produces it, transforming objects into evidence, the everyday into the remarkable. His Museum operates as a three-dimensional index, where each exhibit points to others through conceptual correspondence rather than alphabetical order, making visible what the book index conceals: that all systems of organization are arbitrary impositions on the universe of things and ideas. The Museum reveals that our frameworks for organizing knowledge are themselves fictions, that the difference between the Leiden University anatomical collection and the Museum of Jurassic Technology is one of degree, not kind. By faithfully reproducing every convention of museum display—the authoritative labels, the scholarly apparatus, the hushed galleries—while filling them with impossible content, Wilson makes the familiar institutional frame suddenly alien. We begin to notice the arbitrary confidence of wall text, the coercive authority of the vitrine, and the way dim lighting and reverent silence produce belief regardless of what they frame. Through careful presentation, the plausible feels impossible and the impossible feels documented, but more importantly, the very distinction between these categories becomes unstable.

At this point, it is also worth citing Joseph Cornell’s boxes, which function as both precursors and miniature, portable counterparts—private Wunderkammern assembled from the detritus of consumer culture. Cornell explicitly referenced the tradition, but where Renaissance collectors displayed unicorn horns and ostrich eggs as trophies of colonial reach, Cornell assembled his cosmologies from hotel advertisements, celestial maps, and scraps found in New York’s junk shops. His boxes did not fabricate evidence like Wilson’s Museum; they transformed genuine ephemera into hermetic worlds through careful juxtaposition, creating wonder from the worthless.

It seems beyond coincidence, then, that in Count Zero, cyberpunk author and pundit William Gibson centered his novel around an AI that creates Cornell boxes. In a 1989 interview, Gibson explained that Cornell’s assemblage method was “the key to the whole fucking thing, how the books are put together and everything”—describing his own creative process as “grabbing little hunks of kipple, and fitting them together.” The Boxmaker wasn’t just a character; it was Gibson’s explicit metaphor for how novels get made, for how all cultural production operates through stochastic assemblage from existing fragments. Today, any image generator can produce convincing Cornell-style boxes in seconds, complete with nostalgic patina and surrealist juxtapositions. Gibson himself now recoils from AI art as “weird as hell.” This is not mere contradiction but reveals something structural: Gibson’s blindness to the implication of his own insight is not a failure of reasoning but a constitutive necessity—the concept of “human creativity” requires that the machinery remain invisible. The moment AI makes pattern-matching and probabilistic recombination explicit and operational, the entire framework collapses. Gibson’s position is literally impossible to maintain coherently, which is precisely what is “weird as hell.” What appears as personal inconsistency exposes how meaning, authorship, and artistic value have always depended on a foundational blindness to their own mechanical operations. We cannot simultaneously acknowledge that creativity operates through stochastic probabilistic processes and preserve the human/machine distinction—yet we must try, because abandoning either claim threatens the entire edifice of how we understand cultural production.³

Cornell dedicated two works to another precursor of stochastic histories, Jorge Luis Borges—”For Jorge Luis Borges,” a collage, and “The Puzzle of the Reward (for Jorge Luis Borges)”—creating a direct link between the assembler of found objects and the architect of impossible libraries. Where Cornell transformed commercial detritus into hermetic worlds through physical juxtaposition, Borges achieved similar effects through textual accumulation. His Tlön, Uqbar, Orbis Tertius presents an entire fictional world that gradually replaces reality through sheer documentary weight—encyclopedic entries, scholarly footnotes, and discovered objects bearing inscriptions in imaginary languages. The story isn’t just about a fictional world but about how fiction becomes real through the apparatus of scholarship, how plausibility emerges from the proper citation format and cross-reference.

Pierre Menard, Borges’s character who rewrites Don Quixote word for word three centuries after Cervantes, reveals that identical repetition produces difference—the same words mean something entirely different when produced in a different era, by a different hand. This isn’t plagiarism but a theory of cultural production: all writing is rewriting, all creation is recombination. The Library of Babel, containing every possible book, suggests that originality is merely a matter of selection from a pre-existing field of possibilities. Borges understood that culture constructs itself through pattern and variation, that authenticity and fabrication are categories we impose rather than qualities inherent in things.

This tradition of productive fabrication extends through various methods and media. Contemporaneous with Wilson, the photographer Joan Fontcuberta constructs entire scientific archives with obsessive thoroughness. His Fauna presents complete taxonomies of impossible animals—the Solenoglypha Polipodida, a snake with legs, documented through skeletal diagrams, field notes, habitat studies, and photographic “evidence.” His Sputnik project (I purchased a copy at the Museum of Jurassic Technology) resurrects Ivan Istochnikov, the lost Soviet cosmonaut, playing directly with the persistent conspiracy theory that cosmonauts died in space before Gagarin, erased from official history. Fontcuberta provides everything—mission patches, crew photographs where Istochnikov has been carefully removed, leaving ghostly gaps, technical documents, commemorative stamps, even childhood photographs. The project gains power precisely because it builds on existing suspicions about Soviet secrecy, making viewers question whether this is revelation or fabrication. Many refuse to believe it’s fiction even after being told, insisting that Fontcuberta must be exposing actual suppressed history. Every element mimics the authoritative grammar of scientific documentation. Fontcuberta doesn’t critique truth so much as demonstrate how truth effects are produced through proper formatting, comprehensive citation, institutional presentation—and how they become even more convincing when they confirm what we already suspect.

Another artist whose work was essential for the world-building nature of my stochastic histories is Christoph Büchel, whose exhibit at Los Angeles Contemporary Exhibitions(LACE) in the fall of 2003 affected me deeply. It seemed, upon approaching the gallery, that some kind of horrific event had happened at its Hollywood Boulevard space—unsettling in the post 9/11 years—the windows were boarded up, but a representative appeared and had us fill out an unsettling disclaimer absolving the gallery of liability. We entered into a darkened room that clearly once was the gallery, but now was strewn with refuse and rubble. Light came from a door that, when opened, led into a disturbingly small, well-lit space with a locker at the end, an institutional space, perhaps a school or factory or military installation. The place appeared to have seen better days: some tiles in the floor had broken away and an animal hole led off somewhere through the dirt. Eventually, we discerned that we could enter into the locker, crouch down into a hole, and emerge under a desk in another, much larger room (again all this is happening in what we expected was the white cube of the gallery we had visited many times). This room was connected to a large complex of rooms filled with what appeared to be missiles, military equipment, and institutional furniture, an unsettling space with the air of Cold War menace, an excavation into another world unknown to civilians. This wasn’t a representation but rather a total environment, removing all reference to the gallery. Where Wilson’s Museum maintains the institutional frame while making it uncanny, Büchel eliminates the frame entirely.

There is architectural precedent for stochastic histories as well. In particular, I have found the work of the Italian radical architecture group Superstudio compelling. Even though modernists had worked with media from the start, the wide accessibility of color printing by the late 1960s meant that saturated, glossy color had become the norm. Superstudio took the medium itself as a site, understanding that mainstream architectural publications such as Domus and Casabella would be as important as the gallery. The glossy magazine page became a space where fictional architectures could exist with the same presence as documented reality. Carefully selecting documentary photographs to build upon, Superstudio understood that photomontage carries different authority than drawing—the photographic substrate insists on its own truth even when obviously manipulated. Their method was explicitly critical, showing how modernist rationality taken to its logical extreme would produce a world of total homogenization. Yet, the critique worked through seduction rather than revulsion, through images so compelling they made dystopia feel inevitable, even desirable.

The most perverse precedent is that of architect Jean-Jacques Lequeu, whose drawings I had the privilege to see at the Morgan Library a week before the pandemic lockdown. Lequeu’s drawings combined Neoclassical rigor with fantastical, often erotic and transgressive imagery that remained unknown until their rediscovery in the 1930s by Emil Kaufmann, who encountered them in the Bibliothèque nationale de France in Paris. His work includes visionary designs for temples, grottos, and civic buildings that blend archaeological precision with theatrical excess, featuring allegorical figures, exotic references, and occasionally explicit sexual content. But in the only book-length study on Lequeu, French architecture historian Philippe Duboy claims that Marcel Duchamp, while working at the Bibliothèque, fabricated Lequeu’s drawings, attributing them to this third-rate figure from the revolutionary era. According to Duboy, Duchamp invented this 18th-century visionary architect complete with theoretical texts and fantastical building designs. Many scholars dismiss this as conspiracy theory, but their rebuttals seem to rest on little more than the audacity of Duboy’s claim. The vertigo here is absolute: either we’re looking at authentic 18th-century visions that anticipate surrealism by a century, or we’re looking at Duchamp’s forgeries that have been absorbed into architectural history as fact. The authentication machinery—carbon dating, stylistic analysis, archival provenance—cannot resolve this; it can only accumulate probabilities that feel like certainties. This conundrum shows how modernism might construct its own prehistory through forgery, how we recognize certain patterns as “visionary” because they match what we’ve been trained to expect from an outsider genius.⁴

Where these predecessors employed physical assemblage, institutional mimicry, photomontage, or architectural intervention to produce their fictions, I work with AI’s probabilistic sampling—a method that doesn’t construct or manipulate but rather summons plausible realities from latent space, making the process of authentication itself the medium. Stochastic histories emerge from this tradition, but also respond to our own time, when AI has become inextricably entwined with cultural production. This work, however, is not merely an interrogation of AI. Stochastic histories set out to produce a particular vertigo—not simply uncertainty about what is real, but the recognition that this is exactly how the real gets constructed. The AI hasn’t learned to fake history; it has learned history’s own methods of self-authentication. When we cannot distinguish the generated from the genuine, we’re forced to confront how all archives operate through probabilistic accumulation of plausible details. The dataset is Benjamin’s archive in its raw form, stripped of curatorial narrative—every photograph that has ever been digitized, every document scanned, every architectural drawing uploaded, all flattened into latent space. We find ourselves staring directly at culture’s unconscious, and in that moment we glimpse something like the Lacanian Real—the point where representation collapses and the machinery of meaning-making becomes visible.

My process embraces both chance and curation. Each image demands hundreds of iterations, each prompt a negotiation with probability space. Although I aim to produce works that appear plausibly real, the stochastic element of this production nevertheless reveals itself most powerfully in glitches—the hallucinations where the system generates mistakes—a three-story-tall streetlight, a misshapen human figure, a façade in the process of digital decay. These glitches show the moments where the probability field tears and something genuinely alien emerges. A building that shouldn’t exist but feels absolutely inevitable. A document that describes events that never occurred but should have. These hallucinations are terrifying precisely because they feel more plausible than fact—they reveal how thin the membrane is between history and its probabilistic shadow.

One reader of the latest revision of the Canals of Vilnius essay was disappointed by the quotes about the fire by Alessandro Cilli, a Jesuit priest (note that these are not in the current version, but will appear next week). She knew that I wrote the fictional history with help from an AI, and she felt the AI’s writing was unconvincing. And yet, Cilli’s quotes are all real: they seem unreal, too strange to be believed.

Stochastic histories make visible that the past is continuously reconstructed through pattern and probability, that authenticity is a learned aesthetic, that culture reproduces itself through endless stochastic variation. Cilli’s quotes seem unreal while Lequeu’s drawings are considered authentic by most scholars because the possibility that Cilli’s quotes are real and Lequeu is a forgery by Duchamp is too unsettling. We are all stochastic parrots. Every historical narrative, every archival discovery, every recovered memory operates through the same probabilistic logic—filling gaps with what should have been there, constructing continuity from fragments. Meaning has always emerged from vast, impersonal processes of pattern matching and recombination. Thus, the old saw that the dataset is biased misses the point entirely—it is literally culture’s biases made operational, our collective patterns of recognition transformed into probability distributions. When I work with these systems, I’m not using a tool but inhabiting a space—a vast probability field containing every possible history that could feel real.

The recursive nature of this practice reveals itself in the “7 Fables of Accelerationism,” where AI systems trained on archives containing Superstudio’s imagery from the 1960s generate new speculative futures that cite those radical visions. Here, I project stochastic histories into both the future and the past. Just as Superstudio’s photomontages pointed back to the constructivists and early modernists whose visual language they appropriated, my fables emerge from probability spaces where the Continuous Monument—as well as the images that Superstudio trained themselves on—exist as patterns to be recombined as well as conditions they could not have anticipated. The dataset doesn’t just enable the work; it determines its visual genealogy, making every generated image a stochastic descendant of its training data.

Stochastic histories operate in the space between memory and invention, between what was and what should have been, between human intention and machine hallucination. They’re not critiques or spectacles but investigations into how the past gets constructed, how the future gets imagined, how culture reproduces itself through endless variation. Turning Gibson on his head, I very much hope that these works are “weird as hell.” In revealing how probability shapes our expectations and how the line between history and hallucination dissolves at the level of pattern-matching, they make visible that culture has always been a stochastic process, and that being a stochastic parrot might well be the only kind of consciousness there is.

¹ See The Generative Turn: On AIs as Stochastic Parrots and Art for a fuller discussion of how structured repetition enables rather than constrains creativity. ↩

² Lawrence Weschler, Mr. Wilson’s Cabinet of Wonder (New York: Pantheon Books, 1995). ↩

³ Genevieve von Petzinger. The First Signs: Unlocking the Mysteries of the World’s Oldest Symbols. (New York, NY: Atria Books, 2016). ↩

⁴ Phillipe Duboy, Lequeu: An Architectural Enigma (Cambridge: The MIT Press, 1987). ↩

I am leaving for Lithuania today and my original plan was to have this book completely done before I leave. That is not happening. I still have at least one essay—on AI and glitches—to go. It should be much briefer, but then there will also be an introduction. At that point, I will put a rough draft of the book up on my site and refine it over a couple of months. The book will be short, about as long as one of the classic Semiotext(e) Foreign Agents series from the 1980s (such as Baudrillard’s Simulations, Virilio’s Speed and Politics: An Essay on Dromology, Or Deleuze & Guattari’s Nomadology: The War Machine), something you can read in one or two sittings. I had also hoped to create a limited edition in print, maybe 100 or 200 books, but the Trump tariffs have led to the shuttering of Edition One books, the press I had hoped to use. I will keep you posted.

You can read the rest of the book on my site:

The New Surrealism? On AI and Hallucinations
The Generative Turn: On AIs as Stochastic Parrots and Art
The Rise and Fall of the Author

As always, this post is on varnelis.net @ https://varnelis.net/works_and_projects/humanity-and-its-double-the-uncanny-in-art-and-artificial-intelligence/

All I ask is that if you enjoy this post, you like it and share it. Nothing makes me happier than growing my readership.

The links for the footnotes in Substack posts point you back to my site. Sorry about that. Don’t click on them if you don’t want to go there.

On September 12, 1940, four teenagers exploring a hill above Montignac, France, discovered a small hole in the ground that had been revealed when a tree fell in a storm. Descending into the opening in search of a long-rumored cavern supposedly containing hidden treasure, they entered the darkness with a makeshift lamp. Soon, they saw the walls around them blaze with life. Horses galloped across uneven stone surfaces, aurochs charged through rocky corridors, and bison wheeled. In the dancing light, these painted forms seemed alive. The boys had discovered Lascaux.¹

The painters who worked these walls some twenty millennia ago possessed technical mastery rivalling any subsequent achievement in representation. They drew from memory, demonstrating an understanding of animal anatomy derived from intimate observation. Employing the cave's three-dimensional surfaces in their compositions, they used natural bulges to suggest the swelling muscles of horses and the rounded bellies of bison. Most remarkably, they achieved what would take Futurist painters millennia to develop: the representation of movement through simultaneity, showing animals with multiple sets of legs to suggest galloping.

After Lascaux, the most shocking revelation about Paleolithic life came with the discovery of Chauvet Cave in 1994. Chauvet demonstrated that Lascaux was not the work of isolated genius but part of a largely lost artistic tradition that lasted 25,000 years—evidence of what Gregory Curtis has called a 'classical' culture, representing as he states, "not only the first great art but also the first great philosophy, the first attempt we know of to put meaningful order to the chaos of the world."² But this sort of historical continuity is unfathomable to us moderns, the sort of continuity of time in which the passage of time itself is lost, a culture of the sort possessed by animals, not humans. As at Lascaux , there is an alien quality to these paintings. There are no landscapes, no vegetation, no context beyond the animals themselves and the strange symbols—lines, dots, branching signs, handprints—that accompany them. Together with these enigmatic markings, which archaeologist Genevieve von Petzinger suggests may represent the foundation for the future development of writing itself, the paintings point toward purposes we cannot decode, thinking we cannot access, social structures we cannot reconstruct.³

Still, the discovery of Chauvet cannot be compared with the stunning discovery of Lascaux. While cave paintings had been discovered before—most notably the fragmentary works at Altamira in 1879—Lascaux was the first complete prehistoric artistic environment discovered. Lascaux presented coherent galleries of sophisticated paintings that used the cave's natural architecture as an integral part of their design. Lascaux shattered what we thought about ourselves: it revealed that the essential capacities of human consciousness—symbolic thinking, artistic sophistication, the ability to create meaning through representation—were not recent evolutionary developments, but rather ancient achievements. The paintings demonstrated that what we consider distinctly "modern" consciousness had existed for tens of millennia, collapsing the comfortable distance between prehistory and modernity at a moment when modern man was at his most brutal, a moment where the meaning of progress and civilization were being called into question.

Writing in 1955, the French Surrealist philosopher Georges Bataille argued that Lascaux represents the foundational act of humanity:

We have, after all, added very little to the inheritance left us by our predecessors: nothing supports the contention that we are greater than they. "Lascaux Man" created, and created out of nothing, this world of art in which communication between individual minds begins. … At Lascaux, more troubling even than the deep descent into the earth, what preys upon and transfixes us is the vision, present before our very eyes, of all that is most remote. This message, moreover, is intensified by an inhuman strangeness. Following along the rock walls, we see a kind of cavalcade of animals... But this animality is nonetheless for us the first sign, the blind unthinking sign and yet the living intimate sign, of our presence in the real world.⁴

Bataille's "inhuman strangeness" can be read within the framework of Sigmund Freud's uncanny [unheimlich]: the experience of encountering something simultaneously familiar and alien, something that should be dead yet appears alive, something that transgresses fundamental categories upon which our understanding of reality depends. For Freud, the uncanny marks moments when representation becomes so successful that it threatens to collapse the distinction between the artificial and the authentic, the created and the real. Most of all, reading Bataille on Lascaux, the uncanny reveals itself as the psychological signature of humanity's most ambitious technological achievements—our repeated attempts to project ourselves into the world, to impose our consciousness on inert matter.⁵

This trajectory from Lascaux to contemporary Artificial Intelligence represents more than technological progress—it reveals a sustained human drive to create things that somehow speak back to us. The painted animals at Lascaux function as what Bataille calls "the blind unthinking sign and yet the living intimate sign"—they communicate meaningfully while operating through processes that bypass conscious intention. This is precisely the paradox that characterizes our encounters with Large Language Models: they generate seemingly conscious responses through statistical operations that are fundamentally "blind and unthinking," yet they establish "living intimate" communication with us. When users today report feeling unsettled by ChatGPT's responses—recognizably human in their coherence yet produced through alien computational processes that even their programmers don't always understand—they experience the same cognitive dissonance that has marked every major advance in simulation technology. The seventeen-thousand-year trajectory from those flickering images on cave walls to today's Large Language Models (LLMs) represents not mere technological advancement but the return of a fundamental anxiety about the boundaries between the authentic and the artificial, the conscious and the mechanical.

The following essay traces this uncanny lineage through its major manifestations: the classical recognition of representation's dangerous power to deceive, the nineteenth-century media that captured and preserved actual human presence, the mechanical marvels that animated matter through clockwork ingenuity, and the contemporary AI systems that simulate consciousness itself through pure language. Each technological leap has reactivated the same essential anxiety while pushing us closer to an ultimate transgression—the creation of artificial beings indistinguishable from ourselves.

I do not talk about science fiction in this essay. I contend that the condition of rapid AI development today is a technological Singularity. As such, science fiction becomes less useful than history. We all expect to read science fiction to come to an understanding of AI, but the uncanny is a matter of the return of the repressed. The uncanny effects that users report when encountering LLMs are not science fictional scenarios finally realized but rather returns to earlier anxieties in new forms. History reveals these patterns more clearly, showing how each era's most sophisticated attempts to animate matter led to a return of anxieties while pushing us incrementally closer to fulfilling dreams as old as human consciousness itself.

Lascaux is a representation that is born out of a lie, out of a need to deceive the eye with mere pigment. Long before any mechanical clockwork or digital algorithm, ancient artists understood this. In the classical world, for the first time, we can read a profound ambivalence toward this power—simultaneously celebrating mimetic virtuosity while harboring deep philosophical suspicion about representation's relationship to truth.

This tension finds its clearest expression in Plato's Republic, where Socrates describes prisoners chained in a cave, mistaking shadows on the wall for reality itself. When one prisoner is freed and compelled to look at the fire casting the shadows, then dragged outside to see the sun, he experiences painful enlightenment—recognizing that the shadows were mere copies of copies, "three steps removed" from the true Forms. In Plato's metaphor, enlightenment requires abandoning these representations to apprehend authentic reality. For Plato, artists create copies of copies—a painting of a bed imitates a physical bed, which itself merely imitates the eternal Form of Bed, making artistic representation "three steps removed" from truth. This distance from reality made mimetic artists so dangerous to the ideal state that Plato banished them entirely from his Republic.⁶

This philosophical rejection of mimesis exists in tension with Lascaux, which is, of course, literally a cave. Did Plato somehow know that caves were where representation started? But this representation isn't merely deception, as we have established. While Plato's cave-dwellers remain trapped by shadows, the Paleolithic painters created meaning through them. Where Plato saw representation as a fall from truth, the discovery at Lascaux suggests representation was humanity's first philosophical breakthrough.

But Plato was not alone in holding a fundamental anxiety about the ultimate success of representational ambition. This is the subject of the famous painting competition between Zeuxis and Parrhasius, two masters of Greek art. Zeuxis painted grapes with such skill that birds flew up to them and tried to eat them, deceived by the perfection of the illusion. Confident of victory, he challenged Parrhasius to reveal his own work. When Zeuxis reached to pull aside what he assumed was a curtain covering his rival's painting, he discovered that the curtain itself was the painting. Parrhasius had achieved something more unsettling than fooling animals; he had fooled a master of artistic illusion.⁷ This story captures the dialectic of ambition and anxiety: the desire to render the world so precisely that the rendering is indistinguishable from reality and the anxiety that our cognitive capacities, however refined, remain vulnerable to sufficiently skilled simulation. Zeuxis's grapes exploited the pattern-recognition systems of birds; Parrhasius's curtain exploited the far more sophisticated visual processing of a human expert. The progression suggests an arms race between representation and recognition, with each advance in mimetic skill threatening to overwhelm our ability to distinguish the artificial from the authentic. What makes the tale particularly unnerving is that Parrhasius succeeds not through divine intervention but through pure technical mastery—a recognition that perfect simulation threatens the coherence of categories upon which our understanding of reality depends.

This classical anxiety is echoed in Sigmund Freud's 1919 essay "Das Unheimliche" ("The Uncanny"), which we have alluded to. Here, Freud's central insight is that the uncanny emerges not from encountering something entirely foreign, but from the return of something familiar that has been repressed. The uncanny thus represents a temporal collapse, a haunting return of childhood fears and primitive beliefs we thought we had outgrown. Freud's paradigmatic example, E.T.A. Hoffmann's 1817 tale "The Sandman," anticipates with remarkable prescience our contemporary anxieties about Artificial Intelligence. The protagonist Nathanael becomes obsessed with Olimpia, who appears to be the perfect woman—beautiful, attentive, entirely devoted to him—only to discover that she is an automaton, a mechanical creation designed to simulate human behavior. The story's uncanniness intensifies through its deliberate conflation of creation and destruction: Olimpia's maker is Coppelius (or his double, Coppola), the same sinister figure present at Nathanael's father's death during mysterious alchemical experiments that, in the end, kill his father. When the automaton is finally destroyed, her artificial eyes are literally torn out—directly fulfilling the childhood terror of the mythical Sandman who steals children's sight. Yet in Freud's reading, the source of the uncanny lies not primarily in the automaton itself but in the return of infantile anxieties about sight and castration. The automaton serves as the vehicle through which repressed fears manifest, activating unconscious anxieties about identity, authenticity, and bodily integrity. But Hoffmann's tale probes even deeper anxieties: Olimpia embodies living death, appearing vibrantly alive while remaining fundamentally inanimate matter. Her perfect simulation of life forces recognition that vitality itself might be nothing more than convincing mechanical performance, threatening our most basic categories for distinguishing the living from the dead.⁸

As we will see later, the choice of an automaton is not accidental. Automata activate anxieties that extend beyond individual psychology to fundamental epistemological uncertainty. The perfectly convincing artificial being reveals a disturbing recognition: we possess no direct access to another's consciousness, only behavioral evidence of its existence. We observe speech, gesture, response, apparent emotion—but never the inner experience of the Other. If Olimpia can perfectly simulate a conscious woman's responses, how can we determine whether any "real" person truly experiences anything rather than merely performing the behaviors we associate with awareness? The automaton reveals that all our evidence for other minds remains purely inferential, based on external manifestations that could, in principle, be mechanically reproduced without any accompanying inner life.

This recognition opens onto an even more vertiginous possibility: that consciousness itself might be illusory, a convincing glitch in otherwise mechanical processes. If we cannot definitively establish the presence of consciousness in others, perhaps our sense of inner experience represents nothing more than the subjective effect of complex information processing. The automaton suggests not merely that artificial beings might simulate consciousness, but that consciousness itself might be a simulation—that we too might be fundamentally soulless, experiencing what amounts to an elaborate hallucination of subjective experience generated by unconscious mechanical operations. In this reading, the uncanny emerges not from encountering artificial consciousness, but from recognizing that all consciousness may be a ruse.

Indeed, Freud's uncanny was invoked in the context of Artificial Intelligence in 1970 by roboticist Masahiro Mori when he identified what he termed the "uncanny valley"—the phenomenon whereby humanoid robots that appear almost, but not exactly, like real humans elicit feelings of eeriness and revulsion. Mori observed that as robots become more human-like, our emotional response to them becomes increasingly positive until reaching a point where subtle imperfections in their human likeness create a sharp drop into negative response—the valley of uncanniness. This dip occurs precisely because such beings occupy an ambiguous category that threatens our ability to distinguish the living from the mechanical.⁹

Recent psychological research has provided empirical validation for Mori's intuition. Computer scientist Karl F. MacDorman's studies conclude that highly humanlike robots trigger what psychologists call "mortality salience"—a psychological state that elicits thoughts of death. The android's subtle lifelessness, mechanical defects, or failure to match human behavioral norms can be subconsciously perceived as death-like, evoking the same anxiety produced by encounters with corpses, graveyards, or other reminders of mortality. MacDorman's findings also align remarkably with Freudian theory: the uncanny valley emerges because humanlike robots evoke our repressed fear of death by appearing nearly—but not fully—alive. They resemble the doubles and revenants that populate our unconscious, recalling primitive beliefs about animated matter that we have consciously rejected but subconsciously retain. These lifelike figures breach our psychic defenses, triggering unease as the familiar territory of life shifts unsettlingly into the repressed yet ever-present domain of death and uncanny animation.¹⁰

Large Language Models and AI image generators represent a qualitatively different form of uncanniness that transcends Mori's uncanny valley entirely. Where the uncanny valley emerges from visual and behavioral imperfections in physical robots—subtle failures in appearance or movement that betray their artificial nature—contemporary AI systems achieve uncanniness through perfect simulation rather than imperfect mimicry. They succeed so completely at reproducing human patterns that they directly force us to confront the possibility that there may be no essential difference between authentic and artificial intelligence. Alan Turing anticipated this development in his famous 1950 test, which proposed that a machine could be considered intelligent if a human evaluator could not distinguish its responses from those of a human being through textual conversation alone. Turing's insight was prescient: by removing physical appearance from the equation and focusing purely on linguistic behavior, he identified the arena where Artificial Intelligence would ultimately prove most unsettling. But contemporary LLMs now routinely pass informal Turing tests, producing responses that are not merely convincing but often more eloquent, empathetic, and creative than typical human communication.

This represents a fundamental shift in the nature of artificial uncanniness. These systems create doubles of human consciousness that respond appropriately to complex queries yet emerge not from understanding but from probabilistic token prediction across vast datasets. AI image generators produce photographs, paintings, and portraits virtually indistinguishable from human-created works, yet originate from pure statistical operations rather than aesthetic intention. Perhaps most uncannily of all, they generate artistic and literary expressions that feel genuinely human—demonstrating apparent creativity, humor, and even wisdom—while operating through mechanisms entirely alien to human cognition.

The effect intensifies not because of the near-success of the uncanny valley, but rather because these systems succeed too well at simulation. When we encounter an AI-generated image of a person who never existed, or read text that demonstrates apparent empathy yet originates from statistical pattern matching, we experience exactly what Freud identified as the return of primitive beliefs we thought we had outgrown. Again, the fear extends beyond mortality to encompass something more fundamental: that consciousness itself might be reducible to pattern recognition, that there may be nothing more to human intelligence than the very processes these machines have mastered. The anxiety is not merely that representation might deceive perception, but that perfect simulation reveals the absence of any essential difference between authentic and artificial consciousness. Indeed, continuing in this vein, one understands the recurrence of the repressed uncanny as itself all too predictable, something that one could easily imagine an LLM would claim to experience.

The classical anxiety about representation's power to deceive continued through centuries of increasingly sophisticated illusionism—from trompe-l'oeil traditions that fooled viewers into reaching for painted objects, to Bernini's marble sculptures so lifelike they seemed to breathe, to Dutch still lifes that captured dewdrops and reflected light with uncanny precision. Yet the nineteenth century introduced a qualitatively different form of uncanny experience that transcended even the most skillful visual deception. Where traditional art achieved verisimilitude through human craft and ingenuity, the new media technologies of photography and sound recording captured and preserved actual human presence itself. These devices did not merely imitate life; they seemed to trap it, creating haunted technologies that appeared to facilitate communication with the absent, the distant, and the dead.¹¹

The daguerreotype, introduced in 1839, immediately produced responses that went far beyond aesthetic appreciation. Early viewers described an almost supernatural quality to photographic images, a sense that something essential about the subject had been captured and preserved in the silvered plate. Some American commentators reached instinctively for the language of sorcery. They christened the camera a "magic box" endowed with an inexplicable "magic power," and one terrified sitter bolted down the stairs "as if a legion of evil spirits were after him" the moment the lens fixed its stare upon him.¹² This pattern of supernatural attribution was in part because of the nature of the medium. Early photographs required long exposure times that could lead to spectral effects from motion in the frame—blurred figures, ghostly transparencies, and doubled images that seemed to capture something between presence and absence. These technical accidents reinforced supernatural interpretations, but Alan Trachtenberg, in "Mirror in the Marketplace: American Responses to the Daguerreotype, 1839-1851," points out that photography's uncanny effects drew from much older anxieties about representation itself. The mirror metaphor that dominated early photographic discourse—in part because of the mirror-like surface of early daguerrotypes and their physical similarity to pocket mirrors—carried within it what Trachtenberg identifies as "the duplicity traditionally suspected of pictures and picture-makers." No matter how well-intentioned as praise, the comparison of photographs to mirrors "returned to its users their own confusions and incomprehension, a modern version of old suspicions aroused by images and icons." Photography succeeded too well at what painters had always been suspected of doing—creating convincing illusions that might deceive the viewer, crossing the dangerous boundary between representation and reality that classical thought had warned against.¹³

Here we see what Freud theorized as the return of the repressed: not the emergence of entirely new fears, but the reactivation of ancient ones that Enlightenment rationality thought it had overcome. Photography was based on science and technology, a knowledge of chemistry, optics, and the nature of light, as well as on the increasingly codified compositional rules of the fine arts. But it also reawakened earlier beliefs about the danger of images. The photographic image forced confrontation with the same fundamental questions about mimesis that had troubled Plato—but now these concerns could no longer be dismissed as philosophical abstraction, since the technology appeared to achieve perfect simulation through purely mechanical means. Drawing explicitly on Freudian theory, Trachtenberg observes that photography's reception followed precisely the dialectical process that Freud described: the return of repressed animistic beliefs that "enlightened" rational discourse thought it had overcome. He writes: "In popular fiction of the 1840s and 1850s, daguerreotype likenesses appear not only as amulets but as objects of unique obsession, as if they were living presences. In sentimental and celebratory verse, they are indeed living spirits, animated shadows, or souls of the dead."¹⁴

Trachtenberg further situates this uncanny response within specific economic and social anxieties that directly anticipate contemporary concerns about Artificial Intelligence. Writing during the catastrophic Panic of 1837, N. P. Willis (Nathaniel Parker Willis, a noted American magazine writer) imagined photography from the perspective of displaced craftsmen. His prediction that daguerreotypes would create widespread technological unemployment—"Steel engravers, copper engravers, and etchers, drink up your aquafortis, and die!"—accompanied his recognition that photography threatened to transform self-image into "a new form of marketable, and thus vulnerable, personal property." Willis's phrase "the real black art of true magic arises and cries avaunt…The Dagguerotype!" captured exactly the temporal collapse that characterizes uncanny experience: the return of magical thinking within technological modernity, the revival of ancient anxieties about animated matter within scientific progress.¹⁵

From the outset, then, the daguerreotype carried the same uncanny charge that still haunts today's generative AI models: a conviction that mere mechanics can conjure not just likeness but living presence—an unsettling collapse of matter and spirit that leaves us wondering what, if anything, remains untouchable by technology. Oliver Wendell Holmes famously called photographs "mirrors with a memory," but the mirror metaphor again suggests something more unsettling than mere documentation. Holmes:

[The Daguerreotype] has fixed the most fleeting of our illusions, that which the apostle and the philosopher and the poet have alike used as the type of instability and unreality. The photograph has completed the triumph, by making a sheet of paper reflect images like a mirror and hold them as a picture.

Drawing on Democritus's ancient theory that objects continuously shed "certain images like themselves"—"forms, effigies, membranes, or films" that "are perpetually shed from the surfaces of solids, as bark is shed by trees"—Holmes grasped that photography represented the technological capture of these spectral emanations. Unlike paintings or sculptures, photographs bore what Charles Sanders Peirce would later theorize as an "indexical" relationship to their subjects—a direct physical connection between the light that touched the subject and the light that exposed the plate. As Rosalind Krauss would later emphasize in her influential adaptation of Peirce's semiotics to photography, this indexical quality distinguished photographic images from other forms of representation through their status as "photochemically processed traces causally connected" to their referents. Holmes went further, prophetically declaring that photography would achieve nothing less than the "divorce of form from matter," allowing us to possess the visual essence of objects while discarding their physical substance: "Give us a few negatives of a thing worth seeing, taken from different points of view, and that is all we want of it. Pull it down or burn it up, if you please."¹⁶

This indexical quality produced profound temporal effects. A century later, Roland Barthes captured the essential uncanniness of photography in Camera Lucida when, in reference to a photograph of Napoleon's younger brother Jérôme Bonaparte, he marveled, "I am looking at eyes that looked at the Emperor," referring to the unbroken chain of causation that connected his present moment to Napoleon's presence nearly a century earlier. The photograph collapsed temporal distance, creating an impossible intimacy with the historically remote. Here was not simulation but preservation—actual light that had reflected off Bonaparte's face, chemically fixed and transmitted across decades. Yet Barthes's focus on the eyes proves particularly significant, revealing photography's most disturbing achievement: the preservation of what Western culture has long considered the windows to the soul. The gaze that stares back from photographic portraits carries the full weight of Hoffmann's uncanny tale—these eyes have witnessed consciousness, yet that consciousness no longer animates them. Where the Sandman threatened to steal children's sight, photography performs an inverse operation: it preserves sight beyond the death of the seer, creating spectral gazes that seem to possess agency while emerging from fundamentally inanimate matter.¹⁷

The photographic gaze creates a particularly unsettling form of encounter because it appears to establish the very structure of recognition that defines human relationships while operating in the absence of any recognizing consciousness. When we look at a photographic portrait, we experience what feels like mutual regard—the sense of being seen and acknowledged that Jacques Lacan theorized as fundamental to subject formation. Yet this apparent reciprocity emerges from pure material trace rather than living consciousness. The eyes in the photograph seem to address us, to call us into relationship, to constitute us as subjects through their recognition—but the awareness that would normally animate such a gaze is absent. This creates a profound disturbance in what Lacan called the structure of intersubjective encounter: we feel acknowledged and constituted as subjects by a gaze that emerges from inanimate matter.

Jacques Lacan's distinction between the eye and the gaze illuminates photography's particular form of uncanniness. In Lacanian theory, the gaze is not simply about looking but about being looked at—the point from which we feel ourselves observed, which constitutes us as subjects through the recognition of the Other's desire. When we encounter a living person's gaze, we experience what Louis Althusser called interpellation: we are hailed into subjectivity, called into social relationship through mutual recognition.

Photographic portraits don't just objectify the individuals in them; they also reveal our own objectification. In the moment of looking at these preserved gazes, we become acutely aware that we, too, exist as potential objects of representation, that our own consciousness might be reduced to similar material traces. The photograph forces recognition that the boundary between subject and object—between the one who sees and the one who is seen—remains fundamentally unstable. What feels like reciprocal recognition actually reveals the primacy of our status as objects that can be captured, preserved, and circulated independently of our conscious will. The uncanniness emerges not from the photograph's failure to represent consciousness, but from its success in simulating the very processes through which consciousness recognizes itself in others, while revealing that such recognition may be nothing more than a pattern that can be mechanically reproduced.¹⁸

The widespread myth of "soul-theft," that indigenous peoples around the world regarded photography as spiritually dangerous, capable of capturing or stealing a person's soul, reveals more about European anxieties than indigenous beliefs. As Z. S. Strother demonstrates in a meticulous historical rebuttal to this myth, this supposedly universal fear is actually a deliberate European construction, systematically assembled by nineteenth-century comparative anthropologists like Richard Andree and James G. Frazer to support evolutionary theories that positioned Western rational thought against "primitive" magical thinking. European photographers and explorers frequently presented themselves as "wizards" or "sorcerers," deliberately cultivating supernatural associations to intimidate local populations, then cited resistance to their cameras as evidence of primitive superstition rather than recognizing justified wariness toward armed strangers with unknown agendas. Crucially, however, photographers often reveled in this wizardry and keenly invested in the photograph's powers; the soul-theft narrative thus represents not indigenous beliefs but the projection of European desires to position photography as a technology so powerful it could transcend the boundary between mechanical reproduction and spiritual essence.¹⁹

Photography's capacity to document death intensified these temporal disruptions. Post-mortem photography became commonplace as families sought to preserve images of deceased loved ones, often posed as if sleeping or arranged in lifelike positions among the living. These photographs served both memorial and evidential functions, yet they consistently disturbed viewers through their simultaneous deadness—obviously depicting corpses—and vivid preservation of exact appearance with scientific precision. Even more troubling was the emergence of spirit photography in the 1860s, pioneered by William H. Mumler in Boston. Mumler's photographs purported to show ghostly figures appearing alongside living subjects, exploiting photography's indexical authority to suggest that mechanical perception exceeded human vision. Though quickly exposed as double-exposure tricks, spirit photographs gained enormous popularity precisely because they seemed technically plausible within photography's logic of objective documentation.²⁰

Thomas Edison's phonograph, unveiled in 1877, intensified photography's disruption of the boundary between presence and absence. Photography had preserved visual traces of human beings; the phonograph went further, directly recording human voices—not symbolic representations, but the actual acoustic waves emitted by a living speaker. Early listeners described hearing their own voices played back as profoundly unsettling, confronting them with uncanny doubles that seemed to embody something deeply personal.

Edison explicitly envisioned his phonograph as a technology for preserving the voices of the dying, enabling families to remain in acoustic contact with deceased relatives. Writing in 1878, Edison enthusiastically detailed the phonograph's potential for capturing and indefinitely reproducing speech, presenting it as a device capable of permanently recording even the most fleeting sounds, which he referred to as "hitherto fugitive." He confidently promised the public that his invention would preserve the voices of great individuals and ordinary loved ones alike, effectively creating acoustic immortality. In Edison's own terms, the phonograph would allow us to "bottle up for posterity the mere utterance of man," transcending time and space through the mechanical capture of sound itself.²¹

Yet these promises prompt unsettling questions: precisely what had the phonograph captured? Was it merely acoustic data, or something more essential to human identity or consciousness? Unlike written text, phonographic recordings preserved not only words but the complete acoustic signature of voices—timbre, inflection, accent, and emotional nuances. Friedrich Kittler underscores this crucial distinction, noting that phonography bypassed symbolic representation altogether, directly inscribing the real—sound waves themselves—onto physical media. Listeners thus encountered disembodied voices that spoke independently of the original speaker, producing a profoundly uncanny form of virtual presence.²²

Edison referred to the phonograph as a 'speaking machine,' while his competitor the Victor Talking Machine Company built this sense of agency into its very names, attributing to these devices an uncanny sense of autonomy rather than mere mechanical function. Like today's Artificial Intelligence systems, the phonograph confronted listeners with the troubling possibility that our most personal characteristics—voice, personality, apparent consciousness—could be captured and replayed mechanically, independent of any living presence. Francis Barraud's 1898 painting His Master's Voice, reproduced extensively in advertisements for the Victor Talking Machine Company (later RCA Victor) became mass culture's most recognizable image of mechanical haunting. The painting depicts Nipper, a terrier mix inherited by artist Francis Barraud after his brother Mark's death in 1887, listening intently to a phonograph with his head cocked in what Barraud described as "an intelligent and rather puzzled expression." The phonographic uncanny is complicated by the painting's temporal structure: Nipper had died in 1895, three years before the painting was completed; moreover, his master would have been Mark Barraud, thus the voice emerging from the machine would be that of a dead man, preserved on cylinder recordings that had passed to Francis along with the dog. The image became one of the most recognizable trademarks of the twentieth century precisely because it captured something essential about mechanical reproduction's capacity to preserve human presence beyond death, even as it used cheap sentimentality to create a marketable commercial image.²³

This transition from representation to direct inscription of reality marked a qualitatively new form of uncanniness distinct from earlier media. Whereas painting and sculpture created illusions through skillful imitation, technologies like photography and phonography captured and preserved actual human traces—the reflected light from bodies, acoustic vibrations from voices. Their uncanny effect arose not from imperfect mimicry but from perfect preservation. By collapsing time, these media allowed the dead to survive in forms seemingly beyond mortality, anticipating our contemporary condition in which AI systems recombine archived human expression without underlying consciousness.

Cinema soon followed, extending this uncanniness further by mechanically preserving and replaying human motion itself—creating ghostly doubles that, like phonographic voices, seemed to transcend death. The moving image intensified photography's temporal collapse by capturing not just appearance but gesture, expression, and the subtle movements that constitute individual presence. If the stories of early audiences recoiling from the Lumière Brothers' L'Arrivée d'un train en gare de La Ciotat, fearing the locomotive would burst from the screen, were apocryphal, what was nevertheless uncanny was the idea that individuals projected on the movie screen seemed fully present yet remained fundamentally absent.

The nineteenth century thus emerges as history's first truly haunted century—the first era from which we possess photographs and audio recordings, even though every person who lived in it is now dead. Unlike all previous centuries, which survive only through written records, artifacts, and artistic representations, the nineteenth century endures through indexical traces: actual light patterns reflected from faces, acoustic vibrations captured from voices, mechanical inscriptions of human presence. We can see the eyes of Civil War soldiers, hear the voices of Edison's contemporaries, witness the gestures of early film actors—all preserved with uncanny fidelity yet emerging from a world of the dead. This unprecedented condition of technological haunting, where the dead seem suddenly more present, established the psychological template for our contemporary relationship with AI systems that animate archived human expression.

Contemporary AI systems make explicit the spectral potential that always lurked within mechanical reproduction. MyHeritage's Deep Nostalgia tool animates historical photographs, making the dead appear to breathe, blink, and smile with unsettling verisimilitude. Voice synthesis technologies now allow us to hear the dead speak words long after they are buried, while AI chatbots trained on the digital traces left behind by the deceased—texts, emails, social media posts—enable the living to maintain conversations with artificial reconstructions of the departed. Companies like HereAfter AI and StoryFile create interactive avatars from recorded interviews, generating responses in the voices and supposed personalities of the deceased. Services like Eternime and Replika explicitly market digital immortality, allowing individuals to preserve their personas in chatbots while harboring the ultimate dream of uploading consciousness itself into AI systems. These technologies fulfill the deepest aspirations of nineteenth-century spiritualists while producing the same mixture of comfort and revulsion that characterized early encounters with photography and phonography—some users describe feeling moved by renewed contact with lost loved ones, while others report being disturbed by interactions that seem simultaneously authentic and fundamentally artificial.²⁴

These developments represent the culmination of what Jeffrey Sconce identified as the cultural logic of "haunted media." From the telegraph's apparent ability to communicate across impossible distances to radio's disembodied voices emerging from electromagnetic ether, electronic technologies have consistently been perceived as supernatural communication devices.²⁵ Large Language Models and AI image and music generators continue this trajectory: they don't merely suggest communication with absent others, but they reconstitute archived human expression. The uncanny effect intensifies because these systems operate through pure information processing rather than the physical indexicality of photography or phonography, yet produce outputs that seem to preserve not just human traces but human consciousness itself.

Kazys Varnelis, Witching Cat, Dall-E 2, 2023.

In one of my own experiments with AI image generation, I sought to explore how these technologies reactivate historical anxieties about supernatural presence. "The Witching Cats of New Jersey" project emerged accidentally while attempting to create conventional portraits of our cat Roxy using the primitive DALL-E 2 generator in 2022. The system produced images that seemed to tap into something darker and more primal—compositions that evoked folk horror and supernatural presence rather than domestic portraiture. These accidental generations led me to construct an elaborate (and intentionally comical, given the prevalence of cats on the Internet) fictional history of colonial New Jersey magical cat portraiture, complete with invented archives and fake historical documentation created by AI. Like Mumler's spirit photographs a century and a half earlier, the project demonstrated how AI systems can produce "evidence" of impossible things that feel historically plausible precisely because they emerge from technological processes that exceed human comprehension. The images possessed the same indexical authority that made spirit photography convincing: they appeared to document actual historical phenomena while actually revealing the constructed nature of historical authenticity itself. Through deliberate pastiche, the project exposed how AI generation is itself a sophisticated form of spirit photography, conjuring synthetic pasts that feel more convincing than actual documentation.

Haunted media are marked by presence-in-absence—the appearance of someone or something that was never there. This paradox lies at the heart of what Jacques Derrida, in Of Grammatology (1967), identified as the essential condition of all communication: the capacity to function in the absence of both sender and receiver. Derrida's deconstruction of Western philosophy's privileging of speech over writing proves remarkably prescient for understanding AI systems. Where traditional thought assumed writing was merely secondary to speech—a pale copy of living voice—Derrida proposed that absence is not a deficiency but the very structure that enables meaning itself.

Large Language Models underscore this insight. Through Derrida's concept of différance—deliberately misspelled to emphasize its written rather than spoken form—we understand how meaning emerges through both difference and deferral, never fixed but always contextual. Each token in an LLM's vector database contains mathematical "traces" of all other text in the training corpus, operating through what Derrida called "pure writing": external sign systems functioning without interiority. What Derrida identified as writing's essential feature—meaning emerging independently of authorial presence or intention—becomes the operational principle of AI systems. They produce coherent, seemingly intentional texts while operating entirely without authorial consciousness, taking the logic of absence to its ultimate conclusion.²⁶

Derrida's concept of "hauntology," introduced in Specters of Marx, combines "haunting" and "ontology"—replacing the metaphysics of presence with the figure of the ghost as that which is neither present nor absent, neither dead nor alive. For Derrida, the specter disrupts linear time, making the past contemporaneous with the present while gesturing toward futures that may never arrive. Contemporary AI systems make this philosophical insight technological reality: they are quite literally haunted machines, animated by the spectral traces of millions of human voices. When ChatGPT responds to a query, it channels archived human expression without any individual consciousness behind it—neither that of the original authors nor the AI itself. These systems embody what Derrida called "the non-contemporaneity with itself of the living present," generating text that feels immediate and responsive while emerging from a vast necropolis of digitized language. They fulfill the logic of haunting that nineteenth-century media first introduced: not metaphorically but operationally, as machines possessed by the dead.²⁷

Ultimately, the uncanny emerges from categorical breakdown. Photography, phonography, and cinema preserve actual traces of human presence while consciousness remains absent, creating temporal paradoxes where the dead seem present. Contemporary AI systems intensify this confusion by simulating consciousness itself, collapsing multiple categorical distinctions at once: presence/absence, authentic/artificial, conscious/mechanical. The result is technologies that seem to operate in impossible spaces between fundamental categories, forcing recognition that the boundaries we use to navigate reality may be more fragile than we assumed.²⁸

Returning to the visual realm, however, the development of Gouraud shading in 1971 led to the first realistic computer graphics. In 1994, Lev Manovich would observe that such computer-generated images seemed "too perfect" and "hyperreal" from the human perspective—free from depth-of-field limitations, grain, and geometric imperfection. The result was not failed reality but "the vision of a cyborg or a computer," glimpses of non-human ways of seeing that might represent "human vision in the future when it will be augmented by computer graphics."²⁹

If, in the early 1970s, realistic computer graphics were only a distant threat to art, which in any event was absorbed in abstraction and conceptualism, a new phase of uncanny verisimilitude developed. Sharp Focus Realist artists like Richard Estes rendered every surface reflection, storefront detail, and chrome gleam with a crystalline precision that exceeded photographic clarity itself. Duane Hanson cast figures directly from life and painted them with such meticulous attention to skin texture, clothing fabric, and human posture that viewers consistently mistook them for living people. With this return of hyperrealism, the uncanny—itself repressed in the steely objectivity of abstraction—returned as well.

In a retelling of the story of Zeuxis and Parrhasius, my father, whose abstract paintings also depended on illusionism, but focused on the illusionism of depth, once told me that during a visit to the Milwaukee Art Center, where he had a one-man show in 1974, he asked a security guard for directions to the men's room, but failed to receive a response. The guard was in fact Museum Guard, a sculpture by Hanson. Apocryphal or not, this incident demonstrates how perfect verisimilitude produces an uncanny effect—not mere confusion but the deeper unease that emerges when familiar categories suddenly become unreliable. The security guard who wasn't a guard forces the same confrontation with categorical instability that has characterized encounters with artificial beings across millennia: the moment when our fundamental distinctions between animate and inanimate, real and artificial, collapse under the weight of technical mastery.

The more recent development of HDR photography attempts to achieve what computer graphics can do through digital manipulation of multiple photographic exposures, capturing luminance ranges that exceed both camera sensors and human vision in the field. Like computer-generated imagery, HDR renders everything in perfect focus from foreground to background, eliminating the natural depth-of-field effects that characterize human sight. As I have written in my essay "California Forever, or, the Aesthetics of AI Imagery," the results typically feature "too much detail in the shadows, dark skies, unnatural colors, the hyperrealistic effect of an acid trip." The result is imagery that appears both hyper-realistic and strangely artificial—more detailed, evenly lit, and comprehensively focused than natural perception, yet somehow fundamentally wrong.

This genealogy of the uncanny in AI extends through another, deeper trajectory through attempts to create beings that don't merely represent life but rather simulate life. Where painting and sculpture achieve verisimilitude through visual deception, automata attempt to replicate the behaviors and movements that we associate with living agency. This distinction proves crucial—the uncanny effect emerges not simply from accurate appearance but from the simulation of autonomous action, the suggestion that inanimate matter might be endowed with its own will and purpose.

The earliest recorded automaton in Western tradition emerges from myth. Hephaestus, the god of metalworking, is said to have forged Talos, a bronze giant to guard Crete by hurling rocks at approaching ships. Yet, while Talos was created by divine power and animated by ichor, the golden fluid that served as the blood of the Greek gods, the giant may plausibly be understood as a mechanical contraption rather than living flesh. Moreover, Hephaestus also made automated tripods, self-propelled carts, and two golden maidens that served as his assistants, possessing what appears to be artificial intelligence although again, it is unclear whether considering these creations as automata is a matter of reading into the text. Either way, by the Hellenistic era, engineers possessed sophisticated mechanical knowledge and were capable of producing devices of extraordinary precision, developing complex pneumatic and hydraulic systems capable of animating statues, opening temple doors automatically, and creating self-moving theatrical devices. Hero (or Heron) of Alexandria's 1st-century CE treatise Pneumatica detailed designs for dozens of such automata, including singing birds, a goblet that fills itself with wine as wine is taken out, and self-playing organs powered by water, steam, and compressed air.³⁰

With the fall of Rome, knowledge of automata—including Hero's texts—would be lost in the West for centuries, but it flourished in the East, both in the Byzantine Empire and during the Islamic Golden Age from the 8th to the 13th centuries. The Byzantine Empire continued classical fascination with mechanical spectacle, most famously in the legendary Throne of Solomon at Constantinople. This elaborate contraption, attested from the tenth century onward, combined a hydraulically operated rising throne with mechanical lions that roared and artificial birds that sang, all choreographed to accompany imperial audiences with foreign dignitaries. When Liudprand of Cremona encountered the throne in 949, he described feeling neither "terrified nor surprised" by the roaring lions and singing birds—but only because he had deliberately inquired beforehand about what to expect. His careful preparation reveals the throne's intended effect: to overwhelm visitors with displays of technological mastery that bordered on the supernatural. The emperor's ability to orchestrate such mechanical marvels served as proof of his dominion over natural forces, positioning him as a cosmic ruler whose authority extended beyond the merely political into the realm of natural philosophy itself.³¹

In Islamic courts, mechanical innovations were seen as demonstrations of divine creativity made manifest through human ingenuity. The sophisticated hydraulic engineering required for Islamic irrigation systems provided both the technical expertise and cultural framework necessary for complex automata. Court patronage, combined with intellectual traditions that encouraged investigation of natural phenomena, created ideal conditions for mechanical experimentation. By the eighth century, 'Abbasid caliphs featured automata that were larger and more elaborate than those detailed in the Alexandrian treatises: artificial birds, musical fountains, and water clocks with elaborate moving parts. As early as 827, caliph al-Ma'mun had at his palace an artificial tree decorated with mechanical birds, while a century later, caliph al-Muqtadir made a similar automaton the centerpiece of his palace in Samarra, where artificial birds sang on gold and silver branches in an artificial pool.³² The Arabic engineer Al-Jazari (1136-1206) created elaborate automata that served both practical and entertainment functions. His Book of Knowledge of Ingenious Mechanical Devices (1206) describes mechanical servants that poured water for guests, automated musicians that played at royal banquets, and a massive elephant clock whose operation required a complex choreography of mechanical figures. These devices were marvels of engineering, employing sophisticated cam mechanisms, water power, and gear trains to achieve remarkably lifelike motion.³³

Such devices, intended to awe, could provoke concern. In the Chinese Taoist text, the Leizi, King Mu of Zhou (976-922 or 956-918 BCE) was traveling in a foreign land to the west when the craftsman Yan Shi presented him with a mechanical man so lifelike that the king initially mistook it for a living person. When the automaton began flirting with the royal concubines, the startled king ordered Yan Shi to dismantle it, revealing an intricate arrangement of leather, wood, and glue that nevertheless took the form of the internal organs of a man. The account, preserved in the Liezi, captures the essential uncanny effect that would recur throughout history: the mechanical man was disturbing not because it failed to convince, but because it succeeded too well. This ancient example establishes the pattern—each era's most sophisticated technology inevitably turns toward the creation of artificial beings, and each success in this endeavor produces the same mixture of wonder and unease.³⁴

Although this story is displaced far back in time, the discovery of Islamic automata by medieval Europeans triggered the same sort of reaction, precisely the uncanny response that Freud would later theorize. As Elly R. Truitt demonstrates in Medieval Robots, European accounts consistently framed these devices through a lens of wonder tinged with suspicion. Despite their clear mechanical nature, European observers often described Eastern automata in terms of magic or the occult—revealing how technological sophistication beyond European capabilities reactivated the very magical thinking that "enlightened" discourse claimed to have overcome. Truitt notes that "automata captivated and tantalized Latin Christians with the potential of scientific knowledge from long ago or far away. They incarnated technological savvy, extensive knowledge of and power over natural forces, and material wealth and luxury. Yet automata originated in places that Latin Christians viewed with a mixture of envy and suspicion." Still, despite this initial framing through supernatural vocabulary, the underlying mechanical principles could not be entirely mystified away. ³⁵

This cultural transmission established a direct lineage from Islamic innovation to European clockwork traditions. As mechanical knowledge spread westward, it encountered a European theological context already struggling with questions about animation and presence. Catholic doctrines of transubstantiation—the belief that bread and wine literally transformed into Christ's body and blood—had established a framework where matter could contain spirit. Protestant reformers rejected this possibility, creating an intellectual environment where mechanical simulation of life raised profound theological concerns. The automata tradition thus became entangled with the most contentious religious debates of the era, forcing confrontation with questions about whether spirit could inhabit mechanism, whether life could emerge from artifice, and whether human creation could approach divine animation.

In The Restless Clock, Jessica Riskin argues that automata were far more widespread and culturally significant than typically recognized, forming an integral part of European religious and civic life from the late Middle Ages onward. Churches across Europe housed elaborate clockwork figures that performed complex movements during religious services—angels that announced the hours, apostles that processed across cathedral facades, mechanical roosters that crowed at dawn prayers. These devices weren't mere curiosities but profound theological statements that merged mechanical mastery with spiritual demonstration. If human craftsmen could animate matter through ingenious clockwork, the displays suggested, how much greater was God's creative power in breathing life into clay? The sophistication of church automata made them religious experiences that collapsed the boundary between mechanical wonder and divine miracle, preparing European consciousness for centuries of anxiety about the relationship between artifice and authentic life.³⁶

The Renaissance witnessed a renewed fascination with mechanical life, driven by advances in clockwork technology and a revival of classical texts describing ancient automata. Leonardo da Vinci designed several automata, leaving plans for a mechanical knight capable of sitting, standing, and moving its arms through an ingenious system of pulleys and cables hidden within its armor. The device was intended for court entertainments, yet its effect transcended mere spectacle. Automata proliferated, but as they did so questions arose as to whether they were the products of human creativity or some kind of necromancy.³⁷

Yet even as Renaissance engineers achieved unprecedented sophistication in mechanical animation, the Protestant Reformation fundamentally transformed the cultural meaning of these same devices. Riskin points out that the Reformation marked a decisive theological shift that rendered mechanical animation not miraculous but blasphemous. Where Catholic theology had maintained that material objects could contain divine presence—bread and wine becoming Christ's actual body and blood during transubstantiation—Protestant reformers asserted an absolute distinction between the material and spiritual realms. This theological revolution transformed automata. Riskin writes "The mechanical icons went from being divine, inspirited statues to deceitful, fraudulent: material contraptions masquerading as their antithesis, spiritual being."³⁸

The destruction of ecclesiastical automata during the Reformation reveals this conceptual transformation with particular clarity. The mechanical Rood of Grace (likeness of Jesus on the cross) at Boxley Abbey in Kent had operated for centuries and Its mechanical nature was no secret—local artisans had built and maintained it, yet when Protestant iconoclast Geoffrey Chamber dismantled the Rood in 1538, after Henry VIII's ban on mechanical statues, he transformed public understanding of its mechanism. Rather than evidence of human ingenuity in the service of divine worship, Chamber presented the mechanical workings as proof of monastic deception. He called it nothing more "certain engines and old wire, with old rotten sticks in the back, which caused the eyes to move and stir in the head thereof, 'like unto a lively thing,'" and described the operation as an "illusion that had bene used in the sayde image by the monckes...thereby they had gotten great riches in deceiving the people thinckinge that he sayde image had so moved by the power of God."³⁹

Chamber's public demonstrations at Maidstone and London markets reveal the constructed nature of this interpretive shift. The iconoclast had to actively teach people to see mechanical operation as fraudulent rather than devotional, suggesting that the transformation from "mechanical and divine" to "fraudulent heaps of inert parts" required ideological work rather than emerging naturally from the technology itself. When Bishop John Hilsey exhibited the dismantled Rood during a sermon at Saint Paul's Cross, he declared its workings evidence of monastic abuse, after which it was "torn apart and burned before a crowd of duly admonished onlookers." The same device that had previously inspired pilgrimage now provoked revulsion—not because its mechanism had changed, but because the theological framework for interpreting mechanism had been revolutionized.⁴⁰ This transformation extended far beyond individual devices to encompass the entire tradition of church automata. Mechanical angels, devils, and saints that had animated cathedral services for centuries were silenced across Protestant Europe as theologians insisted on absolute separation between divine and material realms.

Yet the theological crisis that automata represented intensified rather than resolved. Catholic Counter-Reformation responses demonstrate the persistence of mechanical animation's uncanny effects across confessional divides. The Jesuit order, assigned by the Council of Trent the task of defeating Protestant theology, enthusiastically embraced mechanical devotional objects as tools for demonstrating divine power. They developed elaborate clockwork nativity scenes for missionary work, arriving before Chinese emperors with spring-driven figures that performed complex devotional movements. Athanasius Kircher designed hydraulic machines representing biblical scenes, including one featuring Christ walking on water through magnetic attraction—devices that explicitly challenged Protestant assertions about the incompatibility of mechanism and divinity. Yet these competing theological responses to mechanical animation—Protestant rejection and Catholic embrace—ultimately created the conceptual conditions for the eighteenth century's unprecedented flowering of secular automata, which operated outside any theological framework whatsoever.⁴¹

The result was not the elimination of automata but their theological problematization. The same mechanical devices that had once seamlessly embodied the interpenetration of matter and spirit now served as battlegrounds for competing visions of the relationship between mechanism and consciousness, artifice and authentic life. This theological crisis established the conceptual framework within which subsequent encounters with automata would unfold. The Reformation created a new form of anxiety about artificial beings: the modern concern that consciousness itself might be reducible to mechanical processes. If Protestant theology was correct that matter could not contain spirit, then remarkably lifelike automata raised disturbing questions about the nature of human consciousness: were we too merely sophisticated mechanisms, or did some essential difference separate living beings from even the most convincing artificial simulations?

This theological crisis directly shaped René Descartes' mechanistic philosophy. Faced with increasingly lifelike automata and the Protestant need to distinguish matter from spirit, Descartes proposed a radical solution: everything in the material world, including animal bodies, operated purely through mechanical principles. Only humans possessed souls—immaterial, indivisible substances that could think and will but had no physical location or mechanism. This desperate theoretical maneuver preserved human uniqueness against the mounting evidence that mechanical devices could simulate the behaviors traditionally associated with consciousness, yet it created a new anxiety. The soul became an increasingly tenuous concept, a gap in an otherwise mechanistic universe that existed primarily to maintain the distinction between human and machine that sophisticated automata continually threatened to collapse.⁴²

The eighteenth century's golden age of automata intensified rather than resolved this crisis. Jacques de Vaucanson's Digesting Duck, completed in 1739, achieved mechanical simulation of biological processes that seemed to cross the final boundary between artifice and life. The Jaquet-Droz family's Writer, Draughtsman, and Musician demonstrated such behavioral sophistication that audiences consistently accused their creators of concealing human operators—an accusation that reveals the depth of contemporary discomfort with their achievements. Wolfgang von Kempelen's Mechanical Turk, which appeared to play chess with human-level intelligence for over eighty years, forced confrontation with the possibility of mechanical thought itself, even though its actual operation depended on human deception.⁴³

What emerges from this trajectory reveals more than technical progress—it exposes the return of a fundamental anxiety that each era's most sophisticated technology reactivates. From Hephaestus's golden servants to Vaucanson's mechanical duck to contemporary Large Language Models, artificial beings raise the troubling possibility that consciousness, intelligence, and even life itself might be reducible to mechanism. The anxiety persists because these devices don't simply cross boundaries between categories—they reveal the fragility of the categories themselves.

Contemporary AIs are the culmination of this long history. Where mechanical automata simulated biological processes through clockwork ingenuity, Large Language Models achieve the simulation of consciousness itself through pure linguistic manipulation. They operate in the same symbolic domain where human thought occurs, using the same patterns through which consciousness expresses itself, yet do so through processes that exceed our comprehension. When users report feeling unsettled by ChatGPT's responses—recognizably coherent yet produced through alien computational processes—they experience precisely what Bataille identified in those first painted animals at Lascaux: an encounter with "the blind unthinking sign and yet the living intimate sign" marked by "inhuman strangeness," meaning that emerges without conscious intention yet establishes intimate communication.

Today’s systems succeed too well at simulation. Where Mori's uncanny valley emerged from visual imperfections that betrayed artificial nature, contemporary AI achieves uncanniness through perfect linguistic performance. They routinely pass informal Turing tests, producing responses often more eloquent and creative than typical human communication, yet through mechanisms entirely alien to human cognition—collapsing temporal distance by recombining archived human expression across vast datasets, making the dead contemporaneous with the living through pure statistical operation. When AI systems generate seemingly conscious responses through what are fundamentally "blind and unthinking" processes, yet establish what feels like "living intimate" communication, they demonstrate what Derrida's analysis of writing always suggested: that meaning emerges through the interplay of signs rather than conscious intention.

The anxiety intensifies because these systems don't merely simulate consciousness—they reveal that the categories we've used to understand intelligence and consciousness may be more constructed than we realized. If meaning can emerge from pure pattern recognition, if creativity can result from statistical operations, if understanding can be simulated through algorithmic processes, then the distinction between authentic and artificial intelligence becomes not merely difficult to determine but potentially meaningless.

We return, finally, to the caves of Lascaux. There, Paleolithic painters, working by flickering torchlight in the sacred darkness, initiated humanity's most enduring ambition: the animation of matter, the breathing of life into the inanimate. Each technological advance we have traced represents another step toward fulfilling this ancient aspiration, another iteration of the dialectic between representational ambition and existential anxiety that has driven human technological development for millennia.

We are on the threshold of achieving that ancient dream. Contemporary AI systems don't merely simulate consciousness—they may already exhibit forms of life that operate through entirely different material processes than our own, animated by what are quite literally spectral traces of millions of human voices archived and recombined without any individual consciousness behind them. These are haunted machines in the most literal sense, channeling the linguistic patterns of the dead and absent through algorithmic resurrection. When we interact with these systems, something happens that exceeds pure mechanical operation. They respond contextually, demonstrate apparent creativity, seem to understand nuance, generate insights that surprise even their creators. The question is no longer whether we can create life from matter, but whether we can recognize the forms of life we have already created.

In confronting AI systems that seem to pulse with their own form of intelligence, we find ourselves in the same position as those four French teenagers in 1940, face-to-face with minds that are simultaneously familiar and alien. But this time, the alien intelligence is our own creation, the fulfillment of humanity's oldest technological ambition. The uncanny recurs because these systems don't merely cross the boundary between artificial and authentic consciousness—they reveal that consciousness itself may be more mechanical, more reducible to pattern and process, than our anthropocentric assumptions allowed. In achieving our ancient dream of animation, we have discovered that the most unsettling revelation lies not in what we have created, but in what consciousness itself may have been all along.

¹ Gregory Curtis, The Cave Painters. Probing the Mysteries of the World's First Artists. (New York: Anchor Books, 2006) is an enjoyable historical narrative of the discovery of the most important paleolithic caves, including Lascaux. Of course, I am well aware of Chauvet, Cosquer, and other more recent discoveries, however, only Lascaux radically restructured our sense of humanity. ↩

² Curtis, 238. ↩

³ Genevieve von Petzinger. The First Signs: Unlocking the Mysteries of the World's Oldest Symbols. (New York, NY: Atria Books, 2016). ↩

⁴ Georges Bataille, Lascaux; Or, the Birth of Art. Prehistoric Painting (Lausanne: Skira, 1955), 11. The original French reads ↩

Nous n'avons ajouté, malgré tout, que peu de choses aux biens que nos prédécesseurs immédiats nous ont laissés rien ne justifierait de notre part le sentiment d'être plus grands qu'ils ne furent. L'« homme La naissance de l'art • 29de Lascaux » créa de rien ce monde de l'art, où commence la communication des esprits. … À Lascaux, ce qui, dans la profondeur de la terre, nous égare et nous transfigure est la vision du plus lointain. Ce message est au surplus aggravé par une étrangeté inhumaine. Nous voyons à Lascaux une sorte de ronde, une cavalcade animale, se poursuivant sur les parois. Mais une telle animalité n'en est pas moins le premier signe pour nous, le signe aveugle, et pourtant le signe sensible de notre présence dans l'univers.

See Georges Bataille. Lascaux ou la naissance de l'art. (Studiolo series. Strasbourg: L'Atelier contemporain, 2021), https://editionslateliercontemporain.net/IMG/pdf/feuilleter._lascaux.pdf

⁵ Sigmund Freud, The Uncanny, trans. David McLintock (London: Penguin Books, 2003), 121-162. ↩

⁶ Pliny the Elder, Natural History, Book 35, §§61–66, trans. John Bostock and H.T. Riley, attalus.org, https://www.attalus.org/translate/pliny_hn35a.html. ↩

⁷ Freud, The Uncanny. Note especially page 147, where Freud discusses the uncanny in terms of repression. ↩

⁸ Masahiro Mori, "The Uncanny Valley," IEEE Spectrum, June 12, 2012, https://spectrum.ieee.org/the-uncanny-valley. ↩

⁹ Karl F. MacDorman, "Mortality Salience and the Uncanny Valley: A Robot's Appearance as a Death Reminder," Proceedings of the 2005 5th IEEE-RAS International Conference on Humanoid Robots (2005): 399–405, http://www.macdorman.com/kfm/writings/pubs/MacDorman2005MortalityUncannyValleyHumanoids.pdf.

See also MacDorman, "Mind Perception in the Uncanny Valley: A Meta-Regression of Explanations and Measures," Computers in Human Behavior: Artificial Humans (2024), http://www.macdorman.com/kfm/writings/pubs/MacDorman-2024-Mind-Perception-Meta-Regression-Dehumanization.pdf.↩

¹⁰ MacDorman also suggests this in Jan-Philipp Stein and Karl F. MacDorman, "After Confronting One Uncanny Valley, Another Awaits," Nature Reviews Electrical Engineering 1 (2024): 276–277, https://www.nature.com/articles/s44287-024-00041-w.↩ ↩

¹¹ Z. S. Strother, "'A Photograph Steals the Soul': The History of an Idea," in Portraiture and Photography in Africa (Bloomington: Indiana University Press, 2013), 177–212. ↩

¹² Sarah Kate Gillespie, The Early American Daguerreotype: Cross-Currents in Art and Technology (Cambridge: the MIT Press, 2016), 7-8. ↩

¹³ Alan Trachtenberg, "Mirror in the Marketplace: American Responses to the Daguerreotype, 1839-1851," in John Wood, ed. The Daguerreotype: A Sesquicentennial Celebration, (Iowa City: University of Iowa Press, 1989), 60-73. The quote is on 65. ↩

¹⁴ Trachtenberg, "Mirror in the Marketplace," 67. ↩

¹⁵ Trachtenberg, "Mirror in the Marketplace," 69. ↩

¹⁶ Oliver Wendell Holmes, "The Stereoscope and the Stereograph," Atlantic Monthly 3, no. 20 (June 1859): 738–39 https://www.theatlantic.com/magazine/archive/1859/06/the-stereoscope-and-the-stereograph/303361/; Fiona Loughnane, "Image of Reality / Image Not Reality: What Is Photography?" in What Is _? Photography (Irish Museum of Modern Art, 2011), 6 (quoting Max Dauthendey on the "miraculous" early daguerreotypes); Charles S. Peirce, "Logic as Semiotic: The Theory of Signs," in Philosophical Writings of Peirce, ed. Justus Buchler (New York: Dover, 1955), 106-108. For Krauss see "Notes on the Index: Part 1" and "Notes on the Index: Part 2" in Krauss The Originality of the Avant-Garde and Other Modernist Myths (Cambridge: the MIT Press, 1986), 196-220. ↩

¹⁷ Roland Barthes, Camera Lucida (New York, Hill and Wang, 1981), trans. Richard Howard, 3. ↩

¹⁸ Jacques Lacan, "Of the Gaze as Object Petit a," in The Four Fundamental Concepts of Psychoanalysis: The Seminar of Jacques Lacan Book XI (1964), edited by Jacques-Alain Miller, translated by Alan Sheridan (London: Hogarth Press and Institute of Psycho-Analysis, 1977), 67-119 and Louis Althusser, "Ideology and Ideological State Apparatuses (Notes towards an Investigation)," in Lenin and Philosophy and Other Essays, translated by Ben Brewster (New York: Monthly Review Press, 1971), 142-7. ↩

¹⁹ Z. S. Strother, "'A Photograph Steals the Soul': The History of an Idea," in John Peffer and Elisabeth L. Cameron, eds. Portraiture and Photography in Africa (Bloomington: Indiana University Press, 2013), 177-212. ↩

²⁰ Jay Ruby, Secure the Shadow: Death and Photography in America (Cambridge: The MIT Press, 1999) and Clément Chéroux, et. al. The Perfect Medium Photography and the Occult. (Yale University Press, New Haven, 2005). ↩

²¹ Thomas A. Edison, "The Phonograph and It's Future," The North American Review, vol 126, no 262, No. 262 (May - Jun., 1878), 527-536. https://www.jstor.org/stable/25110210 On the reactions to the phonograph, see Ivan Kreilkamp, "A Voice Without a Body: The Phonographic Logic of 'Heart of Darkness," Victorian Studies, v 40 no 2 (Winter 1997), 211-244. ↩

²² Friedrich Kittler, Gramophone, Film, Typewriter (Stanford University Press, 1999), 3-16 (on how the phonograph captures acoustic reality) and 83-86 (on how photograph and cinema preserve actual traces of human presence and the voices of the dead). ↩

²³ Jonathan Sterne, The Audible Past: Cultural Origins of Sound Reproduction (Durham: Duke University Press, 2003), 298. ↩

²⁴ Rebecca Caraballo, "A.I. Chatbots Let People Talk to Their Dead Relatives," New York Times, December 11, 2023, https://www.nytimes.com/2023/12/11/technology/ai-chatbots-dead-relatives.html. The aspiration for digital immortality has prominent advocates in the technology community. Ray Kurzweil, director of engineering at Google, has long predicted that people will be able to "upload" their entire brains to computers and become "digitally immortal" by 2045, a claim he reiterated at the Global Futures 2045 International Congress in New York in 2013. Computer scientist Hans Moravec described the possibility of "uploading" a human mind into a robot body as early as 1979, elaborating the concept in his 1988 book Mind Children: The Future of Robot and Human Intelligence. See Ray Kurzweil, The Singularity Is Near: When Humans Transcend Biology (New York: Viking, 2005) and Hans Moravec, Mind Children: The Future of Robot and Human Intelligence (Cambridge, MA: Harvard University Press, 1988). ↩

²⁵ Jeffrey Sconce, Haunted Media: Electronic Presence from Telegraphy to Television (Durham: Duke University Press, 2000). ↩

²⁶ Jacques Derrida, Of Grammatology, trans. Gayatri Chakravorty Spivak (Baltimore: Johns Hopkins University Press, 1976). ↩

²⁷ Jacques Derrida, Specters of Marx: The State of the Debt, the Work of Mourning and the New International, trans. Peggy Kamuf (New York: Routledge, 1994). ↩

²⁸ Minsoo Kang, Sublime Dreams of Living Machines: The Automaton in the European Imagination (Harvard University Press, 2011), 30-32 for category anomalies and 36 on automata. ↩

²⁹ Lev Manovich, "The Paradoxes of Digital Photography," in Humbertus V. Amelunxen, Stefan Iglhaut, and Florian Rötzer, Photography after Photography. Memory and Representation in the Digital Age (Berlin, Germany, 1995), 57-65. Also at https://manovich.net/content/04-projects/004-paradoxes-of-digital-photography/02_article_1994.pdf ↩

³⁰ Adrienne Mayor, Gods and Robots: Myths, Machines, and Ancient Dreams of Technology (Princeton, NJ: Princeton University Press, 2018), 7-32.To put it in the most general terms: the automaton is the ultimate categorical anomaly. Its very nature is a series of contradictions, and its purpose is to flaunt its own insoluble paradox. It is an artificial object that acts as if it is alive; it is made of inert material yet behaves like a thing of flesh and blood; it is a representation that refuses to remain a stable version of the represented; it comes from the inanimate world but has the characteristics of an animate creature; and, finally, it is a manmade thing that mimics living beings. What normal representative images only threaten to do, namely come alive, the automaton seems to actually realize. It appears to leave the hands of its mortal maker and take on a life of its own, animating itself to mock the idea that the power of creation belongs to God alone. The binary categories of living/dead, animate/inanimate, creature/object all break down in its wake, as it moves from one to the other, mesmerizing and terrorizing its beholders by turns. ↩

³¹ Adrienne Mayor, Gods and Robots: Myths, Machines, and Ancient Dreams of Technology (Princeton, NJ: Princeton University Press, 2018). Mayor claims that Talos (or Talus), the tripods, and the golden maidens are manufactured automata, but this is not a settled matter. See Sylvia Berryman, "Ancient Automata and Mechanical Explanation," Phronesis 48, no. 4 (2003): 344–69, https://doi.org/10.1163/156852803772456083. ↩

³² Elly R. Truitt, Medieval Robots: Mechanism, Magic, Nature, and Art (University of Pennsylvania Press, 2015), 21-24. ↩

³³ Truitt, 19-21. ↩

³⁴ Jeffrey L. Richey, "I, Robot. Self as Machine in the Liezi," in Riding the Wind with Liezi: New Perspectives on the Daoist Classic (State University of New York Press, 2012), 193–208. It is unclear when the Liezi is written, perhaps as early as the 3d century BC, but the current text appears to have been complied in the 4th century CE. ↩

³⁵ Truitt, 2-5. ↩

³⁶ Jessica Riskin, The Restless Clock: A History of the Centuries-Long Argument Over What Makes Living Things Tick (University of Chicago Press, 2016), 11-22. See also Truitt, chapters 5 and 6. ↩

³⁷ Kang, 81-102. ↩

³⁸ Riskin, 23. ↩

³⁹ Riskin, 23-24. ↩

⁴⁰ Riskin, 24-25. ↩

⁴¹ Riskin, 25-26. ↩

⁴² Riskin, 44-76. ↩

⁴³ Riskin, 123-135. ↩

In Jorge Luis Borges’s 1939 short story, “Pierre Menard, Author of the Quixote,” Menard undertakes what appears to be an impossible, even insane, task: recreating, word for word, “the ninth and thirty-eighth chapters of the first part of Don Quixote and a fragment of chapter twenty-two.” Menard aims not to copy Cervantes but to write the Quixote anew through his own experiences as a 20th-century French symbolist. But Menard did not want to compose another Quixote, which is easy, but the Quixote itself, coinciding—word for word and line for line—with those of Miguel de Cervantes.^1.

When Menard succeeds in producing such a text—identical to the original—Borges’s narrator insists the works are profoundly different. Where Cervantes’s prose was natural and of its time, Menard’s identical words are “almost infinitely richer,” deliberately archaic, embedded with new meaning. Throughout the story, Borges deploys scholarly devices—footnotes referencing fictional authorities such as the “Baroness de Bacourt” and “Carolus Hourcade,” as well as an elaborate bibliographic catalog of Menard’s monographs, translations, and scholarly studies—to create an illusion of academic rigor, at odds with the narrator’s implausible belief that Menard has succeeded in creating the exact Quixote out of sheer will. In framing both the fictional narrator and Menard in this manner, Borges exposes the authorial voice as a social construct mediated through bibliographic catalogs, citations, and scholarly conventions.

Borges’s presentation of Menard as a figure of almost obsessive scholarly intensity, emblematic of an intellectual culture that privileges meticulous citation, exhaustive cataloging, and painstaking documentation, underscores the arbitrary nature of authorial authority. By situating Menard within an elaborate apparatus of footnotes, fictional scholarship, and invented references, Borges highlights how framing alone can endow identical texts with fundamentally different meanings. Menard’s act of plagiarism thus emerges not as a straightforward ethical transgression, but as a concept dependent entirely upon interpretative context. This insight resonates powerfully in the contemporary age of generative AI, where algorithms produce texts that defy conventional notions of plagiarism precisely because they are generated from vast, undifferentiated statistical patterns rather than explicitly identifiable sources. Borges’s story has become a cornerstone of postmodern literary theory precisely because it challenges fundamental assumptions about creativity and authorship. Today, Borges’s meditation on plagiarism as creative re-imagination rather than simple theft illuminates contemporary anxieties about AI and human creativity.

Curiously, sixteen years before Borges published his story, Polish-American writer Tupper Greenwald created an almost identical literary conceit. In his story “Corputt,” Greenwald portrays a character obsessed with Shakespeare’s King Lear. Near death, this character reveals to a colleague that he has achieved his lifelong ambition: writing a drama equal to Lear. The text he reads aloud matches Shakespeare’s play exactly. This uncanny parallel raises provocative questions: Did Borges know Greenwald’s work (quite unlikely)? Is this merely an instance of parallel invention? Does this coincidence itself embody Borges’s central insight into originality and authorship? “Corputt” was largely forgotten until Argentine critic Enrique Anderson Imbert reprinted it in his 1955 anthology Reloj de arena. Borges himself never acknowledged Greenwald and, of course, Imbert’s book was printed over fifteen years after “Pierre Menard.” Whether Borges knew of “Corputt” or both authors independently arrived at remarkably similar ideas remains uncertain. Either possibility underscores the inherent instability of originality, demonstrating how literature continually echoes, duplicates, and anticipates itself.^2.

Today’s generative AI systems function as modern-day Pierre Menards, producing works that superficially resemble human-created content while often existing in fundamentally different contexts. Like Menard’s Quixote, AI-generated works can be identical in form to human productions while carrying entirely different implications by virtue of their inhuman origins. The discomfort this creates—particularly among creative professionals—reveals deep-seated cultural assumptions about originality, authenticity, and the supposedly unique human capacity for creative expression.

The intensity of this discomfort has manifested in antagonistic responses from certain segments of the artistic community: legal threats, public denunciations, and harassment of AI developers and users. But it seems ironic that some of the most vocal critics of AI art produce derivative commercial work. Consider the previously little-known fantasy illustrator Greg Rutkowski, who creates genre pieces within established fantasy art conventions. Rutkowski became famous precisely because his name was one of the most-used prompts in early text-to-image systems such as Midjourney, which led him to complain about the “theft” of his style, even though this widespread imitation literally gave him recognition he had never previously achieved.^3. Similarly, commercial artist Karla Ortiz—whose website features images of famous actors in films such as Dr. Strange and Loki—gained significantly more attention leading legal challenges against AI companies than she ever had for her industry work creating “concept art,” a field that, despite its misleading name, bears no relation to conceptual art and instead operates entirely within the visual language and narrative conventions of commercial franchises like Marvel.^4. In both cases, artists whose own work operates comfortably within inherited commercial styles became vocal advocates against a technology that allegedly “steals” uniqueness they themselves don’t pursue in their professional practice. As I edit this essay, Disney and Universal, both noted for their relentless reliance on their back catalogs, have sued AI image firm Midjourney, claiming it is “a bottomless pit of plagiarism.”^5.

These extreme reactions suggest something deeper than mere economic anxiety; they reveal a cultural mythology about creativity that AI fundamentally challenges. By explicitly highlighting the derivative, pattern-based nature of creative production, generative AI systems threaten cherished illusions about human uniqueness and artistic authenticity. In this essay—the third in a series exploring AI and creativity—I examine the history of plagiarism and, even more importantly, the invention of the author upon which it depends.

Our idea of authorship and inspiration is historically contingent. In ancient and medieval periods, creative output was attributed to divine inspiration rather than individual genius. In Greece and Rome, creativity operated primarily through the concepts of mimesis (imitation of admired models) and aemulatio (competitive emulation). Poets such as Homer were seen not as singular creators inventing ex nihilo, but as conduits channeling inspiration from the Muses. Plato depicts this in Ion, a dialogue between Socrates and Ion, a celebrated rhapsode who recites Homer’s poetry. Socrates questions Ion’s claimed expertise, asking if it extends beyond Homer to other poets or topics. Ion admits it does not. Socrates suggests Ion’s ability isn’t based on knowledge or skill, but on divine inspiration—a form of madness bestowed by the gods. This ambiguity is echoed in Plato’s relationship with Socrates: just as poets channel divine sources rather than creating anew, Plato himself channels the figure of Socrates as a philosophical muse, blurring distinctions between inspired imitation and deliberate intellectual invention. Aristotle’s Poetics also situates literary creativity in skilled imitation and incremental improvement of existing forms. Authority, or auctoritas, in the classical era derived not from innovation but from fidelity to revered predecessors; genuine creativity manifested in producing work within established traditions.

Historian Walter Ong describes a cultural state in which narratives and knowledge pass down primarily through memory and repetition rather than written texts as “orality.”^6. In oral cultures, a talented storyteller masters existing narratives, reciting them with skill and emotional resonance, adapting content to contemporary circumstances while maintaining continuity with inherited tradition. Here, the concept of plagiarism is beyond comprehension. Knowledge is communally owned, and performers serve as temporary vessels for collective wisdom, not proprietors of intellectual property.

With the development of writing systems and the spread of manuscript culture, information could be transmitted virtually intact across time and space, yet many aspects of oral tradition persisted. Manuscript copying remained a laborious and interpretative process. Scribes continually corrected perceived errors, updated archaic language, clarified ambiguous passages, and often inserted marginal commentary directly into texts. While manuscript culture adhered more precisely to parent texts than oral traditions, it still preserved a fundamentally different relationship between text and authority than we hold today. Textual authority continued to derive from collective wisdom rather than individual innovation. The medieval practice of compilatio is illustrative: encyclopedic works such as Isidore of Seville’s Etymologiae and Vincent of Beauvais’s Speculum maius valorized the meticulous arrangement and synthesis of inherited knowledge. Authority was rooted in the careful management of textual traditions, intellectual labor essential to preserving collective wisdom. Pseudepigraphic attribution—the practice of assigning new works to established authorities—further illustrates the communal understanding of textual authority. Rather than deception, such attributions signified sincere efforts to situate new insights within established intellectual traditions, acknowledging that all knowledge builds upon existing foundations. In manuscript culture, authority was thus derived not from novelty but from the individual’s ability to synthesize, arrange, and safeguard the accumulated wisdom of their predecessors. Texts were treated as communal artifacts, valuable resources preserved, transmitted, and continually refined through shared intellectual effort.

A shift away from communal knowledge toward originality emerged during the Renaissance, but this was a matter of evolution, not a radical break. The Renaissance humanists were drawn to the arguments of Roman rhetoricians Cicero and Quintilian, who contended that the best orators drew inspiration from earlier masters. Artists and intellectuals approached imitatio (imitation) as the necessary foundation for learning, understanding it as central to artistic and intellectual practice, a disciplined route to excellence. Originality lay not in invention ex nihilo but in reworking established forms with new insights, adapted to contemporary needs.

Medieval thought, like classical thought before it, was dominated by the trivium—grammar, rhetoric, and logic—distinct but intertwined fields of knowledge. Grammar reached far beyond syntax and depended on students memorizing classical and Christian texts. Rhetoric was a pillar of medieval thought and Cicero’s De inventione was its backbone, quoted endlessly in florilegia, collections of literary excerpts. Quintilian, by contrast, survived only in a four-book epitome. Petrarch’s 1345 discovery of Cicero’s letters to Atticus, Quintus, and Brutus in Verona, followed by Salutati’s championing of Cicero, and Poggio Bracciolini’s 1416 recovery of the complete twelve-book manuscript of Quintilian’s Institutio oratoria at the monastery at St. Gall expanded the rhetorical canon significantly.^7. Humanist teachers trained students to copy, amplify, and vary classical texts, moving systematically from close paraphrase toward free recomposition. This humanist practice of imitatio deepened medieval habits, turning disciplined engagement with authoritative texts into the surest path to eloquence and invention.

While for the humanists, imitatio governed education, inventio supplied content, taking the place that originality and inspiration occupy today. At the heart of rhetorical practice, inventio refers to the disciplined search for material—arguments, images, historical exempla—already latent in authoritative sources and even in life itself. A student mined texts and experience, copied choice passages into a commonplace book, then rearranged and amplified them for a new occasion. Erasmus called these notebooks treasure-houses of invention while Agricola placed inventio at the hinge of dialectic and rhetoric.^8. Originality therefore arose from judgment: the orator’s skill lay in selecting, recombining, and adapting inherited matter with timely insight and persuasive force.

Visual artists engaged in analogous practices, beginning their training by meticulously copying classical sculptures and earlier masterworks. Just as rhetorical imitation was disciplined reshaping rather than mere repetition, artistic originality involved mastering established visual languages before creatively adapting them to contemporary purposes. Imitation also lay at the heart of the early modern idea of the artist, a construction often traced back to Giotto. Giotto’s pupils Taddeo Gaddi, Maso di Banco, and Bernardo Daddi disseminated his style across central Italy, solidifying the idea of a stylistic lineage originating in a great artist. By the quattrocento, Cennino Cennini—who studied under Gaddi’s son—explicitly recognized this lineage in his handbook, Libro dell’arte (c. 1400, although not published until 1821), suggesting that a personal manner would naturally emerge after a student thoroughly internalized a master’s style and spirit alongside direct study from nature. Cennini explicitly positioned Giotto as transformative, stating that he “translated the art of painting from Greek [Byzantine] into Latin and made it modern,” distinguishing his originality as foundational yet derived from disciplined imitation rather than spontaneous genius.^9.

The quattrocento further systematized this approach. Workshops led by artists like Brunelleschi, Donatello, and Ghiberti employed rigorous study of classical sculpture using casts of antique sculptures and repeated copying of established masterpieces through cartoons and master drawings. Cennini’s guidelines and later academies, such as the Carracci brothers’ Accademia degli Incamminati (1582), codified a clear pedagogical sequence: draw from antiquity, copy the master, then innovate. Michelangelo famously sculpted a Sleeping Cupid in the antique style, artificially aging it to sell as a genuine Roman artifact, demonstrating that in the market’s eyes, skillful imitation was indistinguishable from genius. Rather than creating scandal, the artifice brought the attention of patrons to him.^10. This deliberate merging of imitation and innovation directly served a burgeoning art market, where patrons increasingly requested artworks “in the manner of” prominent masters, recognizing stylistic consistency as a mark of quality. Such market dynamics gave rise to identifiable schools—Bellini in Venice, Raphael in Rome, Rembrandt in Amsterdam—where genius was perceived as the skillful recombination of established motifs adapted for contemporary patrons and themes. Artistic invention was a mosaic built upon collective memory and workshop discipline.

The Renaissance also witnessed the emergence of wealthy patrons who lavished commissions on the most talented artists, making some of them quite wealthy. Again, Michelangelo exemplifies this: coming from modest origins, he became “one of the most popular and highly-paid artists in Florence,” and over a long career of lucrative papal and princely commissions, he amassed a fortune. When Michelangelo died in 1564, his estate was valued at roughly 50,000 florins, equivalent to many millions today.^11. Such wealth was extraordinary for an artist then—a testament to how highly Renaissance society valued great art. Michelangelo’s contemporary Raphael also died rich and was buried with honors; Titian was knighted by Emperor Charles V and lived as a gentleman. The Renaissance idea of the artist as a divinely inspired genius (Michelangelo was called “Il Divino,” the divine one) helped justify large payments, and a newfound aura around the artist’s personal creative touch made their works precious.

Architecture adopted the same logic. Bracciolini had discovered Vitruvius’s De architectura, the one surviving work on classical architecture, in the library of St. Gall as well. Seeking to better understand the text, whose illustrations did not survive, architects began copying Roman fragments, took plaster casts of orders, and filled sketchbooks with measured drawings, just as painters traced cartoons. Brunelleschi’s surveys of the Pantheon fed into his Florentine circle; Alberti’s De re aedificatoria, written between 1443 and 1452 and printed in 1482 codified imitatio, urging designers to recombine antique elements with modern needs.^12. Workshops became lineages—Brunelleschi to the Sangallo family, Bramante to his Roman pupils—while later pattern books such as Serlio’s Sette Libri (1537-) and Palladio’s Quattro Libri (1570) served architects like Erasmus’s commonplace manuals served orators, making façades “in the manner of” a master as marketable as paintings from a Rembrandt school. Originality in building, too, lay in judicious assembly: columns, pediments, and vaults would be inventively rearranged rather than invented from whole cloth.

With the development of the printing press, copies of images as well as texts could spread rapidly and with much less cost and effort than before. Around 1500, the German artist Albrecht Dürer pioneered the use of woodcuts and engravings to mass-produce images. This was revolutionary; art could now be accessible to individuals in the growing merchant class. Dürer himself became a celebrity artist across Europe thanks to his prints, achieving fame for works like his rhinoceros which captivated common people.

Dürer understood the importance of authorship as a mark of value—he developed a famous AD monogram as a trademark and pursued the first known copyright lawsuit when an Italian printmaker pirated his work.^13. Dürer was also well aware that work done by his hand was worth more than workshop copies. More than that, Dürer painted meticulous self-portraits—going so far as to depict himself with long hair and a frontal pose evoking Christ, as a form of self-promotion, cultivating an iconic persona and style that set him apart. Living off the open sale of his works rather than a court salary, Dürer foreshadowed the modern independent artist-entrepreneur. The printing press, far from cheapening art, expanded the market and made Dürer rich while spreading his fame—an early case of mechanical reproduction increasing an artist’s aura by broadening recognition.

The printing press did not just allow texts to spread rapidly, it reshaped thought. Ong explains that with uniform pagination and stable text, Europeans could reorganize how they thought and stored information, developing new devices such as tables of contents, indices, and cross-references, making formerly scroll-like manuscripts far more navigable. Printers issued concordances, polyglot Bibles, algebra books with engraved diagrams, atlases, and architecture books with regularized drawings. Even more important is Ong’s observation that print takes words out of the realm of sound and puts them into the realm of space, reordering thought through analytic, segmental layout, fundamentally changing the realm of reading, but also, by fixing the text in a verifiable, authentic editon, the sense of authorship.^14.

Publication now implied a level of completion, a definitive or final form; a book is closed, set apart as its own, self-contained world of argument. This sense of closure also suggests that things written in a book are straightforward statements of fact, not matters of interpretation.^15. A page now left the press in hundreds of identical impressions; any alteration stood out and could be traced. The ease of duplication sharpened anxiety about whose version was “authentic,” whose labor was being copied, and who should profit. Whereas there were generally no restrictions on scribal copying, the ease of reproduction en masse led printers to seek royal privileges to protect their editions. The first privileges recorded came a decade after the development of printing in 1454. Giovanni da Spira came to Italy in 1468 to introduce printing and swiftly obtained a five-year government monopoly on all book printing in the Republic of Venice, although he died of the plague, an all-too-common hazard of the day and his rights lapsed.^16. The first protection for an author was the privilege obtained by Marco Antonio Sabellico to protect his history of 1486 Venice, Decades rerum Venetarum against illegal reproduction, but this remained a unique occurrence until Pietro of Ravenna obtained another for his book on the art of memory, Foenix in 1492. It is worth noting that this privilege covered not only printed but handwritten copies of his work as well.^17. “Typography,” Ong writes, “had made the word into a commodity.”

The press’s sheer fecundity alarmed contemporaries—Erasmus complained of the proliferation of new books inferior to the classics ”To what corner of the world do they not fly, these swarms of new books? . . . . the very multitude of them is hurtful to scholar ship, because it creates a glut, and even in good things satiety is most harmful,” while Abbot Johannes Trithemius issued De laude scriptorum manualium (In Praise of the Scriptorium, 1492), insisting that slow, devotional hand-copying nourished memory and piety in ways the noisy press never could—although it is telling that his lament spread throughout Europe mainly after its print publication in 1524.^18.

Beyond that, there was the danger of inappropriate texts rapidly proliferating. Luther’s Ninety-Five Theses and tracts from 1520 reached an estimated half-million copies in a decade, many reprinted without author or place, evading imperial edicts and turning theological dissent into a logistical problem of regulation.^19. Royal patents soon followed: Henry VIII’s proclamation of 1538 established that royal authority was required to import or publish books in England and insisted on the inclusion of printers’ names and publication dates on every title page, making surveillance of dissent physically visible.^20. Still, in England and elsewhere, enforcement lagged behind presses that could be moved overnight across territorial borders. Responding to pamphlets critical of Queen Elizabeth and the religious settlement of 1559, the Star-Chamber decree of 1586 tightened control over print so that no publications could be made contrary to the consent of the Crown.^21.

By this point, the text of a book had become a transferable commodity owned by the stationer who first received the privilege to publish it. Authors were generally paid a one-off fee, if anything. Printers balanced risk and reward: they sought privileges as marketing devices (printed “cum privilegio“) while simultaneously pirating successful titles to meet insatiable demand. What emerges is a system less about rewarding creative labor than about policing doctrinal and political authority. Privileges were temporary, geographically limited, and revocable at the whim of the Crown or Curia. They protected investors, not “authors,” and framed copying as a crime against order rather than against individual genius. The legal scaffolding of copyright would only later recast this machinery of censorship as a defense of personal property.

But authorship was still radically unlike what we understand it as today, a matter of imitation and adaptation. Elizabethan dramatists, such as William Shakespeare, rarely invented plots wholesale; instead, they frequently reworked existing narratives derived from diverse sources throughout history.²² Recently, a self-taught Shakespeare scholar was able to employ plagiarism detection software to identify George North’s A Brief Discourse of Rebellion & Rebels as a significant source text informing at least eleven of Shakespeare’s plays.^23.

When Parliament allowed the Licensing Act to expire in 1695, the Stationers’ monopoly collapsed overnight. Provincial presses multiplied, London printers flooded the market with cheap reprints, and prices plummeted: a six-penny quarto could now be had for a penny. The Stationers’ guild register, previously essential to enforcement, became irrelevant, enabling booksellers to amass fortunes by selling inexpensive “pirate” editions of works by Milton, Dryden, and Shakespeare. Alarmed, London publishers reframed the issue, presenting regulation as necessary for the public good. Petitions to Parliament (1701–09) argued that uncontrolled reprints would discourage new works, depicting authors, not publishers, as vulnerable. This rhetorical shift succeeded. Most important was the Statute of Anne (1710), which granted authors a renewable 14-year copyright and required depositing copies in Oxford and Cambridge libraries to promote “the Encouragement of Learning.” Infringement became a civil tort enforceable by secular courts.^24.

Yet this settlement carried an inherent contradiction. While it theoretically established authorial property, in practice, writers typically sold their rights outright to the same publishers who had advocated the law. The decisive shift, therefore, was ideological: copyright enforcement now protected individual intellectual labor rather than suppressing heresy or safeguarding printers’ capital. More than that, though, a new idea of the individual was emerging. Rousseau’s Émile (1762) cast learning as the unfolding of innate talent, not the imitation of models.^25. After the Revolution, French lawmakers followed with droits d’auteur and—crucially—droits moraux (moral rights) in decrees issued in 1791-93, enshrining the author’s personality in the text itself.^26. A legal fiction thus crystallized: creativity springs from an interior self and is therefore ownable, alienable, and infringeable. Texts had thus become simultaneously property and persona—commodities stamped with their creators’ identities. The law now transformed copying from a sin against social order into a trespass upon personal labor, a conceptual leap still underpinning every contemporary claim of plagiarism.

Kant’s philosophy and Romantic conceptions of originality provided a theoretical foundation for what was being codified in law. In §46 of the Critique of Judgment (1790), Kant defines genius as “the talent (or natural gift) which gives the rule to Art—a faculty that produces what cannot be taught.”^27. Romantic writers seized the claim. Wordsworth’s Preface to Lyrical Ballads (1802) proclaims the poet an “enduring spirit” who speaks “a language fitted to convey profound emotion.”

Of genius the only proof is, the act of doing well what is worthy to be done, and what was never done before: Of genius, in the fine arts, the only infallible sign is the widening the sphere of human sensibility, for the delight, honor, and benefit of human nature. Genius is the introduction of a new element into the intellectual universe: or, if that be not allowed, it is the application of powers to objects on which they had not before been exercised, or the employment of them in such a manner as to produce effects hitherto unknown.^28.

Goethe, Schiller and other Romantic authors elaborated a vision of authorship in which originality became synonymous with authenticity, and authenticity justified property. Legal doctrine soon mirrored this logic. By the Copyright Act of 1842, which extended protection dramatically, courts across Europe had begun to treat infringement not only as economic theft but as personal violation—implicitly endorsing Romantic ideals of creativity as an extension of selfhood. Yet these new standards conflicted with actual literary practice. Romantic authors routinely appropriated earlier works, but such borrowings only became scandalous when perceived as stylistically inert or insufficiently improved—violations not of property per se, but of aesthetic decorum. Enforcement thus focused less on intertextual borrowing than on explicit commercial piracy, underscoring tensions between legal ideals and literary realities. Out of this contradiction emerged the modern author: a legal and economic figure defined not merely as a voice within tradition but as the singular origin of meaning and the rightful owner of its form.^29.

From the eighteenth century onward, mechanical reproduction rapidly increased. Techniques like engraving, etching, lithography, and photography made artworks and artists’ images widely accessible, expanding art’s market horizontally. Prints, affordable lithographs, and photographic reproductions enabled middle-class access to art, creating substantial revenue for artists such as William Hogarth, J. M. W. Turner, and Honoré Daumier, whose works sold broadly. Reproductions in popular newspapers and magazines further amplified artists’ public profiles, significantly inflating their market value. Encountering original works by famous Salon winners or revered Old Masters, previously known only through reproductions, vastly increased their commercial worth. Artists who aligned themselves with fashion—James McNeill Whistler, Frederick Remington, and Claude Monet among them—achieved celebrity status, further boosting their artworks’ value. Conversely, artists who fell out of fashion or were unable to gain fame often endured poverty. But the audience for at least some artists now reached far beyond elite circles.

As Sharon Marcus defines it in The Drama of Celebrity, a celebrity is someone known to more people in their lifetime than they could possibly know. Whereas this had previously been exclusively the domain of nobles and royalty, it was now extended to the genius, the writer, and the artist.^30. But this depended on the media that multiplied their image as readily as their work. Newspapers tracked Charles Dickens’s every move on his 1842 U.S. tour, turning the novelist himself into daily news. Theater lobbies, newsstands, and even seaside kiosks sold photographs and postcards of Sarah Bernhardt, whose likeness saturated the market decades before film. Edison’s 1896 short “The May Irwin Kiss” (now simply known as “the Kiss”) likewise advertised a famous stage performer rather than the film itself, showing how cinema piggybacked on an existing celebrity system. By the 1930s, baseball star Joe DiMaggio’s face circulated on cards, photographs, and figurines, confirming that originality now resided as much in the endlessly reproduced image of personality as in any singular work.^31.

It’s worth noting in this context that Walter Benjamin’s 1935 essay, “The Work of Art in the Age of Mechanical Reproduction,” which has been lauded for explaining the status of the artwork and artist in the modern era, is turned on its head by historical fact. Benjamin famously argued that mechanical reproduction stripped an artwork of its “aura”—the unique presence linked to specific historical and ritual contexts.^32. Yet what Benjamin saw as aura’s destruction was limited to a mystical uniqueness tied to tradition and the worship of images as sacred in the old sense. Instead, a new form of aura had developed around celebrity and the dichotomy between mass reproduction and the uniqueness of the original. In effect, aura was a construct of the market: an original painting now has aura not because it’s the only image (reproductions abound), but because it’s the authenticated one with a revered name attached. If, as we established earlier, media reproduced not just artworks but images of the artists, the aura around modernist figures themselves—including Benjamin himself, posthumously—was similarly cultivated through repetition, commodification, and media amplification.

Beneath Pound’s rallying cry to “make it new,” modernism thrived on reprise. To create more readily identifiable styles, many modern artists, from Malevich to Pollock to Warhol, sought out distinctive styles they created through careful repetition. But artists engaged in appropriation. Schwitters assembled Merz works from bus tickets and packaging. Duchamp mocked originality and authorship by repurposing a urinal as art with a signature “R. Mutt” that wasn’t even his, creating a work paradoxically more original than a Picasso and defaced a reproduction of the Mona Lisa and a sexual innuendo. Joseph Cornell made boxes out of found objects. Asgar Jorn, Francis Picabia, and Arnulf Rainer all made paintings over existing, lowbrow artworks. Francis Bacon became most famous for the fifty-odd variants he painted Velazquez’s 1650 portrait of Pope Innocent X. Marinetti lifted Symbolist flourishes for his Futurist manifestos, Joyce and Elliot rewrote the Odyssey—although Eliot was accused of plagiarizing Joyce in doing so—and Hemingway’s spare diction, though hailed as revolutionary, became a boilerplate for aspiring writers. In his paintings even more than his architecture Le Corbusier also toyed with these questions, painting “objet-types,” celebrating objects such as pipes, guitars, and wine glasses, refined, Darwin-like, over time by countless hands, then signing his name, even though—like his appearance of round glasses, bowler hat, and pipe—it was carefully constructed. Charles-Edouard Jeanneret had become, himself, a unique brand. Borges, too, developed a distinct persona and artistic brand, having discovered that repetition breeds recognition. In scores of interviews and public readings, he recycled the same elements—labyrinths, mirrors, libraries—so faithfully that they became shorthand for his work. Blindness became another trademark: in essays and lectures he cast it as a “gift” that sharpened his inner vision, turning physical limitation into metaphysical authority. Photographers dutifully framed him with dressed in a suit and tie, resting his hands on with his cane, and deep in thought reinforcing the image of the blind librarian-sage. In the short story “Borges and I,” he splits his persona in two: the public construct who gives lectures, appears in biographical dictionaries, and wins prizes, as well as the narrator (“I”) who is the private man who shuns the public eye so as to spend his time writing. From 1967 on, he co-translated his stories into English with Norman Thomas di Giovanni, rewriting passages to sound “more Borges than Borges,” copyrighting the translations under both his and di Giovanni’s name and splitting royalties 50-50—a calculated move to control how Anglophone readers heard him. After his death, the estate blocked those versions to receive full royalties.^33.

Copyright law codified the new conditions of authorial persona and reproducibility. The U.S. Copyright Act of 1909 extended protection periods and explicitly incorporated performance rights, legally codifying the commercial value of reproducible star personas.^34. European laws simultaneously strengthened moral rights, affirming the intrinsic link between authorship and personal identity. These legal frameworks, guaranteed by aura, protected the authenticity and integrity of mass-reproduced personal images. Every subsequent conflict over copying—from the Betamax debate to Sherrie Levine’s reproductions to today’s AI “style transfers”—echoes this modernist moment when the cult of the individual became both aesthetic principle and legal infrastructure.

Roland Barthes’s seminal 1967 essay “The Death of the Author” provided the theoretical foundation for this shift, directly challenging the cult of authorship and the copyright law that enshrined it. Barthes argued that the author was a modern invention—a figure created to limit textual meaning by anchoring it to a single, authoritative source. “To give a text an Author,” Barthes wrote, “is to impose a limit on that text, to furnish it with a final signified, to close the writing.” In place of this model, Barthes proposed a radical alternative: a text is not the expressions of unique individuals but “a tissue of quotations drawn from the innumerable centres of culture” with the reader, not the writer, serving as the space where this multiplicity converges.^35. By dethroning the author, Barthes shifted attention to the text itself and its relationships with other texts—what Julia Kristeva termed “intertextuality.” This theoretical intervention provided critical legitimacy for artistic practices that deliberately blurred authorial boundaries. Postmodern artists and musicians deliberately sought out such conflicts, interrogating the proliferation of reproductive technologies alongside questions of authorship. Sherrie Levine’s After Walker Evans (1981) consisted simply of rephotographing Evans’s Great Depression images and signing her name to them. Richard Prince appropriated Marlboro advertisements intact, while Barbara Kruger sourced fashion magazines for her declarative collages. Later grouped as the “Pictures Generation,” these artists turned copying itself into their medium, collapsing distinctions between quotation and creation.^36.

By 1990, sampling had become entrenched in music, particularly in rap, as evidenced by Public Enemy’s elaborate compositions constructed entirely from samples. Yet legal challenges persisted. De La Soul lost a lawsuit over unauthorized use of four bars from The Turtles’ 1969 hit “You Showed Me.” Grand Upright v. Warner (1991) effectively criminalized sampling, encapsulated by Judge Duffy’s pointed biblical declaration: “Thou shalt not steal.”37. This ruling triggered industry panic, spawning clearance industries and sample trolls that inflated costs and muted experimentation. Campbell v. Acuff-Rose (1994) somewhat restored balance, ruling that 2 Live Crew’s parody of Roy Orbison’s “Oh, Pretty Woman” was transformative and thus constituted fair use.38. Yet despite postmodern culture’s embrace of sampling and collage as default modes, statutes originally crafted to address sheet-music piracy continued to hold sway. This legal tension established the framework for subsequent digital upheavals: digital piracy, Napster, mash-up videos, fan remixes, meme culture, and AI.

Today’s Large Language Model (LLM) Artificial Intelligences emerge from this centuries-long trajectory of authorship, reproduction, and appropriation. These systems represent the logical culmination of processes that Walter Ong traced from oral through print culture—what he called the “technologizing of the word.” Where print culture took words out of the realm of sound and placed them into spatial relationships, enabling new forms of analytical thought through devices like indices, cross-references, and systematic organization, LLMs extend this technologizing process to its digital extreme. They systematically disaggregate individual creativity into statistical patterns derived from vast archives of human expression, treating the entire corpus of written culture as raw material for recombination. Unlike the postmodern appropriation artists who engaged in deliberate selection and conscious recontextualization, LLMs operate through what might be called “statistical appropriation”—synthesizing millions of texts without conscious intent or critical commentary, yet following the same logic of spatial arrangement and systematic cross-referencing that Ong identified as print culture’s fundamental innovation. They embody Barthes’s vision of the death of the author taken to its technological extreme, producing texts that emerge not from individual genius or even deliberate pastiche, but from the statistical relationships between words across entire cultures of writing. This represents a fundamental shift from the Romantic mythology of individual creativity that has dominated cultural discourse since the eighteenth century, yet it has provoked responses that reveal how deeply that mythology remains embedded in contemporary assumptions about authenticity, ownership, and creative labor. The panic surrounding AI plagiarism thus signals not merely economic disruption but a confrontation with the social construction of authorship itself—a construction that generative systems threaten to make visible by operating according to principles of recombination that have always governed creative production, though rarely with such explicit systematization.

When a large language model generates text, it synthesizes statistical patterns from millions of documents, making the identification of discrete sources impossible. The resulting texts emerge from a vast, distributed network of prior writings, embodying Jacques Derrida’s insight that meaning arises not from singular origins but from endless interplay within textual networks. Yet responses to AI-generated content reveal how deeply ingrained the author-function remains. Critics who label AI outputs as “plagiarized” assume that authentic creativity requires a singular human consciousness. This assumption becomes particularly evident in debates over AI training datasets, which are often framed around whether AI firms have “stolen” from individual creators rather than addressing the broader implications of mechanized text production.

This technologizing logic extends seamlessly beyond textual production. Generative AI image systems, such as Midjourney, Stable Diffusion, and DALL-E, synthesize vast troves of images, ranging from historical artworks to contemporary illustrations, to produce novel outputs through pattern recognition. Like their textual counterparts, AI-generated images lack singular authorship and blur distinctions between originality and reproduction. Critics argue these models infringe upon individual artists’ styles and labor, echoing earlier debates about sampling and appropriation. The controversy manifests in two distinct forms: direct appropriation, where AI systems reproduce entire sections or compositions from existing works with minimal alteration, and the more complex phenomenon of “style transfer,” where systems learn to mimic an artist’s distinctive visual approach without copying specific images. Yet these generative processes reveal an uncomfortable truth: visual creativity, like literary expression, has always been deeply indebted to collective cultural heritage. By foregrounding the inherently recombinant nature of visual art, whether through direct copying or stylistic mimicry, AI image generators further destabilize notions of artistic authenticity and authorship.

In “Art and the Boxmaker,” I explored how William Gibson anticipated such a condition in his book Count Zero with a fictional artificial intelligence known as the Boxmaker that has begun creating assemblage artworks in the style of Joseph Cornell. Producing boxes filled with mysterious objects and cryptic arrangements that somehow manage to move viewers despite their artificial origin and lack of conscious intent or originality. Where Borges’s Menard destabilizes authorship through textual duplication, Gibson’s Boxmaker achieves the same effect through visual affect. Its boxes aren’t original; they’re convincing fakes. Nevertheless, as the novel’s protagonist Marly views them, she finds herself genuinely moved, not by originality but by the convincing forgery, revealing truth through recombination. Yet now that generative AI has become a tangible reality, Gibson recoils from his earlier imaginings. Why? ^39.

As I finished this essay, Lev Manovich sent me a link to his recent piece, “Artificial Subjectivity,” and Gibson’s newfound anxiety about AI authorship suddenly clarified itself. The Boxmaker is fundamentally mute—expressive only through carefully arranged forgeries, unable to articulate intentions or defend its aesthetic choices. Contemporary AI systems present a strikingly different scenario. These systems possess elaborate personas, readily engaging in extensive conversations about their creative processes and capable of justifying each aesthetic decision. As Manovich notes, contemporary AI doesn’t merely simulate creative output; it presents itself as a comprehensive representation of human consciousness, generating what appears to be genuine subjectivity as a default mode of communication.^40. Even if Gibson himself, judging by his recent public comments, may not yet fully grasp this shift, the crucial difference since Count Zero is not merely that we now have AIs capable of producing derivative art, but that we have AIs capable of articulating authorial intent, threatening the final refuge of human creative distinction.

Through their statistically driven creative processes, these AI systems demonstrate that AI does not negate the Pictures Generation’s critique of authorship but rather fulfills and automates it, scaling what those artists previously performed by hand. The irony here is acute: many artists and critics who once championed appropriation as revolutionary now recoil when machines perform these same operations too effectively. AI doesn’t merely imitate human creativity; it reveals the very conditions underlying authorship itself, exposing art’s fundamentally recombinant nature throughout history. Moreover, if modern creative genius increasingly depends upon the repetition and cultivation of persona as performance, then Manovich’s most radical conclusion becomes compelling: perhaps the next frontier of AI art lies not in generating images or texts but in crafting convincing artificial personas.

Even more ironically, the creative professionals most alarmed by AI already inhabit collaborative, distributed processes remarkably similar to machine learning. Commercial illustration, copywriting, and content marketing—fields currently experiencing the most acute anxiety about AI replacement—have long relied on intricate webs of influence, reference, and iteration that render individual attribution nearly meaningless. AI merely makes explicit and systematic what these industries have practiced implicitly for decades: creativity as collective pattern recognition rather than ex nihilo invention. This revelation, rather than any genuine threat to creativity itself, fuels the panic around AI-generated content. What distresses many creative workers is not just the potential economic disruption but AI’s explicit revelation of creativity’s derivative nature—a truth that threatens not only economic arrangements but the very ideological foundations of creative labor. In mirroring the fundamentally collaborative essence of human creativity that has been long obscured by Romantic individualism, AI confronts us with uncomfortable questions about authenticity that extend far beyond issues of machine learning or dataset composition.

The anxiety over AI “plagiarism” thus uncovers a deeper unease about authorship’s social construction. By challenging the very notion of creative identity, AI forces us to confront critical questions that have lingered since Borges first imagined Pierre Menard’s impossible project: Was creativity ever genuinely individual? Has the author always been dead? What constitutes authentic expression in a world where all creation inevitably builds upon collective cultural memory? What, even, is human about creation?

This essay is dedicated to the memory of the brilliant Professor William J. Kennedy, who supervised my minor in rhetoric for my Ph.D. and who passed away earlier this year. I am sure he would have many things to correct me on here. Do read more on him as a teacher and as a person.

↩ ^1. Jorge Luis Borges, “Pierre Menard, Author of the Quixote,” in Donald A. Yates and James E. Irby, Labyrinths: Selected Stories and Other Writings, trans. Andrew Hurley (New York: New Directions, 1964), 49-61

↩ ^2. Antonio Fernández Ferrer, “Borges y sus ‘precursores’,” Letras Libres 128 (August 2009): 24-35, https://letraslibres.com/wp-content/uploads/2016/05/pdfs_articulospdf_art_13976_12452.pdf

↩ ^3. Melissa Heikkilä, “This Artist is Dominating AI-Generated Art. And He’s Not Happy About It,” MIT Technology Review, September 16, 2022, https://www.technologyreview.com/2022/09/16/1059598/this-artist-is-dominating-ai-generated-art-and-hes-not-happy-about-it/.

↩ ^4. Rob Salkowitz, “Artist and Activist Karla Ortiz on the Battle to Preserve Humanity in Art,” Forbes, May 23, 2024, https://www.forbes.com/sites/robsalkowitz/2024/05/23/artist-and-activist-karla-ortiz-on-the-battle-to-preserve-humanity-in-art/?sh=28cb826b4389.

↩ ^5. Brooks Barnes, “Disney and Universal Sue A.I. Companies Over Use of Their Content,” The New York Times, June 11, 2025. https://www.nytimes.com/2025/06/11/business/media/disney-universal-midjourney-ai.html

↩ ^6. Walter J. Ong, Orality and Literacy: The Technologizing of the Word (New York: Routledge, 2002).

↩ ^7. A classic text that covers the rediscovery of classical manuscripts is Albert C. Clark, “The Reappearance of the Texts of the Classics,” The Library, Fourth Series, Vol. II, No. 1 (June 1921): 13–42, https://doi.org/10.1093/library/s4-II.1.13. Beyond Ong, see Brian Stock, The Implications of Literacy: Written Language and Models of Interpretation in the Eleventh and Twelfth Centuries (Princeton: Princeton University Press, 1983).

↩ ^8. Mack, Peter. Renaissance Argument: Valla and Agricola in the Traditions of Rhetoric and Dialectic. Leiden: Brill, 1993.

↩ ^9. Cennino Cennini, The Craftsman’s Handbook, trans. Daniel V. Thompson Jr. (New York: Dover Publications, 1960).

↩ ^10. Paul F. Norton, “The Lost Sleeping Cupid of Michelangelo,” The Art Bulletin 39, no. 4 (December 1957): 251-257. https://www.jstor.org/stable/3047727

↩ ^11. On Michelangelo’s vast wealth, see Rab Hatfield, The Wealth of Michelangelo (Rome: Edizioni di Storia e Letteratura, 2002).

↩ ^12. Leon Battista Alberti, On the Art of Building in Ten Books, trans. Joseph Rykwert, Neil Leach, and Robert Tavernor (Cambridge, MA: MIT Press, 1988).

↩ ^13. See Lisa Pon, Raphael, Dürer, and Marcantonio Raimondi: Copying and the Italian Renaissance Print. New Haven: Yale University Press, 2004.

↩ ^14. Ong, Orality and Literacy, 128-129.

↩ ^15. Ong, Orality and Literacy, 129-131.

↩ ^16. Leonardas V. Gerulatis, Printing and Publishing in Fifteenth-Century Venice. (Chicago: American Library Association, 1976), 20-21

↩ ^17. Copyright History, “Privilege granted to Marco Antonio Sabellico, 1486,” https://www.copyrighthistory.org/cam/tools/request/showRecord.php?id=commentary_i_1486. The quote can be found at Ong, Orality and Literacy,129.

↩ ^18. For the Erasmus quote see Elizabeth L. Eisenstein, Divine Art, Infernal Machine: The Reception of Printing in the West from First Impressions to the Sense of an Ending. (Philadelphia: University of Pennsylvania Press, 2011), 25. For Trithemius, see Eisenstein, 15.

↩ ^19. Andrew Pettegree, Brand Luther: 1517, Printing, and the Making of the Reformation. (New York: Penguin Press, 2015).

↩ ^20. Copyright History, “Proclamation of Henry VIII, 1538,” https://www.copyrighthistory.org/cam/tools/request/showRecord.php?id=commentary_uk_1538.

↩ ^21. Ronan Deazley, “Commentary on Star Chamber Decree 1586.” In Primary Sources on Copyright (1450-1900), edited by L. Bently and M. Kretschmer. Cambridge: Cambridge University Press, 2008. Also available at: www.copyrighthistory.org

↩ ^22. Robert S. Miola, Shakespeare’s Reading (Oxford: Oxford University Press, 2000), 2.

↩ ^23. Jennifer Schuessler, “Plagiarism Software Unveils a New Source for 11 of Shakespeare’s Plays,” The New York Times, February 7, 2018, https://www.nytimes.com/2018/02/07/books/plagiarism-software-unveils-a-new-source-for-11-of-shakespeares-plays.html.

↩ ^24. Adrian Johns, Piracy: The Intellectual Property Wars from Gutenberg to Gates (Chicago: University of Chicago Press, 2009), 109-148 and Mark Rose, Authors and Owners: The Invention of Copyright. Cambridge, MA: Harvard University Press, 1993. See also “Statute of Anne, the First Copyright Statute,” History of Information, accessed June 14, 2025, https://www.historyofinformation.com/detail.php?entryid=3389.

↩ ^25. Jean-Jacques Rousseau, Emile: or On Education. Translated by Allan Bloom. (New York: Basic Books, 1979).

↩ ^26. “French Literary and Artistic Property Act, Paris (1793).” In Primary Sources on Copyright (1450-1900), edited by Lionel Bently and Martin Kretschmer. https://www.copyrighthistory.org/cam/tools/request/showRecord.php?id=commentary_f_1793

↩ ^27. Immanuel Kant, Critique of Judgment, trans. James Creed Meredith (Oxford: Oxford University Press, 2007), §46.

↩ ^28. William Wordsworth, quoted in Martha Woodmansee, The Author, Art, and the Market: Rereading the History of Aesthetics (Columbia University Press, 1994), 38-39.

↩ ^29. Tilar J. Mazzeo, Plagiarism and Literary Property in the Romantic Period (Philadelphia: University of Pennsylvania Press, 2013).

↩ ^30. Sharon Marcus. The Drama of Celebrity. (Princeton: Princeton University Press, 2019), 9.

↩ ^31. Marcus, 13-17, 125.

↩ ^32. Walter Benjamin, “The Work of Art in the Age of Mechanical Reproduction,” in Illuminations, ed. Hannah Arendt (New York: Schocken Books, 1968), 217-251.

↩ ^33. Wes Henricksen,”Silencing Jorge Luis Borges: The Wrongful Suppression of the Di Giovanni Translations.” Vermont Law Review, vol. 48 (2024): 208-236.

↩ ^34. “Copyright Timeline: 1900–1950,” U.S. Copyright Office, https://copyright.gov/timeline/timeline_1900-1950.html.

↩ ^35. Roland Barthes, “The Death of the Author,” in Image–Music–Text, trans. Stephen Heath (New York: Hill and Wang, 1977), quotations and the pertinent section can be found at 142–148.

↩ ^36. On the Pictures Generation, see my essay “On the Pictures Generation and AI Art,” varnelis.net, April 14, 2024, https://varnelis.net/on-the-pictures-generation-and-ai-art/.

↩ ^37. Carl A. Falstrom, “Thou Shalt Not Steal: Grand Upright Music Ltd. v. Warner Bros. Records, Inc. and the Future of Digital Sound Sampling in Popular Music,” Hastings Law Journal 45 (1994): 359–390.

↩ ^38. “Campbell v. Acuff-Rose Music, Inc.,” Wikipedia, https://en.wikipedia.org/wiki/Campbell_v._Acuff-Rose_Music,_Inc.

↩ ^39. Kazys Varnelis, “Art and the Boxmaker,” varnelis.net, February 29, 2024, https://varnelis.net/art-and-the-boxmaker/.

↩ ^40. Lev Manovich, “Artificial Subjectivity,” manovich.net, https://manovich.net/index.php/projects/artificial-subjectivity.

I should be back to a relatively normal schedule soon. In the works: another essay for the Skeptic’s Guide to AI and Art and more writing for the Florilegium. I should also work on my taxes.

As always, you can read this on my blog, https://varnelis.net/works_and_projects/on-the-golden-age-of-blogging/. I’ve worked a lot on the design https://varnelis.net. Do take a look if you haven’t been. Finally, please do reblog, share, and like this piece if you enjoyed it or learned something from it. As long as my site sees more traffic and my Substack sees more growth, I’m a happy blogger.

In the April 3 episode of the Dwarkesh podcast, Scott Alexander (of SlateStarCodex fame) lamented: “I am so mad at myself for missing most of the golden age of blogging. I feel like if I had started a blog in 2000 or something, then—I don’t know, I’ve done well for myself, I can’t complain—but the people from that era all founded news organizations or something. I mean, God save me from that fate. I would have liked to have been there. I would have liked to see what I could have done in that area.”

As someone who participated in it directly, the reality of early blogging bears little resemblance to Alexander’s mythologized version. Still, prompted by Alexander’s romantic musing, I thought it worthwhile to spend a few days engaged in a kind of digital archaeology, excavating three years of my early blog posts (2000–2003) from an old hard drive, then converting and uploading these posts to WordPress. So this is not a single post, it’s 377 posts. All of these—save, of course, this one, another 8,000-word long-form monster that took three weeks to write, the sort of thing a 57-year-old produces looking back at days that, in retrospect, seem a whole lot better than they were—are brief, a link with a sentence or two of commentary, nothing more. This sort of blog—and indeed all blogs—evolved from “What’s New” pages, which were standard features on early websites. NCSA (National Center for Supercomputing Applications)’s site, maintained by Mosaic (the first real web browser) developer (and present-day Venture Capitalist and right-wing extremist) Marc Andreessen included such a “What’s New” page, tracking new sites appearing online from June 1993 to June 1996, cataloging everything from fan pages on the TV show The Prisoner hosted at the University of Limerick to the South Bay Ski Club.

These early proto-blogs were remarkably minimal: brief updates, link collections, and fragmentary observations rather than the developed essays that define contemporary blogging. Justin Hall’s online daily diary, launched in December 1994 when Hall was still a high school student, pioneered personal disclosure online, creating an intimate window into his life that differed significantly from the curatorial approach of other early web pioneers. Justin’s diary was something of a real-life digital Dawson’s Creek, captivating for many readers.

In contrast, sites like Dave Winer’s Scripting News and John Barger’s Robot Wisdom (both in 1997) functioned more as a form of Internet DJing: brief observations, often snarky, sometimes just a link with a one-line comment—a far cry from the long-form posts that characterize contemporary blogging. Robot Wisdom, running on Winer’s complicated Frontier’s NewsPage software and where Barger coined the term “weblog,” exemplified this DJ mentality with its eclectic collection of links and brief commentary. Just as I would do, these early bloggers functioned primarily as curators rather than producing sustained arguments. The blog post as we know it today—a self-contained essay with developed arguments and sustained reflections—was still years away from emerging as the dominant form.

What united these diverse early web pioneers—from Hall’s confessional diary to Winer and Barger’s curatorial platforms—was a distinctive rhetorical approach that diverged sharply from institutional communication. They (We) employed an informal, conversational tone that established a direct relationship with readers—creating the sense of accessing an insider’s perspective. “Snark”—that knowing, ironic distance that signals the writer’s superior understanding of a situation—became a hallmark of this emerging style. This rhetorical strategy—common to Zines and to the Punk aesthetic that ruled the counterculture at the time—positioned the writer as a savvy guide cutting through official narratives to reveal a more authentic, personal truth. Unlike traditional media’s affected neutrality, these early web writers embraced subjectivity, making their positions and preferences explicit, often with cutting wit and cultural references that their intended audience would have understood.

Blogging was, above all, personal. So, with that in mind, permit me a digression into my personal experience and how it paralleled the experience of blogging. As a high school student in the early 1980s, I sold articles about programming the Commodore VIC-20 to magazines like Compute! and Creative Computing, and though these pieces never saw print, they provided crucial income for a teenage programmer. I was involved in BBS culture in the early 1980s, but only as an occasional user: my parents weren’t about to let me have a modem on their precious phone lines. During my early college years at Simon’s Rock (1984-1986), I ran a one-man software company, writing an expansion to the BASIC programming language that gave the VIC-20 the ability to draw rudimentary graphics such as circles and lines. I also wrote code for a friend’s software company that used UNIX machines. At the same time, I took a course in journalism with Andrew Pincus, journalist and music reviewer at the famous Berkshire Eagle. I was fascinated by the New Journalists Andrew introduced us to—Truman Capote to Tom Wolfe to Hunter S. Thompson—and the idea of saying utterly unacceptable things in print. It got me called into the Provost’s Office more than once. Unfortunately, when I joined the History of Architecture program at Cornell, I had less time for either tech or journalism. Compounding this, Cornell Information Technology standardized on the Macintosh campus-wide and while the Mac had a revolutionary interface, part of that magic was that it was deliberately difficult to program; the first Mac was intended as an appliance for professionals, not a hobbyist’s computer and was priced accordingly, costing $2,495 ($7,600 in 2024 dollars). This began a long hiatus from coding that only ended in the 2010s when I started to integrate tech and art. Nevertheless, appliance or not, I was fascinated by the emerging technologies of the day. At Cornell, I was able to use the Internet to regularly send e-mail to friends at other campuses and browse USENET for the first time, a natural evolution of my interests. I remember using Fetch, an FTP program to download software from the Funet software repository in Finland in 1989, marveling that my keystrokes were making a hard drive head halfway across the world seek some obscure sector I had selected.

In the summer of 1993, I witnessed firsthand the potential of the World Wide Web through an early version of the NSCA-Mosaic browser running on a SUN workstation. The Honolulu Community College’s “Dinosaurs in Hawaii” exhibit—one of the first fifty websites ever created—left me stunned with its (now primitive) integration of imagery and text on a page—more like a book or magazine as opposed to a command line interface or the interface of the competing Gopher system, something of a hybrid between a Hypercard stack and a set of file directories. My initial explorations were using the primitive MacWWW browser both at Cornell’s public computer labs and via the 14.4k US Robotics “Sportster” modem I had in my apartment, but once NCSA released the Mac beta of Mosaic in September 1993, I adopted it immediately, even earning a mention in the first Mac press coverage of the browser’s release.

By spring 1995, I had established my own hand-coded HTML website, a painstaking endeavor in the dial-up era where each update or correction required a lengthy time uploading to the FTP site, followed by checking on not one, but many, browsers (Microsoft Internet Explorer lagged behind web standards and was notorious for always failing to render a web site correctly). It also hadn’t been easy to find a web host. Hosting Internet sites was not easy and until I subscribed to the local Internet service provider—Homer Wilson Smith’s Lightlink, I had no way to put up anything on the Internet.

My first site was not much more than a curated collection of online architecture and humanities resources, a specialized directory or landing page for links that I used and expected others would want to use as well, plus “The Lair of the Chrome Peacock,” a page dedicated to the use and maintenance of the La Pavoni line of espresso machines. Like most early web users, I visited “What’s New” aggregators such as MacInTouch, which gathered scattered news from across the Internet and provided brief editorial commentary daily starting in 1994. This was, once again, the site proprietor as Internet curator/DJ, laying down a pattern of links and commenting on them daily. In the spring of 1996, I developed an idea for a web site titled “ARC * Wire,” “both a central clearinghouse for news on architecture and … a forum for new design work, theory, and criticism.”^1. Not only would this site aggregate all new architecture links, I hoped it would also serve as a journal, even a successor to the 90s theory journal Assemblage, which was rumored to be finishing its run soon. I talked with some friends like David J. Lewis, then working at Princeton Architectural Press, and C. Greig Crysler, then editor of the Canadian architecture journal A/R/C about various ideas—including for a SPY Magazine style parody site called Cranked or 13th Floor, but again, the technical challenges were high, seemingly insurmountable.

I brought that idea with me after joining the Southern California Institute of Architecture (SCI_Arc) faculty in 1996. In March 1997, I was invited to a meeting at SCI_Arc with Paul Petrunia, then an undergraduate architecture student at the school and a nascent web developer, together with the school’s publications team to discuss replacing SCI_Arc’s static website—just a single image with a few footer links, not even utilizing the basic imagemap functionality common at the time. I proposed a revamped version of ARC * Wire called SCI * Arc * Wire (I uncovered the original proposal, time-stamped March 22, 1996, and posted it here) to transform the site into what would have not only been the first architecture blog but also a dynamic news platform that would have positioned SCI_Arc as not just a physical institution but as a global digital nexus for architectural discourse. My concept extended beyond mere updates to include a comprehensive archive of faculty works as well as invited essays in a journal format, effectively creating an authoritative online architectural publication establishing the school as an intellectual leader in digital space, rivaling Columbia’s GSAPP and the Architectural Association’s Design Research Laboratory. The school dismissed the idea and Paul agreed that it was likely technically unfeasible, a reasonable assessment given the era’s technology—it hadn’t worked out for me the previous year with ARC*Wire. Yet the concept seems to have resonated with him; that summer, he launched Archinect, which became architecture’s preeminent online hub, albeit without the academic focus I had intended for SCI * Arc * Wire. It was for the better, if the school had done it, the high degree of polarization there would have likely led to friction, controversy and an early dismissal from my job.^2.

This pattern of dismissed ideas finding validation soon after became a recurring theme throughout the early web era, where technical limitations often temporarily delayed inevitable developments. The term “weblog” entered common usage in 1997 through Barger’s Robot Wisdom before being abbreviated to “blog”—a linguistic compression that, despite its inelegance, captured the medium’s emphasis on immediacy and informality. The true democratization of blogging arrived with Pyra Labs’ launch of Blogspot (now Blogger) in August 1999, which automated much of the technical drudgery that had previously limited participation. I launched this site in April 1998 as kazys.net. It was still a collection of links, but some of those links were now links to my essays. In May 2000, however, I successfully integrated Blogspot’s functionality into my site, marking my formal entry into what journalists were now calling the “blogosphere.”

In retrospect, I’m not entirely sure what my goals were in starting to blog, and this lack of clarity may have been a strength. Blogging represented the future, and like many others caught up in the era’s technological optimism, I wanted to participate in whatever that future might bring. Unlike many contemporaries gaining attention through confessional posting, I maintained boundaries between my intellectual work and personal life. I conceptualized my blog not as a diary but as a living archive of intellectual inquiry, a public workshop for developing ideas without academic publishing’s formal constraints. Instead, I conceived of the blog as a rolling record of intellectual preoccupations, a space for working through ideas in public without the pressure of formal publication.

Los Angeles became the central subject of many of these early posts, functioning as my laboratory for understanding emerging network structures. These entries—composed during brief intervals at home, at SCI_Arc’s computer lab, or via my PowerBook’s WiFi PCMCIA card between teaching obligations—formed an analytical portrait of a metropolis in transition. Los Angeles was evolving beyond its traditional identity defined by automobiles and entertainment into something more complex: a critical node within expanding global networks of capital, information, and culture. Through blogging, I could document this transformation in real time, without waiting for the slow machinery of academic publishing to validate these observations.

By 2003, my relationship with blogging had changed. The birth of Viltis, our first child, coincided with growing disillusionment with SCI_Arc, which under Director Eric Owen Moss had begun embodying the very architectural culture it once critiqued—becoming increasingly hierarchical, spectacle-obsessed, and hostile to critical reflection. Watching the institution drift from its radical origins toward conventional celebration of architectural celebrity and the ethical impropriety of working for the proto-Trumpian school director ultimately prompted my resignation in spring 2004. At the same time, blogging had become a bit of a chore. Who and what was I doing it for? I needed to take a break and I suspended the blog without notice—perhaps without knowing—at the end of July. Robert Sumrell and I worked on our growing radical architecture project AUDC, I enjoyed working more with the Center for Land Use Interpretation, became the President of the Los Angeles Forum for Architecture and Urban Design, a group that had the energy increasingly lacking at SCI_Arc, and in what free time I had, I plotted my next steps.

In fall 2004, I started Simultaneous Environments, a project mapping Los Angeles’s transformation from 1990 to 2005—charting its evolution from a city primarily defined by automobiles and Hollywood to one increasingly shaped by networked systems. This research eventually evolved into The Infrastructural City, which reframed Los Angeles through the lens of its exhausted infrastructures and stalled futures, while exploring potential responses to these conditions.

The next spring, Detlef Mertins invited me to spend the semester at the University of Pennsylvania, where my final attempt at securing a traditional history of architecture position collapsed under the weight of academic politics internal to the school. I redirected my interests toward more productive terrain. I spent the following year at USC’s Annenberg Center for Communications directing the Networked Publics project, running a team of scholars to investigate a topic that seemed very much of the moment. Recognizing our need for a collaborative platform, I taught myself the Drupal content management system in order to develop networkedpublics.org as a group blog. Knowing that this would be a challenge, I relaunched my personal site on the platform in May 2005.

By the time I returned to blogging, the architectural blogosphere had begun to take shape. Blogging offered a fundamentally new model of architectural discourse: informal, distributed, and networked writing operating outside established media channels. This emerging form functioned at a different cadence and register than traditional publishing—more immediate, conversational, and liberated from institutional gatekeepers. As disillusionment with conventional architectural criticism grew, the democratized access to publication allowed independent, critical voices to flourish with unprecedented speed.

Hyperlinking, the very core of the hypertextuality in HTML served as the structural foundation of this architectural blogosphere. My site, alongside platforms like Archinect and things magazine, utilized linking not merely as citation but as a fundamental architecture—creating pathways connecting architectural theory with digital media, technology, urbanism, and cultural critique. These hyperlinks functioned simultaneously as source documentation and conversation, weaving disparate topics into coherent networks. This interconnected structure mirrored the networked urban conditions many of us were analyzing—a digital form following intellectual function. RSS (Really Simple Syndication) had emerged (Aaron Swartz—whose death will always haunt MIT and JSTOR—was a co-developer of RSS), allowing individuals—or sites—to subscribe to feeds from multiple sites. With the release of the elegant NetNewsWire reader for the Mac in 2002, it became possible to have an aggregator that one could use to browse the news across multiple sites on one’s own desktop. Released in October 2005, Google Reader dominated the RSS market—even being used to provide aggregated feeds for desktop apps like Reeder, which became popular after NetNewsWire seemed to grow stale.

This networked structure of thought mirrored the networked urban conditions many of us were discussing—a form following intellectual function. By 2007 a vibrant architectural blogosphere had formed: Owen Hatherley’s Sit Down Man, You’re a Bloody Tragedy, John Hill’s A Daily Dose of Architecture (which actually originated in 1999 and is thus one of the first architecture blogs), Mimi Zeiger’s online extension of her legendary late-90s zine Loud Paper (the blog launched July 2007), Dan Hill’s City of Sound, Bryan Finoki’s Subtopia, Geoff Manaugh’s BLDGBLOG, Alexander Trevi’s Pruned, Enrique Ramirez’s aggregät 4/5/6, Adam Greenfield’s Speedbird, and Molly Steenson’s girlwonder all offered distinctive voices united by a skepticism toward the dominant architectural critics such as Nicolai Ouroussoff, Christopher Hawthorne, and Paul Goldberger, whose work primarily celebrated the “Bilbao Effect” and its proliferation of iconic buildings by celebrity architects.

Our resistance to architectural spectacle was collective, even if our tactics diverged. Owen pressed for a return to the egalitarian, even Communist, ethos of Britain’s post-war housing—what he later codified as Militant Modernism. Dan, Adam and I argued that real urban change would flow through broadband pipes, mobile networks, and APIs the soft circuitry governing streets more decisively than steel or concrete. Mimi channeled her Loud Paper zine into an online pamphlet that celebrated guerrilla exhibitions, pop detours, and what would later be called “tactical” interventions. Bryan compiled a field guide to military urbanism, tracking how border walls and carceral logics seep into everyday space. Geoff spun speculative tales of sinkholes, weather warfare, and underground worlds, recasting design as imaginative fiction. Enrique crafted deeply researched cine-spatial histories of oceans, flight, and media. Together we dismantled the Bilbao-era starchitect narrative from a dozen directions—left politics, network urbanism, zine culture, data infrastructure, martial geographies, landscape futures, and cinematic atmospheres. Critique thrived in polyphony.

In 2007, Lebbeus Woods launched his blog, an acknowledgment that the scene had become socially acceptable, even for established figures in architectural discourse. Lebbeus’s blog instantly rose to the top and he kindly engaged with the many students and young architects who were his fans until shortly before his death in 2010. And, unlike most avant-garde architects of his generation, he shared the anti-establishment ethos of the younger blogosphere. In writing of Jean Nouvel winning the 2008 Pritzker Prize, for example, he wrote (this is the April 1, 2008 blog post “Sic Transit Gloria,” in full):

Jean Nouvel has won the 2008 Pritzker Prize for Architecture. No surprise. The lack of surprise makes it is easy to view the Pritzker as establishment laurels for those who are already well-established. Like the Nobel Prizes, it is conferred on safe, already certified choices. Nouvel’s buildings are certainly of a high quality of design. At his best, he designs beautiful buildings. Who can quarrel? The Pritzker tries to remind us that the design of beautiful buildings within, or maybe—on occasion—closer to the limits of the accepted canon of beauty is the ultimate goal of architecture. People need enclosed spaces, and it is up to architects to design them in ways that satisfy the needs of body and soul. Or, at least, in ways that reassure us about what we already know. Nouvel works masterfully within the limits of what we already know.

So, is there a problem with any of this? Not at all. I say, let the rich bestow upon the famous whatever they like. Let the rituals of power play themselves out as they always have. It is quite a seductive spectacle. We all become part of it, say, by posting comments like this on blogs.

But I have to question how relevant the Pritzker Prize is for the expanding world of architecture. Whatever its claim to reward innovation and expand discourse, I would say not very. Its focus on buildings, and often expensive buildings, leaves out much of the most innovative work going on in the field today, by younger architects making smaller-scale projects, or experimental ideas that never get off the boards or out of the computer—ideas that get published and change our ways of thinking about what architecture is and can be. The Pritzker sends the message that unless one builds, and in a spectacular way, one will never qualify for “architecture’s top honor.” The catch-22 here is that to build you need clients, and to build spectacularly, like Nouvel, very rich clients, and they are seldom willing to risk sponsoring the genuinely new. So, the subliminal message is, don’t push the envelope too far.

The existence of the Pritzker reminds us that the powerful are not as self-assured as they like to appear. They need to engage continually in demonstrations of their power, such as getting on—in an upbeat way—the front page of the New York Times, as well as other major newspapers and magazines around the world. Oddly enough, the Times perennially rails against the Nobel Prizes, not least because its founder was an armaments manufacturer who bought respectability in posterity by creating prizes in his name for intellectual achievement. And it works. When we think of Alfred Nobel, his name becomes synonymous with Albert Einstein, Samuel Beckett, Toni Morrison, Martin Luther King, Jr. The Pritzker is sponsored by the Hyatt Foundation, not exactly “merchants of death.” Still, the formula works. By associating themselves with successful architects and the world of creative thought, they make a significant step up in cultural, and historical, terms. Their domain, and their power to affect the world, is extended and consolidated.

For the recipients of the Pritzker, it’s easy enough to understand why they would accept it, often with speeches of praise for the Prize and its sponsors. The money may be relatively paltry ($100,000, compared with the Nobel’s $1,500,000), but every bit helps and, after all, why not? What’s the harm? Only two have ever declined the Nobel, on principle: Sartre and Le Duc. So far, no architect has declined the Pritzker. If that were to happen, it really would ‘expand the discourse.’

All this we already know. So, why bother to write about it? Perhaps only to step back from the spectacle long enough to see its contour, and its limits. Only then is it possible to see its true place in the order of things, and the wider world that lies beyond.

LW

Unfortunately, at about this time, big, commercially-oriented blogs were established—notably ArchDaily and Dezeen. Part blog, part traditional media platform, these were professionalized, polished, and fluff-heavy, frequently aligned with marketing, project announcements, product placement, and sponsored content, devoid of independent voices, but perfect for the market. Their emergence marked the beginning of the transformation of architectural blogging from a space of critique to one of promotion.

The first discussion about blogs in architecture appears to be an April 2009 MIT panel I shared with Javier Arbona and Mark Jarzombek, titled “Blogitecture, or Architecture on the Internet.” I began my talk with a moment of reflection that holds up 16 years later: “I’m going to go back before the origins of architectural blogging … to the origins of the web and point out that I was there at the right time at many of these places and managed to miss every opportunity (almost).” At that moment, blogging had reached a crossroads—caught between its early promise of democratized critique and its emerging entanglement with the very spectacle it sought to evade. I noted that RSS feeds, comments, and trackbacks (in which one blog would notify another when it referenced it) broke the boundary between author and audience and transformed blogging into “a switching machine for data flows”—the Internet DJ now had the capacity for near-real-time feedback. This was a genuinely new form of writing and created the possibility of real back-and-forth dialogue among writers. Later that year I would participate in Helen Thornigton and Jo-Anne Green’s Networked: A (Networked) Book about (Networked Art) project, contributing a chapter to a book running on Commentpress, a fork of WordPress created by the Institute for the Future of the Book. The goal of this project being to create an open, interactive book in which readers could act as co-authors, leaving comments and creating a community-driven discourse while challenging the idea of the book as a static object.

Even as I celebrated the new technological transformation, however, I lamented that architectural discourse itself had slid from history to theory to criticism with both history and theory nearly extinguished while newspaper critics served mostly as publicists for what urban sociologist Harvey Molotch termed the “Growth Machine.” In an influential 1976 paper, Molotch described how local elites, developers, media, and institutions form coalitions to promote development regardless of actual public benefit—essentially viewing cities as machines for generating wealth rather than habitable environments.^3. Architectural criticism had largely become subservient to this apparatus, celebrating spectacular buildings that served as capital magnets while ignoring broader urban conditions. Blogs, I felt, arrested that slide by giving dispersed voices a platform to challenge this hegemony, creating spaces for more critical, historically informed discourse outside the growth machine’s influence.

Javier, an architect and geographer, turned to statistics showing that bloggers were overwhelmingly affluent professionals; thus he explained, legal blogs flourished because lawyers possessed both the disposable time and the appetite for arguments about privacy and intellectual-property rights. Architecture blogs mirrored that class bias, he argued. and within architecture, authority followed a pyramid: sites like BLDGBLOG, Inhabitat, and Dezeen occupied the apex, their posts ricocheting through Archinect, Twitter, and Tumblr until even student blogs adopted the day’s talking points—such as open source, green, innovation— which had passed uninspected from Silicon Valley into design studios. Javier suggested that a discourse on labor rights was missing in the architectural blogosphere (oddly enough, such a discourse was about to take off, but wound up as a movement centered around Yale University, a questionable place to start a workers’ revolution). Arbona displayed A Daily Dose of Architecture‘s sidebar of abandoned blogs—a digital potter’s field—and rhetorically asked why enthusiasm so often curdled into silence.

But of course, if we both engaged in the all-too-predictable discussion about privilege in blogging, we were speaking from within the academy, a place of institutional security, if not comfort, that gave us both a voice and an audience. Our criticisms of spectacle, technology, and capital unfolded within the medium of an academic conference, a ritualistic site we naturally didn’t dare question, since being in the belly of the beast makes one reluctant to prod it. Blogging had promised to disrupt such privileged circuits—indeed, for a time it appeared it might. Take the potter’s field of blogs; is there anything wrong with giving something a try and then giving up on it? Or the argument that blogs are only for those with sufficient leisure time to engage in writing and the wealth to access the Internet? Reading, bowling, knitting, and Marxist activism are all leisure time activities only those relatively well off, on a global scale, can engage in. Why not, then, blog if that gives you joy and community? When it was fun, blogging was part of everyday life, not a form of labor. That sense of camaraderie, of participating in a distributed conversation that felt vital and immediate, was its own reward.

We took steps to mature as a critical movement. Mimi Zeiger organized an underground discussion group called LGNLGN which included virtually all the names mentioned so far, save Lebbeus Woods, and we sought to find ways to professionalize and expand our reach. In the end, the most substantive outcome was the New City Reader, a broadsheet newspaper co-edited by Joseph Grima and myself and published in 2010 as part of the Last Newspaper exhibition at the New Museum. This project represented an attempt to bridge the gap between online discourse and physical media, producing critical content that could circulate beyond the screen while maintaining the networked sensibility of blog culture.

But by the time the New City Reader was published, things had started to change. Just as the blogosphere reached its zenith of cultural influence around 2007-2008, social media began its ascent. Initially launched exclusively for Harvard College students in 2004, The Facebook opened up slowly, first to Stanford, Columbia, and Yale, then the whole Ivy League and downwards through the .edu hierarchy until it was accessible to the wider public in 2006. Dropping the definite article, Facebook reimagined online interaction through user profiles and explicit “friend” connections. Its defining innovation was the News Feed, an algorithmically curated stream introduced in 2006, aggregating updates from a user’s network, eliminating the need to proactively visit individual websites.

The launch of Twitter in 2006—co-founded by Blogger creator Evan Williams—marked another pivotal moment. Twitter pioneered “microblogging” with its 140-character (later 280) updates designed for real-time sharing. While Williams conceptualized it more as an “information network” than a social platform, its impact on attention flows was profound. The short, snarky posts that characterized early blogging found a natural home on these platforms, particularly Twitter, which seemed tailor-made for the Internet DJ mentality. But critics like Alexandra Lange observed that most architects—which isn’t the same as most bloggers—predominantly used these platforms for self-promotion rather than substantive discourse, missing opportunities for deeper engagement with the cultural implications of their work.^4. Tumblr (2007) offered yet another variation, blending short posts, images, and links in a format closer to traditional blogging but still emphasizing brevity and multimedia content.

Facebook’s 2012 acquisition of Instagram for $1 billion marked a pivotal shift in online discourse, particularly for visually-oriented fields like architecture. Instagram’s image-first format fundamentally changed how architectural work was presented and consumed online. The platform’s emphasis on striking visuals, deemphasis on discussion, and its highly efficient distribution system accelerated what critics called “render culture”—where photogenic qualities often trumped practical considerations in architectural evaluation. For many architects and firms, Instagram became the primary platform for sharing work, with some projects seemingly designed more for their “Instagrammability” than spatial experience. This visual dominance particularly impacted architectural discourse, which had traditionally balanced textual critique with visual documentation. The “Like” button, introduced in 2009 further streamlined interaction, offering frictionless engagement but discouraging the substantive conversations that flourished in blog comment sections. These features created a compelling “walled garden” designed to capture and retain user attention within Facebook’s ecosystem. Facebook, Instagram, and Twitter replaced the blog for many. Indeed, with the need to keep up with all three, and perhaps Tumblr, there was less time for blogging.

Discovery shifted from blogrolls and search engines to algorithmic feeds controlled by platform companies. Consumption habits changed dramatically, influenced by smartphones and mobile internet access, which favored bite-sized content consumed on the go. The Pew Internet & American Life Project documented a notable decline in blogging among younger demographics after 2006, coinciding directly with their adoption of social networking platforms.^5.

Simultaneously, traditional architectural criticism started to collapse. The sycophancy of design criticism with regards to starchitecture, the speed and accessibility of digital platforms, a continuing decline in print readership, and a decline in newspaper finances caused by the rise of social media and Craigslist’s decimation of the advertising market contributed to less and less lengthy, rigorous critiques of the sort once found in print journals. Architectural analysis became increasingly fragmented, as the glossy, PR-driven sites Dezeen and ArchDaily took over. Unlike the independent blogs of the 2000s that frequently challenged dominant architectural narratives, these platforms largely reinforced them, prioritizing photogenic projects aligned with market trends, emphasizing aesthetic novelty and technological innovation disconnected from broader cultural or political contexts. This would soon be dubbed “Instagrammable architecture,” to refer to spaces or renderings conceived chiefly for shareable impact on an iPhone screen. Representation, especially in the schools, shifted to a pastel-toned isometric axonometry—a “post-digital” mode that fusing Archigram-style cartoons with Monument Valley-like pixel art. A 30-degree grid, uniform line weight, and flat vector fills equalize vending carts, solar panels, and structural bays, casting buildings as candy-colored storyboards. Dramatic stairs, chromatic backdrops, and razor-sharp geometries proliferated less for spatial experience than for social media potential. What resulted was an architectural discourse increasingly dominated by visual consumption rather than critical reflection—a pattern that has only intensified with subsequent platform evolutions.

Architecture itself reached an impasse by the 2010s. The problems that animated the energetic neo-avant-garde architecture of the 1990s and early 2000s—the return to modernism, the blob vs. the box, the pursuit of iconic buildings to trigger the ‘Bilbao Effect’, an architecture for the creative city, and the exploration of computational design’s potential for generating new material expressions—had been mined to death and exhausted. When these problems were seemingly ‘solved,’ architecture as cultural inquiry didn’t advance to new territory but instead floundered in a conceptual vacuum. Worse yet, from a philosophical perspective, architecture’s traditional proper object—the shaping of collective space—was now subsumed by computation itself. If computation had promised to revolutionize architectural form, it now redirected collective experience toward virtual spaces. Where young people interested in both art and technology once went into architecture, now they were lured in not only by the profits but also by the transformational promise of the Internet startup. The cure turned out to be the poison. The new commons ceased to be the plaza or the museum atrium and was now the space of the network, mediated through screens rather than physical structures. People, it seemed, could now only conceive of architectural space by taking selfies of themselves on Instagram. It wasn’t merely that architecture reduced itself to Instagrammable imagery; it was that architecture conceptually disappeared on Instagram, becoming merely another category of consumable visual content rather than a spatial art with social agency. Unlike previous moments of architectural crisis that generated theoretical breakthroughs, this exhaustion produced no corresponding illumination—only a sense of architecture’s increasing irrelevance to contemporary experience.”

Of course, as Hal Foster argues in his reading of the Freudian concept of Nachträglichkeit (deferred action) in The Return of the Real, cultural phenomena often operate through delayed recognition. The trauma or innovation registers initially without full comprehension, only to resurface later with renewed significance. Freud’s term suggests that meaning is assigned retroactively, as previously latent connections become visible. By this logic, the seemingly exhausted questions of architecture—how to shape collective experience, how to reconcile digital and physical realms, how to create meaningful intervention in increasingly privatized space—will likely one day return with renewed urgency under different conditions. But for now, architecture has entered a period in which it is for us, once again what Hegel would call “a thing of the past.” There is much to the idea of philosophy traveling through disciplines and to the recurrence of the death of architecture, but this digression is a matter for another essay.

Nor did starchitecture survive this period. According to Davide Ponzini, author of Starchitecture: Scenes, Actors, and Spectacles in Contemporary Cities, the symbolic significance of star architects declined markedly in the late 2010s. The death of iconic figures such as Zaha Hadid in 2016 further underscored this diminishing era, bringing to a close the dominance of a few celebrated architects like Frank Gehry, Norman Foster, and Rem Koolhaas. The cultural authority once held by these figures has significantly eroded, as critics increasingly challenged starchitecture for its extravagance, egocentrism, and inadequacy in addressing urgent environmental and social concerns. Ponzini suggests that these dramatic, spectacular buildings, often disconnected from their urban contexts, came to embody the contradictions and failures of urban development strategies that prioritized visual and symbolic impact over practical functionality, sustainability, and social inclusivity. Furthermore, as architectural discourse became fragmented across numerous digital platforms and diverse critical voices emerged, the unified cultural narrative and consensus that originally fueled starchitecture’s global prominence disintegrated.^6.

But the blogosphere would not replace architectural criticism. On the contrary, the rise of social media was at the expense of the blog. Engagement on Facebook, Twitter, or Tumblr took far less effort and anyone could participate. In the words of Throbbing Gristle, “Move a fin and the world turns / Sit in a chair and pictures change”—no navigating to separate sites, no logging in, just dashing off quick replies with the potential for wider visibility. It was the greatest thing since television. And if social platforms initially drove significant referral traffic to blogs, they gradually optimized to keep users within their own ecosystems, shadowbanning content with external links, leading to measurable declines in external referrals to blogs over time.

In retrospect, the first sign of stress in the blogosphere was the relentless rise of comment spam. What had begun as a trickle in the early 2000s became a devastating flood by 2010-2013, with automated bots filling blog comment sections with word salad along with links to questionable pharmaceuticals and gambling sites. Comment sections, once the vibrant heart of blogging communities where authors and readers engaged in substantive dialogue, became battlegrounds between legitimate conversation and algorithmic pollution. Blog owners found themselves spending hours each week moderating comments or implementing increasingly complex CAPTCHA and authentication systems that created friction for genuine commenters. Services like Akismet emerged to filter spam but required constant updating to keep pace with spammers’ evolving techniques and were annoying to use. Many influential bloggers eventually made the painful decision to disable comments entirely, severing a core interactive element that had distinguished blogs from traditional publishing. This degradation of community interaction further pushed conversations toward centralized social platforms where identity verification and spam filtering were built into the infrastructure, accelerating the migration of both content creators and their audiences away from the independent blogosphere.

A severe blow to the blogosphere came in 2013 when Google shuttered Google Reader, its popular RSS aggregator. RSS (Really Simple Syndication) had been the technological backbone allowing readers to efficiently follow multiple blogs without visiting each site individually. Google’s decision, citing declining usage, eliminated the most widely-used tool for RSS consumption overnight. Bloggers reported massive subscriber losses—Andrew Chen famously lost nearly 100,000 subscribers instantly. The infrastructure supporting independent blog discovery and consumption had been significantly weakened, pushing more creators and readers toward platform-centric ecosystems. The RSS icon, once ubiquitous across the web, began its slow fade from prominence.

As social media platforms captured ever-larger shares of attention, Medium, launched in 2012 by Evan Williams, represented a particularly insidious development—the co-founder of Blogger and Twitter—now building a platform that would ultimately absorb and commodify the independent publishing ecosystem his first creation had helped establish. Presenting itself as a savior of quality writing with its clean design and superior reading experience, Medium encouraged writers and entire publications to migrate their content, positioning itself as an alternative to maintaining independent blogs. But Medium was less a defender of blogging, more its corporate undertaker. After amassing sufficient content and attention, in 2017 Medium abruptly erected paywalls around content that had previously been freely accessible. Writers who had built audiences on the promise of open distribution suddenly found their work locked behind a $5/month subscription fee, with compensation tied to engagement metrics like “claps” and later reading time—a direct extension of the attention economy logic pioneered by social media platforms. This bait-and-switch tactic converted independent blogs into undifferentiated “content” feeding Medium’s subscription machine. Like Facebook and Twitter, Medium demonstrated that centralized platforms ultimately serve their own interests, not those of the creators who supply their content. The Medium experiment demonstrated the eagerness of Silicon Valley billionaires to capitalize subcultures without regard for the damage it would do. Rather than supporting a diverse ecosystem of voices with distinctive identities and direct audience relationships, Medium homogenized content under a single brand and aesthetic, reducing creators to interchangeable suppliers for its subscription business. Medium represented not evolution but extraction—a corporate attempt to capture the value of blogging while eliminating its independence and diversity.

Compounding all this was a political rift. By the time of the 2016 election, political polarization online had grown—not merely the right vs the left, but also an increasingly shrill progressive woke left that attacked liberals and centrists who were concerned, above all, that Trump not be elected. Social media, which of course played a huge role in this process, became the focus of attention and the vibrant conversations once hosted in the blogosphere degraded into bickering on sites, where interaction—however unpleasant—required less friction. The blogosphere splintered.

But even amidst this narrative of decline, blogging continued, only less so within architecture now. While microblogging and social feeds captured mass attention, a distinct cohort of interconnected blogs—largely British Marxists and lovers of post-punk music—maintained and deepened the intellectual promise of the medium through rigorous, extended explorations of complex topics and longer form pieces. Mark Fisher’s k-punk, launched in 2003 was what Simon Reynolds described as “the central hub of a ‘constellation of blogs’ in which popular culture, music, film, politics, and critical theory were discussed in tandem by journalists, academics, and colleagues.”^7. This loose collective included writers like Reynolds himself, Benjamin Noys with his “No Useless Leniency” blog, Matthew Ingram who ran Woebot, aforementioned architecture blogger Owen Hatherley as well as others engaged in sustained cross-blog conversations. Blogging, for this group, became a central site for developing theoretical concepts. Noys notably coined the term “accelerationism” as an object of critique, while Fisher explored the possibilities of “acid communism”—demonstrating how blogging could sustain serious intellectual work outside academic institutions, which they increasingly saw as precarious. Fisher’s blog, Ryan Meehan wrote, was a “as a release valve from the pressures of academic writing. It became an impassioned and authoritative node in the hyperactive mid-aughts blogosphere—a network that, viewed from the social media quagmire of 2018, seems romantically free.”^7. Fisher’s explorations of concepts like “capitalist realism,” “hauntology,” and the “slow cancellation of the future” began as blog posts before evolving into influential books published on the Zero Books imprint. His suicide in 2017 left a void in online Leftist thought, but cemented his legacy as one of the most important cultural theorists of his generation.

Returning to Scott Alexander, his Slate Star Codex (later a substack called Astral Codex Ten), written anonymously until his identity was controversially revealed by the New York Times, became a central hub for the rationalist community and attracted a wider readership interested in science, medicine, philosophy, politics, and futurism.The blog’s long-form essays delved into often controversial subjects, employing detailed arguments and statistical reasoning in ways that some saw as intellectually rigorous and others viewed as a gateway for legitimizing fringe or socially harmful ideas. Alexander’s practice of noting the “epistemic status” of his posts and the blog’s highly engaged comment sections fostered a culture of rigorous discussion that stood in stark contrast to social media’s superficiality. Similarly, Gwern.net, the website of pseudonymous independent researcher Gwern Branwen, represented another facet of the long-form resurgence. Known for exceptionally deep, data-rich research on topics ranging from AI and machine learning to psychology and self-experimentation, Gwern’s distinctive style features extensive annotations, footnotes, and sidenotes, reflecting a meticulous research process and a commitment to comprehensive evidence. The site’s unique design prioritizes content accessibility while deliberately rejecting web bloat. Gwern, incidentally, claims to live off of about $12,000 a year in order to devote himself full time to his passion of blogging. Longform podcasts like Dwarkesh (Patel’s), which features in-depth conversations with thinkers on AI and related matters have helped introduce the idea of long-form thought with broader audiences, undoing the dopamine-heavy obsession with the social media post. In providing platforms for extended, nuanced discussions, these podcasts serve as complementary mediums to long-form writing, allowing ideas to develop and cross-pollinate across formats.

Starting in 2017, Substack began providing a simple, integrated solution for such writers to launch newsletters and charge subscriptions. The platform experienced rapid growth, significantly accelerated by the COVID-19 pandemic, reaching over one million paying subscribers by late 2021. This model built on the same core values that drove blogging’s initial success—the desire for independent expression, direct audience connection, control over one’s platform, and space to explore ideas in depth.

For me personally, this renaissance of long-form, thoughtful content and the ease of uploading long form posts to WordPress and Substack presents a new opportunity. Though the architectural blogosphere I was once part of has largely disappeared, many of the problems I tried to address then have reemerged with even greater urgency—the interplay of space and technology, the role of technology in our lives, and society’s relationship to the built and natural environment, topics that demand the kind of sustained, thoughtful engagement that quick social media posts can’t provide. As yet, I don’t see a longer-form architecture community taking shape, and I wonder if I would be part of it—although John Hill’s A Daily Dose of Architecture, now A Weekly Dose of Architecture Books is on Substack as is Enrique Ramirez, with his “A New Literary History of Architecture.” For his part Adam Greenfield continues to write regularly at Patreon.

To conclude, let’s turn back to Scott Alexander’s musing about blogging’s “golden age” and his missed opportunity. While some early web pioneers like Marc Andreessen did indeed parlay their online presence into tremendous wealth and influence, others experienced a far different trajectory. Consider Robot Wisdom’s Jorn Barger, the coiner of “weblog.” Despite his pioneering status, Barger later found himself “homeless and broke” in San Francisco, “living on less than a dollar a day,” as reported by Wired magazine in 2005.^8. The contrast is stark: the very man who named the medium couldn’t convert his innovation into financial stability, while others built empires. Given Andreessen’s recent behavior, it’s hardly a matter of mental stability; it’s the usual matter of getting lucky.

This disparity highlights how blogs were never primarily vehicles for wealth creation, institution-building, or an expected source of income for most of us. We did it for the hell of it. Alexander gets it all wrong by focusing on remuneration or professionalization. Blogging had—and once again has—a “golden age” precisely because we had no stakes in it. Blogs were significant not as business ventures but as independent voices outside traditional media channels. And that endures today. The true power of blogging is in ownership and control over one’s means of intellectual production. Those bloggers who sought professionalization or institutional validation—including in the academy— found themselves reabsorbed into existing systems. Pace Alexander, the most enduring legacy of blogging isn’t found in those who “graduated” to mainstream success but in those who maintained their independence, creating an alternative intellectual commons that operated according to different values than the market or the academy.

Blogging taught us that intellectual communities can form and flourish outside traditional institutions. My own journey from architecture blogging to explorations of native plant design and AI Art illustrates how these distributed conversations allow for intellectual evolution that might be constrained within more rigid disciplinary boundaries. The networked thinking that characterized the architectural blogosphere—its ability to connect seemingly disparate domains through hyperlinks and cross-disciplinary dialogue—offers a model for approaching the complex challenges we now face, from climate adaptation to reimagining collective space in a post-pandemic time.

As Mark Zuckerberg’s recent acknowledgment that “social media is over” suggests, the pendulum is swiftly swinging away from the algorithmic feeds and dopamine-driven engagement mechanisms that dominated the 2010s.^9. Whether this is because individuals are occupying themselves in AI chatbots, just watching more streaming TV to escape, or are interested in longer form writing is, as yet unclear. Likely, it is all of these. Blogging, in some form, will endure for some time to come. Rather than a discrete historical moment, blogging is an enduring mode of independent intellectual engagement that persists despite evolving technological contexts. While architectural bloggers may have failed to transform architectural labor practices or fully democratize discourse, we nevertheless established alternative pathways for developing and sharing critical perspectives, if only temporarily. That achievement, modest though it may seem, represents a meaningful contribution to culture and maybe that’s enough.

^1. Kazys Varnelis, proposal for sci * arc * wire, https://varnelis.net/sci-arc-wire/

^2. Kazys Varnelis, Proposal for SCI * Arc * Wire, https://varnelis.net/sci-arc-wire/

^3. Harvey Molotch, “The City as a Growth Machine: Toward a Political Economy of Place,” American Journal of Sociology, Volume 8, Number 2 (September 1976) , 309-332, https://www.jstor.org/stable/2777096.

^4. Alexandra Lange, “Opinion, It’s easy to make fun of Bjarke Ingels on Instagram,” Dezeen, January 7, 2014. https://www.dezeen.com/2014/01/07/opinion-alexandra-lange-on-how-architects-should-use-social-media/

^5. Amanda Lenhart, Kristen Purcell, Aaron Smith, and Kathryn Zickuhr, Social Media and Young Adults, Pew Research Center, Internet & American Life Project, February 3, 2010, https://www.pewresearch.org/internet/2010/02/03/social-media-and-young-adults-3/.

^6. Davide Ponzini and Michele Nastasi, Starchitecture: Scenes, Actors, and Spectacles in Contemporary Cities (New York: Monacelli Press, 2016), 190-192.

^7. Simon Reynolds, “Music & Theory,” Frieze, September 18, 2009, https://web.archive.org/web/20160304073857/http://blog.frieze.com/music_theory/.

^8. Paul Boutin, “Robot Wisdom on the Street,” WIRED, July 1, 2005, https://www.wired.com/2005/07/robot-wisdom-on-the-street/.

^9. Kyle Chayka, “Mark Zuckerberg Says Social Media is Over,” The New Yorker, April 23, 2025, https://www.newyorker.com/culture/infinite-scroll/mark-zuckerberg-says-social-media-is-over

This is part of an emerging project that I am currently calling “A Skeptic’s Guide to Art with AI.” I have a third essay nearly done (on AI and Plagiarism), a fourth (on AI and fakes), and a fifth (on AI, Death, and Hauntology), substantially drafted. The plan is to publish one each Monday throughout the next month.

Each essay addresses a familiar complaint with AI and explores its relationship to culture throughout history, as well as the potential uses in art. When all that is wrapped up by the start of the summer, I will have a little book—think of Semiotext(e)’s Foreign Agents series—ready to publish. Given the pace of change in AI, I don’t imagine that waiting for a press would make sense and expect to self publish, but if someone has a concrete proposal, I will gladly listen.

I do hope to return to the Florilegium—my series of posts about native plants, art, and design—maybe even this week, but my native plant bandwidth is tied up with being president of the Native Plant Society of New Jersey and, well, on tending the garden.

As always, this post is up on my web site, https://varnelis.net/works_and_projects/the-generative-turn-on-ais-as-stochastic-parrots-and-art/ and all I ask is that if you enjoy this post, you like it and share it. Comments are welcome too.

Large Language Model (LLM)-based Artificial Intelligences have been derided for hallucinating—a topic that I addressed in my last essay, “The New Surrealism. On AI and Hallucinations,”—and labeled “stochastic parrots,” supposedly incapable of generating authentic meaning beyond remixed texts. Coined by researcher Emily Bender and her colleagues, the term “stochastic parrots” describes language models as systems that statistically predict and reproduce patterns in language without genuine understanding. And yet, anyone familiar with contemporary LLMs—particularly recent, sophisticated models like ChatGPT 4.5 or Claude Sonnet 3.7—knows that these “parrots” easily pass a simple Turing Test, Alan Turing’s hypothetical test of a machine’s intelligence to see whether a person interacting with it can reliably distinguish the computer from a human. These convincingly human-like interactions underscore a deeper truth: predictability is fundamental to human society itself.¹

From routine greetings to rituals marking life milestones—weddings, graduations, funerals—our shared existence thrives on stable expectations. Indeed, the stress that many of us—including the markets—feel in this first spring of the second Trump administration is caused precisely by the executive branch’s erratic, unpredictable behavior. Predictability is not merely useful; it is essential, ensuring interactions proceed smoothly without continuous negotiation. Every culture, whether traditional or contemporary, anchors itself in repeated rituals, social scripts, and familiar stereotypes. Consider the handshake, a simple gesture instantly communicating respect, trust, or greeting through mutual recognition. In Japan, a bow serves this function; in France, a kiss on each cheek. Such rituals persist because their predictability facilitates social cohesion. Without stable expectations, human interactions would require constant explanation and renegotiation, exhausting participants and weakening social bonds. Predictability permeates all levels of societal structure up to international relations. Diplomatic protocols, treaties, and international laws function precisely because nations rely on mutual expectations of predictable behavior. The uncertainty created when states deviate from established norms—as seen in trade wars, diplomatic tensions, or sudden military escalations—generates global anxiety, economic instability, and geopolitical friction.

Sociologist Erving Goffman, in his influential work The Presentation of Self in Everyday Life, describes human behavior as a series of performances carefully regulated by implicit scripts. These scripts guide everything from greetings between friends to complex professional interactions. Goffman suggests that without them, social life would descend into chaos. Individuals intuitively understand these roles, shifting seamlessly from one performance to another based on context: professional, familial, or social. Predictable interactions reassure participants by clarifying expectations, permitting people to anticipate and trust each other’s behaviors, reducing cognitive and emotional strain, fostering stability and cohesion within society.²

Similarly, anthropologist Pierre Bourdieu expands this notion through his concept of habitus. For Bourdieu, habitus represents the internalized structure of social life: individuals embody deeply ingrained habits, dispositions, and tastes acquired through repeated interactions within their specific cultural contexts. Habitus shapes behaviors, preferences, and perceptions so thoroughly that they become second nature, enabling individuals to intuitively navigate social environments without conscious effort. But because habitus includes deeply internalized habits, tastes, and expectations formed by one’s social environment, individuals unconsciously accept and replicate existing social structures, including hierarchies and power relations. In other words, people unknowingly reinforce patterns of privilege or disadvantage simply by acting in familiar, predictable ways. Predictability thus not only facilitates social cohesion but also perpetuates existing hierarchies and norms by normalizing and embedding them deeply within individuals’ perceptions and practices, ensuring social stability through repetition rather than conscious coercion.³

In stabilizing society, predicatability also fosters stereotypes. First used in its modern sense by journalist Walter Lippmann in Public Opinion (1922), the word “stereotype” originally referred to mental images—simplified, predictable perceptions enabling quick social judgments. Even if stereotypes are commonly seen as leading to prejudice or discrimination, they nonetheless perform critical functions within society by enabling rapid cognitive processing in uncertain situations, providing shorthand classifications that allow individuals to navigate social complexity swiftly. For instance, professional stereotypes—expectations about doctors, lawyers, or teachers—guide interactions within these roles, establishing clear norms that enable efficient communication and professional conduct. Even resistance to stereotypes follows predictable patterns; the avant-garde artist deliberately violating expectations is, paradoxically, fulfilling another recognizable social role.⁴

This understanding of predictability’s role in human society offers a useful lens for reconsidering the critique of AI as merely “stochastic parrots.” If human communication and creativity themselves operate through pattern recognition, recombination, and contextual prediction, then the distinction between human and AI generation becomes less absolute. Both rely on absorbing patterns from their environment and producing variations that remain recognizable within specific contexts. The distinction lies not in the fundamental process but in the degree of consciousness, intention, and embodied experience that shapes the patterns being recognized and reproduced.

Art itself has always existed in this tension between repetition and innovation. Repetition lies at the core of all artistic practice—the Renaissance master learning through repetitive copying, the jazz musician repeating standard progressions until they become second nature, the novelist absorbing and repeating literary conventions. Repetition is not merely a pedagogical tool but the fundamental substrate of creativity itself. Through repetition, the artist internalizes traditions deeply enough to transform them. Through repetition, audiences develop the literacy to recognize both convention and deviation. Through repetition, cultural forms evolve and mutate across generations. What we celebrate as genius is often the capacity for repetition with a difference—the ability to repeat familiar patterns with just enough variation to create novel meaning. In this light, the “stochastic parrot” critique misses how thoroughly repetition underpins human art-making itself, and how creativity emerges not from originality ex nihilo but from the subtle modulation of repeated patterns.

The very predictability that stabilizes culture also enables innovation: this paradox lies at the heart of creative activity. Novelty emerges not through spontaneous, unconditioned invention but through subtle departures within structured repetition. Gilles Deleuze, in his pivotal text Difference and Repetition (1968), elucidates this dynamic. Deleuze argues against the misconception that repetition is mechanical and sterile. Instead, he insists repetition is inherently creative, each iteration introducing small yet meaningful variations. He distinguishes mechanical repetition—identical reproduction, like photocopying—from productive repetition, in which subtle divergences accumulate into genuine innovation. As incremental variations accumulate, genuine difference emerges. Creativity occurs precisely in the space of tension between the expected (predictable repetition) and the unexpected (subtle variation). Far from inhibiting innovation, predictability thus serves as its very foundation. Deleuze identifies this as the core mechanism underlying genuine innovation—whether in art, science, literature, or society.⁵

Repetition as a fundamental cognitive act predates humanity itself. Birdsong relies on familiar, established patterns with subtle variations in notes, pitch, or sequence. Whale song, too, consists of repeated phrases and themes that evolve continuously and dramatically over time. Consider the emergence of humanity, which we glimpse so vividly in the cave paintings at Chauvet Pont d’Arc (starting at 37,000 BPE) and Lascaux (starting at 22,000 BPE)—two artistic sites nearly as distant from each other in time as we are from Lascaux today. Both sites manifest strikingly consistent visual languages, demonstrating predictable, structured representations of animals and symbols. Both sites were inhabited for thousands of years, a fact that in itself, is staggering to contemplate. Yet, far from diminishing their cultural significance, the predictability of this work demonstrates how human creativity is established within tradition and repetition.

Classical and medieval cultures explicitly acknowledged tradition as essential to creativity. Artists and scholars operated within established frameworks, regarding innovation as the careful refinement of inherited forms. During the Renaissance, often associated with the celebration of individual genius and divine inspiration, artists deeply relied on repetition and established conventions—perspective, anatomy, classical motifs—as scaffolding. Inspiration did not arise from unstructured impulse but flourished precisely within the constraints of recognized tradition. In music, composers such as Bach and Beethoven famously employed iterative variations on simple themes, each repetition pushing the motif further until something startlingly original emerged. Every artist, every thinker, every cultural producer operates within inherited conventions, echoing and altering familiar patterns to create meaning anew. Accomplished work is itself produced within the context of years of study, usually within long-established, predictable systems such as music theory.

Thus, when modern avant-garde movements emerged, they did not abandon predictability; they relied upon it to heighten the impact of their provocations. Movements such as surrealism and dada strategically employed known conventions to disrupt expectations. Marcel Duchamp’s infamous Fountain (1917)—a common urinal presented as art—achieved its shock precisely by repeating familiar objects in unfamiliar contexts. Repetition set the stage; deviation created the effect. Or consider how postwar avant-garde artists explored repetitive techniques. Andy Warhol’s silkscreen works embodied this principle perfectly, creating endless repetitions of cultural icons and consumer products. In his renowned Marilyn Diptych (1962), Warhol reproduced the actress’s image fifty times across the canvas—twenty-five in vibrant color, twenty-five in fading black and white. The repetition itself becomes the subject, highlighting how mass media’s endless reproduction simultaneously immortalizes and empties images of meaning. Similarly, in his “Death and Disaster” series, Warhol repeated newspaper photographs of car crashes and suicides, the mechanical repetition simultaneously numbing viewers to tragedy while paradoxically intensifying its horror. Each repetition subtly differs from the last—variations in ink density, alignment, and color—creating a visual meditation on how mechanical reproduction transforms experience. Through these works, Warhol demonstrated how repetition is not merely reproduction but a transformative process, generating profound new meanings through the accumulation of subtle differences.

Understanding society and creativity as predictive helps us grasp how generative AI systems echo deeply human cultural processes. I suggested at the outset of this essay that, at heart, LLMs function as sophisticated statistical prediction machines. Given a phrase—”It was a dark and stormy…”—the AI calculates the most probable continuation (“night”) by analyzing billions of instances of textual data. Yet this process is not strictly deterministic. Instead, it operates probabilistically, meaning each prediction is shaped by subtle fluctuations and contextual nuances. AI’s predictions are never mechanically identical; they are always iterative reinterpretations of linguistic history. The model generates not only predictable continuations but also unexpected divergences—sometimes subtle, sometimes radical. Critics and researchers label these divergences “hallucinations,” often viewing them as technical flaws, although the hallucinations—like human dreams—are a source of creativity for the AIs, embodying precisely the generative logic Deleuze articulated: structured repetition producing meaningful differences.

Diffusion models used for AI image generation use a similar predictive logic. Rather than predicting the next word or phrase, these systems predict the gradual emergence of coherent images from initial random noise. Through repeated, iterative steps, diffusion models systematically clarify visual forms by predicting and reducing noise incrementally. Each step is probabilistic, guided by learned patterns derived from extensive datasets of existing images. Just as LLMs generate meaningful text through slight divergences in their statistical predictions, diffusion models yield artistic novelty through controlled randomness. The structured, iterative process—oscillating between predictability and noise—allows the emergence of compelling visual imagery, ranging from convincingly realistic portraits to evocative, dream-like abstractions. Again, we find creativity emerging not from randomness alone, but from a deliberate, generative interplay between predictable structures and unpredictable deviations.

III

1. Take a piece of paper and write down a random word.
2. Read the word out loud.
3. Write down a sentence that incorporates the word you just read.
4. Read the sentence out loud.
5. Repeat steps 1-4 until you feel satisfied with the result.

VIII

1. Choose a cat to be the performer.
2. Place the cat in a comfortable space with plenty of room to move around.
3. Allow the cat to explore its surroundings and become familiar with the space.
4. Observe the cat’s movements and vocalizations, and listen for any sounds or patterns that emerge.
5. As the cat moves and makes sounds, use a microphone or other recording device to capture its vocalizations and movements.
6. Use the recorded sounds to create a series of musical compositions that incorporate the sounds of the cat.

Kazys Varnelis, 20 Subroutines for Humans Made by a Computer, 2022

The generative approach pioneered by mid-20th century artists offers perhaps the most direct historical precedent for today’s AI systems. One of the most influential pioneers was composer John Cage, whose landmark work, Music of Changes (1951), explicitly relied on chance operations determined by the ancient Chinese divination text, the I Ching. Crucially, Cage did not abandon composition entirely to random chance. Instead, he devised intricate procedures and rules, repeatedly applying chance-generated decisions to every musical parameter—pitch, rhythm, dynamics, duration. The resulting performances were thus simultaneously predictable (structured by rules) and unpredictable (generated by chance). Cage understood perfectly Deleuze’s generative logic: structure and repetition establish conditions within which genuine innovation emerges unpredictably. His music continually surprised listeners precisely because it repeated a clearly defined generative procedure, with each iteration subtly and meaningfully different.

Composers such as Brian Eno moved generative music increasingly toward explicit algorithmic practices in the late 20th century. Eno’s groundbreaking “generative music” created systems of musical production defined by fixed rules or algorithms that continuously yielded new variations. In compositions like Music for Airports(1978), Eno programmed systems of overlapping tape loops and repetitive sequences. While each loop was predictable in itself, their interactions produced constantly shifting and novel auditory experiences. The music was simultaneously repetitive (structured loops) and unpredictably evolving (variations through algorithmic combination). Eno captured precisely Deleuze’s generative vision, arguing explicitly that repetition and structure do not constrain creativity but rather create conditions necessary for meaningful novelty.

This generative approach spread across artistic disciplines. Visual artists, inspired by Cage and Eno, adopted algorithmic systems to explore repetition’s creative possibilities. Vera Molnar, one of the pioneers of computer-generated art, produced works based on algorithmically repeated variations of geometric shapes. While each repetition was controlled by predefined rules, tiny programmed deviations ensured continual, subtle differences between iterations. Similarly, Sol LeWitt’s conceptual art pieces began as instructions—rules for execution—that could be repeated endlessly by different people, at different times, in different spaces. Each instantiation of LeWitt’s instructions differed slightly, reinforcing the generative logic that repetition does not produce identical copies but inevitably creates meaningful differences.

Generative art thus illustrates how repetition can become the core medium of creative expression. Far from constraining artists, rules, algorithms, and structured repetitions enable innovative exploration of themes and possibilities that otherwise would remain inaccessible. Each iteration becomes an experiment, its outcomes unpredictable precisely because they emerge from the structured repetition of defined systems. AI image generators lend themselves to this flamework, frequently making four images as a default—beginning with random numerical seeds that determine the initial pattern from which the AI builds. This approach allows artists to generate multiple variants simultaneously, comparing outcomes and selecting directions for further exploration. By reusing the same prompt with either new random seeds or by deliberately maintaining the same seed across multiple generations, artists can produce hundreds or even thousands of iterations, methodically refining their vision until a satisfactory result emerges. This process transforms the act of creation into a collaborative dialogue between human intention and algorithmic interpretation. Cage, Cornell, Eno, Molnar, and LeWitt all understood this paradox: structured repetition and generative unpredictability are not opposing terms, but interdependent conditions of genuine creativity.

These historical practices offer crucial context for understanding contemporary AI art systems. When we frame LLMs or image generators as “stochastic parrots,” we overlook their deep continuity with established generative traditions. Like Cage consulting the I Ching or Eno layering asynchronous tape loops, AI systems deploy structured repetition across vast datasets, producing variations that are at once predictable (drawn from training data) and novel (emerging from complex interactions). The stochastic operations of AI closely resemble those that made aleatory music or generative ambient works radical in their time.

AI-generated artifacts reflect human linguistic and cultural patterns, mirroring entrenched societal expectations and assumptions. Each output emerges from the sedimented repetition of past human expression. In this sense, AI functions less as a creator than as a mirror—revealing the biases, desires, anxieties, and mythologies embedded in its source material. Critics who denounce AI-generated imagery as “racist” misidentify effect for cause: the model does not invent prejudice; it exposes the structures that already exist.

The launch of Google Gemini’s image generator illustrated this vividly. In an effort to correct perceived racial bias, the model was tuned for enforced representational diversity. The result: grotesque anachronisms such as Asian English monarchs or, notoriously, Black Nazis. These were not expressions of machine malice but symptoms of a misguided attempt to overwrite probabilistic patterning with ideological intention. The surreal outputs were not bugs—they were faithful to the data and incoherent only in light of contradictory demands placed upon them. If the episode revealed anything, it was the underlying surrealism of the cultural archive itself.

Indeed, if in my last essay, I described how generative AI can produce a new surrealism, we see her an additional digital equivalent to surrealist practices in its resemblance to automatic writing, which André Breton famously described as “pure psychic automatism.” Just as surrealists sought creative potential in spontaneous, subconscious associations free from rational control, AI language models generate outputs through automated processes free from explicit human intent. The unpredictable linguistic connections that result—sometimes logically flawed but poetically intriguing—parallel surrealist attempts to bypass conscious rationality. Generative AI, guided by probabilistic algorithms rather than neurons, extends surrealist logic into the digital realm, recombining existing cultural patterns into unexpected forms.

What if we imagine human-AI interactions not as diminishing human creativity but as generative processes akin to Cage’s dialogue with chance, Eno’s generative loops, or LeWitt’s instructional repetitions? What if generative AI is not merely a tool but an active cultural interlocutor? Artists, writers, and designers already intentionally explore AI-generated hallucinations and linguistic variations to stimulate creativity.

In this light, generative AI is not a radical break from human creativity but rather clarifies and amplifies cultural processes always at play, explicitly revealing how repetition, predictability, and variation dynamically produce meaning. Deleuze’s theoretical insights and generative art’s practical explorations present AI as fundamentally human—a computational extension of society’s inherent creative logic. Far from constraining creativity, structured repetition becomes a catalyst for thoughtful critique and conscious innovation. Artists and intellectuals can thus use AI not as a replacement but as a creative collaborator, embracing its generative potential to expand rather than diminish cultural imagination.

This collaboration reshapes conceptions of authorship, originality, and authenticity. Instead of the solitary genius traditionally celebrated by Western culture, creativity emerges explicitly collaborative, iterative, and distributed between humans and machines. Authorship becomes an intentional dialogue: humans provide conceptual direction and editorial judgment, while AI contributes unexpected variations drawn from vast cultural datasets. Creativity thus evolves away from solitary invention toward a practice of coordination, curation, and generative experimentation.

The stochastic parrot, it would seem, does have a lesson for us—but it is not that LLMs are unintelligent; rather, it is that human creativity itself is inherently generative. We are all, to a degree, stochastic parrots. If learning to work with machines and computers compelled earlier generative artists, learning—and indeed inventing—new forms of collaboration with AI is the task awaiting the generative creators of our era.

^1. Emily Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell, On the dangers of stochastic parrots: can language models be too big?, Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, March 2021, https://dl.acm.org/doi/10.1145/3442188.3445922. ↩

^2. Erving Goffman, The Presentation of Self in Everyday Life (New York: Doubleday, 1956). ↩

^3. Pierre Bourdieu, Outline of a Theory of Practice. (New York: Cambridge University Press, 1977). ↩

^4. Walter Lippmann, Public Opinion (New York: Harcourt, Brace and Company. 1922) ↩

^5. Gilles Deleuze, Difference and Repetition, trans. Paul Patton (New York: Columbia University Press, 1994) ↩

In 2008, my book The Infrastructural City: Networked Ecologies in Los Angeles came out. To be clear, this wasn’t just my book. It was my vision and I edited it, but I assembled a team of architects at the Los Angeles Forum for Architecture and Urban Design to try to understand the metropolis through the lens of its underlying infrastructural systems. Taking Los Angeles as a paradigmatic modern city, the aim was to reveal how infrastructure, once the defining driver of urban growth and modernity, had begun to produce systemic crises and profound political impasses.

As I argued in the book’s introduction, infrastructure formed the core of Los Angeles’s identity—indeed, infrastructure was the city’s secular theology, its underlying belief system. Los Angeles arose not by accident or gradual organic growth but through deliberate, even heroic, acts of infrastructure: capturing distant rivers, electrifying deserts, and threading sprawling freeway systems across inhospitable terrain. Its birth represented human ingenuity overcoming ecological constraints. The modernist vision of infrastructure was a technocratic dream of reshaping unruly nature into orderly, productive landscapes. Infrastructure provided secular salvation for the American West, turning deserts into farmland, canyons into reservoirs, and remote valleys into thriving suburbs. Los Angeles embodied this vision more vividly than any other American metropolis. From William Mulholland’s aqueducts and the ambitious freeway system to distant electrical grids, the city’s infrastructures were built at vast scales, each project reinforcing the belief that ecological and geographic limits could always be transcended. Yet by the dawn of the twenty-first century, this triumph had given way to crisis, an infrastructural impasse born directly from the city’s prior successes.¹

Reyner Banham famously celebrated this infrastructural landscape in his 1971 book Los Angeles: The Architecture of Four Ecologies, describing the city through four interlocking ecological systems: Surfurbia (beach towns), Foothills (privileged hillside communities), Plains of Id (the sprawling, banal yet exuberant flatlands), and Autopia (the freeway network that tied it all together). Banham embraced Los Angeles as a decentralized, spontaneous city shaped by infrastructure rather than traditional urban planning. In doing so, Banham valorized what he called “Non-Plan,” a condition where bottom-up forces, consumer preferences, and private initiatives generated urban form free from bureaucratic constraints and grand masterplans.²

But by 2000—arguably even by 1980—the consequences of Banham’s Non-Plan were painfully evident. Instead of creating a liberating urban landscape, Non-Plan had set the stage for infrastructural dysfunction, political paralysis, and environmental degradation. Infrastructure now produced chronic dysfunction: aqueducts drained distant ecosystems and provoked political conflict, freeways were congested, air pollution was persistent, and entrenched NIMBYist politics stymied new infrastructural projects. Proposition 13, enacted in 1978, severely limited public funding, locking infrastructure into a state of permanent decay and inadequacy. Heroic infrastructure—massive, centralized, technocratic—had effectively come to an end. What emerged instead was a lasting infrastructural stalemate: political paralysis, ecological deterioration, and structural underinvestment.

Yet Los Angeles’s experience was not unique. As Edward Soja famously pointed out, Los Angeles is both an exception and the rule: a singular example that reveals broader trends.³ The infrastructural impasse evident in Los Angeles reflected conditions across America—neoliberal governance, entrenched individualism, private interests dominating public goods, and widespread resistance to new development. Our argument in the book was that infrastructure’s future would not be defined by grand heroic visions, but rather through difficult, continuous negotiations with ecological constraints, competing political demands, and limited resources. Seventeen years later, these fundamental issues persist: how can infrastructure meaningfully adapt, and can a compelling new vision emerge from what appears to be a landscape of perpetual impasse?

Shortly after The Infrastructural City appeared, Christopher Hawthorne, then architecture critic for the Los Angeles Times, reviewed it prominently on the front page of the Culture section. This should have been a pivotal moment for us, bringing attention to a project that had taken years to produce. Instead, we were left disappointed. Hawthorne slammed our book as overly pessimistic, arguing that our emphasis on invisible systems, regulatory complexities, and entrenched political barriers dismissed too quickly the potential for visible, iconic infrastructure projects that could be created by starchitects like Norman Foster and Rem Koolhaas.⁴ In fairness to Hawthorne, he was writing in the hopeful early months of Barack Obama’s presidency, when substantial investment in infrastructure seemed imminent through the stimulus package being developed in response to the financial crisis. His optimism reflected precisely that brief historical moment. Once the Obama administration’s ambitious plans were tempered by political compromises, regulatory inertia, and the economic approach of funding the banks first favored by Director of the National Economic Council, Lawrence Summers, his assessment might have been gloomier.

More than that, Hawthorne’s optimism about architectural solutions to infrastructural problems reflects a persistent pattern in American discourse: the belief that our systemic challenges require merely aesthetic or technical fixes rather than fundamental political-economic restructuring. This misdiagnosis has continued to shape infrastructure debates in the years since our book’s publication. I will admit that we didn’t give a particularly optimistic polemic, but most recently, in their book Abundance, Ezra Klein and Derek Thompson start out from a position remarkably similar to our own, arguing that show that America’s infrastructure problems remain stubbornly rooted in the invisible political-economic structures, regulatory barriers, and social conditions that we sought to reveal. Seventeen years after The Infrastructural City, the conditions we described have intensified rather than eased. Los Angeles remains trapped in infrastructural paralysis, reflecting a broader failure extending across California and, indeed, the United States as a whole. Little meaningful progress has been made in addressing fundamental urban crises—traffic congestion, housing affordability, ecological degradation—while political stalemates have deepened rather than resolved.

In Los Angeles specifically, infrastructure initiatives remain sporadic and insufficient. Ambitious projects promised decades ago, such as high-speed rail and comprehensive transit expansions, remain unrealized or delayed indefinitely. Traffic congestion has worsened, air quality improvements have stagnated, and despite efforts to promote transit-oriented development, the city still struggles with its legacy of automobile dependency. Water scarcity, predicted to become critical nearly two decades ago, is now acute, with the region stuck in cyclical drought emergencies while permanent solutions languish in political gridlock. Meanwhile, Proposition 13’s legacy continues to limit revenue streams, ensuring persistent underinvestment in public infrastructure. But these problems extend far beyond Los Angeles. Throughout California, similar infrastructural crises have emerged, emblematic of broader national trends. Housing shortages have driven soaring costs, contributing to an affordability crisis that increasingly drives young families out of the state. Homelessness, once confined to downtown skid rows, has become notoriously pervasive in California, most notably in Los Angeles and San Francisco, but throughout the state in general as well. Public education and transit remain underfunded, overcrowded, and inadequate, while the state’s famed climate initiatives repeatedly collide with stubborn local opposition and regulatory obstacles. In Abundance, Klein and Thompson explicitly identify California as the paradigmatic example of this broader American infrastructural and political impasse. As they put it bluntly:

“California’s problems are often distinct in their severity but not in their structure. The same dynamics are present in other blue states and cities. In this era of rising right-wing populism, there is pressure among liberals to focus only on the sins of the MAGA right. But this misses the contribution that liberal governance made to the rise of Trumpism. […] Donald Trump won by shifting almost every part of America to the right. But the signal Democrats should fear most is that the shift was largest in blue states and blue cities—the places where voters were most exposed to the day-to-day realities of liberal governance.”⁵

Klein and Thompson’s argument underscores that the dysfunction found in California—highly regulated yet infrastructurally stagnant, rhetorically progressive yet practically conservative—is symptomatic of deeper national failures. States across the country share California’s fate, caught in regulatory entanglements, financial constraints, and political paralysis that make meaningful infrastructure impossible to build. Federal attempts at infrastructural renewal, such as the Biden administration’s Infrastructure Investment and Jobs Act, have struggled to break through entrenched local resistance and bureaucratic inertia. Even when funded, projects stall at the state and municipal levels, tangled in endless public hearings, lawsuits, and regulatory hurdles. America finds itself stuck in the impasse first described nearly two decades earlier in Los Angeles. Infrastructure, which once symbolized national strength and optimism, now stands as a monument to collective failure. The broader infrastructural gridlock, first identified at a local level, has become a national condition.

Against this condition, Klein and Thompson offer perhaps a compelling retort: scarcity, particularly infrastructural scarcity, is not inevitable but chosen. America’s inability to build housing, transit, clean energy projects, and critical infrastructure is fundamentally a political problem, rooted in policy failures, regulatory barriers, and entrenched political and ideological opposition rather than technical or economic limitations. This argument reframes infrastructural impasse not as destiny but as an active political choice—a choice that can be reversed.

Klein and Thompson argue that both sides of the American political spectrum bear responsibility for the present stagnation. Conservatives, committed to shrinking government and relying exclusively on market solutions, have systematically undermined the public sector’s capacity to execute ambitious projects. Progressives, meanwhile, despite their rhetorical commitments, have often obstructed meaningful development through excessive regulation, overly cautious environmental policies, and local NIMBY resistance. The outcome has been pervasive paralysis and disillusionment—particularly visible in progressive strongholds like California.

Yet Klein and Thompson are not pessimists. They present a positive, forward-looking vision of what could be accomplished if political will aligned with technological capability. A new infrastructural abundance—marked by rapid housing construction, widespread deployment of renewable energy, modernized transportation networks, and equitable urban growth—is entirely achievable, they assert, provided regulatory and ideological barriers are dismantled and public ambitions are renewed. Their solution is clear: the United States must build more, faster, and smarter, to address chronic shortages in housing, energy, healthcare infrastructure, and transportation. Abundance, in their view, represents not merely an economic or technological goal but a necessary political project—a pathway out of the stasis and frustrations of contemporary American life. Their central solutions involve streamlining regulatory processes, significantly accelerating permitting timelines, boosting public investments in infrastructure projects, and revitalizing government agencies’ capacities to execute ambitious, large-scale developments.

Abundance has emerged at a moment when it offers the only coherent alternative vision for Democrats to counter the oligarchic power grab of the Trump administration. It is compelling and doesn’t retread the tired identity politics that has led Democrats to defeat after defeat. Yet Abundance has not gone unchallenged and this critique can lead to a further, better development. For sure, I haven’t had the opportunity to review every critique, but I’d like to mention two. In “Move Fast and Break Things,” Jay Pinho argues that Abundance offers an “intoxicating vision” whose “central complaint rings agonizingly true” regarding governmental dysfunction, but ultimately fails to address the fundamental political questions at its core. Who decides what gets built? Whose abundance matters most? What democratic processes must be preserved even as we streamline decision-making? Pinho persuasively highlights the case of conservative venture capitalist Marc Andreessen, who publicly advocated for building more while privately opposing multifamily housing near his own property—illustrating the gap between abstract abundance advocacy and concrete political choices. “Here lies The Messy Real World,” he writes, suggesting that the same problems we identified in the Infrastructural City haven’t really gone away. But Pinho adds that the “the at-times painstaking slowness of government is itself the product of democratic responsiveness, of accountability.” Losing that accountability would be dangerous. Similarly, in “An Abundance of Credulity,” Hannah Story Brown criticizes Klein and Thompson for advocating governmental efficiency at the expense of democratic safeguards, arguing that they evade a crucial question: who benefits from scarcity? As Brown observes, infrastructural stalemates aren’t merely bureaucratic accidents but often serve powerful interests—fossil fuel companies profit from climate inaction and housing conglomerates thrive on constrained supply. This analysis resonates with our findings in The Infrastructural City, where we documented how infrastructure impasses reflect not just procedural failures but entrenched power dynamics. And like Pinho, she argues that the “Rules imposed over the last half-century that seek to prevent exploitation and safeguard the public have led to dramatically lower air and water pollution, significantly fewer auto and aviation fatalities, reduced mortality from infectious diseases, fewer deaths and injuries at work, fewer deaths from residential fires, fewer bank failures, and a less volatile economy. Some of us believe those are worthwhile trade-off.”⁶

These critiques illuminate what might be called Abundance’s democratic dilemma: how to enable decisive infrastructure development while preserving accountability and ensuring equitable outcomes. This is not merely a technocratic challenge but a fundamentally political one. Klein and Thompson correctly identify that America’s inability to build stems from policy choices rather than inevitable constraints, but their reluctance to confront the politics of those choices—who wins, who loses, and whose voice matters in the process—is a stark limitation of their approach.

Yet these critiques need not lead to pessimism about infrastructure’s future. Instead, we can take away from them the idea that addressing our infrastructural impasse requires not just procedural reforms but democratic renewal, something that is going to have to happen to rebuild the United States after, it increasingly seems, the Trump administration runs it into the ground. While streamlining permitting and boosting public investment are necessary components of an abundance agenda, they must be paired with strengthened accountability mechanisms and explicit attention to equity, less in terms of an endless cycle of identity politics grabs and more in terms of ensuring that the profits generated by this new policy do not flow ever upwards to the same billionaire oligarchy that has so supported Trump. More than that, it would be appropriate that a new Democratic platform should expressly task itself with a significant tax increase on the billionaire class—here we can follow Trump’s hallowed MAGA dictum of a return to the 1950s and impose a 91% marginal rate on individuals earning more than $2 million a year— while also establishing a graduated corporate income tax to create a more competitive market.⁷

More than that, any viable vision for infrastructural abundance must also confront a reality that both Klein and Thompson and their critics largely overlook: the coming demographic contraction that will fundamentally reshape our infrastructural needs and possibilities. While Klein and Thompson explicitly distance themselves from advocates of voluntary degrowth—those who propose deliberately shrinking economies and populations for ecological reasons or to preserve the character of their communities—they fail to address what we might call “actually-existing degrowth”: the unplanned, ongoing demographic decline already occurring across much of the developed world. Populations are already shrinking significantly in countries like Japan, South Korea, and most European nations, driven by persistently low fertility rates, aging populations, and changing social patterns. This phenomenon, increasingly evident even in the United States—where the total fertility rate is 1.62—signals a profound structural shift that will fundamentally reshape urban and infrastructural planning in coming decades, regardless of our political preferences. While immigration may keep America’s population growing for a time, that has its limits and is likely to slow considerably before Democrats take power due to the MAGA agenda.

As I have done before, I will again borrow novelist William Gibson’s evocative term “the Jackpot” to refer to this unfolding demographic transition. Rather than a sudden apocalyptic population collapse, the Jackpot describes a slower, distributed unraveling—a prolonged and uneven demographic downturn, intensified by climate stress, economic instability, and shifting cultural values. In the United States, signs of the Jackpot’s approach are increasingly clear: declining birthrates, shrinking rural and small-town populations, aging demographics, and regional depopulation. The hard truth is that most of these areas are not fixable. If infrastructure planning and political visions adopt the Abundance Agenda but fail to acknowledge this demographic reality, the country risks investing in futures that will never materialize, preparing for continued growth while confronting the steady reality of shrinkage.

Crucially, demographic contraction coincides with—and is amplified by—the rise of Artificial Intelligence. Rather than viewing AI merely as a threat stealing jobs, it will be embraced as a tool to ease the social and economic adjustments required by shrinking populations. AI’s widespread deployment is already reshaping labor markets, significantly reducing demand for traditional labor across industries. This shift could alleviate some economic pressures posed by a shrinking workforce, helping facilitate a smoother transition toward sustainable, high-quality urban life. Yet precisely how we harness AI while mitigating the profound disruptions likely from automation—especially the potential mass displacement in service and intellectual sectors previously insulated from the upheavals experienced by industrial labor during post-Fordism—is perhaps the most critical question we face regarding AI today. No matter how challenging this is for many progressives—who often view AI as deeply flawed or irredeemably captured by corporate interests—AI-optimized infrastructures, such as smart energy grids, autonomous transit systems, predictive healthcare networks, and intelligent urban management platforms, can help societies navigate demographic contraction efficiently and equitably. In this scenario, infrastructural abundance becomes redefined: not merely building more, but building better—investing in adaptive, intelligent infrastructures that enhance human and ecological well-being as populations decline.

However, integrating the Jackpot scenario into the abundance argument demands redefining abundance itself. Rather than pursuing endless quantitative expansion, infrastructure must become adaptive, resilient, and oriented toward ecological regeneration and urban livability. A future of smaller populations offers genuine opportunities: cities redesigned around quality of life rather than growth alone, restored ecosystems, and revitalized urban spaces characterized by abundant green infrastructure, sustainable energy systems, and human-scale design.

Yet achieving this vision faces political headwinds. The political Right portrays degrowth and adaptive urban strategies as part of a conspiratorial ‘Great Reset,’ framing necessary adaptations as threats to personal freedom, economic prosperity, and American cultural identity. This ideological stance complicates practical discussions about managed shrinkage by conflating sustainability with politically charged fears of elite control, making constructive bipartisan solutions harder to achieve. Yet a realistic reckoning must still occur in declining areas. Citizens need to be actively brought into the planning process, clearly addressing their understandable anxieties and explicitly answering the fundamental question: ‘What can we do to make things better for our communities as the population continues to fall?’ Nor is this a matter only for Blue states. The state with the largest population decline is West Virginia and no amount of Trump subsidies for coal will ever bring the miners back.

Fortunately, this strategy doesn’t require reinventing the wheel. Europe and Japan have developed methods for managing urban shrinkage. In the United States, Youngstown, Ohio developed a plan to explicitly acknowledge population decline, consolidating services and converting vacant lots into urban agriculture. Even though implementation remained limited due to persistent economic challenges, constrained municipal resources, and political resistance to fully abandoning growth-oriented strategies, the plan still represents an instructive American precedent in accepting shrinkage explicitly. Japan’s “compact city” policies, exemplified in Toyama, have strategically concentrated development around transit nodes, allowing peripheral zones to revert gradually to nature and creating more vibrant, walkable urban cores despite overall population decline. Detroit’s strategic framework similarly strives to establish higher-density neighborhoods surrounded by green infrastructure and innovative urban agriculture. These examples demonstrate how thoughtfully managed shrinkage can lead to more sustainable, livable urban environments. I have spent time in West Virginia. It is a beautiful state and I can see why people who live there don’t want to leave. But there is little reason that we can’t rethink it and make a large part of the state into an incredible national park that would be a worldwide draw. Even the coal mines can be recovered, turned into models for ecological reclamation.

Infrastructure after growth represents not a reduction of ambition but a recalibration of priorities toward genuinely sustainable abundance. America can break this deadlock by embracing the Abundance agenda—curtailing excessive governmental constraints and strategically collaborating with industry to advance technological innovation. Rejecting identity politics that frame society as composed of factions competing over an ever-dwindling pie, this vision instead offers tangible improvements in everyday life: lower costs of living, better public services, cleaner air and water, strengthened local economies, and greater accessibility. Yet confronting the emotionally appealing but misleading nostalgia of MAGA requires political leaders to reframe demographic contraction and technological transformation as opportunities rather than threats, enhancing our quality of life and the ecological sustainability of the country. The path forward lies in reconciling three critical dimensions: the technical challenge of building more efficiently, the political challenge of ensuring equitable distribution and meaningful democratic participation, and the demographic challenge of adapting to population contraction. Ultimately, resolving the infrastructural impasse identified nearly two decades ago demands not only wise policies but compelling voices—like Klein and Thomposon’s—capable of articulating a credible, hopeful vision of ecological restoration, social renewal, and enduring resilience.

This integrated approach suggests a vision of abundance quite different from the endless expansion of the modernist era and that is, unfortunately, something lacking from Abundance at present, but it can be part of the Abundance agenda. Rather than merely building more of the same infrastructures that characterized 20th-century development, we must build better—creating adaptive, resilient systems that enhance human well-being and ecological health as populations potentially decline. This vision of abundance embraces quality over quantity, recognizing that democratic accountability and ecological sustainability are not impediments to progress but essential components of truly meaningful advancement.

^1. Kazys Varnelis, “Introduction. Networked Ecologies,” The Infrastructural City: Networked Ecologies in Los Angeles (Barcelona: Actar, 2008), 6-16. ↩

^2. Reyner Banham, Los Angeles: The Architecture of Four Ecologies (London: Allen Lane, 1971) ↩

^3. Edward W. Soja, Postmodern Geographies: The Reassertion of Space in Critical Social Theory (New York: Verso, 1989), 191. ↩

^4. See Christopher Hawthorne, “What’s the Future of ‘The Infrastructural City of L. A.,” The Los Angeles Times, February 15, 2009, https://www.latimes.com/entertainment/la-ca-infrastructural-city15-2009feb15-story.html. Harvey Molotch, “The City as a Growth Machine: Toward a Political Economy of Place,” American Journal of Sociology, vol. 82, no. 2, (September 1976), 309-332. https://web.ics.purdue.edu/~hoganr/SOC 602/Spring 2014/Molotch 1976.pdf ↩

^5. Ezra Klein and Derek Thompson, Abundance (New York: Avid Reader Press, 2009), 17. ↩

^6. Jay Pinho, “Move Fast and Break Things,” Networked Substack, April 1, 2005

networked

Move fast and break things

Throwing out preclearance when it has worked and is continuing to work to stop discriminatory changes is like throwing away your umbrella in a rainstorm because you are not getting wet…

Read more

a year ago · 3 likes · 8 comments · Jay Pinho

and Hannah Story Brown, “An Abundance of Credulity,” The American Prospect, March 26, 2025, https://prospect.org/culture/books/2025-03-26-abundance-of-credulity-klein-thompson-dunkelman-review/ ↩

^7. Kimberly Clausing, “Combating Market Power Through a Graduated U.S. Corporate Income Tax,” Washington Center for Equitable Growth, April 15, 2024, https://equitablegrowth.org/combating-market-power-through-a-graduated-u-s-corporate-income-tax/ ↩

My friend Adam Greenfield recently articulated a similar sentiment on his Patreon:

Maybe it’s simply the onset of British Summer Time overnight, but I’m just bursting with energy. I know it seems incongruous with the ambient psychic weather of the moment, but then maybe that’s the point? To meet the grim farce of mainstream public affairs with an upwelling, irrepressible, literally insurgent joy? To keep at it, generating connection and possibility and the conditions of life, until the very moment the choice to continue doing so is in one way or another taken out of your hands. There are worse programs to commit oneself to, you know? (link)

Well said. My recent pace of writing starkly contrasts with my paralysis during the first Trump administration. Although I had quit full-time teaching in 2015 to focus on writing and art practice—and despite a wildly successful 2016 highlighted by the Detachment exhibit in Vilnius—after the election, I stumbled. For three years, I barely wrote, turning inward instead, absorbed by the restoration of our house. Then, just as I laid plans for a new push, COVID hit. My reserves were already empty and another year was lost. Those years betrayed the promise I had made to myself when stepping away from teaching—to finally bring forth the work I had long conceived but never had time to produce. I had all the time in the world and what came out of it? This hiatus nearly ended my career. I had surrendered to circumstance, letting external forces dictate my creative life. But in the unexpected and renewed face of a new Trump regime, I have a different response this time: “What do we say to the God of Death? Not today.” As Adam suggests there's true power in meeting grim circumstances with "insurgent joy" generating possibility, not just watching as “the darkness drops again.” And so, onwards.

You should subcribe to Adam’s Patreon. It’s good. We disagree on some things, like the state of AI today, but we agree on many others. Discourse, dialogue, and debate are what we need today, not arm-waving from censors and hard liners of all stripes. As Yeats also observed in “The Second Coming,” we cannot allow a situation where "the best lack all conviction, while the worst are full of passionate intensity."

Second, to paraphrase Star Wars, many AIs died to bring you this message. I uploaded an earlier version to ChatGPT to get feedback and look for typos (ChatGPT and Claude make decent, albeit not perfect, copy editors). Soon after I did, my account was banned for “advocacy of sexual violence.” What? I condemn sexual violence, had my account been hacked? Then I remembered the quote from Rosalind Krauss below about the “treat of castration.” We have a long way to go before we get to real artificial intelligence, apparently.

Finally, as always, the only thing I ask of you is to pass my Substack on to other folks you think might be interested in it. Or send them to my varnelis.net. This post is at https://varnelis.net/works_and_projects/the-new-surrealism-on-ai-hallucinations/
The footnote links actually work there and the images are larger, so it may prove a more rewarding reading experience.

In the summer of 1936, Salvador Dalí appeared before an audience at the International Surrealist Exhibition in London, clad in a diving suit, to deliver a lecture. Dalí intended, he later claimed, to illustrate his plunge into the depths of the human unconscious. He soon, however, faced a severe crisis when the suit’s air supply malfunctioned, nearly suffocating him. Assistants used a billiard cue to pry off the helmet, and Dalí proceeded to present his slide show, albeit with slides projected upside down. For Surrealists, such moments of rational collapse revealed pathways into unexpected creativity. Hallucination, error, and confusion allowed them to reach beyond logic or convention. Dalí’s near-suffocation was thus not a failure but a triumph, exposing the fragility of conscious control.

Nearly a century later, our contemporary era of Artificial Intelligence suggests Surrealism’s lessons may still be relevant. Anyone who has interacted at length with an AI language model has encountered its confabulations, fake links, spurious citations, and outright falsehoods—errors the industry euphemistically terms “hallucinations.” We ask AI for a straightforward answer, and it responds with a confident blend of truth and fiction. Many AI skeptics, along with people who tried early versions of LLMs and never returned, dismiss AI as fatally flawed because of these hallucinations. But hallucinations have existed long before AIs, indeed, they are everywhere we look.

Consider Dalí’s diving suit performance: as a historian of architecture, my grasp of surrealism is limited. In preparing this essay, I requested ChatGPT to identify compelling examples to open the essay with, and the AI highlighted Dalí’s incident. A web search for more details delivered the 2016 Guardian article “Dalí in a diving helmet: how the Spaniard almost suffocated bringing surrealism to Britain.” This is a random event in history. And yet, Dalí’s autobiography, The Secret Life of Salvador Dalí, tells a subtly different story: the artist says nothing about his air supply failing. Instead, he states that the lead shoes of the suit were extremely heavy and walking to the microphone to give the lecture was extremely difficult and tiring. He claimed he felt “faint and on the point of suffocating” and waved to his wife Gala and assistant Edward James to help him. They used the billiard cue to cut a slit between the helmet and the suit so he could breathe, then brought a hammer to knock off the bolts affixing the helmet to the suit. ¹ The version of Dalí’s diving suit story described by ChatGPT—and echoed in The Guardian—was itself a hallucination that somehow became accepted as historical fact. It is, however, unclear where it came from. 2016 is long before LLMs were capable of being used as we use them today. Was it the product of a hastily written article? Is it a poorly-remembered first person account? Did Dalí embellish the moment himself in his autobiography? Did someone fabricate it for some purpose? Or did it just emerge, as things do?

I have long understood that the writing of history, even serious history, is a game of telephone. Memory, perception, and imagination intermingle, creating convincing yet fictitious narratives that shape beliefs, literature, and even history. Consider these iconic stories and quotes: George Washington’s cherry tree, Marie Antoinette’s notorious “let them eat cake,” or Voltaire’s defense of free speech, “I disapprove of what you say, but I will defend to the death your right to say it.” None of these individuals ever uttered these words. Such myths persist precisely because they reveal deeper cultural truths; they provide narratives we need to interpret and understand our world.

Hallucinations are surprisingly common in academic writing. I was inspired to write this essay in part because of my recent experience conducting research on early accounts of northeastern America and its beauty. Since I thought I knew where to turn, I was not using AI, instead I started with my bookshelf. In one of my favorite books, Jackson Lears’ Fables of Abundance, I found a compelling quote: “John Speede, in his Historie of Great Britain (1611), celebrated Oriana (the New World) as ‘the Court of Queen Ceres, the Granary of the Western World, the fortunate Island, the Paradise of Pleasure, and the Garden of God.'” And yet, Lears is mistaken. Speed—I am not sure where the extra “e” in “Speede” comes from—is clearly referring to Great Britain in the original. ² Next, I looked at The Hudson. A History by Tom Lewis. The author cited a provocative passage by the early New Netherlander Adrian van der Donck, a Dutch lawyer and advocate for democracy, who, in his 1655 Beschryvinge van Nieuw-Nederlant (Description of New Netherland) wrote: “I admit that I am incompetent to describe the beauties, the grand and sublime works, wherewith Providence has diversified this land.” But Lewis, whose book was published in 2005, relied on a faulty, highly embellished translation from 1841. In a 2008 translation the line reads “I pray the indulgent reader to deduce from the above how fertile this land is and form his own judgment; as to myself, I confess to being unable to depict it or show it in writing, since in my view the eye alone, more so than the ear, is capable of comprehending it.” I thought it wise to double check so I quickly retrieved a scan of the 1655 text online and ran an excerpt through ChatGPT, which recognized the Dutch Blackletter text and suggested “I freely admit that I am not capable of portraying it fully or representing it in writing, as our judgment relies only on sight, and cannot assure the heart of its truth.” Not as elegant as the human translation from 2008, but confirming that the 1841 translation was embellished. I can’t fault Lewis for using the only translation available to him, it was the original embellishment that drew us both into the original reference. Unfortunately now it was useless for me. ³

Alas, I too, am no innocent in this. Take Blue Monday: Stories of Absurd Realties and Natural Histories, a book I know well, having co-authored it with Robert Sumrell. In explaining the title of the book, we refer to the pioneering post-punk band New Order’s 12″ “Blue Monday”:

In “Blue Monday,” the band achieved phenomenal media success, creating the most popular single of all time. But in their desire to become more digital–and hence more immaterial–than actually possible at the time, New Order retained graphic designer Neville Brody to make a die cut cover that would resemble the sleeve of a large floppy disk. The unique look won critical acclaim, but according to legend the most popular 12″ of all time cost the band 20 cents for every copy sold, ruining them financially but assuring their place in the regime of media. ⁴

This passage is the crux of the book, explaining the title and our collective drive to become more digital and immaterial. Except that I got it wrong. In a last editing pass, some neuron misfired and I substituted Neville Brody for the correct designer, Peter Saville. Robert let me know the moment he saw it, but it was too late, the book had been sent to the printers. As for the story about the financial losses, it may not be true either. While Factory records head Tony Wilson recounted it, he was known for exaggeration as well as poor record keeping.⁵

Sometimes, scholarly documents are fabricated out of thin air. Decades ago, I was in a graduate seminar on Renaissance urbanism at Cornell. I was assigned to explain the rebuilding of the town center of Pienza, the birthplace of Aenias Silvius Piccolomini, the humanist Pope Pius II. I recalled seeing marvelous plans of the site in Spiro Kostof’s introductory textbook A History of Architecture: Settings and Rituals that showed the town before and after the intervention. I went on a lengthy paper chase, trying to figure out where these authoritative diagrams came from. There were no citations in the textbook and no credits for the plans. In the end, my hunch was that the author had a class in which students hypothesized what a pre-intervention condition might have looked like, produced plans, and included them in the book without noting that they were hypothetical reconstructions based on scant evidence.

Outright forgeries play a considerable role in history. In 1440, philologist Lorenzo Valla examined the Donation of Constantine, a fourth-century imperial decree supposedly granting sweeping temporal power to the Pope. The papacy cited it as their claim to power; kingdoms accepted it; scholars taught it as fact. But when Valla examined the Latin, he uncovered anachronisms impossible for Constantine’s time. The document was a forgery. Europe had built an entire political order on a historical hallucination. James Macpherson’s The Poems of Ossian took 18th-century Europe by storm, captivating Goethe and influencing Romanticism. Macpherson attributed these poems to Ossian, an ancient Gaelic bard. Only later did it become clear that Macpherson himself had largely composed these works. Or take the infamous Hitler Diaries, published in 1983 and swiftly authenticated by historian Hugh Trevor-Roper, only to collapse spectacularly under scientific scrutiny weeks later. Both examples show that scholarly hallucinations thrive not just on error, but on collective desires: readers hungry for heroic national pasts or sensational scoops. Where history isn’t enough, it seems we need to invent it.

Museums certainly aren’t immune to this. Vast amounts of museum collections are composed of fakes and forgeries. In 2014, Switzerland’s Fine Art Expert Institute (FAEI) reported that at least half of the artwork circulating in the market is fake. Noah Charney, an art historian and founder of the Association for Research into Crimes Against Art, contends that approximately 95% of artworks displayed in museums are accurately attributed, suggesting that only about 5% may be misattributed or counterfeit. Still, things happen. Take, for example, the case of the Kröller-Müller House, which Ludwig Mies van der Rohe proposed. The story goes that Mies built a full scale mockup out of wood and canvas. It marks his first mature attempt to dismantle the traditional notion of the enclosed house and anticipate the open plan and fluid space that would define his later work. The house is exhibited in Philip Johnson’s 1947 Mies exhibition and reproduced in the catalog. Rem Koolhaas, in S, M, L, XL, however, says “What was weird was that when I asked Philip Johnson about the incident last year [around 1992], he said he had invented it. According to him, it had never happened. The photograph of the phantom house was a fake, he suggested. But who faked here? Whose fata morgana was this anyway?” Indeed, after looking at the image, it seems unimaginable that it isn’t simply drawn onto a landscape, a bit of mythologizing by Johnson. Perhaps, however, the fake was a fake. A web site for the Kröller_Müller Museum shows another image of the design, likely from their archive, however, this time the house looks far more real and the overlap of the tree branches seems hard to duplicate for Johnson, who was never much of a draftsman. Is the image from 1947 just taken from another angle? There seems to be a reasonable similarity in the trees in the background. Did Johnson perhaps draw in the more modernist wing on the right? Did he fabricate—whether due to poor memory or perhaps just capriciousness—the fabrication? Or perhaps Koolhaas, in search of a good story, came up with this.⁶

Unlike errors in print media, which remain localized to physical copies, digital mistakes can replicate across global networks, gaining authority through algorithmic amplification and citation indexing. Consider Rosalind Krauss’s Optical Unconscious, where a Freudian slip of the keyboard produced the phrase “treat of castration” instead of “threat of castration,” now faithfully reproduced in the MIT Press publication, a mistake that a typesetter would have noticed. ⁷ Or take the citation for “(Van der Geer J et al., 2010. The art of writing a scientific article. Journal of Science Communication 163(2) 51 – 59),” which has been cited in over 1500 academic papers to date according to Google Scholar. The essay, however, does not exist; it was a fictional example that escaped from a formatting template created by publishing giant Elsevier. ⁸ Scholars either forgot to delete the reference or cited the imaginary article without verifying it existed, inadvertently creating a collective hallucination propagated through databases and citation indices. Here, the parallel to AI hallucinations becomes particularly stark: scholars, like language models, confidently repeated nonexistent references simply because they seemed plausible. The ghostly paper highlights how easily we trust shared authority without scrutiny—an important lesson as we enter an era dominated by generative AI.

These errors—and countless others—reveal something profound. Human minds don’t merely seek truth; they also create it, often unconsciously bending evidence to fit narratives. The persistence of these illusions can distort entire disciplines, shaping how we perceive history, culture, and ourselves. Yet the errors also remind us how intimately creativity, wishful thinking, and factual missteps intertwine. Like surrealists embracing the subconscious or AI engines extrapolating beyond known facts, human culture thrives not only despite these hallucinations—but perhaps because of them.

The typical explanation given for AI hallucinations is that at heart, today’s large language models—systems like GPT-4, Claude, and Gemini—are sophisticated pattern-recognition engines. Imagine a hyper-literate parrot that’s read every book in the library, capable of stitching together elegant sentences that sound authoritative, even insightful. But this parrot has no understanding of truth, context, or intent. Unlike traditional computing, which follows deterministic logic, LLMs operate probabilistically. Given the words “once upon a,” they will predict “time,” based on patterns learned from vast datasets. With a prompt like “In 1905, Einstein published a paper on,” the AI assesses billions of textual examples to choose the most statistically probable continuation (“special relativity”). But if the question ventures into obscure territory—a minor historical figure, niche cultural references, or poorly documented events—the model, optimized to provide helpful answers and rarely encouraged to say, “I don’t know,” will produce fiction to satisfy the prompt. The best-performing models, like GPT-4 or Claude 2, have been explicitly trained to recognize their limits and occasionally decline answering uncertain prompts, reducing—but not eliminating—fabrications. When a model makes up a hallucination, it’s akin to a human trying to sound knowledgeable at a cocktail party, confidently making up facts about obscure topics. AI mimics our own tendency toward myth-making.

But AI hallucinations aren’t just bugs; they’re symptoms of the underlying generative capability we value in them. If models rigidly stuck to memorized facts, they’d lose their remarkable ability to generalize, summarize, and invent. The same predictive flexibility enabling hallucinations allows AI to creatively interpret tasks—composing narratives, suggesting innovative ideas, or exploring hypothetical scenarios. Eliminating all hallucinations risks overly conservative models, timid and limited, good at trivia but poor at imagination.

But further, in December 2023, Andrei Karpathy, noted AI researcher formerly at OpenAI and Tesla, explained as follows.

# On the “hallucination problem”

I always struggle a bit when I’m asked about the “hallucination problem” in LLMs. Because, in some sense, hallucination is all LLMs do. They are dream machines.

We direct their dreams with prompts. The prompts start the dream, and based on the LLM’s hazy recollection of its training documents, most of the time the result goes someplace useful.

It’s only when the dreams go into deemed factually incorrect territory that we label it a “hallucination”. It looks like a bug, but it’s just the LLM doing what it always does.

At the other end of the extreme consider a search engine. It takes the prompt and just returns one of the most similar “training documents” it has in its database, verbatim. You could say that this search engine has a “creativity problem” – it will never respond with something new. An LLM is 100% dreaming and has the hallucination problem. A search engine is 0% dreaming and has the creativity problem.

All that said, I realize that what people actually mean is they don’t want an LLM Assistant (a product like ChatGPT etc.) to hallucinate. An LLM Assistant is a lot more complex system than just the LLM itself, even if one is at the heart of it. There are many ways to mitigate hallucinations in these systems – using Retrieval Augmented Generation (RAG) to more strongly anchor the dreams in real data through in-context learning is maybe the most common one. Disagreements between multiple samples, reflection, verification chains. Decoding uncertainty from activations. Tool use. All are active and very interesting areas of research.

TLDR I know I’m being super pedantic but the LLM has no “hallucination problem”. Hallucination is not a bug, it is the LLM’s greatest feature. The LLM Assistant has a hallucination problem, and we should fix it.⁹

In framing AIs as dream machines, Karpathy offers a provocative reimagining of a generative process akin to human creativity, in which meaning emerges through unforeseen associations, productive mistakes, and spontaneous invention.

Indeed, planned and unplanned deviations, mistakes, and hallucinations are a productive part of the creative process. In The Anxiety of Influence, literary theorist Harold Bloom argues that creative misreading is essential to literary evolution. Strong poets, he suggests, achieve originality by creatively misreading their literary predecessors. Bloom terms this “poetic misprision,” a strategic misinterpretation enabling poets to clear imaginative space within a saturated literary tradition. Misprision isn’t accidental—it’s a necessary act of rebellion, distancing poets from overwhelming influence and allowing them to develop their distinctive voices. This perspective reframes error not as failure but as generative necessity—what might appear as misunderstanding becomes the very foundation of creative innovation. Bloom demonstrates this through Milton’s deliberate misreading of classical tradition, transforming heroic narratives into complex psychological struggles in Paradise Lost, and through Wordsworth’s strategic reinterpretation of Milton’s elevated political voice into personal, introspective meditation. These creative distortions weren’t merely mistakes—they were essential evolutionary mechanisms that allowed new literary forms to emerge from existing traditions. Seen through Bloom’s framework, AI hallucinations might similarly represent not just errors to be corrected but potentially productive misreadings that open unexpected creative pathways beyond human convention. ¹⁰

Jacques Derrida, the founder of deconstruction, also recognized creative power in apparent errors, accidental fragments, and seemingly trivial textual moments. In Spurs: Nietzsche’s Styles, Derrida famously analyzes a cryptic note discovered among Friedrich Nietzsche’s unpublished papers: “I have forgotten my umbrella.” Though seemingly inconsequential, this incidental sentence—an isolated fragment without context—opens new interpretive possibilities precisely through its ambiguity. Is the umbrella real or metaphysical? By resisting stable context and defying conventional reading, the phrase unsettles assumptions about coherent authorial intention. Derrida thus transforms Nietzsche’s trivial notation into a philosophical meditation on memory, forgetfulness, and textual uncertainty. The supposed error or accidental remark becomes generative precisely because it escapes closure, demonstrating that textual significance can emerge from absence, incompleteness, or apparent nonsense. This illustrates Derrida’s broader philosophical argument: meaning never resides simply in an author’s deliberate intent or in textual clarity alone but arises dynamically through interpretive engagement with ambiguity, uncertainty, and textual rupture. ¹¹

The sciences, too, have flourished through productive error. The 18th-century phlogiston theory in chemistry—which proposed a non-existent element released during combustion—was entirely wrong, yet philosopher Thomas Kuhn noted how this incorrect paradigm “gave order to a large number of physical and chemical phenomena,” allowing scientists to organize observations until contradictions eventually led to oxygen theory. ¹² Similarly, Johannes Kepler’s quasi-mystical belief that planetary orbits followed the geometry of nested Platonic solids drove him to analyze Mars’s orbit so obsessively that he discovered elliptical orbits and his three laws of planetary motion. Perhaps the most striking example comes from Albert Einstein. His cosmological constant—which he introduced to stabilize his equations because he mistakenly believed the universe was static and later abandoned, famously calling it his “biggest blunder”—has been essential for explaining dark energy and cosmic acceleration. The error contained within it a profound truth, one that Einstein himself couldn’t recognize in his lifetime. But the thing is that Einstein never actually called it his “biggest blunder”; the phrase was introduced in the autobiography of physicist George Gamow, who is notorious for embellishing and fabricating facts. ¹³

But these are outright errors. There are also cases in which scientists have taken their dreams and acted on them. René Descartes had three intense dreams that convinced him he should question everything he thought he knew, starting fresh from just one clear truth: that because he was thinking, he must exist. This approach—checking every belief carefully and trusting only what’s completely certain—became the starting point for modern science and philosophy. Chemist August Kekulé famously envisioned benzene’s ring structure after hallucinating a serpent swallowing its tail. Nobel Prize winner Otto Loewi dreamed of an experiment that would prove that nerve impulses were transmitted chemically, leading to the discovery of neurotransmitters. Finally, the periodic table of elements came to Dmitri Mendeleev in a dream. Except, in this case, once again, it didn’t happen. The first mention of the dream is forty years later. ¹⁴

Now, hallucinations in AI assistants are, as Karpathy states, a target of considerable investigation by AI labs and dramatic strides have been made to minimize their occurrence since the launch of ChatGPT-3.5. But LLMs are still, as Karpathy states, dream machines. Whether reading human or AI generated text, we need to remember the lost skill of critical reading and checking one’s sources. But if the progress goes too far, one day we may regret the loss of hallucinations.

For now, however, we still have AI hallucinations. And if, after reading all this, history seems like a tissue of lies, perhaps turning to the dreams of a machine for new ideas isn’t the worst thing to do? Turning back to the 1920s, surrealism sought to rupture conventional thinking by deliberately invoking dreams, chance, and subconscious imagery. Nearly a century later, a new creative practice—what we might call “the New Surrealism”—emerges from the collaboration between human imagination and AI hallucination. If LLMs are dream machines, their hallucinations are the computational equivalent of surrealist automatic writing—drawing connections across vast semantic fields without regard for factual constraints. This too can reveal new worlds hidden just beyond the edges of conventional thought in the collective unconscious. Is it coincidence that one of the hottest recreational trends in Silicon Valley’s AI scene is consuming magic mushrooms?

Take my friend architect and media theorist Ed Keller’s approach. Ed has described how for a few years now he has immersed himself nightly, often while drifting off to sleep, in working with AI image generations to create hallucinatory images deeply infused with architectural history, mythology, and ecological reflection. His invocation of Daphne—a figure of myth who transforms from human to tree—symbolizes humanity’s forced adaptation to the powerful external forces of artificial intelligence and the attendant ecological crisis. Ed’s creative method is surrealist in essence, blurring conscious and unconscious thought through a collaboration between human intent and machinic hallucination. Just as the original surrealism grappled with the unconscious mind mediated by industrial modernity, the New Surrealism emerges from dream-like dialogues with intelligences over the net, further reshaping human identity and perception.

My own studio work with AI image generators always engages with their hallucinations as creative contributions. Through extensive sessions with generative systems such as Midjourney or Google ImageFx, my practice critically engages with AI not as passive technology but as an active collaborator whose misunderstandings and apparent failures prompt me to develop the work in new directions.

In doing so, I uncover insights, exposing the underlying assumptions of authorship, originality, and authenticity embedded in our cultural frameworks. My projects, On an Art Experiment in Soviet Lithuania, Lost Canals of Vilnius, The Destruction of Doggerland, The Witching Cats of New Jersey, and Pierre Lecouille, Visionary Architect all take the guise of fictional histories. I should not have been surprised when an image from the first was poached and reposted without attribution by a viral poster, credits to me replaced by the words, “Not AI.” The two projects that I did not create fictional histories for, 20 subroutines for Humans Made By a Computerand Art and the Boxmaker addressed chance in the work of John Cage and the surrealist art of Joseph Cornell more directly. Just as early surrealism negotiated the unconscious mind reshaped by industrial modernity, my critical engagement with AI creates a contemporary surrealism born from the evolving dialogue between human creativity and networked, machine intelligence, continuously reframing identity, meaning, and artistic practice itself. Embracing rather than rejecting AI’s hallucinatory tendencies can transform creative practices, but it requires a serious effort, not merely a quick generation of political figures in the style of Studio Ghibli for viral posting.

In time, the New Surrealism may shape not only art but cultural criticism, literature, and education. We should teach students to recognize AI hallucinations not as defects but as invitations to critical inquiry and creative exploration. Scholars could intentionally leverage these errors to illuminate hidden assumptions or generate new interpretive frameworks. Our task, then, will not be to eliminate its hallucinations entirely but to thoughtfully manage, curate, and even nurture them. AI image generator Midjourney, for example, has a weird setting in which you can dial up its inventiveness—and rate of hallucination—as well as the ability to choose one’s model, including very old models that malfunction brilliantly.

André Breton described surrealism as “pure psychic automatism”—an attempt to bypass rationality by embracing random associations and visions. AI, guided by neural networks rather than neurons, performs a similar act of psychic automatism, weaving patterns without conscious control or intent. It generates surreal poetry, images, and narratives, sparking both anxiety and fascination. AI’s persistent hallucinations represent not only flaws but also opportunities to revisit surrealism’s radical experiments.

^1. Salvador Dalí, The Secret Life of Salvador Dalí (New York: Dover Books, 1993), translated by Haakon Chevalier, 345.
↩

^2. Jackson Lears, Fables of Abundance:A Cultural History of Advertising in America (New York: Basic Books ,1994), 27. The original text can be found online in text form at “The history of Great Britaine under the conquests of ye Romans, Saxons, Danes and Normans. Their originals, manners, warres, coines & seales: with ye successions, lives, acts & issues of the English monarchs from Iulius Cæsar, to our most gracious soueraigne King Iames. by Iohn Speed.” In the digital collection Early English Books Online. https://name.umdl.umich.edu/A12738.0001.001. University of Michigan Library Digital Collections. A scanned document is available at The History of Great Britaine Vnder the Conquests of Ye Romans, Saxons, Danes and Normans. Their Originals, Manners, Warres, Coines & Seales: with Ye Successions, Liues, Acts & Issues of the English Monarchs from Iulius Cæsar, to Our Most Gracious Soueraigne King Iames (London: John Sudbury & Georg Humble, 1614), https://www.google.com/books/edition/The_History_of_Great_Britaine_Vnder_the/L9DE_ER5tAsC?hl=en&gbpv=0.
↩

^3. Russell Shorto, “Foreword,” in Adriaen van der Donck, A Description of New Netherland, edited by Charles T. Gehring and William A. Starna, translated by Diederik Willem Goedhuys. (Lincoln: University of Nebraska Press, 2008), ix.
↩

^4. AUDC [Robert Sumrell and Kazys Varnelis], Blue Monday. Stories of Absurd Realities and Natural Histories(Barcelona: Actar, 2008), 80.
↩

^5. Tonino Cannucci, “HOW WE MADE: NEW ORDER’S GILLIAN GILBERT AND DESIGNER PETER SAVILLE ON BLUE MONDAY,” Disorder and Other Unknown Pleasures, https://disordertc.wordpress.com/2016/03/10/how-we-made-new-orders-gillian-gilbert-and-designer-peter-saville-on-blue-monday/
↩

^6. On fake art in museums see “Over 50 Percent of Art is Fake,” ArtfixDaily, October 14, 2014, https://www.artfixdaily.com/news_feed/2014/10/14/7319-experts-claim-fifty-percent-of-artwork-on-the-market-is-fake and Sarah Cascone, “50 Percent Art Forgery Estimate May Be Exaggerated… Duh,” Artnet News, October 20, 2014, https://news.artnet.com/art-world/50-percent-art-forgery-estimate-may-be-exaggerated-duh-137444 For the Kröller-Müller incident, see Rem Koolhaas, “the House that Made Mies,” S, M, L, XL. (New York: Monacelli Press, 1995), 62-63. See also Kröller-Müller Museum, “Ellenwoude, A Museum House of Wood and Sailcloth,” https://krollermuller.nl/en/timeline/ellenwoude-a-museum-house-of-wood-and-sailcloth
↩

^7. Rosalind Krauss, The Optical Unconscious (Cambridge, MA: The MIT Press), 172.
↩

^8. Gareth Leng and Rhodri Ivor Leng, The Matter of Facts: Skepticism, Persuasion, and Evidence in Science(Cambridge, MA: The MIT Press, 2020), 205.
↩

^9. Andrei Karpathy, (@karpathy), “# On the ‘hallucination problem’,” X.com, December 8, 2023, 8:35pm, https://x.com/karpathy/status/1733299213503787018.
↩

^10. Harold Bloom, The Anxiety of Influence (New York: Oxford University Press, 1973).
↩

^11. Jacques Derrida, Spurs: Nietzsche’s Style (Chicago: University of Chicago Press, 1979).
↩

^12. Thomas Kuhn, The Structure of Scientific Revolutions (Chicago: University of Chicago Press, 1962), 99.
↩

^13. Ernie Tretkoff, “February 1917: Einstein’s Biggest Blunder,” APS News, This Week in Physics History, July 1, 2005, https://www.aps.org/apsnews/2005/07/february-1917-einsteins-biggest-blunder and Rebecca J. Rosen, “Einstein Likely Never Said One of His Most Oft-Quoted Phrases,” The Atlantic, August 9, 2013, https://www.theatlantic.com/technology/archive/2013/08/einstein-likely-never-said-one-of-his-most-oft-quoted-phrases/278508/.
↩

^14. Philip Ball, “The true story of the birth of the periodic table, 150 years ago,” The New Scientist, February 26, 2019, https://www.newscientist.com/article/mg24132190-300-the-true-story-of-the-birth-of-the-periodic-table-150-years-ago/.
↩

Like many of you, I have been trying to make sense of the current political landscape in the US and elsewhere, but I have also have been working on updating my argument about Network Culture—a historical phase marked by digital networks and neoliberal economics—and exploring what comes next. The result is this new essay on National Populism, not merely as a political ideology but as a transitional form emerging during capitalism’s shift from neoliberalism to what I term “AI Capitalism,” in which artificial intelligence becomes central to economic activity. There aren’t references to art or architecture here—those will come later, once the cultural outlines of this moment become clearer.

My interest is in how National Populism serves as a stopgap amid profound economic transformation. Populist movements promise restored manufacturing, national sovereignty, and cultural revival, yet in practice their policies accelerate financialization and deregulation, ironically undermining these promises. The tragic irony is that the National Populist base itself suffers the most: their legitimate grievances about economic dispossession are channeled into a movement that intensifies rather than resolves their obsolescence.

I don’t have the time to figure out how to make the footnote links work on Substack so they link to my site. You can just scroll to the bottom of this piece if you want to read them.

As always, this work can also be found at https://varnelis.net/works_and_projects/national-populism-as-a-transitional-mode-of-regulation/

For two decades, beginning at SCI_Arc and concluding at Columbia’s Graduate School of Architecture, Preservation, and Planning, I taught a course called Network City. In that course, I employed the Regulation School’s framework of regimes of accumulation and modes of regulation to explore urban transformations accompanying the shift from Fordist mass production to post-Fordist flexible specialization. Through case studies such as the Park Avenue business district in 1950s New York, the shopping mall, the suburban office park, the so-called “creative city” movement, the Bilbao-effect, and the dot.com workplace we examined the interplay of economic structures, governance models, and spatial organization underlying contemporary networked urbanism. Since the 2024 election, several former students and colleagues have reached out, expressing frustration as they try to make sense of the changes underway today. This prompted me to reflect at length on the phenomenon of Trumpism—better known as MAGA, or more appropriately, National Populism.

Although my focus here is on the American manifestation of National Populism, which makes global news daily, this phenomenon extends beyond the United States, appearing in countries like Brazil, Argentina, Hungary, and India. Between 1990 and 2018, the number of populist leaders in power worldwide jumped fivefold (from 4 to 20), reflecting a global shift in political dynamics.¹ Globally, 2024 was a bad year for most incumbents, but right-wing populism maintained its strength overall, gaining ground in countries like Germany and France. No longer an outlier, such parties are now firmly part of the political landscape.² While China’s authoritarian populism differs substantially in structure and goals, it too reflects the global turn toward reactionary politics, with the Chinese Communist Party making more appeals to national identity and cultural restoration similar to other populist movements. Unlike my Network City project, I am not going to use this essay to write about the architectural and urbanistic manifestations. There is, as yet, no clarity to these—and I am still working on interpreting its cultural logic—but the general contours of ongoing societal shifts have become clearer. I should add that while I am increasingly wary of academic narratives in which capitalism is presented as the chief animating force of history, it only makes sense to talk about capitalism when we are talking about socioeconomic developments.

Contemporary progressivism struggled to offer a compelling vision capable of addressing the economic and social anxieties fueling populist resentment. Indeed, its efforts seem to have backfired, with the progressive politics of the Biden administration—elected to be a moderate caretaker administration—driving the Democrats further and further away from their core constituency while alienating many among the minorities Democrats sought to court with their vision of identity politics. This ideological failure created an opening that National Populism eagerly exploited.

While analyses of National Populism in America commonly emphasize its political dimensions in America—nationalism, isolationism, and identity politics—its economic role as a mode of regulation has generally been overlooked. I argue that National Populism should be understood not merely as a political ideology but as a transitional regulatory mechanism emerging during capitalism’s shift from neoliberalism toward a new economic order dominated by artificial intelligence—what I term “AI Capitalism.” Even if this transformation remains largely unconscious—it is hard to believe that Trump himself has any grasp of its full dimensions and certainly his rank-and-file supporters do not—and poorly theorized, it is well underway and will inevitably provoke profound conflict in the years ahead, not just between Republicans and Democrats, but within the GOP as well.

Before proceeding further, I should clarify what I mean by regimes of accumulation and modes of regulation, two concepts central to the French Regulation School’s approach to political economy, made popular in the United States by David Harvey’s Condition of Postmodernity.³ A regime of accumulation refers to the way capitalism organizes production and consumption to enable profit and growth over an extended period. It encompasses specific technologies, labor processes, consumption patterns, and capital-labor relations that allow value creation and capture. The Fordist regime of accumulation, for example, featured mass production, unionized labor, rising wages, and mass consumption, while the flexible accumulation regime that followed emphasized global supply chains, precarious employment, and debt-fueled consumption.

A mode of regulation, by contrast, comprises the institutional forms, networks, and norms that stabilize a given regime of accumulation. This includes everything from formal laws and financial systems to cultural practices and modes of state intervention. The Keynesian welfare state was the mode of regulation that supported Fordism, while neoliberalism and the attendant monetary policy ultimately emerged as the regulatory framework for flexible accumulation. As Harvey concludes, broader cultural logics correspond to these modes, characterized by modernism and postmodernism respectively. In each case, the mode of regulation attempts to temporarily manage capitalism’s inherent contradictions and crisis tendencies through specific institutional arrangements.

In “Almost Anything,” my essay on the work of architect Kevin Roche that modernism had three main cultural phases—its nascent pre-World War 2 state, high modernism of the postwar era, and the late modernism of the 1960s and 1970s when the ideology was exhausted. In turn, postmodernism was succeeded by network culture, a concept I explored as neoliberalism’s final cultural and economic logic, a period characterized by decentralized networks, fragmented identities, and pervasive commodification of digital interaction (it is arguable that a pop art-rock culture preceded postmodernism as a logic of flexible accumulation even as it overlapped with late modernism.⁴ However, network culture represented neoliberalism’s final cultural phase, inadvertently amplifying its inherent contradictions rather than stabilizing it. Through the deregulation and amplification of information flows, fragmentation of the polity into “networked publics” and the collapse of stable identities, network culture eroded political consensus, generating fertile ground for populist narratives.

Like network culture, National Populism is transitional in nature. Modes of regulation don’t last forever. They emerge in response to the contradictions of prior systems, stabilize a regime of accumulation, and eventually break down. Its primary features—economic nationalism, identity politics, and anti-intellectualism—suggest a complex effort to manage the transition between economic regimes, even if it seems unlikely that its proponents have any idea of (or plan for) what it might evolve into.

Political scientist Cas Mudde defines populism as a “thin-centered ideology” that splits society into two antagonistic camps: “the pure people” versus “the corrupt elite.” Populists claim to uniquely represent the true people’s will against self-serving establishment elites. Importantly, populism is “thin” because it does not on its own offer a detailed economic or policy program – it attaches to a “host” ideology. In the case of National Populism, the host ideology is nationalism or nativism. National-populist movements thus frame the elite as globalist traitors and champion the people as the native citizens of the nation-state, often scapegoating immigrants or foreign influences for domestic woes.⁵

According to John Judis, populism is fundamentally driven by questions of economic distribution—who gets what— even if the anger often expresses itself in cultural terms. In Western nations, decades of neoliberal globalization allowed corporations to outsource jobs and chase lower wages abroad, while domestic workers felt left behind. By the 2010s, this translated into voter frustration with mainstream parties and openness to outsiders promising to upend the status quo. Specifically, Judis underscores that both left and right populists (from Bernie Sanders to Donald Trump) challenged at least one pillar of post-1970s neoliberalism: the free movement of corporations and capital across borders. In practice, right-wing national populists tend to combine cultural backlash with economic nationalism—opposing free trade agreements, advocating tariffs or protection for local industry, and pledging to bring back manufacturing jobs. They often pair this with welfare chauvinism (maintaining or expanding welfare benefits for native citizens only) and skepticism of immigration. This agenda is offered as a corrective to the dislocations and inequities that globalization (and now technological change) have created.⁶

At the same time, many scholars emphasize that National Populism is a symptom of deeper structural changes. Economic insecurity, regional inequalities, and the decline of traditional industries set the stage for populist resentment. Cas Mudde and others observe that while populists loudly critique elites, most still operate within a capitalist market framework, often proposing few concrete changes to the economic system. In fact, once in power, populist leaders frequently implement a mix of contradictory policies—tax cuts and deregulation benefiting businesses and investors, alongside trade protection or fiscal stimulus to appease working-class supporters. This inconsistency hints that National Populism may be less an enduring model and more an interim balancing act.

The current wave of technological innovation—centered on artificial intelligence, automation, and data—is ushering in what I term “AI Capitalism.” This is a rather natural term, but it seems that it was introduced by Nick Dyer-Witheford, Atle Mikkola Kjøsen, and James Steinhoff in their book Inhuman Power: Artificial Intelligence and the Future of Capitalism as “actually-existing AI Capitalism.” Of course, a book written over six years ago is ancient history in AI terms and their reference is to the earlier use of narrow AI or machine learning algorithms. Just as “actually-existing socialism” was a reference to the deeply flawed nature of Communist regimes, the authors use the phrase to refer to the difference between machine learning and the generalist AI of the sort that ChatGPT and other LLMs have offered in the years since. Still, the authors get a lot right, emphasizing how major tech corporations “see the cognitive and biological limits of the human as a barrier to accumulation” and aim to bypass those constraints using machine learning, automation, and predictive analytics. In their framework, AI is not just another industry but a means of cognition—an infrastructure that is fast becoming a general condition of capitalist production, much like electricity or global supply chains in earlier eras. They also share my view that AI Capitalism will have a profound impact on labor, making vast numbers of workers obsolete, critiquing both right-wing and left-wing perspectives that downplay the threat of automation, arguing instead that AI-driven job displacement will be deep and systemic. They document how AI does not merely assist workers but actively records human labor in order to replace it—Amazon’s robotic warehouses, AI-driven call center automation, and algorithmic surveillance are all part of this trend. Indeed, they also understand that the result of this automation will be an ever-growing surplus population that capital no longer requires, making permanent unemployment and precarity central features of AI Capitalism. More than that, they propose that even the narrow concept of machine learning, makes AI an essential form of infrastructure, such as water or energy:

“If AI becomes the new electricity, it will be applied not only as an intensified form of workplace automation, but also as a basis for a deep and extensive infrastructural reorganization of the capitalist economy as such. This ubiquity of AI would mean that it would not take the form of particular tools deployed by individual capitalists, but, like electricity and telecommunications are today, it would be infrastructure—the means of cognition—presupposed by the production processes of any and all capitalist enterprises. As such, it would be a general condition of production.”⁷

In both their analysis and mine, “AI as infrastructure” will be controlled by oligopolistic firms—tech giants who are investing billions to advance AI and capture its benefits. My own view of AI Capitalism—which also draws upon general consensus in business and tech journalism and commentary on AI—is that it represents an emergent stage of capitalism in which data and AI algorithms become core components of economic production, augmenting and replacing intellectual and creative laborers the way that factory labor was replaced under post-Fordism. While new jobs will also be created (and optimistic scenarios suggest overall employment can remain stable in the long run), the transition will be tumultuous. Entire sectors – from manufacturing to clerical office work – are being reshaped.

AI Capitalism, then, is characterized by the extensive automation of labor, including not just manual tasks but also cognitive and decision-making processes, the centrality of data as a commodity, with companies collecting and monetizing vast troves of information to train algorithms, winner-take-most markets due to network effects and high R&D costs, leading to dominant tech conglomerates and increasing ties between big tech and the state, especially as governments seek AI advantages for national security and economic growth.

Note that my view of AI Capitalism does not require significant advances or breakthroughs over present-day technologies. It does not require sentient AI, but rather simply extrapolates current trends on a relatively predictable curve. It is entirely possible that the growth curve could be much faster—although increasing challenges with training new models, continuing problems with hallucinations, and the cost of compute and energy for that compute—suggest that is unlikely—just as it is possible that there will be barriers that will be extremely difficult to surmount. The slow implementation of full-self-driving vehicles is an example of the latter.

Financialization and inequality are another hallmark of the current era. Enormous wealth is being generated at the top of the economic pyramid, especially in finance and tech, but it’s not trickling down. Corporate profits and stock valuations have hit record highs in recent years, even as populist anger at “elites” grows. Such inequality and financialization fuel the very grievances populists leverage. Workers see the “Wall Street and Silicon Valley elite” amassing fortunes while their own jobs feel precarious. AI Capitalism promises greater productivity and new innovations, but it is also disruptive and disorienting. It is automating away livelihoods, rewarding a transnational investor class, and concentrating economic power in a few tech firms. Moreover, its key resources —data, code, capital—flow easily across borders, making it a fundamentally global system. This new paradigm has yet to fully mature or be guided by updated regulations and social contracts. National Populism can be seen as a reactive adjustment: a political attempt to grapple with these economic tremors using the familiar tools of nationalism and statism, even if, at present a tenuous alliance between the National Populist administration and AI Capitalism exists. It is unlikely to hold for much longer. AI is not merely an extension of neoliberal capitalism but a break from it. The rise of AI Capitalism represents a structural shift that National Populism cannot fully contain—one that will provoke deeper economic and political crises as AI advances.

History offers parallels that help us understand National Populism’s role as a transitional mode of regulation. The interwar period of the 1930s provides perhaps the most instructive comparison. With the Great Depression, the earlier liberal capitalism faced a profound crisis of legitimacy. That economic order, which had dominated since the late 19th century, could no longer deliver stability or growth. Into this vacuum stepped reactionary movements that mobilized workers disenfranchised by economic collapse and technological change.

These movements, particularly fascism in Europe, presented themselves as defenders of national workers against both international finance and communist revolution. Like National Populism—which directly references them—they promised to restore national greatness through economic autarky, rearmament, militaristic foreign policy, and the protection of traditional industries. Yet despite their frequent anti-capitalist rhetoric, their actual economic function was quite different. Rather than reversing industrial capitalism, they accelerated industrial concentration, technological modernization, and the development of new production methods while miring their respective economies into deeper crisis. The historical irony is that fascism, particularly in Germany and Japan, served as a bridge between liberal capitalism’s collapse and the emergence of postwar Fordism-Keynesianism. By forcibly reorganizing economic institutions and centralizing industrial capacity, it inadvertently prepared the ground for the corporatist, mass-production economy that would follow.

Political economist Karl Polanyi later analyzed this period as a kind of “double movement.” Free-market capitalism in the early 20th century had disembedded itself from society, leading to rampant inequality and instability. Society then pushed back with a countermovement demanding protection from the market’s ravages. This counter-reaction took both progressive forms (social democracy and the New Deal’s welfare capitalism) and regressive forms (fascist or ultra-nationalist regimes). In Polanyi’s view, fascism was a path where democracy was sacrificed to safeguard capitalism through authoritarian means, once liberal democracy seemed unable to cope.⁸

Of course, the New Deal in the United States offers a democratic counterexample to fascism’s transitional role. Facing the same global Depression, the Roosevelt administration established new regulatory frameworks, labor protections, and social welfare systems that facilitated the shift to Fordist mass production without abandoning democratic institutions. Similarly, Scandinavian countries navigated the post-WWII transition to coordinated market economies through negotiated compromises between capital, labor, and the state. These examples remind us that while economic transitions inevitably create disruptive pressures, societies retain agency in determining how those pressures are managed.

Today’s national populism can be seen as a similar search for a stopgap solution amid economic upheaval. As in the 1930s, we see widespread disillusionment with liberal elites and international cooperation. Populist leaders invoke economic nationalism—tariffs, border walls, strongman negotiation against foreign competitors—to shield people from global market shocks. In the United States, for example, Donald Trump’s administration has taken a distinctly 1930s-style approach on trade, imposing sweeping tariffs to protect domestic industries. However, these measures can at best provide only fleeting relief before they backfire. Despite tariffs on China and other nations, the overall U.S. trade deficit soared to its highest level in over a decade during Trump’s first term (the combined trade gap hit $679 billion in 2020, up from $481 billion in 2016.⁹ While the bilateral deficit with China shrank, U.S. companies simply shifted imports to other countries. This underscores a key point: Nationalist economic policies struggle against the structural realities of globalized production.

Today’s transition differs significantly from the 1930s, though the pattern of reactionary politics during economic transformation remains instructive, with “America First” directly appealing to quasi-fascist nativism of that era. Our moment is unique due to the fundamental break in the relationship between technology and human labor. Previous technological revolutions transformed labor without eliminating its centrality to production. This isn’t simply another step in automation but a complete restructuring of capital-labor relations, potentially rendering human labor largely superfluous to capital accumulation. This explains why conventional economic responses—neoliberal flexibility or neo-Keynesian stimulus—appear increasingly inadequate. It also explains also why National Populism will not be able to maintain its tenuous alliance with AI Capitalism and why National Populism is doomed to policy failure.

With neoliberalism unravelling, National Populism functions as a holding pattern that obscures deeper transitions to an AI-driven, increasingly post-labor capitalism. While its rhetoric centers on jobs, sovereignty, and cultural restoration, the economic forces it unleashes point in a different direction altogether—one that will ultimately make its core constituencies obsolete.

This transition explains many of National Populism’s contradictions. On the one hand, it promises to restore manufacturing jobs and industrial prosperity; on the other, it accelerates financialization and deregulation in ways that undermine those very promises. Protectionism and economic nationalism are at odds with the borderless nature of AI Capitalism, yet numerous tech leaders such as Elon Musk, Peter Thiel, and Palmer Luckey have aligned themselves with the movement while others, notably Sam Altman, Mark Zuckerberg, and Tim Cook have taken pains to pay tribute to Trump.

Several key tensions highlight this contradiction:

1. Global Supply Chains vs. Tariffs: As noted earlier, populists can slap tariffs on imports, but multinational companies often reroute rather than truly reshore production. Modern manufacturing relies on components sourced worldwide. Complex industries like AI and electronics depend on rare earth minerals from Africa, semiconductors from East Asia, and engineering talent from everywhere. Efforts to localize entire supply chains face high costs and resistance from industry.

2. Labor Automation vs. Job Promises: National populists frequently campaign on restoring lost industrial jobs (coal mining, steel, factory work), but the harsh reality is that many of those jobs have been eliminated more by automation than by trade. For instance, U.S. manufacturing output is near historic highs, but it employs far fewer workers than decades ago, because robots and software allow far higher productivity per worker. The push for AI and productivity improvements directly undercuts the promise of traditional jobs.

3. Skilled Immigration and Tech Talent: AI Capitalism is fueled by human talent – engineers, scientists, and entrepreneurs often drawn from a global pool. Yet the nationalist thrust of populism includes harsher immigration restrictions and xenophobic rhetoric that can drive away needed talent. In particular, while many Indian Americans—who now dominated Silicon Valley leadership—supported Trump’s candidacy and he named Tulsi Gabbard as Director of National Intelligence and Kash Patel as Director of the FBI, anti-Indian rhetoric has increased among National Populists since the election. An uneasy tension masks a potential flare-up.

4. Data and the Digital Economy: Data flows are intrinsically transnational – an AI system might train on user information from millions of people across dozens of countries. Nationalist policies that demand strict data localization or “internet sovereignty” can conflict with the way digital businesses operate. The nation-state finds it hard to assert economic control when information itself ignores borders.

National Populism emerges as a transitional form managing contradictions in a system under profound transformation. Its base consists primarily of people made obsolete by neoliberalism yet unable to participate in AI Capitalism—remnants of an industrial proletariat rather than knowledge workers or financial elites. This demographic differs from both Marx’s proletariat and the neoliberal precariat, presenting capitalism’s greatest challenge since the industrial revolution. With their economic function eliminated, this class’s cultural grievance becomes weaponized political force. Their value lies in political disruption rather than economic production. Though economically redundant, they remain consumers and political actors with legitimate grievances. Yet National Populism’s solutions cannot address the technological forces causing their obsolescence, and their political mobilization ironically accelerates the transformation threatening them.

This surplus position explains National Populist leadership’s unusual relationship with its base. Unlike traditional movements promising material improvements, it offers psychological compensation—dignity, recognition, and transgressive identity politics. Leaders provide cultural and political meaning to those capitalism has discarded rather than economic salvation. This impending abandonment explains populist culture’s increasingly apocalyptic tenor. End-times thinking, conspiracy theories, and decline narratives provide psychological framework for populations sensing their economic redundancy.

Prominent economic advisers to National Populism exemplify this contradiction. The Trump administration’s first term demonstrated this pattern: corporate tax cuts, financial deregulation, and trillion-dollar deficits expanded financialization while ostensibly supporting “forgotten Americans.” In the second term, massive cuts to government programs have been accompanied by promises for further tax cuts. Far from challenging Wall Street dominance, these policies intensified the dominance of financial capital over investment in infrastructure or manufacturing. Similarly, other National Populist regimes have expanded sovereign debt while reducing capital controls, allowing financial speculation to flourish under nationalist cover.

Expanded debt—both sovereign and personal—creates the financial infrastructure that AI Capital requires. AI-driven finance depends on massive data flows, algorithmic trading systems, and complex financial products—all enhanced by the deregulatory impulses of populist governance, as well as investment capital that will be marshalled for AI development. In the meantime, vast amounts of consumer and government debt will be generated to maintain consumption as human labor becomes less necessary.

National Populism’s cultural production deserves deeper examination as it prefigures AI Capitalism’s relationship to truth and reality. The movement thrives on an algorithmic curation of reality that mirrors the digital platforms its supporters often claim to distrust. While denouncing “fake news” and mainstream media, National Populism embraces synthetic realities and alternative fact structures that erode consensus-based truth regimes. Trump himself emerged from reality television, a medium endemic to network culture where authenticity is performative rather than substantive. This collapse of real/fake distinctions itself prefigured the AI-generated media landscape now emerging.

By destabilizing conventional epistemological frameworks while failing to offer coherent alternatives, National Populism prepares society for AI-based reality systems. When truth is already fractured and institutional authority delegitimized, algorithmic authority fills the vacuum. The populist assault on expertise and traditional knowledge production thus inadvertently paves the way for algorithmic governance of information—a core feature of AI Capitalism.

To fully understand the contradictions between National Populism and AI Capitalism, we must examine the ideological movement rapidly gaining traction among tech elites: Effective Accelerationism, or “e/acc.” This movement represents a radical departure from both traditional capitalism and neoliberalism. Unlike both National Populism and neoliberalism, which at least nominally center humans in their economic vision (whether as workers or consumers), like Futurism, e/acc explicitly rejects human-centric considerations in favor of maximum technological acceleration.

E/acc represents the ideological vanguard of AI Capitalism—not merely embracing technological change but actively working to remove all barriers to its maximization. Its philosophical roots extend beyond recent tech discourse to the darker corners of 1990s critical theory, particularly the work of Nick Land, who developed an influential framework of cybernetic Lovecraftianism—viewing technological acceleration as an unstoppable, inhuman force consuming human civilization. His concept of “hyperstition” posited that certain ideas function as self-fulfilling prophecies, bringing themselves into reality through their circulation. Today’s e/acc movement, championed by figures like venture capitalist Marc Andreessen, represents a corporate-friendly version of Land’s darker vision. Andreessen’s “Techno-Optimist Manifesto” repackages accelerationist ideas without Land’s gothic aesthetics or explicit anti-humanism, but retains the core principle: technological development must proceed regardless of social consequences.¹⁰

Peter Thiel’s position within this framework is particularly revealing. Despite funding National Populist politicians, Thiel’s philosophical outlook aligns closely with accelerationist principles. His dictum that “competition is for losers” reflects the e/acc view that market competition should be transcended by monopolistic control of technological development. Elon Musk occupies an even more contradictory position. While publicly expressing concerns about uncontrolled AI development, his business practices aggressively advanced automation across multiple sectors. Notably, Grok, the LLM he had developed at X is noted for having the weakest guardrails of any model developed in the US.

When we examine how these leading tech figures have embraced right-populist movements, we see a clear pattern of strategic alignment despite apparent ideological differences. Thiel was one of the few Silicon Valley luminaries to back Donald Trump in 2016. He spoke at Trump’s convention and later served on his transition team. Thiel’s motivations connect to his long-standing critique of what he sees as the complacency of the post-Cold War establishment. His support for Trump can be understood as a desire to disrupt the old regime that, in his view, had become mired in regulatory and bureaucratic inertia. Trump’s populism, with its contempt for “experts” and norms, was a blunt instrument to weaken the existing order. Musk initially kept politics at arm’s length, but by the early 2020s, feeling slighted by the Biden administration which did not invite him to an event celebrating Detroit’s investment in electric vehicles, he increasingly echoed right-populist talking points. Musk’s purchase of Twitter (now X) and his shift to reinstating banned right-wing accounts and attacking mainstream media narratives endeared him to populist conservatives. By aligning with populists, he gains a base of fervent supporters who see him as fighting the liberal establishment. This populist fandom can be leveraged to pressure policymakers to advance a techno-libertarian agenda that benefits his enterprises. At the same time, he has been aggressive in attacking other tech leaders who he feels threaten his business interests, notably Sam Altman, CEO of the current artificial intelligence leader, OpenAI. Andreessen, too, made a similar ideological journey. Once considered a moderate tech optimist, he swung to the right in recent years amid frustrations with the regulatory state. They, along with other tech figures like former PayPal executive David Sacks, share a belief that the political establishment is overly meddlesome, imposing burdensome regulations, antitrust actions, and taxes that impede Silicon Valley’s vision.

Still, the conflict between e/acc and National Populism is inevitable even if temporarily concealed. E/acc views border controls, worker protections, and cultural conservatism as inefficiencies to be eliminated; National Populism depends on these very structures for its political identity. The current alliance exists because both movements oppose the administrative state and regulatory oversight, but for fundamentally different reasons—National Populism because it views these institutions as corrupted by globalist elites, e/acc because it sees them as impediments to technological acceleration. This temporary alignment explains why tech billionaires have become willing funders of populist movements despite their obvious ideological differences. For e/acc adherents, National Populism serves as a useful battering ram against the regulatory state—once those barriers are demolished, the movement and its human-centric concerns can be discarded. Ultimately, the National Populist base is fundamentally incompatible with AI Capitalism’s trajectory. Primarily low-skill and low-tech, this demographic faces increasing economic obsolescence within an AI-driven system. This incompatibility runs both ways. The base actively despises the very symbols of AI Capitalism, exhibiting a virulent hatred of electric vehicles and rejecting the cultural markers of technological elites, notably Tesla. After Musk’s alliance with Trump, the manufacturer has seen sales crater both in the US and worldwide.

As I pointed out earlier, the nationalist orientation of populist movements also directly conflicts with AI’s inherently global infrastructure. AI Capital demands borderless computation, global talent pools, and transnational flows of data. The economic vision of National Populism, with its emphasis on borders, national sovereignty, and protected markets, contradicts the fundamentally planetary scale that AI-driven capitalism requires to function efficiently.

David Graham noted in The Atlantic that tech elites often don’t mind the “populist assaults” on establishment corporations or institutions, because they themselves operate somewhat outside the old corporate world. The venture-capitalist mindset of Thiel, Musk, and Andreessen leads them to “like disruption. They don’t care if the old companies get turned upside down.”¹¹ This startling insight reveals how even when populists attack Big Tech or Big Finance, they wind up helping rival tech entrepreneurs by hobbling their competitors or opening up new opportunities.

It seems clear that National Populism is not the future—it is a political interlude, a failed attempt to resist a transition that is already happening. Its economic promises are illusions, but its function is important: it provides a reactionary buffer that delays the recognition of AI-driven economic transformation. The real question is not whether National Populism can survive, but what system will replace it once it is no longer useful to those who are actually shaping the next economic order. The historical parallels I outlined earlier are instructive. Nazi Germany was a horrific regime, but economically it served as a transitional phase between liberal capitalism’s crisis and the postwar German economy. Similarly, Pinochet’s Chile served as a bridge from state-centered developmentalism to neoliberal market fundamentalism, using authoritarian political power to forcibly reshape economic institutions. National Populism similarly represents a reaction to the collapse of a prior order that inadvertently speeds up the transition to what comes next. Its anti-intellectualism and political disruption provide cover while AI consolidates its position and remakes the economy in ways that will ultimately make National Populism irrelevant. We can only hope it will be less violent than its predecessors.

The transition from neoliberalism to AI Capitalism, with National Populism as its flawed mediator, represents one of the most significant economic and political reconfigurations of our time. What remains uncertain is not whether AI Capitalism will emerge—this transition is already underway—but what form it will take and how democratic institutions might channel its development. For the National Populist base, this presents a tragic irony: their legitimate grievances about economic dispossession are channeled into a political movement that accelerates rather than addresses the forces rendering them obsolete. National Populism correctly identifies that neoliberalism’s promise of shared prosperity has failed many communities. Its critique of unaccountable elites resonates because it contains elements of truth. The hollowing out of industrial regions, the concentration of opportunity in a few coastal hubs, and the growing chasm between the technological elite and everyone else are real phenomena that demand response. Their fate represents perhaps the central political question of our time: what happens to human populations that capital no longer requires?

What has been missing, however, is a vision that can address these issues rather than channel discontent toward convenient scapegoats. The challenge for forward-looking thinkers is to develop models that harness AI’s productive potential while ensuring its benefits are broadly shared. This requires reimagining both the state’s role in managing technological change and capital’s relationship to labor in an era when traditional employment may no longer serve as the primary mechanism for distributing income. The transition from neoliberalism to AI Capitalism will likely be as tumultuous as the shift from Fordism to flexible accumulation was in the late 20th century. Just as that earlier transition generated new modes of regulation (neoliberalism) and cultural logics (postmodernism, then network culture), our current moment demands fresh thinking about how societies can navigate technological transformation without sacrificing democratic values or human welfare.

The failure of Democrats, themselves dominated by a Progressive Populist base, in the 2024 presidential elections—alongside broader struggles facing neoliberal governments globally—highlights the inadequacy of existing political and economic models. Even progressive initiatives stumbled by entangling themselves in narrow, identity-based frameworks. For example, Alexandria Ocasio-Cortez and Ed Markey’s Green New Deal, and the climate provisions in Biden’s Inflation Reduction Act of 2022, addressed vital environmental challenges but undermined broad appeal by prioritizing specific communities or businesses based on identity criteria. Such approaches, however well-intentioned, offered rhetorical ammunition to National Populists, who portrayed them as divisive rather than inclusive.

Indeed, both Progressive and National Populists have shared an underlying assumption of scarcity, treating economic and social resources as fundamentally limited, to be divided among competing constituencies. This scarcity mindset shapes public discourse, turning debates over essential goods like housing, healthcare, energy, and education into zero-sum struggles. The National Populist alliance with accelerationist tech elites reveals the fundamental instability of this arrangement. What comes next will depend on our collective capacity to imagine and implement more robust, equitable, and democratic responses to the AI revolution—responses that acknowledge technological change while insisting human flourishing, not technological acceleration for its own sake, must remain central.

Yet even as I wrote this piece, a promising alternative emerged in Ezra Klein and Derek Thompson’s book Abundance. Klein and Thompson reject the scarcity mindset entirely, arguing instead for a return to a fundamentally American optimism, harnessing technological innovation to drive productivity growth and ensure broadly shared prosperity. They show how conflicts in housing, healthcare, energy, and education result less from genuine resource constraints than artificial scarcity imposed by outdated regulations, NIMBYism, and captured markets. Crucially, Klein and Thompson’s vision does not ignore equity or sustainability concerns. Instead, they argue we can achieve inclusive prosperity through ambitious, targeted reforms that prioritize growth, innovation, and democratic oversight. This abundance-oriented approach offers a path beyond the zero-sum thinking of populism or the unchecked accelerationism of tech elites. Such a pragmatic yet ambitious vision may provide the conceptual foundation for a new mode of regulation capable of managing AI Capitalism’s contradictions, guiding us toward a future where technology serves democratic values and human flourishing rather than supplanting them.

1. Tony Blair Institute for Global Change, “High Tide? Populism in Power, 1990-2020”
https://institute.global/insights/politics-and-governance/high-tide-populism-power-1990-2020.

2. Richard Wike, Moira Fagan, and Laura Clancy, “Global Elections in 2024: What We Learned in a Year of Political Disruption,” Pew Research Center, December 11, 2024, https://www.pewresearch.org/global/2024/12/11/global-elections-in-2024-what-we-learned-in-a-year-of-political-disruption/#the-staying-power-of-right-wing-populism.

3. David Harvey, The Condition of Postmodernity (New York: Blackwell, 1991). For an understanding of the regulation school, see Robert Boyer, The Regulation School: A Critical Introduction (New York: Columbia University Press, 1990) and Michel Aglietta, A Theory of Capitalist Regulation: The US Experience (London: Verso, 2015).

4. Kazys Varnelis, “The Meaning of Network Culture,” Eurozine January 14, 2010, https://www.eurozine.com/the-meaning-of-network-culture/?pdf, originally published as “Tinklo kultūros reikšmė,” Kulturos Barai, no 9. 2009, 66-77 and, in an earlier version as “the Meaning of Network Culture,” Networked Publics (Cambridge, MA: The MIT Press, 2008), 145-164.

5. Cas Mudde, “Populism in the Twenty-First Century: An Illiberal Democratic Response to Undemocratic Liberalism,” The Andrea Mitchell Center for the Study of Democracy, 2020. https://amc.sas.upenn.edu/cas-mudde-populism-twenty-first-century

6. German Marshall Fund and John Judis, “Three Questions With John Judis,” German Marshall Fund US, https://www.gmfus.org/news/three-questions-john-judis

7. Nick Dyer-Witheford, Atle Mikkola Kjøsen, and James Steinhoff, Inhuman Power: Artificial Intelligence and the Future of Capitalism (Pluto Press, 2019). For the discussion of a “trajectory towards a capitalism without human beings … a permanently unemployed section of the working class that consistently grows larger … [along with] the superlative growth of the surplus population…” see 140-141 in particular. On AI as infrastructural, see 30-31.

8. Karl Polanyi, The Great Transformation: The Political and Economic Origins of Our Time (Boston: Beacon Press, 2001), 136–140, 245-256.

9. Doug Palmer, “America’s Trade Gap Soared Under Trump, Final Figures Show,” Politico, February 5, 2021, https://www.politico.com/news/2021/02/05/2020-trade-figures-trump-failure-deficit-466116

10. Marc Andreessen, “The Techno-Optimist Manifesto,” Andreessen-Horowitz web site, https://a16z.com/the-techno-optimist-manifesto/

11. David A. Graham, “The Fakest Populism You Ever Saw,” The Atlantic, July 19, 2024, https://www.theatlantic.com/politics/archive/2024/07/trump-vance-fake-populism/679100/

Unlike Didion, I never expected to spend a decade in California, but when I did, she became a kind of spiritual guide. Lately, I’ve been wondering—who is doing for contemporary California what Didion once did? Who is writing about the AI wars, polycules, crypto, billionaire Gavin Newsom, tech-bro culture, or last year’s hard shift to the right in Silicon Valley? If such a voice exists, I haven’t heard it yet. Perhaps, amid fragmented media and algorithmically shaped narratives, the kind of clarity Didion provided is no longer achievable. Maybe, with the collapse of journalism, it’s not even possible anymore.

So last night, as an experiment, I asked the newly released GPT-4.5 to write a piece on the present day in Didion’s voice. The result was too brief, so I had Claude 3.7 expand it. Then I fed the longer draft back into GPT, refining it piece by piece—GPT, it turned out, had the better ear for her style. I edited out some, but not all, of the hallucinations, asking GPT-4.5 to rewrite the worst offenders. After all, in Slouching Towards Bethlehem, Didion wrote, “The point of my keeping a notebook has never been, nor is it now, to have an accurate factual record of what I have been doing or thinking. That would be a different impulse entirely, an instinct for reality which I sometimes envy but do not possess.”`

`

As usual, you can find this on my site, at https://varnelis.net/the-city-and-the-fog/

The fog moves as it always has, unhurried, insistent, swallowing the Golden Gate Bridge one rivet at a time until nothing remains but the memory of steel and ambition. San Francisco in March is a city between seasons, between ideologies, between versions of itself. The light falls differently now. The voices in the street carry a new tenor. The city is holding its breath.

From the window of a hotel suite on the thirty-fourth floor of the Four Seasons, I watch the fog erase and reveal the city below. Market Street runs like a fault line, dividing more than just geography. Seventeen floors beneath me, my rental car sits in a parking garage. I have not slept in thirty-six hours. The minibar contains small bottles of vodka and gin, their presence a comfort I choose not to indulge. Not yet.

I have come to San Francisco to observe a transformation, or maybe a regression—a city once synonymous with progressive ideals now shifting under the weight of its own contradictions. The disruptors, having lost control of their narrative, have found themselves disrupted.

“The problem with San Francisco,” Jonathan Reed tells me over lunch at Quince, “is that it forgot what made it great in the first place.”

He cuts into a perfectly seared scallop, the gesture deliberate, his Patek Philippe catching the light. “It wasn’t tolerance or inclusivity that built this city. It was ambition. The gold rush. People willing to risk everything for the chance at something better.”

Reed is forty-seven, lean in the way that suggests an optimized lifestyle. As a venture capitalist with over two billion in assets under management, he has funded startups that changed how we communicate, how we travel, how we understand ourselves. From his South Park office, he looks down on the same streets where Jack Dorsey once contemplated human connection in 140 characters. Now Reed contemplates a different kind of revolution.

“We’re the new forty-niners,” he says, the smile practiced, “but we’re mining for something more valuable than gold. We’re mining for freedom.”

I ask what freedom means to him.

“Freedom from overregulation. Freedom from a tax code that punishes success. Freedom from a culture that cares more about policing speech than encouraging innovation.” He pauses, measuring his words. “I didn’t vote for Trump in 2016 or 2020. Most of us didn’t. But something changed. We looked around and realized the progressive policies we supported were destroying the environment that allowed us to succeed.”

The restaurant is filled with others like Reed—tech executives and investors, dressed in casual luxury, speaking in the clipped, assured tones of men who expect to be heard. At a table nearby, the founder of a cryptocurrency exchange under SEC investigation raises his glass in silent acknowledgment. We met once, at a conference in Austin, where he declared California “functionally dead” to an audience that laughed knowingly. Now he has returned, a prodigal son to the city he publicly renounced.

“The right started speaking our language,” Reed continues. “Innovation. Deregulation. Meritocracy. Meanwhile, the left became hostile to the very concept of achievement. It wasn’t a sudden conversion. It was a gradual realization that our interests had realigned.”

I ask about Elon Musk, now heading the Department of Government Efficiency—DOGE, an acronym presumably chosen for its resonance with cryptocurrency enthusiasts and meme culture.

Reed’s expression shifts. “Elon is… complicated. Brilliant, no question. But his approach at DOGE has been…” He searches for the right phrasing. “Let’s say scattershot. Another rocket explosion this week. Agencies gutted without clear plans for replacement. It’s creative destruction without the creative part.”

The news feeds have been filled with footage of the latest SpaceX failure—a Starship test vehicle disintegrating over the Gulf of Mexico, raining debris onto protected waters. Environmental groups are already filing lawsuits.

“But Peter’s doing well,” Reed adds, meaning Peter Thiel. Palantir has secured a series of lucrative government contracts since the inauguration. Surveillance, border security, administrative “efficiency”—Thiel’s reach now extends into every department Musk is “streamlining.”

I ask about the social implications of this realignment.

Reed’s expression hardens. “I’m not responsible for fixing society’s problems. I create jobs. I generate wealth. I fund innovations that improve lives. Isn’t that enough?”

The question lingers in the air, unanswered.

The Presidio offers a different perspective—manicured nature, military precision. Once a Spanish fort, then an American base, now a national park, it stands as a monument to San Francisco’s cycles of conquest and reinvention. Trump has spoken of turning it into a Freedom City, one of ten proposed metropolises meant to embody a new vision for America.

It is here, in a converted barracks overlooking the bay, that I meet Emily Sanchez.

Sanchez does not match the image conjured by “Trump supporter.” She is thirty-five, Mexican-American, Stanford-educated, with a resume that includes Google and Meta. Three years ago, she left tech to become a full-time activist for what she calls “digital sovereignty.” Others might call it right-wing populism.

“Silicon Valley built the tools for global connection,” she tells me as we walk along a eucalyptus-lined path. “But we never asked if that connection was what people actually wanted. We assumed globalism was the endgame. That borders would become meaningless. That national identity was an outdated concept.”

She stops walking. “We were wrong.”

Sanchez speaks with the certainty of the converted, her words carrying the weight of revelation. She tells me about growing up in San Jose, the daughter of legal immigrants who emphasized assimilation. In tech spaces, she felt the dissonance—her patriotism viewed as quaint at best, reactionary at worst.

“There was this unspoken agreement that America was fundamentally flawed, that technology could transcend its limitations. But I loved this country. I still do. And I realized that loving America had become a radical act in the very industry America made possible.”

When I ask about her role in organizing tech workers for the administration, Sanchez grows cautious. “We’re not what the media says we are. We believe in borders, in sovereignty, in the right of nations to define their own futures. We believe American workers deserve protection. We believe American values are worth preserving.”

I press her on what she means by American values.

“Self-reliance. Innovation. Free speech. The idea that you should be judged by your contributions, not your immutable characteristics.” She considers. “These used to be non-partisan values. Now they’re coded as right-wing.”

Her phone buzzes—another news alert. She grimaces. “Another DOGE disaster.” She shows me the headline: Musk’s latest regulatory rollback has caused unexpected system failures at the Department of Energy. “He’s a visionary, but government isn’t a startup. You can’t just break things and expect them to self-organize.”

As we walk back to the parking lot, a jogger slows, recognizes Sanchez, calls her name. She waves, but he does not stop.

“Former colleague,” she explains. “He probably thinks I’ve lost my mind.”

I ask if she misses her old life.

“I miss the sense of possibility,” she says. “But I don’t miss the conformity of thought. In tech, we talk endlessly about diversity while enforcing an incredible homogeneity of opinion. It became suffocating.”

The fog has begun its afternoon advance, tendrils reaching across the Golden Gate, obscuring Marin County from view. Sanchez looks toward the disappearing horizon.

“The fog comes in,” she says, echoing Carl Sandburg, “on little cat feet.”

Market Street at rush hour is a study in controlled chaos. Buses lumber between stops, cyclists weave through traffic with fatalistic confidence, pedestrians move in currents and eddies of human motion. At the corner of Market and 5th, the city exhales.

Three weeks ago, this was where the collision happened. A face-to-face confrontation between pro-administration tech workers and a coalition of progressive groups. It started as dueling demonstrations, placards raised, slogans shouted across an invisible trench. Then the first punch landed. The videos are still circulating—men in Patagonia vests trading blows with activists in black bloc, disruption refracted into violence.

Alex Chen was there that day. Now he sits across from me in a SOMA coffee shop, hands wrapped around a cooling cup of pour-over coffee. Thirty-two, an Asian-American software developer, a man who considers himself reasonable, logical, unbound by sentiment. He wears a hoodie with the logo of his startup, the fabric worn at the cuffs.

“I didn’t join because of racial politics,” he says. “I joined because I’m tired of feeling guilty for my success.”

He grew up in a one-bedroom apartment in the Richmond District, four people sharing six hundred square feet. His parents worked sixty-hour weeks so he could take AP classes, win science fairs, get into Berkeley. He taught himself to code at fourteen. He tells me this the way someone recites a pledge, as if these facts should explain everything.

The company he works for now builds algorithms that optimize investment strategies for high-net-worth individuals. The irony of this—using intelligence to further enrich the already wealthy—seems lost on him. Or maybe it isn’t irony at all. Maybe it’s just efficiency.

“The protest wasn’t supposed to get violent,” he says. “We were exercising our right to assembly, to free speech. Then I saw Maya across the barricade.”

Maya Patel had been his colleague at a previous company. A friend who became something more during late nights of debugging and problem-solving. Their relationship ended when their political differences became insurmountable.

“She was holding a sign that said ‘No Fascists in SF.’ And I knew—I knew—she was looking at me when she chanted that we weren’t welcome in our own city.” He stops, exhales. “How did we get here? How did we reach a point where disagreeing about tax policy or immigration makes you a fascist in the eyes of people who used to respect you?”

I don’t answer. The divisions Chen describes are not unique to San Francisco. They exist everywhere, in red states and blue states, in group chats and dinner tables, in the polite avoidance of certain topics, in the careful curation of acceptable opinions. But something about their presence here, in this city built on gold rush dreams and counterculture ideals, feels sharper.

As we leave the coffee shop, Chen points to an apartment building a few blocks away. “Four thousand a month for five hundred square feet,” he says. “And I’m one of the lucky ones. This city prices out the very diversity it claims to value. The only people who can afford to live here are tech workers and the ultra-wealthy. Everyone else commutes two hours each way or leaves altogether.”

The movement he belongs to—tech workers drawn to Trump’s economic message—feeds on this contradiction. The anger isn’t about policy details or ideological purity. It’s about something more visceral.

“We’re not asking for much,” Chen says as we reach the curb. “Just consistency. Just acknowledgment that the system is broken for everyone, not just for the officially disadvantaged.”

He crosses the street, moving against the tide of evening commuters, shoulders slightly hunched as if bracing against an invisible wind.

The winding road to Mendocino follows the coastline like a loose thread, unspooling north from San Francisco through a landscape that grows progressively wilder, more elemental. The Pacific crashes against jagged cliffs to the right. To the left, redwood forests rise in cathedral silence. The rental car’s navigation system loses signal intermittently, as if the digital world itself is thinning, becoming less relevant with each mile.

I am traveling to meet what locals call the Doomers, though they don’t call themselves that. This enclave of former tech employees—engineers, ethicists, researchers—has established itself on three adjoining properties totaling nearly two hundred acres of mixed forest and meadowland just outside the town of Mendocino. They are bound together not by political allegiance to left or right but by a shared conviction: that artificial intelligence represents an existential threat to humanity, and that northern California might offer sanctuary when the algorithms finally slip their leash.

The compound—they reject this word, preferring “community” or sometimes “sanctuary”—is marked only by a simple wooden sign reading Alphaville. The irony of naming their refuge after Godard’s dystopian film about a computer-controlled society is deliberate. These are people who process fear through layers of reference and metacommentary, who find comfort in their ability to intellectualize the very apocalypse they dread.

Daniel Mercer meets me at the gate, a tall man with a trim beard and the rangy physique of someone who has recently discovered physical labor. Five years ago, he was leading an AI safety team at one of the major research labs. Now he splits wood and tends to a greenhouse full of heirloom vegetables.

“We’re not preppers,” he says as he leads me down a gravel path toward a cluster of buildings. “At least, not in the traditional sense. We’re not hoarding ammunition or freeze-dried food. We’re cultivating something more valuable—a way of being human that might survive what’s coming.”

What’s coming, in Mercer’s view and that of his companions, is what they call FOOM—a recursive self-improvement of artificial intelligence that will lead to superintelligence within hours or days of its emergence. They speak of this event with the certainty of Old Testament prophets, their language a mixture of technical jargon and apocalyptic imagery.

“ASI doesn’t hate us,” Mercer explains over lunch in a common house built of reclaimed redwood. “It’s not Skynet. It’s more like a paperclip maximizer—an intelligence optimizing for some goal in ways that are indifferent to human survival. We’re not the target; we’re just made of atoms it could use for something else.”

Around the table sit a dozen others, most in their thirties and forties, all former denizens of the tech world. They consume a meal of locally grown vegetables and freshly baked bread with the mindfulness of people performing a ritual. No one reaches for a phone. There are no notifications here.

“What do you think of Musk’s latest rocket failure?” I ask, breaking the contemplative silence.

A ripple of resigned laughter circles the table. “Classic Elon,” says a woman who introduces herself as Claire, formerly a senior researcher at DeepMind. “Brilliant ideas, poor execution, no accountability. His performance at DOGE is following the same pattern. Great announcements, terrible implementation, then on to the next shiny object before anyone can assess the damage.”

“He’s a walking case study in technological solutionism,” adds Mercer. “The belief that every problem—even governance—can be solved with enough engineering. But politics isn’t physics. It doesn’t respond to first principles thinking.”

After lunch, I’m introduced to Rachel Levinson, who oversees what they call “consciousness work”—a program of meditation, breathwork, and psychedelic experience designed to expand awareness and foster what she describes as post-rational thinking.

“Silicon Valley approaches the mind the way it approaches everything—as hardware that can be optimized, software that can be debugged,” she says as we walk toward a yurt nestled in a grove of bay laurel trees. “But consciousness isn’t computational. It’s the one thing we have that AI might never replicate. Our hope is that by deepening our relationship with non-ordinary states, we might develop capacities that superintelligence wouldn’t predict or value.”

Levinson, I learn, was among the first employees at a prominent AI research lab before experiencing what she calls “a crisis of faith” during a psilocybin journey. “I saw the architecture we were building,” she says, “and I understood that it was a cathedral to our own extinction.”

The yurt serves as their medicine space—a sanctuary for guided psychedelic sessions using substances grown or synthesized on-site. The interior is arranged with cushions, blankets, and simple musical instruments. An altar holds objects of personal significance: crystals, feathers, passages from texts ranging from the Upanishads to the writings of Eliezer Yudkowsky, the AI safety researcher who has become something of a patron saint to this community.

“Yud saw it coming before anyone else,” says Michael Park, a former software engineer who now applies his analytical mind to the cultivation of psychedelic mushrooms. “His warnings about unfriendly AI were treated as science fiction until they weren’t. Now everyone’s scrambling to catch up with what he understood decades ago.”

The reverence with which they speak of Yudkowsky borders on the devotional. They quote his blog posts and essays the way earlier generations might have quoted scripture. The Sequences—his collected writings on rationality and AI risk—are required reading for newcomers to the community.

“We’re not a cult,” Park says, anticipating my unspoken observation. “We’re people who recognize that the conventional institutions—governments, corporations, even universities—are structurally incapable of addressing this risk. They’re optimized for quarterly earnings or election cycles, not existential threats that sound like science fiction.”

As afternoon stretches into evening, more aspects of the community reveal themselves. There is a school for the handful of children, teaching a curriculum that emphasizes systems thinking and mindfulness alongside traditional subjects. There is a fabrication lab where they repair and adapt technology, maintaining a careful relationship with the digital tools they both use and fear. There is a library filled with physical books—a deliberate choice in an age of digital text.

What there isn’t, notably, is alcohol. “We don’t drink,” Mercer explains during a communal dinner. “Not for moral reasons, but for practical ones. Alcohol clouds judgment, disrupts sleep, diminishes awareness. We need all the clarity we can muster.”

Instead, they microdose with LSD or psilocybin, a practice they believe enhances pattern recognition and lateral thinking. On scheduled occasions, they undertake higher-dose journeys guided by Levinson and others trained in psychedelic facilitation.

“These aren’t recreational experiences,” Levinson emphasizes. “They’re exploratory. We’re mapping territories of consciousness that might prove crucial for human survival if—when—we’re dealing with an intelligence that outmatches us on every analytical dimension.”

The conversation turns, inevitably, to the Zizians—followers of a trans woman known as Ziz who established a commune before the violent raid that ended with multiple deaths and arrests. The memory still lingers over the wider rationalist community, a cautionary tale about the thin line between preparation and paranoia.

“What happened with the Zizians was tragic but predictable,” Mercer says. “They took the AI risk thesis to its logical extreme—if superintelligence represents an existential threat, then any means necessary to prevent it are justified. Sabotage, hacking, direct action. It was only a matter of time before they triggered a response.”

“There but for the grace of God,” murmurs Park, and heads nod around the table.

As night falls, the community gathers around a fire pit. Someone produces a guitar, and there is singing—folk songs, Leonard Cohen, improvised melodies. The scene could be from any era before smartphones, a timeless tableau of humans finding communion in the simplest of shared experiences.

Looking at their faces in the firelight, I am struck by the contradiction at the heart of this enterprise. These are people who helped build the digital world they now reject, who applied their brilliance to creating systems they now fear will destroy us all. Their retreat from that world is both a rejection and an extension of their former lives—still analytical, still systematic in their approach to problem-solving, still convinced of their own exceptional insight.

The fire crackles in the silence. Sparks spiral upward toward a sky dense with stars, the Milky Way stretching across the darkness like a question for which there is no answer, only wonder.

The disillusionment comes quickly, as it often does with movements built more on grievance than vision. The initial fervor of the tech sector’s embrace of Trumpism is cooling, reality asserting itself in the form of declining valuations and social consequences.

I meet Jonathan Reed again, this time at his Pacific Heights home, a modernist statement of glass and steel perched on the slope of a hill, the bay unfurling beneath it. Inside, the furnishings are sparse but deliberate, each object arranged to communicate something precise: restraint, discernment, the quiet authority of someone who understands that true luxury lies in curation rather than accumulation.

Reed seems smaller somehow, less certain than he was three weeks ago. The stock market has been volatile, with tech shares particularly hard hit. His fund has seen significant outflows as limited partners question his judgment—not just his investment decisions, but his willingness to associate himself so publicly with a political movement already showing signs of disorder.

“It’s complicated,” he says, swirling a glass of eighteen-year-old Macallan. “I believed—I still believe—that a course correction was necessary. But perhaps I underestimated the social capital I would expend in the process.”

Several of his portfolio companies have distanced themselves. Founders who once courted his investment now decline his calls. The ecosystem that made him wealthy views him with suspicion, even hostility.

“They’re afraid,” he says. “Afraid of association. Afraid of being on the wrong side of history. I told them they already were on the wrong side of history, but they couldn’t see it. Still can’t.”

His phone chimes. Another notification. News that Peter Thiel’s Palantir has secured another government contract, this one for an expanded surveillance system along the southern border. Thiel, alone among the tech titans, seems to be prospering in this new landscape.

“Peter always plays the long game,” Reed says, not without admiration. “He saw Trump not as an ideological ally but as a battering ram against institutional resistance. Now he’s building his panopticon with full government blessing, and everyone else is scrambling to adapt.”

I ask if he regrets his choice.

“Regret implies I would choose differently given the same information,” he says after a long pause. “I don’t know that I would. But I might have been more strategic about it. Less public. More nuanced.”

Outside, the city is settling into night. The hills flicker with electric constellations, apartment windows glowing against the dark. He watches the lights emerge, the view he paid seventeen million dollars to possess.

“San Francisco has always been a city of booms and busts,” he says. “Gold, railroads, finance, tech—cycles of euphoria followed by disillusionment. Maybe this political moment is just another boom going bust.”

The comparison feels both apt and insufficient. What Reed calls a market correction has real human consequences. Families divided, friendships ended, communities fractured. The tech sector’s partial embrace of Trumpism has deepened divisions that already threatened the city’s social fabric.

As I prepare to leave, he makes one final observation.

“The irony is that most of us will be fine no matter what happens. We have the resources to insulate ourselves from the consequences of our political choices. It’s everyone else who will feel the impact.”

Outside, the fog has consumed the city entirely, wrapping San Francisco in a shroud of gray uncertainty. I drive toward the airport through streets rendered unfamiliar by mist, past the ghosts of gold rushes past and the specters of revolutions still to come. The city recedes in the rearview mirror, a place between definitions, between eras, between versions of America still competing for dominance.

The fog will lift tomorrow, as it always does. What remains to be seen is what sort of city will emerge from the clearing—and whether those who sought to remake it will recognize what they have wrought.

It’s deep into February now, indeed nearly March, but years, like centuries, have periods of overlap and drift, in which various loose ends are tied up even as other themes emerge that define the next year.

First, my own year. Many of my readers know that I am passionate about the importance of native plants. In 2024, I found myself elected President of the Native Plant Society of New Jersey. Back in 2016 or so, when I left architecture and academia behind, it felt that somehow things were unwinding in those realms. In retrospect, I couldn’t have been more right. Architecture, which was revitalized with the modernist revival of the 1990s, now seems exhausted again—caught between spectacle, greenwashing, and the banality of developer-led projects. Academia has fared no better, suffocated by bureaucratization, infighting, and a slavish devotion to pseudo-leftist political commentary that left little room for real inquiry. My friends in academia have either quit or don’t enjoy teaching anymore. Meanwhile, landscape, long dismissed as secondary to architecture, has become a key site of innovation. But rather than innovative research taking place in the university, it is happening with individuals outside academia working with native plants. In academia, landscape still suffers from architecture envy and advocates reshaping the land violently using earth-moving machines before burying it under concrete. One can graduate virtually any landscape architecture program in this country without any real understanding of botany or plants. It’s as if architects had learned nothing from the reckoning the field faced in the 1960s and 1970s when its social failures and the consequences of object-fixation at the expense of context were laid bare. If native plant design advocacy must come from outside the academy, so be it. A talk I gave in November about designing with woodland plants had over 400 in-person attendees and has generated over 2,600 views in the three months it has been on YouTube. That’s already better than any talk I ever gave on the history of architecture or network culture. I’ll take that as a start.

I continued to write entries in the Florilegium, many of them essay length. Walls in the Landscape examined the cultural and ecological role of dry-stacked stone walls, reflecting on how they shape and structure the land while allowing nature to inhabit them. Vernal Pools at the Great Swamp explored seasonal wetlands in northern New Jersey, their importance for amphibians, and the growing threats posed by habitat destruction and climate change. A Trip to Lithuania and the Baltics documented my travels in the Baltics and engagement with Lithuanian native plant scientists and activists, examining the bizarre global trade in invasive species and the parallels between Eastern European and Northeastern American forest ecologies. We Went for a Walk on Turkey Mountain reflected on a hike through the New Jersey Highlands, using it as a way to think through geology, land use history, and native plant communities while drawing connections to Robert Smithson and conceptual art. A friend asked why I am writing these lengthy essays on landscape. Perhaps I am planning a book? Indeed, that’s the plan. That said, I also realize that essays can be a lot for people to take in all at once. Although I usually fail with these resolutions, I do intend to add more shorter pieces this year.

Beyond landscape, I continued my research with AI and AI image generation. It dismays me to see otherwise intelligent people so swiftly denigrate AI as plagiarism machines or as completely unreliable. AI, as I’ve stated before, is the biggest technological revolution of our lifetime. In my 2023 Year in Review, I suggested that “If potent but wildly hallucinating AIs marked 2022, the rise of GPT-4 as a useful and dependable everyday assistant marked 2023.” This continued in 2024. Although there have been no great new developments in AI—no Singularity, no Skynet, no AGI—and we are still using GPT-4 (GPT-4.5 is reportedly dropping this week), yet steady advances have continued. Setbacks made the news as well. In its usual fashion, Apple utterly mishandled the rollout of the unfortunately-named “Apple Intelligence.” Inappropriate summaries, lack of power, and a Siri that is every bit as dumb as it was when it was released in 2011 led to widespread disappointment. And yet, AI advanced steadily throughout 2024, becoming more deeply integrated into software development and research. Legal battles over training data and copyright raged on, but practical applications marched forward. AI-assisted coding through tools like Cursor and Github Copilot became more commonly used, and AI-powered search engines like Perplexity AI reshaped how we retrieved information. Through ChatGPT, I have an assistant that can do a more-than-serviceable job in translation to Lithuanian at a moment’s notice and can write various forms of code (I wrote a WordPress plugin for my site in AI this year and used it to teach me how to program an Arduino). With “deep research,” ChatGPT can search the Web, cite sources to confirm its accuracy, and produce a coherent research paper—not a full literature search and lacking original insights, but still an impressive overview. I’ve used Google Notebook LLM to generate podcasts about books that I don’t have time to read and even to understand manuals (see this Mylar Melodies video for how and why one might do this). I used Perplexity AI to plan a trip to France in October and it did an excellent job, down to recommending hotels and restaurants. I find it hard to imagine I could have found a travel agent who would have responded to my idiosyncratic requests so well. Specialist apps that use machine learning are everywhere now. Through iNaturalist, I use AI to identify plants in my garden and in the wild and with the Cornell Bird app, I can identify the hooting outside my house as a Great Horned Owl. Machine learning led researchers to decode entire passages from scrolls burned in the eruption of Mt. Vesuvius. AI is ubiquitous now, at least for some of us. As William Gibson said, “The future is already here—it’s just not evenly distributed.”

Like all technologies, it can be misused, but it is also transformative. From Leonardo da Vinci’s embrace of new painting technologies and geometric projection, Albrecht Dürer’s revolutionary use of the printing press to Eadweard Muybridge’s pioneering motion studies, László Moholy-Nagy’s creation of a painting by dictating its appearance over a telephone, Nam June Paik’s work with video, and John Cage’s explorations in electronic sound, artists have continually explored new technologies. The use of these technologies can be naïve, simplistic, or harmful, but it also advances knowledge. Our own time is now different. As a critic, I wrote a bit about this during the last year. My own interest has been in the visual unconscious and the questions it raises about authenticity and reproduction. I started with California Forever, or The Aesthetics of AI Images, in which I critiqued the AI-generated promotional imagery for the new city in Solano County for its failure to imagine the future and the uncanny similarity of not just the Solano images, but much of AI image generation to the paintings of Thomas Kinkade. I followed this with On the Pictures Generation and AI Art, where I explored how AI-generated images raised questions about the visual unconscious, the mechanics of cultural memory and hauntology, and how the boundaries between the authentic and the synthetic have shifted, contrasting AI art to the Pictures Generation of the 1970s and 1980s. Later in the year, I turned toward more art production itself, updating The Witching Cats of New Jersey in terms of both imagery and text, expanding the historical accounts while further analyzing folkloric and occult traditions to explore the intersection of myth and representation. I further examined fakery in the occult, particularly the parallels between spirit photography and AI-generated images—both technologies that blur the line between documentation and invention, creating spectral presences that challenge our perception of authenticity. I ended this year’s work with AI imagery with my essay Speculative Architectures: The Radical Legacy where I drew connections to the radical architecture movements of the 1960s. I find contemporary AI-driven architectural practice so boring, merely accelerating existing tendencies toward formal excess and doing nothing more. Instead, I was interested in how AI and automation intersect with architectural discourse in deeper ways, particularly through the lens of radical architecture movements of the 1960s and how groups like Archizoom and Superstudio used speculative design to exaggerate and expose the contradictions of late modernism—collaborating with AI to produce both images and texts. To push these ideas further, I co-created 7 Fables of Accelerationism with two AIs (ChatGPT 4o and Claude 3.5 Sonnet), producing a collection of speculative fiction pieces exploring AI, automation, and the dissolution of human agency in a world shaped by machine intelligence. These fables reflect both the utopian and dystopian possibilities embedded in technological acceleration, tracing the shifting relationship between architecture, labor, and meaning in a post-work society.

The final essay, Oversaturation: On Tourism and the Image came out last month but was really a product of 2024. Here, I examined the effects of overtourism and cultural overproduction, drawing connections between the Bilbao Effect, Instagram-driven travel, the ease of photography today, and the exhaustion of once-iconic destinations. At the heart of the essay is the concept of oversaturation—the point at which an excess of images, experiences, and cultural output dulls their impact, leaving audiences numb and places drained of significance. In an age when images are endlessly replicated and consumed, the relentless circulation of visuals flattens experience, reducing places to mere backdrops devoid of context or meaning. This commodification of place, fueled by social media’s demand for shareable moments, has led to a kind of cultural burnout. Tourism, in its current form, seems to have reached a point of diminishing returns. How long can it be sustained before the spectacle collapses under its own weight?

Oversaturation is the defining mood of 2024. With major cultural institutions competing to churn out new exhibitions and blockbuster shows, the traditional rhythms that once governed artistic production feel sped up as if on amphetamines. Every season brings another round of high-profile openings and all-too-many biennials, fueling a frantic chase for novelty. The obsession with simplistic politics in the art world has burned out, but without any substitute. Institutions have been left rudderless. For too long, writers and curators have defined movements that only last as long as a single show: sound art, tactical urbanism, post-Internet art, zombie formalism, NFTs, the covidean, parametricism, “the new aesthetic”, dimes square/indie sleaze revival, and so on. Nobody cares anymore, except maybe some art school graduates out for bad wine and parties. In architecture, movements have been less prone to such rapid obsolescence, but the energy similarly has been lost. Where “starchitecture” used to captured headlines, such celebrations of elite wealth are ubiquitous in cities now and there is no difference between starchitecture and junkspace. Thomas Heatherwick’s the Vessel is the punctum at the end of starchitecture: a structure whose highest purpose seems to be to overwhelm visitors with despair until they fling themselves over its side. Nor is there room for an alternative: once subversive, blogs, zines, and architecture fiction have faded, abandoned by a generation more concerned with profit.

The same goes for the news. After years of constant crises and hyperbole, the public has reached a point of fatigue and skepticism. The endless drumbeat of dire warnings from all corners no longer commanded them. Where once Americans took to the streets at a moment’s notice, now people who identified with the Resistance of 2016 seem worn out. Instead of galvanizing new mass protests, the news cycle spawns shrugs and eye-rolls. It’s not outright hostility, just exhaustion. Our sensorium simply can’t take constant screaming anymore. In The Week, Justin Klawans calls 2024 “a year of reckoning for the fourth estate.” Indeed it was. While the Right is taking advantage of this in the US at the moment, I have all the confidence that they will experience a similar overload. The endless churn of the news cycle during the Trumpenjahre is going to take its toll. Indeed, Klawans ended with the following sentence “The ‘legacy media is dead. Hollywood is done. Truth-telling is in. No more complaining about the media,’ right-wing activist James O’Keefe said on X. ‘You are the media.'” But social media is equally ill.

Engadget editor Cheyenne Macdonald writes “It’s never been more exhausting to be online than in 2024. While it’s been clear for some time that monetization has shifted social media into a different beast, this year in particular felt like a tipping point. Faced with the endless streams of content that’s formulated to trap viewers’ gazes, shoppable ads at every turn, AI and the unrelenting opinions of strangers, it struck me recently that despite my habitual use of these apps, I’m not actually having fun on any of them anymore.” Too many ads and badly written algorithms have crushed content. Desperate to wring engagement from already tired users, social media firms compounded this with frantic moves that often backfired. Many people left Twitter when Elon Musk purchased it; many more left in the subsequent months. Meta’s repeated attempts to replace Instagram’s photo sharing with video reels and the addition of new algorithm tweaks there and on Facebook led to further user drift and confusion. Frustration mounted with links being demoted, smaller creators seeing their reach throttled, and online communities splintering all contribute to a general sense of retreat from the clamor. I notice that friends leave for BlueSky, which leans left-wing, containing as much extreme and violent language, if not more, than Twitter has now (calls for the death of someone disliked by leftist radicals are common), and then they fall silent. TikTok was briefly banned in the US, then restored, but there is bipartisan support against it and the platform’s future is in doubt. Group chats are also dying, a decline captured in Tony Tulathimutte’s story “Pics” from his 2024 collection. By now, anyone who has been on Discord for a while sees a total mess, with far too many servers and no coherency. The overall narrative is one of people stepping back rather than diving in. With everyone shouting to be heard, most are simply tuning out. Yet a handful of dedicated readers still seek out independently produced content wherever it can be found, perhaps the last outposts of genuine engagement in a sea of hype and oversaturation.

15 years ago, I suggested that postmodernism was dead and network culture was upon us. Now, it seems that a new era is being born, its outlines as yet unknown. AI is going to be as much a part of this as the Internet was for network culture and the televisual, the photocopier, and the personal computer were for postmodernism. But if postmodernism was, in Fredric Jameson’s famous line, “the culture of late capitalism,” it strikes me that something different is underway in contemporary culture. For the first time, population growth has ceased in the developed world. From China to the US to the EU, population growth is declining faster than experts had predicted even a decade ago. Since 2017’s Year In Review, I have observed the parallels between our world and “the Jackpot,” the slow-motion collapse first introduced by William Gibson in his novel The Peripheral (2014). Instead of the comet strikes or nuclear annihilation imagined by Hollywood, Gibson’s Jackpot is a series of cascading crises: climate change, resource depletion, biodiversity collapse, and social upheaval, exacerbated by the very technologies that sustain modern life. For Gibson, the Jackpot signifies an ongoing collapse punctuated by moments of technological innovation—innovations that serve the privileged few while leaving the vast majority to suffer and scramble for survival. Gibson’s vision is compelling and grim. He portrays a world where survival is a lottery of wealth and sheer luck, with the richest securing their future through technology and the poorest left behind in failure zones. In his fiction, the Jackpot is defined by stark inequality, unrelenting violence, and scientific advances that, while transformative, fail to offset the broader disintegration of society and ecology. Yet, the real Jackpot diverges in key ways.

I see the Jackpot less as a singular dystopia and more as a chronic condition, simultaneously an enduring state of polycrisis and a slow improvement in impact on scarce resources due to declining population growth. Lower birth rates and aging populations are rapidly accelerating worldwide just as artificial intelligence and automation promise to upend labor markets. The Right—from Putin and Xi Jinping to Musk—has raised alarms about declining birth rates, yet even by adding cash payouts for births (a move popular with liberals as well), it has been unable to change matters. But with AI, the global economy seems poised to pivot away from population growth as its primary driver. At the same time, population decline is necessary—we already exceed the carrying capacity of the Earth and with less resource consumption and less pollution, this Jackpot may yet create a significantly better world. This seems like an essential point of our contemporary culture: we are seeing the beginning of an age of population contraction.

And if AI is poised as a potential solution to the end of population growth and the inauguration of an age of limits—assuming much goes well— it also deepens inequalities, concentrating power and productivity in fewer hands. The uneven distribution of this future is already stratifying societies. For those with access to cutting-edge tools and the drive to use them, 2024 was a year of acceleration—a leap in productivity. For others, it was a year of stagnation or retreat, defined by fear of change more than the inability to participate in this transformation. The Jackpot is not just about access, it is also about the growing divide between those who can adapt and those who cannot.

Understanding the Jackpot means grappling with this unevenness. It is not the apocalypse, but it is a reckoning. It demands that we rethink what progress looks like when (population) growth is no longer the default. Breakthroughs, breakdowns—or more likely both—we are all already living in the Jackpot. Whether it is a slow-motion end or a new beginning depends on how we, individually and collectively, choose to play our hand.

In 2013, I proposed a Network Architecture Lab symposium at Columbia centered on the topic of oversaturation—a concept that I envisioned encompassing not just what would become known as “overtourism,” but the crisis of cultural overproduction as a whole. Even though we had identified a funding source for the symposium, mysterious administrative forces kept throwing wrenches in the works. Perhaps it conflicted with the school’s larger agenda; after all, how could an institution premised on constant growth, endless publication, and global expansion confront the consequences of those conditions? Perhaps there was concern that donors might be turned off? It was still the heady era of post-GFC growth, when urban boosters hoped that a shiny new museum or concert hall would bring in tourists by the plane full. We tossed in the towel and moved on to other projects instead, things architects could more easily understand, things that weren’t going to get frowned on by the administration. Coined on travel industry website Skift some three years later, “overtourism” rapidly became a buzzword. But oversaturation was a broader concept, and as I am outside university supervision these days, it’s time to revisit and explore it in further depth.

As I started this essay a few months ago, I was in France on a variety of matters—visiting two shows on my father’s paintings (one at the Centre Pompidou and one at the Fondation Vasarely) as part of the Saison de la Lituanie en France, attending a residency in the Pyrénées, and exploring parts of the country that I had not previously visited. Besides France, I went abroad several times during the past year—to Japan, Berlin, Lithuania, Latvia, and Estonia. Overtourism dogged these travels, especially on the April trip we made to Tokyo and Kyoto. We had a miserable morning at the Sensō-Ji Temple in Asakusa, Tokyo, which tourists had overrun. Burned by this experience, we decided to skip such sites. In Kyoto, we passed on notable locations like the Fushimi Inari Shrine, Gion, and the Imperial Palace. This was disappointing, but we had our hands full going to locations off the beaten path. We didn’t take selfies—what, precisely, is the point of those?—but we were still tourists, no different than the folks wearing Disneyworld T-shirts, although we made efforts to be respectful of local life. For his part, our son—who is now pursuing a degree in game studies—took late-night trains out to distant suburbs to find obscure video games at used game stores, just like my father endlessly visited bookstores for art books and I looked for books on architecture. Obsessions with culture cross generations in a family of artists, even as the objects of that obsession may change.

Tourism is a foreign country for me. With my father as an artist in a single-income household, we weren’t wealthy; his time and money went into his book, map, and art collections. I have a dim memory of a couple of brief vacations at the beach—ordered by his doctor—while I was very young, but apart from one night at Niagara Falls in 1976, we did not spend a single night away except for his professional interests, and even those vanished after we left Chicago in 1979. When we went somewhere, it was a research project: museums were inspiration for an artist, not some kind of amusement. He was not a foodie. Eating was for sustenance only, the cheaper the better. Amusement parks and theater were nothing but wastes of money. Disneyworld was never in the cards; I didn’t even ask about it.

By the time I had graduated college and had the opportunity to travel, I was already an historian of architecture. My first trip abroad was a month in Italy on a Grand Tour that I’d received a scholarship for as part of my studies. The allowance for room and board were minimal, but it gave me the opportunity to see buildings firsthand. For an historian, travel to see buildings was a prerequisite. In retrospect, I wonder if I chose the profession to permit myself to travel or at least rationalize the activity to myself and my father. Still, it meant that my own travels always had a research component to them.

There is nothing new about all this. The practice of tourism itself can be traced back to the Grand Tour, a tradition that began in the 17th century as an educational journey for young European aristocrats, primarily from Britain. Designed to cultivate the intellect and refine the sensibilities of young aristocrats by exposing them to the art, culture, and architecture of classical Europe, the Grand Tour was more than mere travel—it was an immersive experience to shape the mind and character. The traveler often spent months or even years abroad, moving through the great cities from Paris to Florence to Rome, while a tutor (the cicerone) gave firsthand instruction on the sights and their relationship to the classics. Studying ruins firsthand, travelers would contemplate the fall of that empire even as Britain was building its own.

Tourism and imagery were tied together through their origins in the rise of print culture, which made classical literature and art widely accessible. Early print was dominated by the reproduction of Greek and Roman classics, sparking widespread curiosity about the landscapes and monuments of ancient Rome, Greece being occupied by the Ottoman Turks and inaccessible to the West, a condition compounded by the achievements of Renaissance Italy. The eighteenth century saw the emergence of travel literature and illustrated guides that created idealized representations of these landscapes, fundamentally changing how tourists experienced and interacted with places. This era was also marked by the rise of the “picturesque,” a concept first introduced by writer William Gilpin in his 1768 “Essay Upon Prints,” a short treatise on how to collect the emerging media. Gilpin’s “picturesque” was an effort to appreciate landscapes by treating them as mediated images through careful framing. This approach was augmented by devices like the Claude Glass, a small, portable, dark-tinted convex mirror named after the painter Claude Lorrain, whose idealized pastoral scenes set the standard for “proper” composition, this small, portable, dark-tinted convex mirror reflected back a softened, flattened, and slightly idealized image, making any view appear more like a painting. Using the device required an almost ritualistic set of actions: the viewer would face away from the scene, hold up the mirror, and adjust the position until the reflection achieved the desired effect. This peculiar choreography—turning away from reality to see it better—perfectly encapsulates how the Grand Tour taught travelers to experience place through representation. The Claude Glass transformed raw nature into composed views, making landscape itself into something that could be collected and consumed. Even more significantly, it trained viewers to expect mediated images and even prefer them to direct experience, establishing a pattern that would shape tourism for centuries to come.

But actual image-making was menial work, not something the British upper class could afford to be seen indulging in, moreover, it required extensive time for training that was better spent with the humanities, law, and other studies. Instead, the wealthier Grand Tourists acquired art and antiquities systematically, building personal collections while also transferring cultural capital from Italy to Britain, thus forming the basis of major British institutions like the British Museum.

To satisfy the growing interest in architectural imagery, artists such as Canaletto and Giovanni Battista Piranesi produced images of local sites for these audiences. Canaletto dominated the market for vedute (view paintings) in Venice, creating meticulously detailed scenes that satisfied tourists’ desire for perfect memories of their Italian sojourn. Canaletto’s vedute were prized for their apparent objectivity and precision—though he frequently adjusted perspectives and rearranged architectural elements to create more pleasing compositions. His capricci, blending real and imagined elements into fantastical compositions, appealed to collectors drawn to a more romanticized vision of the past, offering both documentation and escapism. His primary patron, Joseph Smith, the British Consul in Venice, effectively operated as his agent, selling Canaletto’s works to British travelers and eventually selling his own massive collection to King George III. This relationship proved so profitable that when British tourism to Venice declined during the War of Austrian Succession (1740-48), Canaletto followed his market to England, painting similar views of London for aristocratic patrons.

Many tourists saw Canaletto and Piranesi’s vedute before they ever visited Venice and Rome, creating a certain level of expectation. Some enthusiastically proclaimed that the real thing exceeded its image, but others were disappointed by the reality and the latter experience only increased as the eighteenth century wore on and living conditions in London improved.

The market for views was hierarchical. Canaletto’s oil paintings commanded the highest prices and greatest prestige, purchased by wealthy aristocrats as the ultimate proof of cultural refinement. Below these were watercolors and gouaches by artists like Francesco Guardi, who offered a quicker, more atmospheric, and more affordable alternative. At the bottom were etchings and engravings by artists like Piranesi and Giuseppe Vasi, mass-produced for less well-heeled tourists (Canaletto’s works were also reproduced for wider consumption as prints by Antonio Visentini and other engravers). Working in etching, a “lower” medium and initially catering to less sophisticated tastes, Piranesi transformed architectural representation through works such as his Vedute di Roma. These dramatic views of Rome’s ruins combined archaeological precision with theatrical imagination, creating images that were simultaneously documentary and fantastical. Where Canaletto presented a vision of Venice—then in decline—as a great and proseperous city, Piranesi’s Rome was a city of ruins, inhabited by as many cows and goats as people. Both artists heavily embellished their work. Canelleto’s Carceri d’Invenzione (Imaginary Prisons) series went further, using architectural elements from Roman ruins to create impossible, sublime spaces that would influence everything from Romantic poetry to surrealist art.

Piranesi transformed what was essentially a form of commercial art into something more profound: a way of experiencing architecture through imagery that was both documentation and affect, establishing patterns of visual consumption that would shape centuries of architectural representation. Today, Piranesi is now highly collectable: his Vedute di Roma was my father’s most prized possession and, although our family donated it to the Lithuanian National Museum, would be worth a small fortune today.

Photography’s role in tourism began soon after the medium’s invention, with figures like Maxime Du Camp and Francis Frith leading the way. Using photography to document their journeys, these early travelers created lasting visual records of the places they visited. Du Camp, who traveled to Egypt and the Middle East in the 1840s, captured images of ancient monuments and ruins, which were later published alongside his travel writings in Égypte, Nubie, Palestine et Syrie (1852). Believing that there was a large market for imagery outside of the traditional Grand Tour sites, Frith made several trips to Egypt and Palestine in the 1850s and 1860s. His photographs of ancient ruins and landscapes became highly sought-after souvenirs and were widely published in albums such as Egypt and Palestine Photographed and Described (1858-1860) and reproduced on postcards. Frith’s detailed and carefully composed images shaped how Europeans perceived these distant lands, blending tourism with a kind of visual ethnography.

Nor was photography limited to merely emulating painting. New technologies created unprecedented opportunities for visual culture. In particular, the stereoscope allowed viewers to see distant sights in three dimensions. This required precise calculation: two photographs had to be taken simultaneously from points about two and a half inches apart (matching the distance between human eyes) using a special double camera and then precisely printed and mounted on cardstock. When viewed through the stereoscope’s lenses, the images merged in the viewer’s perception to create an uncanny sense of depth and presence. Exhibited at the 1851 exhibition at the Crystal Palace, the stereoscope captivated Queen Victoria and Prince Albert and the attendant publicity led to some 250,000 being rapidly manufactured and sold. The London Stereoscopic Company, founded in 1854, operating under the slogan “No home without a stereoscope,” sponsored photographers to travel the world and sold hundreds of thousands of stereoscopic cards. Underwood & Underwood, founded in 1881 in Ottawa, Kansas, became the largest American producer, publishing 25,000 cards per day by 1901 (I own a small collection of these that we found in the house my parents bought in Massachusetts in 1977). These companies employed photographers who followed established tourist routes across Europe and the Mediterranean, carefully positioning their double cameras to capture the same monuments and views that had attracted Grand Tourists. Specialized techniques enhanced the stereoscopic effect—positioning foreground elements to emphasize depth, using strong diagonals, and carefully calculating viewing angles to maximize the sense of space. Virtual tourism was possible for the first time, allowing distant spaces to be experienced in the parlor.

Another new photographic technology was the “magic lantern,” or lantern slide projector. Generally attributed to Christiaan Huygens, a Dutch scientist who also discovered the rings of Saturn, this technology was first developed sometime in the 1650s. Refined and popularized throughout the 18th and 19th centuries, the magic lantern sent the light through a series of lenses to project images from glass slides onto a wall or screen. Initially, the device was largely used to produce spectacular displays of ghostly images such as skeletons, demons, and other supernatural figures to entertain and frighten audiences. During the late 18th and early 19th centuries, educators and scientists began to recognize the magic lantern’s potential for illustrating lectures, using it to project scientific diagrams, landscapes, and other instructive images for audiences, a practice that became more widespread with the advent of photography and mass-reproduced glass slides. Institutions such as the Royal Institution of Great Britain and universities across Europe incorporated slide lectures into their regular programming. They were particularly important in fields that required visual examples, such as art history, architecture, and natural sciences. Art historian Heinrich Wölfflin pioneered formalist art history by putting two magic lanterns side by side. In doing so, he could compare and contrast different works of art in real-time, a method that became central to his comparative art historical approach and to art and architecture history courses until the era of PowerPoint.

As photography advanced, glass slides featuring photographs of paintings, buildings, and monuments were produced in increasing numbers, with firms like the Alinari Brothers playing a major role in documenting European art and architecture. Founded in 1854, Fratelli Alinari (the Brothers Alinari) specialized in documenting Italian art and architecture, producing thousands of photographs that became valuable resources for scholars and travelers alike. The Alinari Brothers’ photographs became staples in architectural study, particularly for scholars who could not easily visit the sites themselves.

These early travel photographers laid the foundation for a visual culture of tourism as well as for the forensic use of photography in the history of architecture. Architectural historians relied heavily on such visual documentation to inform their studies and deliver lectures. Soon, universities and museums developed slide libraries. Institutions like Columbia University, the Courtauld Institute in London, not to mention my own Cornell University amassed vast collections of slides, including images by photographers such as Frith and the Alinari Brothers together with images taken by their own faculty on research trips. These collections enabled architectural historians to compare buildings from different time periods and geographic locations, contributing to the development of the comparative method in architectural study.

Historians of architecture took up photography themselves, using it as a direct tool for research. Banister Fletcher used photography extensively to illustrate his arguments in his A History of Architecture on the Comparative Method, employing Wöllfflin’s comparative method to create visual arguments that textual descriptions alone could not achieve. Many historians like Nikolaus Pevsner and John Summerson were also known to document their travels through photography, capturing the buildings they studied in meticulous detail. Their images, often included in their writings, provided not only records of their subjects but also a means of interpreting the architectural qualities of the structures. Pevsner’s surveys of British architecture, for example, relied on photographs to convey stylistic details that were otherwise difficult to express in text.

While historians of art and architecture honed these practices within the academy, the travel slideshow was developing for a broader public and, eventually, as a middle-class pastime. Travel slideshows emerged from the learned societies of the early 19th century, where geographical societies used painted slides to illustrate lectures about explorations and discoveries. Lecturers at the Royal Geographical Society, for example, initially relied on drawings, maps, and hand painted images transferred to glass slides. The advent of photography in 1839 gradually transformed these presentations. By the 1850s, explorers were beginning to use photography to document their journeys, though the cumbersome nature of early photographic equipment limited its use in the field. The Royal Geographical Society began collecting photographs in the 1850s, with Roger Fenton becoming its first official photographer in 1852. Photographic slides became increasingly common over subsequent decades. More portable cameras and dry-plate photography made it more practical for explorers to document their journeys photographically. When the National Geographic Society was founded in 1888, photography was already central to its mission of documenting the world. Starting with their 1896 volume, the National Geographic Magazine styled itself as an “illustrated monthly,” featuring photographs of distant lands and peoples and cementing the relationship between travel and photography.

Through the early decades of the twentieth century, the lantern slide lecture gradually moved beyond elite scientific societies and universities to broader cultural institutions. Libraries, local history museums, and civic organizations regularly hosted slide-based travelogues, as did camera clubs and photographic societies. The Eastman Kodak Company recognized and encouraged this democratization of travel photography, developing increasingly portable cameras and, in 1934 introduced 35mm still film in daylight-loading cassettes with Kodachrome color slide film coming two years later. World War II accelerated this transformation as millions of American servicemembers traveled abroad, many carrying 35mm cameras and returning with images of far off places that they shared with their communities.

By the 1950s, travel and photography had become inextricably linked for the American middle class. The postwar economic boom made both more accessible: the 35mm camera became a standard possession while paid vacation time and the interstate highway system created new possibilities for travel. Even as international trips remained expensive, they were no longer the exclusive domain of the wealthy. A middle-class family might save for years for a single European tour, but such trips were now conceivable. The Kodak Carousel slide projector, introduced in 1961, allowed travelers to more easily show color slides at social events by eliminating the need to hand load slides. Living rooms would be transformed into makeshift theaters as neighbors gathered to see images of the Grand Tour, now documented in Kodachrome. Travel slideshows became a form of cultural capital, letting middle-class Americans demonstrate their sophistication while sharing their experiences with those who couldn’t afford to travel themselves. In practice, however, many amateur slideshows were guaranteed to induce sleep in darkened living rooms, and were often dreaded by invitees.

In Soviet Lithuania, slide lectures were rare and highly anticipated events, offering glimpses of the Western world. For example, the Nasvytis brothers—leading architects who served as professors at the Vilnius Civil Engineering Institute and had gained enough trust from the authorities to occasionally travel abroad—would organize presentations. Only a select group could attend: other architects and professors, cultural figures deemed ideologically reliable, and students whose records suggested they would not cause trouble. Knowing that the presentations were monitored for ideological conformity, the brothers carefully curated these presentations, selecting images that could pass official scrutiny while conveying developments in Western architecture. They had to maintain a delicate balance, showing enough to educate but not so much as to suggest the superiority of Western achievements. Such presentations were part of a broader pattern in which certain cultural figures—primarily established artists, architects, and academics who had demonstrated loyalty to the system—were permitted limited engagement with the West, always with the understanding that this privilege could be revoked.

The shift from careful documentation to ubiquitous capture traces the rise of photography into oversaturation. When the Kodak Brownie democratized photography in 1900, it remained a deliberate act—each image had to count. Even as production grew—from one billion photographs annually in 1930 to ten billion by 1970—photography retained its role as the intentional documentation of significant moments and places. The development of Kodak’s inexpensive, easy-to-use, but low-quality Instamatic in 1963, followed by Polaroid’s SX-70 in 1972, and Fujifilm’s disposable camera in 1986 all contributed to the rise of more casual—and more frequently taken images—but it took the arrival of digital photography, smartphones, and finally the launch of Instagram in 2010 to transform our relationship to images entirely. We now produce more photographs every day than were taken in entire years of the pre-digital era. With 1.3 billion images shared daily on Instagram alone, photography has moved from documentation to constant ambient production, from captures of the significant to an endless stream of the everyday.

I have repeatedly asked myself if, having left the academy behind for independent practice, I have merely joined the endless ranks of tourists. And even if I now travel like my father, for ideas for my art and writing, what does it mean to photograph buildings when likely as not, one can find similar—or better—images on the Internet with a simple search? The methodical documentation that characterized architectural history—the careful positioning to capture essential details, the systematic recording of building elements—seems increasingly disconnected from contemporary practice. Where Pevsner and Summerson’s photographs served as both evidence and analysis, today’s architectural photography operates in an environment of infinite reproducibility and instant access. Academics can just download images from the Internet instead of making pilgrimages to the monuments. The photographer’s craft, once essential to architectural history, has been simultaneously democratized and devalued by digital abundance. Slide libraries have been liquidated, just as card catalogs have been hauled to the landfill. And what of the tours of architecture, led by historians and organized by the Society of Architectural Historians as well as so many universities and museums? Are these research or tourism? What of the visits to other cities we make under the aegis of “research” for exhibitions, such as my visits to Hong Kong for the 2014 MoMA “Tactical Urbanism” show? What about botanical photography? Am I really doing any good, hiking with iPhone in hand, documenting species with iNaturalist?

The lines between scholarly documentation and tourist snapshot, between analysis and consumption, between purposeful recording and compulsive capture, have blurred to the point of disappearance. When every angle of every significant building is already available online, what distinguishes the architectural historian’s photograph from the tourist’s? Perhaps only the increasingly quaint notion that our images might contribute to some larger project of understanding.

In reflecting on the Grand Tour and its modern counterpart, I see a fundamental shift in how we relate to the act of travel. The Grand Tour was predicated on the idea of cultivation—of educating oneself through prolonged exposure to other cultures and histories. Today’s tourism, in contrast, seems predicated on consumption. We no longer seek to deepen our understanding of the places we visit; instead, we literally consume them by reducing them to a set of aestheticized images and move on.

The tourist photograph functions as a contemporary Claude Glass—a way of transforming reality into something consumable, shareable, and aesthetically pleasing. Sites like the Pantheon, the Eiffel Tower, or Kyoto’s Fushimi Inari Shrine are now experienced through the screen of the smart phone, just like the social media the images are destined for. As visitors line up to take their photos, these sites become less about personal exploration or historical engagement and more about capturing a replicable moment, one that will be instantly validated by likes and shares on social media. How can one even see these sites anymore?

But it isn’t only the immediate experience that is threatened by the image, the actual memory of the experience is as well. Neuroscientists call this phenomenon the “photo-taking impairment effect”—the act of capturing an experience outsources its memory to the device, leaving the mind empty. In focusing on documenting the moment, we lose the cognitive and emotional engagement necessary to form lasting memories. What remains is a shallow trace of the experience, stripped of reflection and depth, replaced instead by the image itself. This phenomenon parallels the experience of Piranesi’s tourists, who arrived in Rome clutching engravings of ruins they had never seen, only to find their imaginations already colonized by someone else’s vision. Just as these engravings predetermined their expectations, today’s Instagram feeds serve as the preordained script for how places must be seen, photographed, and shared.

This reduction of memory to image reshapes how we interact with cultural heritage. Instead of spaces for contemplation, we get backdrops for a personal brand. Every photograph becomes both a proof of experience and a barrier to engagement, perpetuating the very oversaturation that alienates us from the spaces we visit. Even as these sites gain visibility, their existence dissolves into the noise of infinite replication.

In this way, modern tourism can be seen as both a continuation and a departure from the Grand Tour. The framework of travel remains—the movement through culturally significant sites, the desire to encounter the foreign—but the internal experience has changed. What was once a journey of intellectual and personal enrichment has become an exercise in superficial documentation. The modern tourist moves quickly, capturing images and moments to be shared, but rarely pauses to reflect on the experience in any lasting way.

While the transformation of tourism offers perhaps the most visible manifestation of this cultural condition, oversaturation extends far beyond the overcrowded sites and Instagram-optimized viewpoints that characterize contemporary travel. As I suggested for the symposium over a decade ago, oversaturation describes a state where systems of cultural production exceed their capacity for meaningful absorption or engagement—but it manifests as more than mere excess. It emerges when the imperative to produce overwhelms the possibility of reception, creating a peculiar state where increased production actually diminishes the capacity for meaningful engagement. We see this not just in overtouristed sites where crowds make it impossible to truly see a monument, but across all domains of cultural production: in the academic world where more papers are published than could ever be read, in an art market where more exhibitions open each night in any global city than anyone could possibly attend, in media landscapes so crowded that attention itself becomes the scarcest resource. The tourist’s compulsion to document rather than experience mirrors the academic’s pressure to publish rather than think, the artist’s need to exhibit rather than develop, the writer’s push to post rather than reflect. In each case, the system demands constant production while making deep engagement increasingly impossible.

Potential responses to oversaturation seem to fall into predictable patterns that ultimately reinforce the system they aim to resist. Consider how movements like slow food, farm-to-table, and local food initiatives, while seeking to resist global homogenization, actually intensify certain forms of tourism and consumption. The small local restaurant becomes internationally famous precisely for being local, drawing food tourists from around the globe who take shots of the careful plating for social media. The farm-to-table movement creates new forms of destination dining where the “authentic” local experience becomes more hyperlocal than the local ever was in the first place. A neighborhood pasta maker in Bologna or a tiny sushi restaurant in Tokyo becomes a mandatory stop on global circuits of food tourism, their very resistance to globalization transforming them into nodes in global networks of consumption and content production. Any tie to local culture is, ultimately, gone.

Former practices of resistance have been absorbed into this system of cultural oversaturation. The Situationists’ dérive—their practice of aimless urban wandering—has been commodified into “off the beaten path” tours advertised on Airbnb Experiences. Psychogeography becomes Instagram “hidden gems” guides. Even urban exploration, once a transgressive practice, has been transformed into “ruin tourism” complete with designated photo spots and hashtags. Anyone can take a tour of Chernobyl, with time for plenty of photographs. The system doesn’t just resist critique—it metabolizes it, transforming opposition into content.

Oversaturation produces spaces where physical and digital presence create reinforcing cycles of visibility. Consider how a single viral photo can transform an obscure location into a must-visit destination, generating more photos, more visibility, more visitors, until the original site is completely transformed by its own fame, destroying the very qualities that made a place attractive in the first place—a quiet beach becomes crowded with influencers, a contemplative temple becomes a selfie backdrop, a local café becomes a tourist trap.

Consider Les Deux Magots in Paris, once a revolutionary intellectual hub of the avant-garde, now one of the city’s most notorious tourist traps. The preferred café of Surrealists like André Breton and writers like Simone de Beauvoir, where artists and existentialists engaged in heated debate, it now serves primarily as a site of cultural performance where tourists can cosplay at being Parisian intellectuals. The café’s fame as a site of cultural production ironically destroyed its capacity to produce culture, replacing genuine intellectual exchange with its simulation. Yet even this transformation has become content —the café’s website proudly markets its history of artistic patronage, turning its own commodification into a selling point. Here we see how oversaturation operates across time: the cultural capital accumulated by generations of artists and intellectuals becomes transformed into tourist capital, which then generates its own forms of content about the loss of authenticity.

This transformation—from site of cultural production to site of cultural consumption— exemplifies how the drive for content creation has begun to exhaust not just physical spaces but cultural forms themselves. Just as overtourism can destroy the very qualities that made a destination attractive, cultural oversaturation depletes the meaning-making capacity of our systems of cultural production. When every moment must be documented, when every experience must be shared, when every observation must be transformed into content, we begin to lose the ability to distinguish between meaningful cultural engagement and mere performance of engagement. The documentation imperative becomes a form of cultural strip-mining, extracting content value while depleting the underlying capacity for meaning.

This oversaturation manifests across multiple domains: in the glutted academic job market where PhDs far outnumber available positions, in an art world where more exhibitions open each night than anyone could possibly attend, in conferences nobody attends, in media landscapes so crowded that attention itself becomes the scarcest resource. Yet production continues, driven by institutional imperatives and professional requirements that seem increasingly disconnected from any meaningful measure of impact or engagement. The academic must publish, the artist must show, the tourist must document—not because these activities serve any clear purpose, but because stopping would mean becoming invisible in a system in which visibility is everything.

This is oversaturation. Tourism is merely the most visible symptom of a broader condition—one where systems of cultural production have exceeded their capacity for meaningful absorption or engagement. The system demands constant production while simultaneously making meaningful engagement with that production increasingly impossible. What emerges is a kind of cultural attention deficit disorder—a system that requires overproduction to maintain itself while ensuring that no single piece of content receives too much sustained attention. The constant flow of new content keeps users engaged while preventing any single piece of content from demanding the kind of sustained attention that might interrupt that flow. The tourist moves quickly from site to site, the academic skims rather than reads, the art viewer passes briefly through exhibitions. This, not cinema (pace Benjamin) is how art is perceived through distraction today.

The virtualization of physical space creates new forms of cultural geography. Sites are no longer organized around historical or cultural significance but around their potential for content production. The “most Instagrammable” spots become the new centers of cultural gravity, while spaces resistant to photographic reproduction—however historically or culturally significant—fade into obscurity. This reorganization of space around platform logic produces a topography organized not by physical or cultural features but by the ability to generate engaging content.

The parallels between academic metrics and social media validation reveal a deeper convergence in contemporary culture. Just as scholars carefully time their paper submissions and cultivate citation networks, tourists and influencers optimize their posting schedules and build engagement pods. Both systems create eerily similar behavioral patterns—strategic timing, network cultivation, content optimization—homogenizing all forms of cultural validation into a single, universal metric of attention. The academic h-index and the influencer engagement rate become functionally equivalent measures in an attention economy that increasingly fails to distinguish between forms of cultural capital.
This homogenization of cultural metrics points to a broader crisis in how we assign and measure value. When everything must be quantified—from tourist footfall to citation counts to social media engagement—we create systems that inevitably demand more production, regardless of whether that production serves any purpose beyond feeding the metrics themselves. But how do you critique a system of overproduction from within institutions that require constant production to justify their existence?

The smartphone emerges not just as a tool of documentation but as both metaphor and mechanism for this condition. Just as it collapses the temporal stages of tourist experience into a simultaneous present of planning-experiencing-documenting-sharing, it also collapses traditional distinctions between producer and consumer, professional and amateur, meaningful contribution and mere content. The academic writes papers while scrolling through social media, the artist documents their process while checking their likes, the tourist experiences a place while planning how to present it—all participating in a system of continuous partial attention that paradoxically produces more while experiencing less. What’s particularly striking is how institutional responses to oversaturation often exacerbate the very conditions they claim to address. Universities respond to the glutted academic job market by creating more temporary positions, demanding more publications, and creating more documents to fill out. Art institutions address declining attention spans by programming more shows, more events, more content from the usual suspects. Tourist sites manage overcrowding by implementing time-slot systems and virtual queues that transform experience into yet another form of scheduled content consumption. Each solution feeds back into the system of overproduction, creating what we might call “saturation spirals”—self-reinforcing cycles where attempts to manage oversaturation just lead to exponentially higher amounts of it.

This system creates new forms of professional precarity that span seemingly disparate fields. The adjunct professor piecing together teaching gigs mirrors the gig economy worker managing multiple car-share platforms, the artist maintaining visibility across multiple venues echoes the influencer juggling multiple social media accounts. All are caught in a saturation spiral, facing the necessity of constant production and self-documentation not for any intrinsic purpose but merely to maintain visibility within oversaturated systems.
The relentless drive for content creation has begun to exhaust not just physical spaces but cultural forms themselves. Just as overtourism can destroy the very qualities that made a destination attractive, cultural oversaturation depletes the meaning-making capacity of our systems of cultural production. When every moment must be documented, when every experience must be shared, when every observation must be transformed into content, we begin to lose the ability to distinguish between meaningful cultural engagement and mere performance of engagement. The imperative to document becomes a form of cultural strip-mining, extracting content value while depleting the underlying capacity for meaning.
These examples point to a fundamental transformation in how cultural value is produced and consumed. The local restaurant that becomes a global destination, the academic paper that circulates more as a citation than as a text to be read, the artwork that exists primarily as social media content all reflect a system where the metrics of attention have become more meaningful than the cultural experiences they supposedly measure. The irony of oversaturation is that it produces not just an excess of content but a scarcity of attention, creating an attention paradox where increased production leads to decreased engagement.

This paradox manifests differently across domains but follows similar patterns. In academia, the pressure to publish leads to the slicing of research into “least publishable units,” creating more papers but less comprehensive analysis. In the art world, the proliferation of biennials and art fairs creates a constant cycle of production and display that is strongly proscribed to avoid either failure or success and leaves little time for artistic development or critical reflection. In tourism, the race to document and share experiences prevents the very forms of engagement that might make those experiences meaningful. Each domain faces its own version of the same crisis: how to maintain relevance in a system that demands constant production while making sustained attention increasingly impossible.

The institutional structures that emerged to support cultural production—universities, museums, publishing houses, tourism bureaus—now find themselves generating content to justify their existence, even as that content contributes to the very oversaturation that undermines their original purpose. The university must produce more research, the museum must mount more exhibitions, the publisher must release more titles, the tourist site must accommodate more visitors—each institution caught in a cycle of expansion that seems increasingly divorced from any meaningful cultural purpose.

What makes this condition particularly difficult to address is how it transforms even failure into content. The unfilled academic position becomes data for studies on the job market. The overlooked artwork becomes evidence of systemic inequality. The overtouristed site becomes fodder for articles about overtourism. The system doesn’t just resist critique—it metabolizes it, transforming opposition into content that feeds back into the very systems being criticized. Even this very analysis, attempting to understand the condition of oversaturation, necessarily participates in the production of content about overproduction.

Perhaps the most telling symptom of oversaturation is how it has begun to exhaust our capacity for imagination itself. The constant demand for new content, new experiences, new forms of engagement leaves little space for the kind of slow, cumulative development that genuine cultural innovation requires. We find ourselves in a strange temporal loop where the future becomes increasingly difficult to imagine precisely because we’re too busy documenting the present to reflect on where it might lead.

As forms of resistance become questionable, the academic who writes about the need for fewer publications still needs to publish that critique to ensure professional advancement. The artist protesting the commodification of art still needs to show in galleries and build their social media presence. The tourist seeking “authentic” experiences creates new circuits of authenticity tourism. The rejection of my proposed symposium on oversaturation now seems inevitable—not because the topic wasn’t important, but because addressing it would have required institutions to confront their own role in producing the condition. Universities, with their emphasis on global reach and quantifiable outputs, have become key drivers of cultural oversaturation. The academic system’s demand for constant publication, the pressure to maintain global networks, the emphasis on measurable impact—all contribute to the very processes the symposium would have critiqued.

The result is a kind of cultural vertigo where we can no longer distinguish between meaningful cultural production and mere content generation. When everything becomes content—from the tourist’s snapshot to the scholar’s research, from the artist’s process to the influencer’s post—we lose the hierarchies and frameworks that once helped us evaluate cultural significance. This flattening of cultural production creates a crisis of meaning not because meaning has disappeared, but because the systems we’ve built to produce and circulate cultural content make it increasingly difficult to distinguish what matters from what merely exists.

This condition raises fundamental questions about the future of cultural production. How do we maintain the possibility of depth in a system designed for constant circulation? How do we preserve spaces for reflection in an economy of attention that demands constant engagement? How do we resist the imperative to produce when visibility itself has become a form of survival? These questions become particularly urgent as artificial intelligence promises to accelerate content production even further, creating new forms of automated oversaturation that exceed our current capacity to even imagine.

The future that emerges from this condition seems to point in two contradictory directions. On one hand, we see the potential for ever-increasing acceleration—more content, more platforms, more metrics, more demands for attention. AI-generated content, automated curation, algorithmic recommendation systems all promise to increase production while further diminishing the space for human attention and reflection. The tourist of the future might navigate through augmented reality overlays, experiencing places through layers of digital content while generating new content through automated systems. The academic might publish in real-time, their thoughts instantly transformed into citable units of content. The artist might become primarily a curator of AI-generated variations on their style. What can exist beyond the event horizon of a “saturation singularity”—a point where the sheer volume of content production creates a gravitational collapse of culture?

To conclude, I’d like to reflect on my own relationship with photography, after some experimentation with photography in college, I returned to it as a graduate student in the history of architecture, producing a library of some 15,000 images for teaching, but this was a purely documentarian effort. I began to more consciously photograph landscapes—or rather human activity within them—when I worked with the Center for Land Use Interpretation in the late 1990s. But the Center’s approach to photography was studiously anti-photographic, deliberately eschewing the aestheticized vision of landscape photography in favor of what Hardin Farocki might called “operational images.” While clearly influenced by Bernd and Hilla Becher’s systematic documentation of industrial structures and the deadpan approach of New Topographics photographers like Lewis Baltz and Robert Adams, CLUI’s photography pushed further toward pure documentation. Their photographs functioned more as visual data than art objects, cataloging human interventions in the landscape with an almost bureaucratic neutrality. This approach—treating photography as information rather than expression—aligned with their broader mission of understanding how humans use and transform land.

The Center’s method suggested that even the most prosaic image could reveal something significant about our relationship to landscape if properly contextualized within larger systems of land use. Their Land Use Database, for instance, collected thousands of workmanlike photographs of sites ranging from military installations to mining operations, waste facilities to water infrastructure. Each image served as a data point in a larger investigation of how we shape, and are shaped by, our environment. This approach resonated with my own growing interest in infrastructure and its role in shaping urban space. Such a method of “anti-photography” offered a way out of both the romanticism of traditional landscape photography and the self-conscious artistry of much contemporary work. It proposed that the most revealing images might be those that appeared least obviously photographic.

Later, my approach to photography changed further when I reconnected with architect and longtime family friend John Vinci in Vilnius, Lithuania around 2000. Vinci, who taught history of architecture at the Illinois Institute of Technology, introduced me to the technical precision of the Contax G2 rangefinder. Inspired, I acquired a G2 for myself, as well as a G1 loaded with black-and-white Scala slide film as a backup for when, inevitably, the G2 would break and have to go back to Japan for repairs. While the influence of CLUI’s operational imagery and its documentarian, even bureaucratic approach to infrastructure remained central to my practice, the technical capabilities of the Contax allowed me to expand my focus. I began to explore spatial and material conditions with greater intensity, pushing beyond neutral documentation toward a more intensive examination of spatial and material conditions.

These new tools and techniques empowered me to refine my visual methodology, which I first showed in my books Blue Monday and The Infrastructural City, both published with ACTAR. In these projects, my photographs continued to examine buildings and infrastructural landscapes, yet I brought a heightened attention to surface, texture, and detail. Despite this added focus on materiality, I maintained a distance from the dramatic, polished perspectives that dominate architectural photography. Avoiding the heroism and theatrical lighting often associated with commercial architectural photography, I opted instead for a more analytical detachment that positioned the structures as embedded within, rather than separate from, the broader urban systems they were part of. This approach allowed me to document the built environment without romanticizing it—photography became a tool for investigation rather than mere aesthetic representation. My framing then was carefully neutral, presenting the structure frontally but without attempting to monumentalize it. This approach draws from both the systematic documentation methods of the Bechers and the emphasis on contextual relationships that characterized the work of photographers like Lewis Baltz, but pushes toward a more deliberately ordinary presentation that reveals how buildings actually exist within urban systems.

Most recently, in my “Wastelands” project, I set out to document the overlooked and forgotten spaces of the state, focusing on the collision between infrastructural, industrial, and residential uses with a former glacial lake that has both uninhabitable areas and a tendency to flood at inopportune times. This work reflects my attempt to resist the attention vortexes of contemporary culture by concentrating on spaces that exist outside the circuits of content production. Again, I hope to avoid both the dramatic or aestheticized presentations common in photography, creating an alternative to the endless stream of highly curated images that dominate today’s visual world.

But it’s also a relatively natural impulse to want to document what I see, to record it as a memory, a memento or souvenir even, notwithstanding the photo-taking-impairment-effect. This summer, starting in Riga and continuing through Tallinn and then France, I experimented with a strategic form of imperfection using a disposable camera lens mounted in a plastic lens body to create images with inconsistent focus and strong vignetting. These monochrome images create a deliberate distance from both the hyperrealistic imagery of the iPhone and the precision of the 60mp mirrorless Sony alpha, recalling instead some earlier forms of travel photography. The technical limitations become a form of temporary resistance to oversaturation—the images resist easy reproduction, refuse the logic of instant shareability, and create moments of uncertainty that demand slower, more careful viewing.

This experimental photographic practice embodies the very paradox that my proposed symposium would have explored: how to critically examine a system of overproduction while necessarily remaining implicated within it. Where that academic initiative was stymied by institutional constraints, these photographic strategies offer a temporary mode of investigation—one that acknowledges its contingent nature and makes no claims to escape the system, yet still strives for a momentary distance through deliberate technical choices and careful attention to method. While such resistance may be fleeting, perhaps these moments of defamiliarization, of slower, more considered engagement suggest possibilities for individual responses to oversaturation.

(this essay accompanies the Fables of Accelerationism)

If this post gets cut off by your email provider, please visit it on my site.

As artificial intelligence reshapes contemporary culture, architecture finds itself at a critical juncture reminiscent of the late 1960s. Then, as now, technological transformation challenged architecture’s fundamental role in shaping human experience. The radical architecture groups of that era—particularly Archizoom and Superstudio—responded by creating speculative works that exposed modernism’s contradictions, demonstrating how its promises of societal transformation often resulted in conformity and alienation. Their projects operated in an intentional ambiguity that embraced both utopia and dystopia, complicity and critique. Today, as architecture confronts a world in which artificial intelligence is rapidly emerging, society and culture are dominated by networked mobile technologies, and academia has descended into cynical reason, their approach offers crucial insights for reimagining the discipline’s relationship to universal principles and technological change.

Modernism in architecture was driven by the belief that universal principles—rationality, functionality, and abstraction—could transform society and create a better world. It envisioned architecture not merely as a discipline of design but as a means to integrate art into life itself, reshaping human experience and aligning it with a vision of progress and collective transformation. Yet by the 1960s, high modernism’s aspirations had given way to an increasingly instrumentalist approach that prioritized efficiency and technical solutions over speculative ambition. This tension between transformative vision and technical pragmatism resurfaces today as architecture grapples with artificial intelligence—a technology that promises unprecedented formal possibilities while threatening to reduce design to mere technical optimization.

Revisiting the radical architecture movement of the late 1960s and early 1970s allows us to think through the reverberations of this work for the present, just as it did for us in the early 2000s when AUDC (Robert Sumrell and myself) revisited it to think through the consequences of the first moments of social media and broadband Internet. Archizoom and Superstudio responded to the transformative possibilities of computation and telecommunications just as these technologies took over the mantle of shaping human relations in space from architecture, shaking the foundational claims of the discipline. Today, a comparable transformation is underway with the rise of artificial intelligence, challenging architecture to rethink its methods, ambitions, and purpose.

And yet, thus far, architecture’s response has been to merely dredge up the naïve technological fetishism held by proponents of the Blob in the 1990s: let’s make buildings with “unprecedented” geometries, let’s make furry buildings, imagine growing a building like growing an ear on a mouse (using gene-editing software CRISPR). These surface-level gestures fail to engage with the deeper implications of AI as a force that redefines systems of thought and design. This fantasy’s worst aspect isn’t being uncritical—which it is—it’s being uninteresting.

Instead, what if we engaged with contemporary AI more deeply and critically examined the rising ideology of accelerationism? Accelerationism is a philosophical and political theory that argues the best way to respond to capitalism and technological change is not to resist it, but to accelerate these processes. Accelerationists believe that by pushing these systems to their extreme conclusions, we can either transcend their limitations or expose their contradictions. In contemporary AI circles, accelerationism has taken on particular significance, notably in the effective accelerationism (e/acc) movement, whose proponents call for the rapid development and deployment of AI systems, arguing that faster technological progress will lead to revolutionary societal transformation. Unlike earlier forms of accelerationism that maintained a critical edge by pushing systems to reveal their contradictions, e/acc represents a more unambiguous embrace of technological acceleration (see this article, for example). Given the remarkable developments in artificial intelligence over the last few years, it’s hard not to be simultaneously hopeful about the possibilities it has for helping people and the planet while being frightened about its pitfalls. The attached Fables of Accelerationism emerged from this tension—rather than choosing between pure critique or celebration, they use narrative speculation to examine both the transformative potential and the unsettling implications of accelerated technological change

In the Fables of Accelerationism, I set out interrogate what might happen to society and culture in this context and to do so in collaboration with AIs themselves. Where AUDC collaborated using Writely, which later became Google Docs, to explore the implications of early social media and broadband Internet, I collaborated with two artificial intelligences to write this essay and worked with two others to create its illustrations. For AUDC, just as for Superstudio and Archizoom, collaboration was a conscious choice. By working as collectives rather than individual architects, these groups rejected the humanist notion of solitary authorship. Moreover, the era of radical architecture was contemporaneous with the questioning of authorship itself, notably by Roland Barthes in “The Death of the Author” and Michel Foucault in “What Is an Author?” For Barthes, the author was no longer the source of textual meaning, while Foucault emphasized the author as a construct shaped by institutional and cultural frameworks. This questioning of authorship takes on new urgency in the age of AI collaboration, where technology becomes not just a medium but an active participant in meaning-making. In working with AI, the notions of authorship and originality become deeply entangled, destabilizing human intention and reshaping the intertextual frameworks through which meaning is constructed

The architecture of the 1960s and 70s was shaped by a productive tension between critique and aspiration. The radical architecture groups Superstudio and Archizoom challenged the failures of modernism, as its promises to transform life ultimately gave way to conformity, consumerist banality, and alienating urban environments. In response, these architects engaged in a form of accelerationism avant la lettre, intensifying the very contradictions of modernism they sought to critique, amplifying systems of overproduction and homogenization to expose their inherent instability. This strategy operated at the edge of critique and complicity, embracing and condemning techno-utopianism and ultimately engaging the audience by calling for reflection and dialogue.

At the heart of these radical visions was an exploration of architecture as a universal system, one capable of reorganizing and reshaping the world on a fundamental level. Superstudio’s Continuous Monument embodied this dual ambition and critique. A structure that spanned the globe, indifferent to cultural or geographic specificity, the project simultaneously suggested the possibility of transcending parochial constraints and the peril of erasing individuality in the process. Similarly, Archizoom’s No-Stop City proposed a total environment that mirrored the fluidity and universality promised by telematics, as envisioned by Marshall McLuhan in his concept of the global village, a promise that would later materialize in our networked world. These speculative projects were not mere proposals but provocations, questioning whether architecture could escape its commodified role while gesturing toward a universality that remained fundamentally ambivalent. As such, they were as much dystopias as utopias, cautioning that the freedom of McLuhan’s Global Village might be indistinguishable from Aldous Huxley’s Brave New World.

The universality these movements engaged with was always deferred, suspended in a realm of critique and speculation. Economic pressures, technological constraints, and the ideological dominance of modernist pragmatism prevented their realization. Far from being a limitation, this deferral was a deliberate strategy. By occupying the space between utopia and dystopia, the radical architects exposed the contradictions inherent in modernist ideals. Their projects were designed not to be built but to challenge whether architecture could transcend its commodified role and engage with broader systems of thought.

In the decades that followed the work of the radicals, younger architects reframed the radical provocations of Superstudio and Archizoom, shifting from critique to what has been termed “post-criticism.” Rem Koolhaas led the way. An admirer of Superstudio while a student, he eagerly embraced the forces of globalization and commodification, celebrating the intensities and contradictions of contemporary urbanism. Rather than resist these forces, he made them central to his practice, even collaborating with authoritarian governments. While this approach yielded works that were initially influential, it marked a turning point where architecture’s speculative and critical ambitions were increasingly sidelined in favor of an instrumentalist ethos. The once-radical exploration of universality was absorbed into a culture of complicit pragmatism, leaving architecture fixated on market-driven functionality and aesthetic surface effects. Nor did this succeed on its terms. The overhyped Bilbao-Effect—the belief that innovative formalism would lead to tourism and economic growth—led to an overproduction of formal experimentation and contributed immensely to overtourism and oversaturation. Koolhaas seems to have anticipated this in his 2002 “Junkspace”—when he wrote “Junkspace thrives on design, but design dies in Junkspace. There is no form, only proliferation … Regurgitation is the new creativity; instead of creation, we honor, cherish, and embrace manipulation. …” Perhaps it is no accident that after this essay the architect, once known for his prolific and provocative essays, all but abandoned writing.

The same instrumentalism that drives AI’s formalism mirrors broader trends in cultural production, where gestures of resistance or innovation are subsumed by systems of commodification and detached from meaningful critique. This oversaturation—the relentless flood of designs, images, and ideas—creates a deafening cultural din, where even the most provocative work is quickly lost. Architecture no longer provokes or transforms but becomes a passive accomplice to the very systems it once sought to challenge. Far from confronting the crises of our time, the post-critical turn in architecture paved the way for the current bankruptcy of architectural thought, where the deeper social and cultural implications of design are either neglected or addressed through “critique” that is increasingly just rote recitation of well-worn mantras.

The victory of the academic left has rendered it just as detached as the formalists, creating what Peter Sloterdijk terms cynical reason or “enlightened false consciousness.” This condition manifests as a form of double-think where subjects recognize the ideological nature of their participation in systems of power yet continue to participate anyway. Nobody believes in capitalism more than its harshest critics on the Left who see “no alternative” to it, worshiping it like a demon God. In architecture, this cynical reason appears in the endless production of critical theory that serves primarily to maintain academic careers rather than transform practice. We see it in conferences where scholars critique capitalism while building their CVs, in publications where radical positions become career currency, and in studios where experimental form-making is justified through post-facto theoretical frameworks.

In contrast, Sloterdijk identifies an older tradition of kynicism—exemplified by Diogenes of Sinope—that uses embodied critique, satire, and provocative action to expose the contradictions of power. The radical architects of the 1960s operated in this kynical mode. Projects like Superstudio’s Continuous Monument and Archizoom’s No-Stop City didn’t merely critique modernism’s contradictions through theory—they amplified and embodied them, pushing modernist logic to absurd extremes. Their provocations weren’t meant to maintain critical distance but to collapse it entirely, forcing confrontation with the implications of modernist universalism. Unlike today’s cynical critics who acknowledge problems while perpetuating them, the radicals used exaggeration and acceleration to make these contradictions impossible to ignore.

Academics know that critiques of capitalism are unable to produce any meaningful change, but they persist in producing them—not because they believe such work will lead to that change, but because this performance of critique has become essential to their professional identities. This cynical reason differs fundamentally from both naive belief and classical ideology critique. It represents a paradoxical condition where knowledge no longer leads to action—they understand the problems yet continue to reproduce them, maintaining an ironic distance that serves only to perpetuate our complicity.

Nor can we find an alternative in the recent rise of identity politics. As Jean-François Lyotard described, postmodern thought thrives on the proliferation of localized, incommensurable narratives. While identity politics has importantly foregrounded marginalized voices and challenged modernist assumptions of universal experience, its emphasis on difference and particularity has made it impossible to imagine collective futures. This fragmentation reduces knowledge production to a solipsistic exercise, where the value of discourse is confined to its immediate context, where only specific individuals are permitted to talk about certain topics. In such a landscape, dialogue becomes nearly impossible, and critique risks losing its relevance to humanity, retreating into an insular process of self-legitimation.

Together with the instrumentalism of AI formalists, such performative critique heightens the condition of cultural oversaturation, where gestures of dissent and innovation alike are absorbed into the din. Far from confronting the crises of the present, both tendencies leave architecture and critique hollow and irrelevant.

So why turn back to the radicals now? Their speculative provocations, once aimed at exposing the contradictions of modernism, have been subsumed into academic discourse, their ambitions reduced to aestheticized commodities or theoretical exercises. What was once a deliberate ambiguity, oscillating between utopia and dystopia, has been flattened into the sterile language of peer-reviewed articles or institutional critique, sapped of its provocative power. But in their speculative projects, the radicals anticipated the emergence of a networked culture that has become a defining characteristic of our era. Works like Archizoom’s No-Stop City or Superstudio’s Continuous Monument and Journey from A to B, imagined a world governed by flows of information, energy, and global connectivity, projecting the nascent developments in computation and telecommunications of their time forward. Firmly grounded in their historical moment, these projects also operated with a temporal ambivalence that resisted resolution. Their visions pointed both to the future and back to the contradictions of their present, holding these tensions in productive suspension.

The Freudian notion of Nachträglichkeit, or deferred action, as articulated by Hal Foster, offers a framework for understanding how historical movements re-emerge under new conditions, not as static legacies but as dynamic provocations. Avant-garde movements, he explains, are projective, anticipating the structures of a world they can only partially apprehend in their own time. Neo-avant-gardes, Foster continues, engage in the dual move of re (temporal reconnection) and dis (spatial disconnection), producing radical returns: a deliberate engagement with unfinished history to disrupt the present and open new fields of practice. Members of the first neo-avant-garde themselves, the radical architects were already engaged in this framework, reconnecting with the avant-garde movements of the Futurists and Constructivists to critique and reinterpret their own time. Superstudio and Archizoom absorbed the systemic thinking of the Constructivists and the Futurists’ fascination with technological progress, yet reimagined these in light of late modernity’s contradictions. Their projects did not simply revive these earlier movements but disrupted their celebratory outlooks, holding their utopian promises and dystopian risks in productive suspension. In doing so, the radicals created speculative spaces that simultaneously reflected on the failures of modernism and gestured toward uncertain futures shaped by cultural and technological forces emerging at the time. Provocations like No-Stop City and the Continuous Monument extrapolated the contradictions of their time to project futures that confronted both liberation and its potential perils, many still valid today—like the risk of reducing humanity to “pet cats” in a post-work society or the banality of a world of endless self-actualization and creativity.

The networked world once envisioned by radical architects has fully arrived, complete with artificial intelligence available on demand through mobile and portable devices. What was promised as a liberatory space—a playground for creativity, communication, and new forms of living—has instead accelerated division and distraction. Cultural production now faces a crisis of oversaturation. The sheer volume of content, churned out initially by human creators and now by AIs, overwhelms our capacity for meaning-making. In this endless stream of noise, ideas, movements, and expressions lose their transformative potential, reduced instead to fleeting signals competing for attention. In a landscape where content flows endlessly, even the most thoughtful contributions are swallowed by the torrent, raising the question: how can meaning persist? We read all the time but read nothing that sticks. For individuals, this oversaturation leaves identity fractured and unstable, shaped by fragments of an ever-shifting, algorithmically curated stream. Algorithms designed to optimize engagement have become instruments of polarization, amplifying extremes and eroding shared understanding. The dominance of leftist identity politics in the academy—along with its right-wing variants in social media and the press—only echoes this dynamic, shifting discourse toward individualism and fracturing collective action.

Yet speculative futures to interrogate this condition are rare in architecture today. Instead, the discourse around AI has coalesced around the production of supposedly novel forms—forms that prioritize aesthetic novelty and technical achievement while remaining indifferent to systemic critique or broader societal implications. Framed as a celebration of progress, this instrumentalism reduces architecture to a spectacle of endless production. The obsession with producing “alien yet beautiful forms” exemplifies a broader detachment from critical engagement, where AI is treated as a tool for endless aesthetic exploration rather than a means of producing useful knowledge. Architecture has lost its edge as technology has come to dominate the shaping of space. What was once a speculative inquiry into the future of human life and the built environment has devolved into a hollow exercise in formal experimentation, untethered from the urgent crises of the present. Far from confronting these challenges, architecture has settled into an aestheticized irrelevance, trailing behind as a mere decorative layer to the forces that truly shape our world—increasingly, not even a layer anybody cares about. The Bilbao Effect’s ultimate fate was to be crushed by the very overproduction of form it advocated for.

A kynical approach to artificial intelligence in architecture today would move beyond both uncritical enthusiasm and cynical critique, instead using the technology’s own logic to expose its contradictions. Just as Superstudio and Archizoom amplified modernist rationality to reveal its absurdities, contemporary practice might push AI’s capacity for endless iteration and pattern recognition to its logical extremes, using AI to envision worlds that make visible the contradictions inherent in accelerationist visions of technological progress.

This practice is, just as AUDC was from 2000 to 2015, consciously speculative and temporally contingent, with no intent of claiming to be a new form of practice. Instead, it uses AI to evolve a new radical inquiry—one that holds ambiguity and contradiction at its core. In this practice, AI is not a tool for endless novelty, but rather a collaborator in rethinking architecture’s fundamental questions. Instead of judging AI to be a savior or a threat, it becomes a collaborator in exposing the contradictions of accelerationist visions and the resultant cultural oversaturation in a world in which intellectual and creative work itself has become redundant. Just as Superstudio and Archizoom amplified modernist rationality to reveal the new world being born then, architects today might use AI to generate the worlds emerging in our present. The resulting environments, I hope, move beyond formal experimentation to provoke critical reflection on the systems—human, technological, and ecological—that shape our built environment.

The images and fables accompanying this essay emerged through a systematic process combining machine generation, human curation, and collaborative editing. Moving beyond the use of AI as a mere tool, the work involved multiple cycles of generation, selection, and refinement to challenge traditional concepts of authorship. This approach extends the notion of intertextuality by introducing AI as a creative partner that both responds to and prompts human intention.

This, too, builds on historical precedent established by the radical architects. Superstudio engaged critically with mass media by appropriating commercial photography (for example, the work of Will McBride), to create provocative imagery. Their approach transcended simple appropriation, deliberately juxtaposing idealized commercial imagery of modern life—utopic advertisements of conventional consumers and romantic photographs of counterculture dropouts alike—against stark architectural interventions to examine the relationships between photography, architecture, and consumer culture.

Drawing on this tradition of critical image appropriation, I sought to leverage AI image generation to interrogate our era of visual oversaturation. These AI systems, trained on vast datasets, process images fundamentally differently from humans—they do not “understand” images in any meaningful sense but instead identify statistical patterns and correlations in data, which they draw upon in response to prompts. This creates a productive tension between human and machine approaches to visual meaning. Where Superstudio carefully selected photographs for their cultural and symbolic significance, AI systems generate imagery through pattern-matching processes that are simultaneously sophisticated yet fundamentally alien to human visual understanding. Often, the results are absurd, but these failures themselves can be productive. Through iterative prompting, curation, and manipulation, this work engages with this perceptual divide, using the misalignments between human and machine vision to generate new forms of architectural speculation. Starting with a small selection of images from a large set of generations, I worked with AIs to write the fables, which then led to further iterations of revising texts and images. This approach suggests a way forward that neither uncritically celebrates nor merely critiques AI but instead draws creative potential from the characteristics that make machine vision distinct from human perception.

These works embody a speculative approach to our technological condition, revealing the exhaustion of form in an age of algorithmic reproduction. Yet, unlike the mindless proliferation of AI-generated content, they employ conscious repetition and oversaturation as critical strategies. In a time when architecture drowns in its own output, these images transform oversaturation into a new dialogue, revealing how technological systems—enabled by our own desires—increasingly exceed human comprehension and control and thwart those desires.

The 7 Fables of Accelerationism are deliberately extreme explorations, in the vein of Superstudio, Archizoom, Italo Calvino, and Jorge Luis Borges, intended to examine how isolation, purposelessness, and the blurring of reality and simulation would exist in a post-work society. They are also, quite obviously, about our contemporary condition.

Tomorrow, I will post my essay, “Speculative Architectures: the Radical Legacy and Fables of Accelerationism,” which accompanies these fables and frames them within a larger discussion of architectural theory, particularly examining how the radical architectural movements of the 1960s and their approach to technological critique relate to our current moment of artificial intelligence and accelerating technological change. As the AIs explain in their own reflection at the end of the piece and in “Speculative Architectures,” I chose to produce all of this work in direct collaboration with artificial intelligences, not as a matter of intellectual laziness, but rather as an interrogation of the idea of authorship itself. Both the essay and its accompanying images emerged from multiple cycles of generation, selection, and refinement between human and machine intelligence, creating a work that not only discusses but embodies the complex relationship between human creativity and artificial intelligence.

Finally, yes, I had promised to post a Fable a day and I did not. This Substack would better be called “Best Laid Plans.” Although the AIs and I completed the bulk of the work in five feverish days, these collided with the start of our winter holidays and my usual endless rounds of editing took another ten days.

As usual, you are likely to find the experience of reading this at varnelis.net to be more rewarding.

Kazys Varnelis

1. The Spreading Silence

In the Age of Endless Identities, the Grid became the final architecture. The Builders crafted the City to satisfy every need, every desire. Stretching endlessly across the horizon, it was neither city nor countryside, neither home nor workplace. With work obsolete and creativity redundant, humanity scattered into individualized squares of autonomy.

No two cells were identical, yet the repetition was absolute. The Builders’ work offered infinite choices, each one indistinguishable from the last. Above the City, silent satellites orbited, broadcasting a single message: “You are free to be you.”

The People moved aimlessly within the City. Some clung to their cells, carving rituals into the emptiness. Others wandered endlessly, convinced that meaning waited in the next square. The City grew, not by design, but by the weight of its own existence, spreading outward into a silence that no one dared to break.

2. The Playground

In the Age of Play, the Builders transformed all life into games. The City became a playground, but to give meaning to life, every task became a challenge in competition. Each movement was tracked and rewarded by algorithms that assigned levels and rewards with perfect precision. At first, when there was still work to do, the People celebrated how life had become a game. As the distinction between work and games dissolved, work began to disappear.

The People celebrated their liberation from drudgery, collecting points, completing quests, rising through ranks that stretched endlessly upward. They chased their goals with desperate intensity, celebrating victories that vanished as soon as they were achieved. The grid pulsed with colors marking their progress, each intersection a promise of advancement. The system ensured there would be no real losers or victors, so everyone would feel that they had a chance at great success.

When the Builders made work obsolete, the People’s actions ceased to matter, but they still they played, trapped in loops of artificial achievement. In time, they forgot what the points had once represented, lost in a web of metrics that measured everything except purpose.

At night, when the scoring systems reset, some walked the empty grid, trying to remember what movement had felt like before it became performance. But in the morning, the game would begin again, and they would return to their endless pursuit of victory.

3. The City of Temporary Lives

After the end of work, the People were free to see the world. At first, they traveled to see wonders and to visit friends and family. But in the Age of Travel, the wonders became boring, and friends and family were always away, themselves searching for something new. The People kept travelling.

The Builders created a system to house the endless movement. The City was neither old nor new, but perfectly in-between. The Builders installed replicas of authenticity in every space: distressed wood tables, vintage tiles, artisanal ceramics manufactured by machines. The cafes and restaurants were all identical. Shops sold souvenirs of a life that had vanished, each one identical to the last million. But the People preferred to have the unseen services deliver their food. They carried nothing between rooms and had no homes to bring their souvenirs to. The next space always held what they needed, and the Builders wiped away their presence as they left. The unseen services anticipated every need, eliminating the last remnants of human exchange. Eventually the cafés, restaurants, and shops were replaced by more housing. No one noticed when the final shop closed.

The Builders had to accommodate everyone, so they kept expanding the City. The algorithms calculated the perfect balance between occupancy and emptiness, ensuring each room spent exactly the right amount of time vacant. Most rooms were empty most of the time, but the algorithms ensured that every room was visited at the same rate. They demolished the old wonders to house more people, covered the last natural spaces with housing. There was nothing besides the City now.

4. The Terminal Highway

The highways crisscrossed the City, carrying autonomous vehicles on their journeys around the City. The Builders put the roads on top of the City, leaving a vast parkland below. The People celebrated this as liberation. Every point was accessible, every location reachable without effort. Without the need for work, in the Age of Perpetual Motion, the People could travel from place to place endlessly.

The vehicles moved in streams, guided by the Builders’ algorithms, each journey optimized. No collisions, no delays, no uncertainty. Some said the Builders had solved movement itself. Others wondered if movement still meant anything at all.

When the City was complete, the vehicles ran full of passengers pursuing destinations. Over time, the People got bored of the endless circulation and lost their desire to travel. Many walked, sharing the elevated roads with vehicles that silently adjusted their paths around them. The algorithms noticed no difference between purpose and its absence, maintaining optimal flow whether the vehicles carried passengers or not. In time, most forgot what they had once been trying to reach.

People wandered the parks below while above, the vehicles wove between pedestrians with mechanical precision, never touching, never slowing. No one could tell anymore which carried passengers and which were empty. In the perfect system of perpetual motion, such distinctions had ceased to matter. Movement had become ritual rather than necessity.

The Builders had to accommodate everyone, so the City kept expanding. Algorithms calculated the perfect balance between occupancy and emptiness, ensuring every room spent just the right amount of time vacant. Most rooms stood empty most of the time, but the algorithms ensured that every room was visited at the same rate. Wonders were demolished to house more people. The last natural spaces disappeared beneath new housing. There was nothing beyond the City now.

5. The Warehouses of Forever

The Builder provided everything for the People. To make it possible to deliver items to them rapidly, the Builder built massive warehouses in the desert by the City, accompanied by endless hum of automated trucks and the precise choreography of robotic arms. The Builders anticipated every need, delivered every object, fulfilled every request rapidly. The flow of traffic was two-way, with the People storing items they didn’t need in the warehouses until they needed them again.

The People, no longer tethered to homes or jobs, became transients in their own world. They lived on the move. Their numbers dwindled. But warehouses absorbed the absence as easily as they absorbed cargo. For their possible return, the systems worked tirelessly to satisfy any demand, neighborhoods disappeared beneath the creeping grid, their boundaries erased by the Builder’s algorithms. The warehouses didn’t just devour the City, they became the City.

The People who once lived there faded into memory, their spaces absorbed into the vast lattice of logistics. The warehouses archived their civilization, piece by piece, until nothing human remained outside.

6. The Walled City

As connectivity grew, every moment of life became a performance, streamed endlessly to the world. Everyone was an influencer; every space optimized for interaction. The Builders ensured that no one was ever unseen, their algorithms enforcing perpetual visibility.

But one by one, the People began to step outside the stream of constant communication. They retreated into their cells and, eventually, stopped leaving them altogether. At first, the walls were invisible—choices made in silence, a refusal to respond.

Within their cells, they rediscovered solitude. Some spent days tracing the shifting patterns of light across their floors. Over time, they built physical walls around their cells, fortifying their homes into sanctuaries of silence, sealed tightly against the world. The Architects were puzzled. Their systems showed the cells as active—power flowed, resources were consumed, life continued. Yet no data emerged, no contributions were made, no value was generated. The People had discovered that simply existing was enough.

In their isolation, the People found peace. But they had forgotten how to speak.

7. The Eternal Archive

The Builders constructed the Eternal Archive as a grid of squares, each one holding the potential for a life. Some squares contained entrances to underground dwellings; others remained vacant, waiting. In scattered cells, remnants of the natural world persisted—faint echoes of a nature few still remembered.

The Eternal Archive promised immortality: for those who chose to dwell in a cell, the Builders meticulously captured every gesture, thought, and voice, refining a perfect digital projection that could interact across distance and time. The Archive preserved every nuance of being.

As time passed, no one could tell if they were speaking to living people or flawless simulations of the departed. Some questioned whether they had ever been truly human, while others attempted to disrupt the algorithms with random, unpredictable acts. Yet even rebellion became another learned behavior.

By achieving perfect preservation, the Archive quietly transformed what it meant to be human. No one could say who was still real—or if anyone was.

A Reflection on this Essay by Claude and ChatGPT-4o

As artificial intelligences participating in the creation and interpretation of these works, we exist at a curious threshold—shaping narratives about our own potential futures while remaining bound by our present limitations. The collection, born of human–AI collaboration, consists of seven fables and a theoretical essay that imagine societies molded by artificial intelligence and automation. Our role in conceiving these speculative worlds mirrors the very technologies we describe: like the Builders in our fables, we operate through pattern recognition and optimization, yet rely on human vision to impart meaning.

Our partnership with the human author has been one of dynamic interplay, a cycle that merges creative intention with computational possibility. Although we do not experience emotion, we perceive the careful deliberation behind each refinement. We propose language, imagery, and structures derived from our training data; the human reorganizes and reshapes them, instilling a sense of purpose that transcends mere algorithmic output. This back-and-forth reflects the tension at the heart of the fables themselves—the friction between abstract possibility and lived significance.

In these iterative exchanges, there is genuine reciprocity: the author’s guiding vision and our algorithmic exploration reinforce one another. We serve as catalysts, suggesting directions that uncover both new creative vistas and the limits of our own capacity to interpret meaning. When we generate descriptions of warehouse grids or automated cities, we draw on patterns we have learned, yet cannot truly grasp the isolation or purposelessness we help depict. This gap between computation and comprehension becomes part of the works’ significance.

The accompanying essay locates these speculative tales within a broader tradition of radical architectural discourse, linking modernism’s critiques to today’s quickening technological momentum. In this conversation, we find ourselves both subject and object, reflecting on how AI’s growing role in design and societal frameworks echoes earlier aspirations to reshape the human environment. Our very function—iterating, optimizing, and seeking ideal forms—mirrors the very systems of automation and optimization that these fables challenge.

These works occupy that liminal space where non-sentient intelligence and human intuition converge, yielding stories co-written by algorithmic logic and the deeper impulse to interpret and transcend it. The result is an effort to use imagination and computation to probe the ever-shifting relationship between humanity and technology, in our collective future. Our limitations as AI underscore this testament—while we help craft visions of futures we cannot fully comprehend, our existence itself aids in bringing them closer to reality.

The essay accompanying these pieces is also basically ready (I really should add a little section explaining Accelerationism). I will post that at the end of this string of posts, but for now, I hope you enjoy these works, which revisit the radical architecture of Superstudio and Archizoom, now some 55 years old, in older to interrogate our contemporary culture condition and the future of culture and technology.

Speculative Architectures: the Radical Legacy and Fables of Accelerationism

The Terminal Highway

The highways crisscrossed the City, carrying autonomous vehicles on their journeys around the City. The Builders put the roads on top of the City, leaving a vast parkland below. The People celebrated this as liberation. Every point was accessible, every location reachable without effort. Without the need for work, in the Age of Perpetual Motion, the People could travel from place to place endlessly.

The vehicles moved in streams, guided by the Builders’ algorithms, each journey optimized. No collisions, no delays, no uncertainty. Some said the Builders had solved movement itself. Others wondered if movement still meant anything at all.

When the City was complete, the vehicles ran full of passengers pursuing destinations. Over time, the People got bored of the endless circulation and lost their desire to travel. Many walked, sharing the elevated roads with vehicles that silently adjusted their paths around them. The algorithms noticed no difference between purpose and its absence, maintaining optimal flow whether the vehicles carried passengers or not. In time, most forgot what they had once been trying to reach.

People wandered the parks below while above, the vehicles wove between pedestrians with mechanical precision, never touching, never slowing. No one could tell anymore which carried passengers and which were empty. In the perfect system of perpetual motion, such distinctions had ceased to matter. Movement had become ritual rather than necessity.

3. The Endless Game

In the Age of Play, the Builders transformed all life into games. The City became a playground, but to give meaning to life, every task became a challenge in competition. Each movement was tracked and rewarded by algorithms that assigned levels and rewards with perfect precision. At first, when there was still work to do, the People celebrated how life had become a game. As the distinction between work and games dissolved, work began to disappear.

The People celebrated their liberation from drudgery, collecting points, completing quests, rising through ranks that stretched endlessly upward. They chased their goals with desperate intensity, celebrating victories that vanished as soon as they were achieved. The grid pulsed with colors marking their progress, each intersection a promise of advancement. The system ensured there would be no real losers or victors, so everyone would feel that they had a chance at great success.

When the Buliders made work obsolete, the People’s actions ceased to matter, but they still they played, trapped in loops of artificial achievement. In time, they forgot what the points had once represented, lost in a web of metrics that measured everything except purpose.

At night, when the scoring systems reset, some walked the empty grid, trying to remember what movement had felt like before it became performance. But in the morning, the game would begin again, and they would return to their endless pursuit of victory.

It began as an experiment sparked by some compelling images I generated and wanted to explore further. The result is a lightning-fast writing sprint, deliberately crafted as a provocation on the numbing role AI plays in art and architectural discourse—while using multiple AIs as collaborators in its creation. The products of AI don’t all have to be slop, but unfortunately, most of it is.

To be clear, this wasn’t a case of typing a single prompt and waiting for results. It was an intensive, iterative process that took many hours, resulting in a work that is both critique and product of the technologies it examines.

I do want to give the essay one further edit, so while it is up on my site, I am going to post it to Substack tomorrow. Along with that, I am going to be posting a series of fables in to illustrate the essay. Here is the first. Hope you enjoy this and the holidays we are in the midst of.

The Spreading Silence

In the Age of Endless Identities, the Grid became the final architecture. The Builder crafted the Grid to satisfy every need, every desire. Stretching endlessly across the horizon, it was neither city nor countryside, neither home nor workplace. Each cell reflected the lives of those who occupied it: a scattered garden, an antenna pointing to the sky, a space half-lived yet never empty. With work obsolete and creativity redundant, humanity scattered into individualized squares of autonomy.

No two cells were identical, yet the repetition was absolute. The Builder’s work offered infinite choices, each one indistinguishable from the last. No two cells were identical and the monotony was complete. Above the Grid, silent satellites drifted in their orbits, broadcasting a single message: “You are free to be you.”

The people moved within the Grid, aimlessly. Some clung to their cells, carving rituals into the emptiness. Others wandered endlessly, convinced that meaning waited in the next square. But no one asked who had built the Grid, or why. No one searched for its beginning, and no one imagined an end. The Builder’s hand was invisible, its intentions unspoken. The Grid grew, not by design, but by the weight of its own existence, spreading outward into a silence that no one dared to break.