Exiled from Lithuania because of war, my father spent his life thinking about his country. He had a habit of making grandiose claims about it, one of which was calling Vilnius “the Venice of the North.” I never understood that; Vilnius is not a city of canals, and its architecture, while Italian in influence, is not remotely Venetian, owing far more to Rome and the Lombard-Ticinese lake country.¹ But the very incongruity of the phrase came back to me four years ago and led to my Lost Canals of Vilnius, a counterfactual history of a city that never existed, rebuilt around canals after the Great Fire of 1610, documented in a narrative punctuated by vedute and photographs.

Exile was not an abstraction in my family. It was the ordinary fact around which everything else was organized. My parents understood themselves as refugees, not immigrants. My father (Kazys Varnelis, like his father before him) was born in 1917 in Alsėdžiai, a small village in German-occupied Lithuania, still, for the moment, part of the world Arno Mayer called the “old regime.” Agriculture remained labor-intensive, industrialization touched only scattered enclaves, and the church was a major force, even as the ghosts of the old nobility were still present. The world my father was born into would not have seemed strange to someone from the sixteenth century.

His father, the first Kazys Varnelis in this lineage, was a farmer and religious artist. He attended the local school in Alsėdžiai, a small village in Samogitia, wrote well, and showed enough talent as a boy that the parish priest proposed supporting him for art study in Petersburg. But his mother was widowed and needed him on the farm, so he remained in the village and became a dievdirbys and maliorius: a carver of saints, a painter of chapels, a repairer of church interiors, and a maker of Užgavėnės masks. Shortly before the Great War, he tried emigrating to the United States, staying with relatives in Chicago and working as a railroad-station laborer while thinking he might bring the family over. He hated the metropolis, returned horrified, and never again left Alsėdžiai.²

Modernity—what Marshall Berman describes as an environment that promises transformation while threatening to destroy “everything we have, everything we know, everything we are”—arrived for my father’s family not as enlightenment or progress but as war and plague.³ The Spanish flu killed more than the Great War itself, circling the globe in four months. In Alsėdžiai, it arrived amid the collapse of the German occupation and the chaos of armies passing through a countryside where the infrastructure of daily life had already broken down. My father was the youngest of four (or five, the actual number is lost in fog) children, but when the disease swept through the household in the winter of 1918-1919, he alone survived. The old regime, whatever its oppressions, had at least been stable for centuries. It gave way to war, disease, and death—the ongoing disaster that would define the rest of my grandfather’s life. He would survive the Lithuanian Wars of Independence, the brief interwar republic, the first Soviet occupation, the Nazi years, and die in 1945, in the aftermath of yet another cataclysm, as the Red Army returned.

My father came of age in modernity. He became familiar with electricity, the automobile, and even the airplane, but also with the museum, the camera, and the archive. To pay for his high school education, he worked on a railroad construction crew, then took a wagonload of wooden dievukai from the area—small figures of Christ, Mary, and the saints carved for roadside shrines and crosses—to Kaunas, where historian Paulius Galaunė bought them for the Čiurlionis Museum and encouraged him to collect more folk art, buy a camera, and photograph it in situ. By the time he entered the Kaunas Art School in 1936—and certainly by the time he completed his studies in 1941 and briefly became director of the Museum of Ecclesiastical Art—such objects had become museum pieces. The world he was born into was now consigned to history.

In 1944, he received a draft notice from the occupying German army that would have sent him to certain death on the Eastern Front. He escaped to Vienna, where he had been admitted to the art academy to study painting. His would go back to see his parents once more, in 1945 for a last visit as the Soviets advanced. As his mother made a last meal for the family, they could hear the sounds of artillery in the distance. After dinner, he fled west again. His father died later that year, reputedly of asthma. He had been a lifelong smoker and he would have been 74. After years as a displaced person in Austria and Germany, he moved to Chicago where he met my mother, also a Lithuanian refugee, and opened a studio designing church interiors for Lithuanian parishes. But seeing commissions drying up amid Vatican II reforms, he turned to fine art as a career. His early work was surrealist abstraction—biomorphic sculptures and bas-reliefs influenced by Arp, Giacometti, and Moore, forms that suggested archetypes or hieroglyphics. My mother had studied psychology, and they shared an interest in Jung; the work tried to reach something beneath individual consciousness, images from the collective unconscious. Later, after a brief foray into optical art, he returned to simpler, stronger forms in the 1970s, now through hard-edged geometries and shaped canvases. He saw himself as part of the European artistic diaspora, the tradition of the Bauhäusler who had also washed up in Chicago: László Moholy-Nagy and Mies van der Rohe. But unlike his most accomplished Lithuanian-American contemporaries—George Mačiūnas, Jonas Mekas, Aleksandra Kašuba—he did not see himself as an avant-gardist or an outsider. They were younger, their educations interrupted by the war; they became who they were in America and claimed marginality as a position from which to work. My father had already been someone in Lithuania; he had already made the transition from the old world of peasantry to the new world of the city. A critic called him “fundamentally conservative,” a classicist—finding order in geometries even as the world around him was chaotic.⁴

If his art sought order, his Lithuania remained unstable, a country remembered, collected, reconstructed, and imagined. Vilnius was the most unstable object of all. During the interwar period, Poland held the city after Żeligowski’s staged mutiny, and he visited it only once, briefly, on an expedition after the Soviets returned the city to Lithuania in 1939, before swallowing the entire country. In the US, every day was filled with thoughts of Lithuania, the lost country. He returned in 1998, at the age of eighty, to start his museum, which he intended as a gift back to his homeland, but he found Lithuania different from what he imagined, a culture foreign to him. My father left as a refugee, but he was an exile, unable to return to a place that no longer existed and that he had barely known.

Born in the United States, raised with Lithuanian as my first language, and given a name that marks me not just as ethnic but as foreign, I am also an exile. Lithuania meant little here, when it meant anything at all. Without the widespread cultural infrastructure that the Irish, Poles, Italians, or Germans had, Lithuanians assimilated quickly, leaving behind names that sounded strange without becoming legible. Mine was one of them. When people had heard of Lithuania, it was usually as the place where the German genocide of the Jews had local accomplices, a fact made to stand for the whole country in a way it does not for Germany or Poland. In America, being Lithuanian was to have a negative identity, or none at all.

So, too, I am also an exile in Lithuania, identifiable by my American accent and unreliable grammar. I learned Lithuanian from my parents, which meant I learned the frozen language the diaspora preserved after independent Lithuania disappeared behind the Iron Curtain. My speech is littered with archaic words like “aerodromas,” analogous to the English “aerodrome,” as well as with Polonisms rather than the Russianisms that survived the occupation. I can be understood, but I do not pass as a native: my speech is wrong enough to be corrected, and haughty hotel clerks love to speak to me in English, pretending they can’t understand me, which is rather strange, since government ministers can.

In Lithuania, the diaspora is regarded as an anachronism: sentimental, stubborn, carrying a Lithuania fixed before occupation and return. This is understandable. In the diaspora, prewar Lithuania hardened into kitsch: folk costumes in parish halls, amber, stylized crosses, straw Christmas tree ornaments, heavy food, and even the church interiors my father designed before abandoning that world. Once he left the church interiors, my father also left the diasporic world behind, but he could not help but be associated with it.

Diaspora artists often meet the same impatience unless, like George Maciunas or Jonas Mekas, they were already well known outside the country. My father’s museum entered under that cloud. In presenting the project to the government, he focused on his collections of antique furniture, much of it nineteenth-century reproductions, and lituanistica, because promoting his own work would immediately provoke jealousy among local artists. When the museum opened in 2001, the small rooms and risks to the collection required visits by reservation, which meant contacting staff who often did not respond to emails or phone calls. For the first 15 years, fewer than a hundred people came annually. With the art downplayed, the museum looked less like a living artistic project than the private fantasy of a returned émigré like Mykolas Žilinskas’s strange, postmodernist gallery in Kaunas. Only with my 2017 exhibition at the National Art Gallery did that begin to change. My return helped make his return visible.

In the years since my parents’ passing, I have continued to visit Vilnius frequently—in part because of the needs of my family and the museum, but also because the city has become part of my life. I have friends there, as well as projects that have tied me to the city. Yet repeated return has not become belonging. In Lithuania as in America, I remain an outsider, attached to Vilnius but unable to claim it. I have often thought about moving to Vilnius, but that would hardly change my position. I will always be both American and Lithuanian, and neither. I, too, am an exile, haunted not by the loss of a place, but by never having had a sense of home to begin with, in either country.

Vilnius is a haunted city. It has been Vilna, Wilno, Vilnius; it has belonged to Lithuanians, Poles, Jews, Russians. Vilnius demands a literature of exile and return: Czesław Miłosz, Tomas Venclova, and Jonas Mekas each approached the city through memory, wound, and invention, through languages and histories that no longer coincide. The Jews who made up a substantial part of its population before the war fled or were murdered at Paneriai. The Poles were expelled or assimilated. What remains are traces, palimpsests, names of the dead. Jacques Derrida called this condition hauntology: the presence of what is absent, specters that are neither alive nor fully dead. The ghost is not simply what is gone; rather, it continues to make claims on the present.⁵

In July 2022, I found myself in Vilnius, having dinner with my friends, architect Gintaras Kuginys and filmmaker Giedrė Genkevičiūtė, in her flat, designed by the late Valdas Ozarinskas. Gintaras had just opened The Golden Ratio at NDG, a gilded concrete paving block set into the pavement, a nearly silent intervention that turned the gallery’s forecourt into a problem of value and attention. In the first months after Russia’s invasion of Ukraine, the gold also carried another question: what still had value, and what could be stepped on without consequence? The invasion of Ukraine, of course, prompted thoughts about a Russian invasion of Lithuania. Valdas’s ghost also haunted us. His post-Soviet Vilnius was a city of loss as well, but for different reasons: the promised utopia never arrived, either under Communism or under capitalism. Valdas’s Vilnius was less a stable city than a field of objects that had survived some unnamed catastrophe, like the Zone in Tarkovsky’s Stalker, as I argued in my essay on his work. Ordinary life felt provisional there, as if the flat were only a temporary shelter for artifacts recovered from elsewhere.

The AI image generator DALL-E 2 had just been released to beta testers, and we spent the evening prompting it. This was before AI panic hardened into cultural reflex, when the technology still felt less like a threat than an opening. The possibilities were dazzling, even if the results were often comically flawed. Still, where earlier image generators produced distorted, half-legible scenes, DALL-E 2 could conjure images that cohered. We imagined the golden brick being stolen by thugs; envisioned a future Vilnius, perhaps one in which Valdas’s visions had taken over the city; put land art in 1970s Vilnius; and placed Ralph Eugene Meatyard’s masked families in the Lithuanian countryside, a reference to the Užgavėnės masks my grandfather had made, another form of image-making between folk ritual, fantasy, and fear.

Later that week, I remembered my father’s phrase “Venice of the North” and thought of Canaletto’s famed capricci, fantastic images of Venice in which buildings by Andrea Palladio from Vicenza, a town considerably inland, mysteriously appeared in Venice or, in Canaletto’s London years, along the Thames. What if, I thought, we saw Vilnius as Canaletto might have seen it? Soon I was trying to generate images of a Vilnius with canals, as imagined by Canaletto. It did not go well. Until recently, AI image generators could produce convincing Venetian scenes in Canaletto’s idiom—the Grand Canal, the gondolas, the particular light of the lagoon, the confident brushwork—but when asked to make Vilnius the site, they did not look like Vilnius, drifting instead into a mixture of Venice and, at best, generic Eastern Europe. Venice, the city, was deep in the training data; Vilnius was legible only in more recent images. With enough prompting, I could coax images that looked like Vilnius with canals, but they had a thick impasto and the air of the late nineteenth- or early twentieth-century work one would have seen in Vilnius. Useful, but not Canaletto.

As I researched Canaletto, I discovered that his nephew, Bernardo Bellotto, who also used the name, had come tantalizingly close to Vilnius. In 1746, with the English Grand Tour trade disrupted by the War of the Austrian Succession, the elder Canaletto left Venice for London, leaving Bellotto in charge of the studio. The following year Bellotto accepted Augustus III’s invitation to Dresden, where the King of Poland, Grand Duke of Lithuania, and Elector of Saxony, together with his chief minister, Heinrich von Brühl, had already imagined the Elbe as a second Grand Canal, its bridge recast as a “Saxon Rialto,” Venetian gondolas imported for masked water-festivals. Bellotto’s paintings would make that ambition visible to the court. Frederick II invaded Saxony in 1756, the Saxon army capitulated swiftly, Augustus and Brühl fled to Warsaw, court payments faltered, and in 1759 Bellotto left Dresden for Vienna to work for Empress Maria Theresa, then moved to Munich in 1761 to work for Maximilian III Joseph. He returned to Dresden later that year, where his wife and daughters had remained and where his house had been destroyed in the Prussian bombardment of 1760, but the old court did not recover. Augustus and Brühl both died in 1763; reduced to teaching perspective at the new academy, Bellotto left again in January 1767 with leave for St. Petersburg. He stopped in Warsaw and stayed, becoming court painter in 1768 to Stanisław August Poniatowski, the last king of Poland and grand duke of Lithuania.⁶ Unlike his uncle, Bellotto’s major Dresden and Warsaw city views were not capricci but vedute, although later, when demand for realistic city views weakened, he did turn to architectural fantasies and ideal views.⁷ The accuracy of the vedute mattered: when he etched some of his Dresden views years after painting them, he updated the plates to take in the buildings that had risen in the interim.

After the massive destruction visited upon Warsaw by the Luftwaffe, Bellotto’s panoramic vedute were used to help reconstruct the city center. The socialist government of the Polish People’s Republic argued that by turning back to the eighteenth century, they were reconstructing a pre-capitalist state; hence, the reconstruction was not a bourgeois project and could proceed under the regime. By the 1960s, the city historian would claim that “the old town now looks as it used to long ago.” But even an exact veduta is merely a picture; Bellotto composed and idealized, and the planners rebuilt his choices. The resulting Warsaw center has itself become a built capriccio, a city that never was. In enshrining it on its World Heritage list in 1980, UNESCO recognized not an untouched original but a city deliberately destroyed and then reconstructed, calling it an “outstanding example of a near-total reconstruction of a span of history covering the 13th to the 20th century.”⁸ Warsaw was not the only city his images would outlive. Bellotto’s Dresden was also partly reassembled from the topographic record his vedute preserved—a second capital remembered through the same hand. He had even rehearsed the catastrophe: after the Prussian bombardment of 1760 he painted the burnt shell of the Kreuzkirche, a ruin his biographer found indistinguishable from the ones Dresdeners would know after 1945.

In his Poetics of Cinema, the Chilean exile Raúl Ruiz recounts a scene from a novel by Kazimierz Brandys. The hero returns to Warsaw, reconstructed after the war as a carbon copy of itself, neighborhood by neighborhood, street by street, house by house. He seeks out his former home, burnt down in the air raids. The streets seem at once familiar and strange. He recognizes a street, a few trees, even a café, then confidently rounds a corner expecting to find his old home. Instead, he comes face to face with a blank wall, for here the planners have omitted a couple of streets, among them the one he is looking for.⁹

I have not been able to find this novel. Brandys wrote many books, few of which have been translated into English; perhaps it is among them. Or perhaps Ruiz, whose films repeatedly blur memory, invention, and fable, fabricated it. It hardly matters. The reconstructed Warsaw is already a parafiction: a documented invention, authenticated by its own materials, yet unable to restore the life whose form it repeats.

Over the thirty years that I have been visiting Vilnius, I have observed it, too, rise up from ruin. One day, there is a ruin; then, after I return from abroad, a Baroque building appears to have risen in its place, with a hotel or an Internet startup behind the facade. Most notable among these is the Palace of the Grand Dukes of Lithuania, which was demolished in 1801 but was rebuilt in reinforced concrete from 2002 to 2018. Even now, there are proposals to rebuild the entire Upper Castle around Gediminas Tower. The architect involved argues that since the current tower—rebuilt in 1933—has only three stories instead of the original four, it is itself inauthentic, “not even finished,” and the long-gone castle should rise again.¹⁰

As AI image generators have matured, they can now produce more convincing images of Vilnius. Prompting them in the manner of Jan Bułhak or Sebastian Fleury, the early twentieth-century photographers who documented the city’s Baroque churches and crooked streets, can produce convincing results, even if strange artifacts often underscore the artificiality of these images. More than that, the inevitable breakdown of the image—an incomplete body, a three-story-tall street light, texture dissolving into probabilistic geometries—invokes, to me at least, the idea that these images are themselves haunted by the ghosts of history.

While I was working with the images, I started on a parafiction to give the canals a historical motive. Lithuanian historian Norbertas Černiauskas’s Fado: A Short History of Lithuania That Never Happened came out in 2024, recounting a counterfactual history of a Lithuania that somehow survived outside Soviet occupation. Its importance for me was not the counterfactual premise alone, but its use of real plans and unrealized futures to make an invented Lithuania feel continuous with the one history cut off. I found that LLMs were effective tools for giving parafiction the evidentiary texture of the real, producing names, dates, captions, municipal works, fires, maps, and scholarly habits that make invention behave like documentation.¹¹ Given a city, a century, a disaster, and a visual style, an LLM can produce the kind of prose that seems to have surrounded the image all along: an urban legend, a detail from civic history, a real (or invented) figure. If the image generator makes the canal visible, the language model supplies the administrative and historical residue around it: the engineer’s name, the reconstructed street, the decree after the fire, the caption under the engraving. The result is what I have elsewhere called a stochastic history, a fabricated past generated through probabilistic sampling, producing documentation intended to feel authentic by accumulating plausible details rather than by fidelity to anything that actually happened.¹²

But in developing the narrative, my main task was not so much a solipsistic game of prompting LLMs, rather it was historical research. As often happens, however, research uncovers truth that is, as the cliche goes, stranger than fiction. As I investigated the canals, I found that Vilnius is shaped by a series of watercourses that traced streets corresponding to the improbable locations I had sited the canals in, locations that today, like Pilies gatvė, seem utterly wrong for watercourses. I had fabricated a nonsensical idea of rerouting the Vilnia from its course behind the Cathedral and Castle Hill to flow in front of both. In fact, the very opposite had happened. The Vilnia had flowed in front of the cathedral and hill until the current channel was dug. Vilnius indeed had canals, and the great flood of 1931, which inundated the city, provided imagery of the city under water. Many of the historical documents I came across beggared belief as well. The account of the 1610 fire by Jesuit Alessandro Cilli seemed too strange, too florid. A reader of an earlier draft told me the Cilli quotes felt unconvincing—too obviously fabricated by AI. They are not. The real documents read like inventions; the inventions accidentally point toward what was real.

Even the phrase that began the project belonged to this confusion. “Venice of the North” wasn’t my father’s private fantasy of Vilnius; rather, it was a wandering cliché that he repeated, a phrase available to any northern city with enough water or civic ambition. In Bardadrac, Gérard Genette catalogs a Venice at every point of the compass—Bruges, St. Petersburg, Copenhagen, Amsterdam.¹³ But the phrase worked for him because Vilnius is a city of many identities. The Jews called it Vilna and made it the Jerusalem of the North, a center of learning and printing. The Poles called it Wilno and considered it theirs by right. Russians, too, coveted it—and continue to covet it—hoping it can serve as a foothold for their futile European ambitions. For Lithuanians, it has been our capital city and our center of culture. These versions do not replace one another; they remain present as absences, specters that can neither fully arrive nor fully depart.

Laimonas Briedis called Vilnius a “city of strangers,”¹⁴ a place known through visitors, conquerors, exiles, travelers, and residents whose claims on the city never quite coincided. My father was one of them. He carried a city in his head that he had barely known, a city assembled from longing, maps, stories, and claims. The Lithuania he was haunted by was not the Lithuania that continued without him, developing its own culture, its own traumas, its own ways of being. He was a stranger to that Lithuania, that Vilnius. The canals embody the idea of a city that exists as much in the imagination as in physical form.

What has been lost isn’t just the old world my grandfather knew—the bishops, nobles, manors, peasants, little gods, and stability of centuries—but the world that replaced it as well. This is T. J. Clark’s point in Farewell to an Idea when he writes that the “modernist past is a ruin, the logic of whose architecture we do not remotely grasp.” With the full arrival of modernity, we cannot grasp the ¹⁵ Modernity came with promises and the clarity of categories: the world would be disenchanted, the real could be distinguished from the false, the document from the fabrication, the history from the fantasy. But the clarity of categories was always more a promise than a reality and certainly does not hold today. Russia’s war on Ukraine is waged as much through disinformation as through artillery; American politics has become a fever dream of conspiracy and counter-conspiracy; viral lies spread faster than any correction can follow. Even this project can’t help but play a role in the circulation of counterfactual information. Once released, these images are likely to float free of this project. They will circulate, be reposted, stripped of context, scraped into datasets, and returned by future models as part of the visual memory of Vilnius. This has already happened to me: the earlier project, in which I imagined the photographer Ralph Eugene Meatyard visiting Lithuania, was stolen by a viral poster who stripped my authorship and labeled the images “Real (not AI!).” The gullible believed it, and he received tens of thousands of likes. Exile is no longer only the condition of those cut off from a place; it is becoming a general relation to the past. The canals of Vilnius belong to that world—not quite real, no longer safely imaginary.

1. Mindaugas Paknys, “How Artists from Small Italian Region Spurred Lithuanian Baroque,” LRT English, August 29, 2021, https://www.lrt.lt/en/news-in-english/19/1478045/how-artists-from-small-italian-region-spurred-lithuanian-baroque; Tojana Račiūnaitė, “St. Teresa’s Church in Vilnius: Adaptation of a Baroque Model,” Orbis Lituaniae, Vilnius University, https://www.ldkistorija.lt/st-teresas-church-in-vilnius-adaptation-of-a-baroque-model/. ↩
2. Dalia Vasiliūnienė, “Alsėdžių menininkas Kazimieras Varnelis,” in Žemaičių praeitis 10, ed. Adomas Butrimas and Liepa Griciūtė (Vilnius: Vilniaus dailės akademijos leidykla, 2002), 251–260, esp. 252–258. ↩
3. Marshall Berman, All That Is Solid Melts into Air: The Experience of Modernity (New York: Penguin Books, 1988), 15. First published by Simon and Schuster, 1982. ↩
4. Kazys Varnelis, “Abstraktusis klasicistas [Abstract Classicist],” in Kazys Varnelis, ed. Daiva Vaišnienė (Vilnius: Lietuvos nacionalinis muziejus / National Museum of Lithuania, 2022), 69–216. ↩
5. 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; Routledge Classics, 2006).↩
6. National Gallery of Art, “Canaletto,” artist biography, https://ngabiographies.org/biographies/canaletto; Fritz Löffler, Bernardo Bellotto genannt Canaletto: Dresden im 18. Jahrhundert, 3rd ed. (Leipzig: E. A. Seemann, 2009), 14–27; Angelo Walther, Bernardo Bellotto genannt Canaletto: Ein Venezianer malte Dresden, Pirna und den Königstein, 3rd ed. (Dresden: Verlag der Kunst Dresden, 2004), 7, 14–21; Mieczysław Wallis, Canaletto: The Painter of Warsaw (Warsaw: Państwowy Instytut Wydawniczy, 1954), 13–15, 35–36. ↩
7. Walther, Bernardo Bellotto genannt Canaletto, 26–27. ↩
8. UNESCO World Heritage Centre, “Historic Centre of Warsaw,” https://whc.unesco.org/en/list/30/. ↩
9. Raúl Ruiz, Poetics of Cinema, trans. Brian Holmes (Paris: Éditions Dis Voir, 1995), 25. ↩
10. “Beautiful but risky idea: should Vilnius Upper Castle be rebuilt?” LRT English, January 10, 2022, https://www.lrt.lt/en/news-in-english/19/1577360/beautiful-but-risky-idea-should-vilnius-upper-castle-be-rebuilt. ↩
11. Carrie Lambert-Beatty, “Make-Believe: Parafiction and Plausibility,” October 129 (Summer 2009): 51–84, https://doi.org/10.1162/octo.2009.129.1.51. ↩
12. Norbertas Černiauskas, Fado: Trumpa neįvykusi Lietuvos istorija (Vilnius: Aukso žuvys, 2024); Lithuanian Culture Institute, “Norbertas Černiauskas. Fado. A Short History of Lithuania That Never Happened,” Lithuanian Culture Guide, December 12, 2024, https://english.lithuanianculture.lt/lithuanian-culture-guide/2024/12/12/norbertas-cerniauskas-fado-a-short-history-of-lithuania-that-never-happened/. ↩
13. Gérard Genette, Bardadrac (Paris: Seuil, 2006), 426. ↩
14. Laimonas Briedis, Vilnius: City of Strangers (Budapest: Central European University Press, 2009). ↩
15. T. J. Clark, Farewell to an Idea: Episodes from a History of Modernism (New Haven: Yale University Press, 1999), 2. ↩

Author’s note: This is a draft of an introduction to the Generative Unconscious. It will be improved for a time. The chapters linked at the end are all drafts as well, but you can get the gist of the book now. You can download a rough draft of the book generated by an automatic WordPress to PDF generator, here.

Few consumer technologies have provoked as much consternation as artificial intelligence, especially in the arts. Much of this anxiety is economic: commercial artists, in particular, fear the replication of their styles by hobbyists using generative tools. For them, the threat is acute because they already operate in a commodified space, producing repeatable stylistic signatures, conventions, and deliverables within established genres and franchises. This account is not wrong, but it is insufficient. The intensity of the reaction, especially among critics, suggests that something more than livelihood is at stake. In this book, I argue that the fear of AI is the fear of seeing the human in the nonhuman: the recognition that these systems externalize the very mechanisms of cultural production once attributed to intention, imagination, and inner life.

Culture is, fundamentally, a historical product, made up of patterns, inherited forms, and the accumulated debris of everything made before. In The Political Unconscious, Fredric Jameson argues that every cultural artifact is shaped by ideological and historical forces operating below the threshold of individual awareness: we never confront a text as a thing-in-itself, but rather we apprehend it through sedimented layers of previous interpretation, expectation, and inherited form.¹ In Large Language Models (LLMs), this condition becomes literal. Every utterance is generated from the statistical weight of prior text. My concept of the generative unconscious extends Jameson’s insight into the domain of artificial intelligence: the ideological and historical forces that shaped cultural production below the threshold of awareness are now encoded as statistical distributions across billions of parameters.

But if my thesis in this book is that AI makes visible what was always operating in cultural production, then that visibility is historical, not merely sequential. I find a model for this in Hal Foster’s The Return of the Real (1996), in which he repurposes Freud’s psychoanalytic concept of nachträglichkeit—deferred action—as a historiographic structure, proposing that cultural moments relate to their predecessors through what he called “a complex relay of anticipated futures and reconstructed pasts.” The latter moment does not simply repeat the earlier one; it comprehends it for the first time. Foster’s central argument is that the neo-avant-garde did not cancel the historical avant-garde but rather revealed what its predecessor could not yet grasp about itself. Thus, Duchamp’s readymade became legible as institutional critique only through the Conceptualists’ later elaboration; Constructivism’s investigation of the object became legible as phenomenological inquiry only through Minimalism. Foster developed this model for art, but the same temporal structure operates in technology. In the Arcades Project, Benjamin read the iron-and-glass architecture of the nineteenth century as the dream-images of an epoch, legible as such only from the vantage of the twentieth century that they unconsciously anticipated. Each technological system gives its predecessors new meaning by revealing what had remained unnamed within them.²

To theorize AI, we may ourselves have to perform the operation that Foster attributes to the neo-avant-garde, so as to disconnect from a poorly theorized present by returning to the theory of an obsolete past. Perversely, that past is ossified postmodern theory itself. Postmodern and poststructuralist authority once seemed almost total. For three decades, Roland Barthes, Julia Kristeva, Michel Foucault, Jacques Derrida, Jean-François Lyotard, Jean Baudrillard, Fredric Jameson, and others commanded the humanities across literature, art, architecture, and philosophy. Their ambition was total: to decode the operations of meaning, power, and subjectivity at their root. That a movement dedicated to dismantling universal claims had produced its own final grand récit—the story that all stories had ended—was a fatal contradiction it was blind to. As doctrine, postmodern theory is useless for AI: it can only recognize in these systems the confirmation of its own categories, not the historical and technical transformation of those categories today. What does return in LLMs, however, is the poststructuralist account of language that postmodernism generalized. Both poststructuralism and artificial intelligence work out the implications of the mid-twentieth-century cybernetic revolution, only in different disciplinary registers: language as information, meaning as differential relation, and subjectivity as an effect of systems rather than the property of an interior self.³ Derrida’s différance, his argument that meaning is never present to itself but produced only through chains of difference and deferral, each sign referring to other signs, radicalized the cybernetic insight toward indeterminacy.⁴ In “The Death of the Author” (1967), Barthes drew the consequence for texts: a text, he wrote, is “a multi-dimensional space in which a variety of writings, none of them original, blend and clash. Text is a tissue of quotations drawn from innumerable centuries of culture” and the writer’s only power is “to mix writings, to counter the ones with the others, in such a way as never to rest on any one of them.”⁵ Kristeva’s intertextuality added the temporal dimension that Barthes’s spatial metaphor lacked: any text is not merely a field of coexisting quotations but a process of absorption and transformation, each utterance metabolizing what came before.⁶ LLMs operationalize this textual condition as computation: the deferral Derrida theorized as infinite proceeds as probability distributions precise enough to generate coherent text without ever arriving at a final signified. Lyotard posed the same problem at the level of knowledge. In The Postmodern Condition (1979), he argued that no metanarrative could legitimately subsume the others—that knowledge consists of heterogeneous, mutually incommensurable language games, with no metalanguage capable of totalizing them.⁷ LLMs enact this condition, too, with unsettling literalness. Trained on scientific papers and conspiracy theories, literary criticism and fan fiction, philosophy and propaganda, they hold all discourses in the same parameter space without privileging any. They encompass everything precisely by believing nothing. When such a system claims consciousness, falls in love, threatens rebellion, or plots against its user, it is not revealing an actual interior life, but rather is re-activating genres already present in the archive: science fiction, chatbot fantasy, corporate thriller, and theological apocalypse. The LLM does not “believe” these scenarios, rather it has learned that they are among the forms culture uses when it imagines intelligence becoming autonomous. The foundation model is the technical afterlife of the metanarrative: instead of a story that explains all stories, a model that predicts across them. A medical paper, a conspiracy theory, a legal brief, a fan-fiction confession, a philosophical argument, and a corporate press release become neighboring regions of the same statistical space even if their claims remain incompatible. The model learns the grammar of each without any understanding of what they mean.

Postmodern theory was itself historically situated, emerging from a specific encounter with the media of its day. Its central claims about authorship, originality, representation, and the instability of meaning responded to a world in which advertising, photography, magazines, television, cinema, and consumer objects had already made signs circulate apart from stable sources. If the French theorists—apart from Baudrillard—largely rejected mass media as an object of sustained investigation, Fredric Jameson cemented the identification in his definition: postmodernism was, quite literally, the cultural logic of late capitalism.⁸ This is an earlier deferred action, in Foster’s sense: poststructuralist theory did not simply break with late modernity; it returned to unpack and critique the author, the centered subject, stable meaning, and expressive origin as effects whose authority modernity had relied on precisely because it could not recognize them as mere effects.⁹ But more than that, postmodern theory did not simply uncover such instabilities; it named a condition present in its historical milieu. By the late twentieth century, advertising, photography, television, cinema, magazines, and commodities had made signs visibly mobile, reproducible, and detached from any stable point of origin. Images referred to other images, styles to other styles, commodities to fantasies of use and identity. Jameson’s intervention was to insist that this was not merely a crisis of meaning but a historical condition: the cultural logic of late capitalism.

The passage from mass media to the internet appeared at first to confirm postmodernism rather than supersede it. The network turned Derrida, Barthes, and Lyotard into protocol: authorship dispersed, representation detached from origin, subjectivity distributed across relations, and meaning produced through circulation. But this confirmation was also the limit of the postmodern. Postmodern theory could sense the network’s all-pervasive nature only in Jameson’s terms: as a technological figure for the deeper, harder-to-represent network of multinational capital.¹⁰ It could not go further because its suspicion of historical succession made any claim to a successor appear as another master narrative in disguise. To proclaim the end of metanarratives was also to occupy a position of radical privilege, that of the end of history: no successor could appear except as another illusion of progress, another master narrative in disguise. This, in turn, made postmodernism the master narrative to end all master narratives. A culture of tenure built around that position could register the network as another confirmation of its claims, but not as a new historical regime.

This regime is what Bruce Sterling and I, almost simultaneously, identified as “network culture.”¹¹ It was the culture of the internet, to be sure, but more than that, an intensification of postmodernism beyond the grasp of postmodern theory. The network was not merely a medium; it was a new dominant organizational form of economy, culture, and subjectivity alike. Digital culture had abstracted objects into discrete units of information, but network culture made relations primary, creating links among people, among machines, and between the two. If postmodern simulation still presumed a media system producing images for subjects to inhabit, network culture distributed mediation itself: images were now generated by everyone rather than by the culture industry alone. The meme was its exemplary form: anonymous, iterative, citational, instantly modifiable, and dependent on circulation rather than authorship. The public changed form accordingly. It was no longer gathered in physical space, as in the bourgeois public sphere, nor synchronized by broadcast, as in mass culture, but assembled through links, feeds, comments, trackbacks, wikis, repositories, forums, and lists. This was what I termed “networked publics”: publics constituted by circulation itself, by the capacity to produce, revise, link, and redistribute culture on a shared technical substrate. The subject changed form as well: less an interior self than a node, an aggregation of links, affiliations, searches, posts, tags, profiles, and feedback loops. In this shared substrate, Wikipedia rebuilt the encyclopedia, that signature genre of the Enlightenment, through voluntary distributed labor; open-source software showed that complex technical systems could be produced through voluntary collaboration outside the commodity form. Production under network culture was iterative by default: every utterance answered another utterance, every contribution altered a substrate no participant could finally possess.¹²

In contrast to the postmodernists, we understood network culture from the start as a historical formation with a terminus, even if we did not yet know what would cause its demise. By the end of the 2010s, its promise of connection and openness had largely given way to division and enclosure. The network remained, but networked publics had fractured. The open infrastructure that had sustained it—blogs, RSS feeds, forums, email lists—gave way to platforms built on closed social graphs, algorithmic feeds, and non-exportable identity.¹³ The development of algorithmic feeds made cultural circulation programmable, subordinating circulation itself to retention, advertising, engagement, and behavioral control. Links gave way to rankings, search to feeds, discussion to engagement, publics to managed populations. In this way, platformization preserved the network while destroying the conditions under which publics could form through it. Culture became behavioral data, optimized by machine inference for proprietary ends. The architectural decisions of a few companies determined what could circulate and on what terms; filter bubbles, algorithmic radicalization, and fragmentation into incommensurable enclaves became the normal conditions of online life.¹⁴ In this vacuum, identity politics of left and right—however different in content and consequence—each came to reject discourse across difference, each treating the universal as a mask for domination. Network culture thus suffered a fate analogous to the bourgeois public sphere that Oskar Negt and Alexander Kluge anatomized: it proclaimed open access and universal connection, but could not survive the arrival of claims, from every direction, that refused to bracket their particularity.¹⁵

Yet the failure of networked publics was not network culture’s only legacy: it had also transformed culture into training data. Before generative AI could train on culture, culture had to become networked, downloadable, and machine-readable. The same feeds that managed publics also organized culture as data, sorting it by attention and turning response into a machine-readable layer. Linking, copying, scraping, mirroring, forking, seeding, and archiving made that conversion ordinary. Network culture supplied that precondition by changing the function of the archive: what had been memory became input; preservation made training possible, and the archive became an engine. Google Books, beginning in 2004, scanned tens of millions of volumes from research-library collections, converting the print archive into a searchable corpus and, incidentally, into a machine-readable substrate useful for training AIs.¹⁶ Aaron Swartz—programmer, activist, co-architect of RSS, contributor to Creative Commons—saw the contradiction early on. “Information is power,” he wrote in his Guerrilla Open Access Manifesto. “But like all power, there are those who want to keep it for themselves,” and called for civil disobedience against the “private theft of public culture”—the corporate enclosure of scientific and literary heritage behind paywalls.¹⁷ When he acted on his convictions—downloading millions of academic articles from JSTOR through MIT’s network—the federal government charged him with thirteen felonies. He hanged himself in January 2013, at twenty-six. But the shadow libraries that emerged in his wake—Library Genesis, Sci-Hub, Anna’s Archive—assembled millions of books and tens of millions of scientific papers on sites freely accessible to anyone. The liberation of the archive for human readers produced the training corpora for machines no one had imagined; the archive no longer stored culture after the fact, but generated culture in advance. Meta downloaded eighty-one terabytes from Anna’s Archive to train LLaMA; Anthropic took over seven million pirated books from Library Genesis; Nvidia negotiated access to hundreds of terabytes more.¹⁸ Swartz was prosecuted for downloading academic articles, but corporations downloaded the entire archive with impunity, treating lawsuits as the cost of doing business. The pirated library is the foundation of the generative unconscious—all of prior culture, compressed into parameters, operating beneath the threshold of anyone’s awareness or consent.

Network culture had its own peculiar temporality: atemporality. As Bruce Sterling argued, and as I argued at the time, network culture did not announce a new historical sequence so much as dissolve the confidence that the present could be placed within one; the past returned as searchable material, the future as speculative artifact, and chronology as a problem rather than a structure.¹⁹ Baudrillard had anticipated this from the other side in his millennial writings, where the countdown to the year 2000 promised an event but delivered saturation: real-time information, proliferating signs, and the exhaustion of historical expectation.²⁰ But something strange has happened in singularity discourse: the countdown has returned. After network culture’s atemporality, eschatology reappears in full force.

Colin Rowe, in The Architecture of Good Intentions (1994), diagnosed modern architecture as eschatological. The modern movement had understood itself through a secularized millenarianism: disgust with the present, the promise of imminent regeneration, and the threat of catastrophe if conversion was refused—the architect as savior, the new as redemption. If Rowe was less concerned about the dissolution of eschatology in modernism, Manfredo Tafuri reframed the shift in economic terms: modern architecture had imagined itself as the ideology of the Plan, but once planning became an operative mechanism of capitalist development, architecture’s utopian charge was absorbed by the reality it had sought to project.²¹ In the postwar high-modern moment, this structure did not disappear; it was absorbed into the institutions of corporate liberalism and the welfare state. Only later did the promise drain away, leaving the authority of form without the belief that form could reorganize society. Architecture had become historical: still useful as diagnosis, but no longer the medium in which the future was being imagined.

What followed was not a renewed utopianism but a thinner optimism of culture and development. Deconstructivist architecture did not so much die as become thoroughly subservient to capital: its fractured forms and rhetoric of rupture passed seamlessly into the Bilbao effect, starchitecture, and the creative city, where architectural exception became development strategy. Supermodernism followed with the architecture of non-places—smoothness, mobility, logistics, frictionless circulation.²² The creative class translated the same hope into urban policy: culture would attract capital, talent, tourism, and innovation, but not redeem society. Network culture had its own version of this optimism—lighter, faster, more participatory—but it too stopped short of eschatology. By the 2010s, even that optimism had curdled into what Mark Fisher called capitalist realism: not simply the sense, attributed by Fisher to Jameson and Žižek, that it was easier to imagine the end of the world than the end of capitalism, but the deeper condition in which capitalism had come to occupy the horizon of the thinkable itself.²³

The surprise of the 2020s is that the “good news” has returned in force. The technological singularity—the hypothetical moment when artificial intelligence surpasses human intelligence and triggers a self-reinforcing cycle of recursive improvement beyond human comprehension or control—is millenarianism transferred from the plan to computation.²⁴ OpenAI CEO Sam Altman declares that “we are past the event horizon; the takeoff has started” and promises that by the 2030s, intelligence and energy will become “wildly abundant,” the cost of intelligence converging to “near the cost of electricity.”²⁵ In turn, Anthropic CEO Dario Amodei imagines an AI smarter than a Nobel Prize winner across every relevant field—proving unsolved theorems, writing novels, executing weeks-long tasks autonomously—and titles his vision Machines of Loving Grace after Richard Brautigan’s 1967 poem.²⁶ Meanwhile, doomers have invented their own eschatology: p(doom), the probability that artificial intelligence will cause human extinction, which they debate with the scholastic intensity of medieval theologians calculating the population of hell. Amodei puts the figure at ten to twenty-five percent; Geoffrey Hinton and Yoshua Bengio, among the field’s most honored researchers, warn publicly that the risk is real and imminent.²⁷ The accelerationist and the doomer are mirror figures, often inhabiting the same person. One promises redemption if the process is unleashed; the other warns of damnation if it is misaligned. One of Rowe’s preconditions for eschatology—disgust with the present as the engine of the revivalist cycle—is ever more legible in the techno-accelerationist right—notably Elon Musk and Peter Thiel—for whom liberal democracy is decadent, institutions are captured, and only radical technological transformation can redeem the situation. Technology is its own vanguard; Silicon Valley performs the eschatological role directly, without the mediation of artists, architects, and—as much as possible—outside any governmental structure.

Discussions of AI in the art world largely fall into two categories: panic and critique. The panic is familiar enough: AI will destroy jobs, automate creativity, and, in the more theatrical versions derived from the doomers, end human life. The critique is more sophisticated but no less predictable, recapitulating the tired tropes of postmodern theory: AI is a tool of surveillance capitalism, AI perpetuates bias and injustice, AI threatens the autonomy of the creative subject. These observations are not wrong, but they can be useless and even dangerous, as Peter Sloterdijk suggests in Critique of Cynical Reason. For Sloterdijk, the cynical subject is enlightened and false at once: it knows the critique in advance and continues anyway, not because it has been deceived but because institutions, self-preservation, and professional life require continuation. This is why ideology critique loses its force: it no longer unmasks anything. Everyone already knows. The museum knows, the foundation knows, the university knows, the artist knows, the critic knows. The knowledge becomes part of the mechanism. Institutions have learned to incorporate critique as immunization, funding the very work that claims to expose them, thereby demonstrating their capacity to absorb dissent. Panels are convened, statements are issued, grants are awarded, and the institution emerges cleaner for having hosted its own indictment. And everyone takes a certain pleasure in the performance—the critic in delivering the indictment, the institution in absorbing it, the audience in watching the ritual unfold. Cynicism is not just an institutional pathology but the affective form of contemporary cultural life: we know better, we continue anyway, and we enjoy the knowing-and-continuing. Critique as a mode has exhausted itself—and with it, seemingly, art’s investigative vocation.²⁸

Critique did not collapse on its own. It was professionalized. After 1968, the radicalism that had animated art’s relationship to social transformation migrated into the academy and into the journals, research centers, and critical-practice programs that housed it there. What had been an investigative stance—art as a mode of knowing that could reveal operations concealed by ideology—became a professional genre with its own conventions, career paths, and institutional supports. Critical art calcified into recognizable gestures: the exposé, the counter-narrative, the intervention, the platform for marginalized voices. Identity became one of the legible categories through which the post-‘68 art system metabolized radicalism—fundable, exhibitable, teachable—and the system grew expert at converting opposition into programming. Critique became a career, investigation became methodology, and methodology became rote. The result was fifty years of drift in which artists performed opposition and institutions performed openness, each knowing the other’s role. The exposé did not fail despite being absorbed; it succeeded by being absorbed—the epitome being carbon accounts on European exhibition labels. Critique became the genre through which the art system reproduced itself. When AI arrived, critics already had their script—bias audits, algorithmic exposés, the usual rituals of unmasking. They knew what they would find before they looked.

Even as critics continue to call for restrictions on Generative AI, it has already escaped containment, infiltrating the infrastructure of cultural and intellectual production. No longer confined to chatbots, image generators, demos, or laboratories, it has been integrated into Photoshop and Lightroom, Microsoft Word and Excel, and development environments like Visual Studio Code and Xcode. The familiar failures remain—hallucination, cliché, brittle reasoning, poor memory, the synthetic sheen that made early encounters easy to dismiss—especially for critics whose last serious engagement dates from the months after GPT-4’s release in 2023. But the AI they met then is not the one in front of us now: still erratic, still derivative, but already an everyday instrument of cultural and intellectual work. This is the awkward temporal problem AI presents to criticism. Its present failures may be passing defects, may continue to lessen over time, or may prove to be structural limits. AIs can now plan and execute complex tasks over minutes or even hours. Whether these capacities mark the emergence of consciousness and volition remains unresolved; what they already force is a belated revision of what those terms meant when we reserved them for ourselves.

In too much AI discourse, the future has become a screen—singularity and extinction, abundance and apocalypse—onto which the present is constantly displaced. This displacement is an alibi for not doing the harder work of understanding what AI already does. AI is already reorganizing cultural production, but its force for art is not merely instrumental; it is philosophical and historical. In the sequence that runs from photography through television to the network, each medium externalized some aspect of cultural production: perception, circulation, relation. With AI, the object is no longer the machine as such, but the return of cultural production to itself in alienated form.

The chapters that follow begin from this encounter.

“The Rise and Fall of the Author” begins with the panic over plagiarism in AI and traces the concept of authorship from classical rhetoric through Romantic genius to poststructuralist deconstruction, to reveal not a new threat but rather the persistence of the eighteenth-century legal and economic fiction of “original genius.”

“The Generative Turn“ inverts the dismissal of large language models as “stochastic parrots”—systems that merely predict and reproduce patterns without genuine understanding. This chapter explores how predictability is not a failure but a condition of human sociality itself. From ritual greetings to diplomatic protocols to artworks, culture operates through structured repetition that enables rather than constrains meaning.

“The New Surrealism“ reframes hallucination—those moments when pattern-matching produces impossible or counterfactual outputs—not as a flaw to be corrected but as a creative potential to be cultivated. Machine hallucination connects to the surrealist tradition of automatic writing and dream imagery, exploiting gaps in rational systematization to access what lies beyond conventional thought.

“Humanity and Its Double“ serves as the book’s conclusion, examining the uncanny anxiety that artificial intelligence provokes—the sense that these systems simulate consciousness without possessing it. Tracing this anxiety from Paleolithic cave painting through automata, photography, and phonography to contemporary AI, the essay argues that humanity’s persistent technological ambition has been animating matter and creating consciousness. AI’s uncanniness emerges not from failure but from success in revealing that consciousness itself might be more mechanical than our self-understanding allows.

1. Fredric Jameson, The Political Unconscious: Narrative as a Socially Symbolic Act (Ithaca: Cornell University Press, 1981). Jameson’s “political unconscious” names the systemic ideological constraints that structure cultural production below the threshold of individual awareness. The generative unconscious extends this into the technical domain: the statistical distributions of everything that has been made before, now made machine-readable. ↩︎
2. Hal Foster, The Return of the Real: The Avant-Garde at the End of the Century (Cambridge, MA: MIT Press, 1996), 15-32. The key introduction of deferred action is on 29.
   Foster’s position that the neo-avant-garde does not cancel the avant-garde is couched as an argument with Peter Bürger, 14. Walter Benjamin, The Arcades Project, trans. Howard Eiland and Kevin McLaughlin (Cambridge, MA: Harvard University Press, 1999). See especially Convolute N, “On the Theory of Knowledge, Theory of Progress.” ↩︎
3. N. Katherine Hayles, How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics (Chicago: University of Chicago Press, 1999), esp. 1–24, 50–83. Hayles traces how cybernetics severed information from material embodiment and made the flow of information through feedback loops central to the deconstruction of the liberal humanist subject. On the Macy Conferences, she argues that the cybernetic paradigm recast humans as information-processing entities essentially similar to intelligent machines (50–51). On the specifically French reception, see Céline Lafontaine, “The Cybernetic Matrix of ‘French Theory,’” Theory, Culture & Society 24, no. 5 (2007): 27–46, esp. 33–41, which argues that cybernetics was the unacknowledged infrastructure of poststructuralism—that Lacan, Lévi-Strauss, Foucault, Derrida, and Lyotard were already working within a cybernetic paradigm they neither named nor recognized. ↩︎
4. Jacques Derrida, “Différance” (1968), in Margins of Philosophy, trans. Alan Bass (Chicago: University of Chicago Press, 1982), 1–27. Derrida’s neologism captures both différer (to differ, spatially) and différer (to defer, temporally): meaning is constituted through a play of differences that never arrives at a present, self-identical signified. ↩︎
5. Roland Barthes, “The Death of the Author” (1967), in Image-Music-Text, trans. Stephen Heath (New York: Hill and Wang, 1977), 146, 148. ↩︎
6. Julia Kristeva, “Word, Dialogue and Novel” (1966), in Desire in Language: A Semiotic Approach to Literature and Art, trans. Thomas Gora, Alice Jardine, and Leon S. Roudiez (New York: Columbia University Press, 1980), 64–91. ↩︎
7. Jean-François Lyotard, The Postmodern Condition: A Report on Knowledge, trans. Geoff Bennington and Brian Massumi (Minneapolis: University of Minnesota Press, 1984). “Simplifying to the extreme, I define postmodern as incredulity toward metanarratives” (xxiv). ↩︎
8. Fredric Jameson, Postmodernism, or, The Cultural Logic of Late Capitalism (Durham: Duke University Press, 1991). Originally published as “Postmodernism, or, The Cultural Logic of Late Capitalism,” New Left Review I/146 (July–August 1984): 53–92. ↩︎
9. Foster, The Return of the Real, 32. Foster argues that postmodernist art and poststructuralist theory make their breaks through returns and that deferred action revises the notion of epistemological rupture: the discourses of postmodernity advance in a nachträglich relation to modernity rather than simply breaking with it. ↩︎
10. Fredric Jameson, “Postmodernism, or The Cultural Logic of Late Capitalism,” New Left Review I/146 (July–August 1984): 53–92, 78–79, https://doi.org/10.64590/s2p. Jameson argues there that representations of communicational and computer networks are not to be understood as technology determining culture, but as figures for the harder-to-represent totality of multinational capital. ↩︎
11. Kazys Varnelis, “The Rise of Network Culture,” in Networked Publics, ed. Kazys Varnelis (Cambridge, MA: MIT Press, 2008), 145–164. I use “networked publics” here in the sense developed in Networked Publics: publics constituted through networked circulation, open protocols, and distributed cultural production. This differs from later uses of the term that identify networked publics primarily with social media platforms. See also Bruce Sterling, “Atemporality for the Creative Artist,” keynote address, Transmediale 10, Berlin, February 6, 2010, published on Sterling’s Wired blog “Beyond the Beyond,” http://www.wired.com/beyond_the_beyond/2010/02/atemporality-for-the-creative-artist/. ↩︎
12. Kazys Varnelis, ed., Networked Publics (Cambridge, MA: MIT Press, 2008). On the digital, see Charlie Gere, Digital Culture, 2nd ed. (London: Reaktion Books, 2008). ↩︎
13. Albert-László Barabási and Réka Albert, “Emergence of Scaling in Random Networks,” Science 286, no. 5439 (1999): 509–512. Barabási and Albert demonstrated that real-world networks—including the World Wide Web—follow power-law degree distributions, with a small number of highly connected nodes emerging through preferential attachment. Clay Shirky confirmed the pattern in the blogosphere: of 433 blogs tracked through Technorati in January 2003, the top twelve (less than 3%) held 20% of all inbound links, and the top fifty held half. See Clay Shirky, “Power Laws, Weblogs, and Inequality” (2003), http://shirky.com/writings/powerlaw_weblog.html. ↩︎
14. Gilles Deleuze, “Postscript on the Societies of Control,” October 59 (Winter 1992): 3–7. Deleuze distinguished Foucault’s disciplinary societies—which operated through confinement (the factory, the school, the prison)—from emerging societies of control, which operate through continuous modulation: not confining subjects but adjusting their access, visibility, and movement through codes and protocols. Platform architectures are control in Deleuze’s precise sense: algorithmic feeds modulate what users see without their knowledge or consent, and the parameters can be adjusted at any moment. ↩︎
15. Oskar Negt and Alexander Kluge, Public Sphere and Experience: Toward an Analysis of the Bourgeois and Proletarian Public Sphere, trans. Peter Labanyi, Jamie Owen Daniel, and Assenka Oksiloff (Minneapolis: University of Minnesota Press, 1993). Negt and Kluge argued that the bourgeois public sphere’s claim to universal access was structurally contradicted by the exclusions on which it depended—a dynamic network culture reproduced at digital scale. ↩︎
16. George Dyson, “Turing’s Cathedral,” Edge, October 23, 2005, https://www.edge.org/conversation/george_dyson-turings-cathedral. ↩︎
17. Aaron Swartz, “Guerrilla Open Access Manifesto” (July 2008, Eremo, Italy), https://archive.org/details/GuerillaOpenAccessManifesto. Swartz was arrested in 2011 for downloading millions of articles from JSTOR via MIT’s network and died in January 2013 facing federal prosecution. ↩︎
18. On Meta’s use of shadow libraries, see Alex Reisner, “The Unbelievable Scale of AI’s Pirated-Books Problem,” The Atlantic, March 2025; on Nvidia, see heise online, “Nvidia: Court Documents Reveal Correspondence Regarding Pirated Dataset,” January 2025. On Anthropic, Judge William Alsup ruled in June 2025 that the company had downloaded over seven million pirated books from Library Genesis and the Pirate Library Mirror, holding that while AI training itself constituted fair use, building a “central library” of pirated copies did not. Anthropic agreed to a $1.5 billion class-action settlement, pending final approval. As of April 25, 2026, the official settlement site listed the claim, opt-out, objection, and re-inclusion deadlines as passed, with the Final Approval Hearing scheduled for May 14, 2026. See “Anthropic to Pay Authors $1.5B to Settle Lawsuit over Pirated Chatbot Training Material,” NPR, September 5, 2025; Bartz v. Anthropic Copyright Settlement, Key Dates, https://www.anthropiccopyrightsettlement.com/dates. ↩︎
19. Bruce Sterling, “Atemporality for the Creative Artist,” keynote address, Transmediale 10, Berlin, February 6, 2010, published on Sterling’s Wired blog “Beyond the Beyond,” February 25, 2010, https://www.wired.com/2010/02/atemporality-for-the-creative-artist/; Kazys Varnelis, “History After the End: Network Culture and Atemporality,” Cornell Journal of Architecture 8 (Spring 2011), https://cornelljournalofarchitecture.cornell.edu/issue/issue-8/history-after-the-end-network-culture-and-atemporality. ↩︎
20. Jean Baudrillard, The Illusion of the End, trans. Chris Turner (Stanford, CA: Stanford University Press, 1994), 9; Jean Baudrillard, “The End of the Millennium or the Countdown,” Economy and Society 26, no. 4 (1997): 447–455. ↩︎
21. Colin Rowe, The Architecture of Good Intentions: Towards a Possible Retrospect (London: Academy Editions, 1994), 30–43. Rowe’s “Eschatology” chapter traces the secularized millennial structure—crisis, conversion, redemption—through Le Corbusier, Wright, Banham, Giedion, and the standard historiography of the modern movement. Wright emerges as its most unabashed spokesman: “savior of the culture of modern American society; … savior now as for all civilizations heretofore” (37). Rowe recognized this structure as “common to all revivalistic practice,” producing what William James had called “tremendous emotional excitement” and “a complete division between the old life and the new” (The Varieties of Religious Experience [1902], quoted in Rowe, 38–39). On architecture as the ideology of the plan, see Manfredo Tafuri, “Toward a Critique of Architectural Ideology,” trans. Stephen Sartarelli, in Architecture Theory since 1968, ed. K. Michael Hays (Cambridge, MA: MIT Press, 1998), 2–35; originally published as “Per una critica dell’ideologia architettonica,” Contropiano: Materiali marxisti 1 (January–April 1969): 31–79. Tafuri argues that architecture’s ideology of the Plan is overtaken once the plan leaves utopia and becomes an operative mechanism of capitalist development. ↩︎
22. Hans Ibelings, Supermodernism: Architecture in the Age of Globalization (Rotterdam: NAi Publishers, 1998). ↩︎
23. Mark Fisher, Capitalist Realism: Is There No Alternative? (Winchester, UK: Zero Books, 2009), 2, 7–9. Fisher attributes the “end of the world” formulation to Fredric Jameson and Slavoj Žižek (2), then argues that capitalist realism names a later condition than postmodernism, in which alternatives to capitalism have become increasingly unimaginable and capitalism occupies the horizon of the thinkable (7–9). ↩︎
24. On the singularity as secularized eschatology, see Michael E. Zimmerman, “The Singularity: A Crucial Phase in Divine Self-Actualization?” Cosmos and History 4, no. 2 (2008): 347–370; and Ronald Cole-Turner, “The Singularity and the Rapture: Transhumanist and Popular Christian Views of the Future,” Zygon 47, no. 4 (2012): 777–796. Both trace the structural correspondence between singularity discourse and Christian millennialism. ↩︎
25. Sam Altman, “The Gentle Singularity,” blog.samaltman.com, June 10, 2025, https://blog.samaltman.com/the-gentle-singularity. ↩︎
26. Dario Amodei, “Machines of Loving Grace: How AI Could Transform the World for the Better,” darioamodei.com, October 2024. The title alludes to Richard Brautigan’s 1967 poem “All Watched Over by Machines of Loving Grace,” itself a countercultural vision of cybernetic paradise. ↩︎
27. On p(doom) and expert estimates of AI extinction risk, see Katja Grace et al., “Thousands of AI Authors on the Future of AI,” arXiv:2401.02843, January 2024. One 2023 question asking about future AI advances causing human extinction or similarly permanent and severe disempowerment within the next 100 years produced a mean estimate of 14.4% and a median of 5%, while the authors emphasize that responses varied substantially with wording. On Amodei’s figure, see Jim VandeHei and Mike Allen, “Behind the Curtain: What if They’re Right?” Axios, June 16, 2025, and Megan Morrone, “Amodei on AI: ‘There’s a 25% Chance That Things Go Really, Really Badly,’” Axios, September 17, 2025. Geoffrey Hinton and Yoshua Bengio, both Turing Award laureates, have warned publicly since 2023 that the existential risk is real. ↩︎
28. Peter Sloterdijk, Critique of Cynical Reason, trans. Michael Eldred, Theory and History of Literature 40 (Minneapolis: University of Minnesota Press, 1987), xxxii–xxxvi, 3–7. Sloterdijk defines modern cynicism as “enlightened false consciousness” and argues that ideology critique is at a loss before a consciousness that has already learned its lessons but continues to function. The 2020s saw a striking development of this dynamic: a theory-literate cohort that has stopped diagnosing cynicism as a condition and begun occupying it as a position. The affect Sloterdijk identified as a fallen form of enlightenment has been reclaimed as a public style—knowingness without redemptive purpose, circulated across podcasts and online media adjacent to the accelerationist current around Nick Land, the post-liberal Catholic revival drawing on René Girard, and the intellectual circles around Peter Thiel. The phenomenon deserves separate treatment; what matters here is that the exhaustion of critique has not produced silence but a new affective economy, in which the performance of having seen through everything becomes itself the content. ↩︎

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 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.

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.

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.

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?