“Graham’s Escape,” Illustration by Henry Lamos for “When the Sleeper Wakes”
by H.G. Wells (1899) Source

In 1938 H.G. Wells famously predicted that “the whole human memory can be, and probably in a short time will be, made accessible to every individual.” But his vision of a so-called World Brain involved more than just organizing information; he was convinced that it would eventually incubate a “widespread world intelligence.”

Wells is best remembered today as a sci-fi writer (and enthusiastic proponent of free love), but he was also a prolific polemicist and social reformer, who thought deeply about the transformative possibilities of new technologies. Today, his vision of the World Brain seems eerily prescient, as we start to encounter tools capable not just of organizing human knowledge but of generating it.

I was reminded of Wells’s work while reading Henry Shevlin’s recent essay Behaviourism’s Revenge, where he ponders the question of machine consciousness. Putting aside the endless debates about AGI, he argues that ultimately, intelligence is in the eye of the beholder. “At a certain point, whether or not a system is ‘really’ conscious becomes less important than how people respond to it in practice.”

Wells arrived at something like this intuition from roughly the opposite direction. His putative machines are not convincing because of their inner workings, but because of the roles they come to occupy in the social imagination.

But what Wells underestimated—though he came close—is the extent to which such a system might tend toward extractive labor practices, and how easily it can shade into a kind of intellectual totalitarianism. In order to work effectively, any such system requires continual care and feeding.

Wells was not blind to the level of work that might be involved in constructing and maintaining the World Brain. He imagined an elaborate machinery of classification, curation, and coordination, carried out by trained experts and supported by new institutional forms. But in his telling, this work manifests as a kind of enlightened bureaucracy—and he declines to ask how such efforts might be compensated in practice, or whether social power imbalances might emerge as a result.

Today, companies are paying workers to train the next generation of “intelligent” machines—folding laundry, setting tables, assembling furniture on factory production lines, or evaluating the quality of AI outputs—what researchers Mary Gray and Siddharth Suri call “ghost work.” These systems do not learn in the abstract; they are often built from painstaking, repetitive acts performed by a human labor force. Such systems do not simply spring into being; they must be continuously maintained.

Wells’s vision of the World Brain thus carries with it a darker undercurrent. He imagined that the World Brain would “hold men’s minds together in something like a common interpretation of reality,” ultimately “pull[ing] the mind of the world together.” But that vision raises an obvious question: whether a single, shared interpretation of reality is inherently desirable—and what kinds of biases might be embedded within it. Such universalist ambitions are never neutral.

More troubling still, Wells suggested that the World Brain would give rise to “a common ideology.” As appealing as such a thing might sound in principle—as long as the ideology happens to align with one’s own—it also implies a narrowing of intellectual life, with limited space for dissent or competing perspectives. In Wells’s World Brain, difference is not eliminated so much as absorbed into a single, overarching framework.

Implicit in Wells’s vision was a particular theory of authority: that knowledge could be centralized, curated, and stabilized by a relatively small group of trained experts.

Beginning with his 1905 work, A Model Utopia, Wells developed a fascination with the problem of information retrieval—the need for better methods for organizing the world’s recorded knowledge. This led him to reject old values and institutional strictures and embrace a mechanistic approach, one founded on Taylorist ideals of scientific management and a belief in the power of science to solve humanity’s problems, and the coming war in particular. Only by improving the flow of information, he reasoned, could humanity restore its collective moral, political and intellectual health.

In 1939, Wells published World Brain, a collection of essays and lectures drawn from his decades of thinking about the possibilities of new technologies for strengthening humanity’s collective intellect. He saw universal access to knowledge as more than just an intellectual boon but as a crucial step toward an uplifted society, one that “foreshadows a real intellectual unification of our race.”

But who would do the heavy lifting towards building this utopian world? His answer to the problem of coordination arrived in the figure of the “Samurai”: a self-selecting class of highly trained individuals who would guide and maintain the system. In Wells’s hands, the messy realities of knowledge work—its tedium, its conflicts, its power dynamics—are transmuted into a kind of ethical calling.

Wells’s Samurai bear a striking similarity to today’s AI artisans: highly skilled, accomplished people wielding an outsized influence over the lives of others.

“The social theorists of Utopia,” Wells writes, “did not base their schemes upon the classification of men into labour and capital. They esteemed these as accidental categories, indefinitely amenable to statesmanship, and they looked for some practical and real classification upon which to base organisation.” The Samurai were also, Wells writes, the custodians of the future: “Except for processes of decay, the forms of the human future must come also through men of this same type, and it is a primary essential to our modern idea of an abundant secular progress that these activities should be unhampered and stimulated.”

The systems we are now building present themselves as autonomous, self-improving, even self-explanatory. Yet beneath that surface lies a vast, distributed, and often invisible workforce: people who label data, train models, evaluate outputs, and moderate content. If Wells’s World Brain imagined a visible class of “Samurai” charged with maintaining the system, our own equivalents tend to remain out of sight—even as they perform many of the same functions. In this sense, Wells’s vision retains a distinctly premodern cast: for all its technocratic aspirations, the Samurai system rests on something like a feudal social order, with a small class of stewards presiding over a much larger—and largely unacknowledged—labor force.

When coherence presents itself as consensus, and consensus as truth, the line between coordination and control begins to blur. The machines may “think,” but what they produce is not a unified world brain so much as the appearance of one—assembled from the efforts of a workforce whose contributions are easy to overlook. In doing so, it carries the weight of authority while obscuring the work required to sustain it. Our present-day World Brain does not just generate answers; it demands our belief.

Notes:

• Portions of this essay appeared in my books Informatica, Cataloging the World, and in my dissertation.

• For an excellent overview of Wells’s work and the dark side of his imagination, see Boyd Rayward’s excellent critical reassessment.
  
  Other sources cited:

• Gopnik, Adam, “The War Inside H.G. Wells,” The New Yorker, November 15, 2021.

• Gray, M. L., & Suri, S. (2019). Ghost work: How to stop Silicon Valley from building a new global underclass. Houghton Mifflin Harcourt.

• Shevlin, Henry. Behaviourism’s Revenge (2026).

• Well, H.G. A Modern Utopia (1905).

• Wells, H.G. World Brain (1936).

As a ten-year-old boy visiting Edinburgh for the first time in the 1970s, my mother took me one afternoon to a quirky tourist attraction: the Camera Obscura. You climbed to the top of a five-story building and stepped into a dark room. Someone opened the skylight, and suddenly you saw what looked like a live video feed of the street life down below. Except there was no electricity involved; the whole thing ran on light and shadows.

I was so taken with the experience that when I got back to my suburban middle school in Virginia, I decided to take a run at building a replica of it for my sixth grade science fair. My hokey attempt was essentially a cardboard box with a couple of holes poked through it. It barely worked. But the thing served its purpose, and somehow I managed to walk away with a second-place ribbon (an achievement that in hindsight feels more like a testament to the prevalence of baking soda volcanoes and balloon rockets than a validation of any particular scientific acumen on my part). Afterwards, I promptly tossed it in the trash and returned to my sixth grade life.

Fast forward a few decades, and I found myself back in the camera’s orbit.

It turned out that the original camera had been created by a fellow named Patrick Geddes in the 1880s (from an original design by Aristotle). Geddes, it turns out, was a close collaborator of Paul Otlet, a man I had already been researching and writing about for several years. Otlet too had once visited the Camera Obscura, and came away sufficiently intrigued that he set out to meet its inventor.

Otlet later recalled his first visit to the Tower: “As I stood, crammed with new knowledge, and upset by such sudden change in my ordinary ways of looking at things, Geddes pulled aside a curtain. ‘In here you may rest awhile.’”

Otlet and Geddes quickly struck up a friendship. They recognized each other as kindred spirits. Both thought deeply about the interconnectedness of human knowledge, and felt that the world was facing an epistemic crisis. There was too much information being produced, and not enough effort expended on trying to synthesize it.

Geddes saw his Outlook Tower as a first step in trying to address that problem. In its original incarnation in the 1880s, the Camera Obscura was just one component in a broader project that Geddes had conceived called the Outlook Tower. The camera sat atop a series of exhibits stacked on successive floors of the building, each featuring a display intended to inform them about the city and its relationship to the wider world.

“The general principle is the synoptic one,” Geddes wrote, “of seeking as far as may be to recognise and utilise all points of view—and so to be preparing for the Encyclopaedia Civica of the future.”

On the floor below the camera was an exhibit about the history of Edinburgh; on the floor below that, the subject was Scotland, followed by Europe, and then, finally, the world. On the ground floor was a darkened chamber outfitted with a single chair: the so-called Inlook Room, in which each visitor could reflect on what he or she had just learned.

The Outlook Tower marked Geddes’s first attempt at realizing what would become an all-consuming vision: to unite the world’s disparate branches of knowledge into a unified whole. Over his long and varied career he pursued that vision through books, speeches, and close collaborations with Otlet—who, like Geddes, believed that museums had an important role to play.

Geddes felt that the best way to shape the social mind would involve a new kind of teaching museum, designed to educate members of the public about their immediate environment and its relationship to the wider world.

He envisioned an institution devoted to presenting a unified overview of the intellectual world, arranged according to a master classification scheme inspired—like Otlet’s Universal Decimal Classification—by Positivist conceptions of a rational order to the sciences.

Embracing that synthetic view of human knowledge, Geddes imagined his new museum along the lines of an encyclopedia, in which the entries would be presented not in book form but mapped on a wall and displayed “as an orderly series of labels”: a kind of 3D interface, closely tied to a particular geographical location.

To ensure that the museum connected with each visitor at a personal level, he grounded each exhibit in locally observed phenomena, by presenting “universal classes of things and facts by displaying locally generated or found exhibits of these classes.”

In this way, the Index Museum would orient each visitor to a spectrum of human knowledge that was broad but always rooted in the immediate experience of the surrounding area.

Geddes’ dreams of a transformed city stood in stark contrast to the gritty reality of Edinburgh, where a massive influx of new arrivals from the country had left the city contending with desperate poverty, high infant mortality rates, and widespread illiteracy.

From high atop his new building, he saw the city he loved becoming engulfed in cheap housing, caked in dirt, awaft in smoke, and stewing in raw sewage.

But most Edinburgh residents steered clear of neighborhoods like the infamous Leith slums, a district of “squalid lanes and closes” that few people of means had any reason to visit.

Geddes, like other civic-minded reformers, took a broader view of the problem, arguing that the lives of all Edinburgh residents were inexorably intertwined, whether they liked it or not, and that the solution to the slum problem involved taking a more integrative and constructive view of city life.

He hoped to educate the public about a reality that many of them seemed disposed to ignore, and perhaps incite them to action.

The display in the tower would play a critical role in this transformation, functioning as an active force for social change rather than a mere popular attraction.

Like Otlet, he had been strongly influenced by Auguste Comte’s ideas about sociology as a tool for effecting change in the world, and of the importance of approaching the other “preliminary” disciplines—like biology, physics, chemistry, and mathematics—in the context of their relationship to human society at large.

These subjects “are not so purely abstract or externally phenomenal as their students have mostly supposed,” but are each and all of them a development of the social process itself.

Geddes’s work throughout his career anticipates what would later come to be known as systems thinking. He was less interested in discrete bodies of knowledge than in exploring the relationships between them. Knowledge, in this view, was not simply something to be accumulated, but something to be organized, navigated, and understood as a whole.

In that sense, he can be seen as a spiritual and intellectual ancestor of later thinkers like Donella Meadows, Gregory Bateson, and even Buckminster Fuller—who likewise emphasized the importance of seeing systems in their entirety rather than solely in terms of their parts. Geddes recognized that the whole is not only the sum of its parts, but that the elusive “whole” is itself a subject worthy of study unto itself.

And what of the Camera Obscura? It survives to this day, reimagined as part of an exhibit that includes a new so-called World of Illusions. It’s now one of those Instagrammable tourist spots—and from what I can tell, a far cry from Geddes’ and Otlet’s original vision of a world museum. Nonetheless, at the heart of the museum the camera still stands, peering out over the surrounding countryside, a lonely synopticon trying to help people make sense of the messy world around them.

Note: This essay is adapted in part from my book Cataloging the World. For a general introduction to Geddes, I also recommend: Philip Boardman’s The Worlds of Patrick Geddes (1978) and Murdo Macdonald’s Patrick Geddes’s Intellectual Origins (2020).

For years I’ve been fascinated by stories of history’s also-rans—the forgotten heroes and otherwise marginalized figures whose work has been left out of the historical canon, but nonetheless left a lasting imprint on the world.

In the past I’ve written about people like Paul Otlet, Wilhelm Ostwald, and Suzanne Briet—European information scientists whose legacies survive in the long shadow of better-known Anglo-American thinkers like Vannevar Bush, Doug Engelbart, and J.C.R. Licklider.

While researching the second edition of my first book, however, I came across an intriguing American who also played a major role in shaping the field of information science. But there was more to his story than that. The methods he developed were absorbed, almost wholesale, into the emerging American intelligence apparatus that became the CIA. Yet today he remains largely forgotten: Herbert Haviland Field.

The reasons for Field’s erasure are complex and multi-layered, but one major factor almost certainly involved his physical disabilities. He had a severe stutter, and was so nearsighted he could barely see. Despite showing remarkable early intellectual promise—earning his PhD at Harvard at age 23—he recognized early on that his limitations would make it difficult for him to pursue a conventional academic career.

In 1893, he startled his family by announcing that he would abandon his chosen field of zoology altogether. He had found a new calling, one that he could pursue from the comfort and privacy of his own desk: organizing the world’s information.

The problem of too much information

By the early 1890s, scholarly publishing had expanded so rapidly that many research disciplines seemed to be coming apart at the seams. A newly industrialized press and expanding communications networks fueled an explosion of journals and monographs worldwide, while catalogs, indexes, and bibliographies failed to keep pace. Scholars were missing critical connections, and at times outright duplicating each other’s work.

Field coined a new term to describe this conundrum: “the science information problem.”

Field was scarcely the only one to recognize the problem. Across Europe and the United States, a small coterie of thinkers had begun to treat “information” as a field of study unto itself: something that could be organized deliberately rather than accumulated haphazardly. Field studied the new numeric classification system developed by Melvil Dewey, and followed the ambitious international bibliographic scheme promoted by Paul Otlet and Henri La Fontaine.

Like Otlet, Field believed the real power of indexing techniques lay beneath the surface of individual texts. A single article might touch on several subjects. A well-constructed index could make those connections visible and usable, enabling scholars to make multi-dimensional connections between sources and findings.

International intrigues

But Field was more than just a bibliographer. He was also a student of government, and the emerging bureaucracies of information management that were becoming essential to the functioning of the modern nation-state.

Early in his career he had traveled to Russia to study the paper filing systems used by its vast bureaucracy. He examined card catalogs under development at the United States Department of Agriculture. In Paris, he consulted with Alphonse Bertillon, whose card-based system for managing criminal records demonstrated how classification could function across large populations.

Taken together, these encounters convinced Field that the future of knowledge lay less in books and journals than in the systems people used to extract information from them. He was a fervent believer in the hidden powers of index cards—the flat-file database technology of the day. They were standardized and portable, could be copied, sorted, and refiled as needed—unlocking all kinds of ontological possibilities.

With modest backing from two zoological societies and more substantial support from his father, Field began to imagine a new kind of bibliographic enterprise. Scholars would subscribe to a steady stream of index cards, mailed out every two weeks, each card describing a newly published piece of research. Subscribers would file the cards locally, integrating them into their own collections. Over time, these cabinets would form a loose but coordinated arrangement. Knowledge would be dispersed across many locations, but remain broadly aligned.

This point mattered to Field. Unlike Otlet, who envisioned a monumental central archive, Field believed no single institution could ever hope to keep pace with modern science. Understanding depended instead on coordination among many partial collections. Each subscriber would possess only a fragment of the whole picture, but together those fragments could add up to a unified whole.

Seeking international legitimacy, Field relocated to Paris and gave his enterprise a new name: the Concilium Bibliographicum. He lived alone and worked obsessively, racing to establish credibility before better-funded competitors like the Royal Society in London could claim the high ground.

Field hired professional catalogers and invested heavily in infrastructure. As his subscription business grew, he purchased printing presses, employed typesetters, and installed an automated card-cutting machine. By 1898, the Concilium was producing more than 1.6 million cards every year. He built an imposing headquarters in Paris, an information palace meant to convey permanence and authority.

For a brief period, the project thrived. Major institutions like the Smithsonian, Harvard, Yale, Princeton, the University of Michigan, and the Library of Congress all subscribed to the service. But the scope of the problem continued to expand. By the turn of the century, the sheer velocity of published information was continuing to spiral out of control. The zoological literature alone demanded tens of thousands of new cards annually. Field’s production costs rose quickly. Revenue did not.

By 1908, subscription growth had stalled. Appeals for institutional funding failed. A large personal loan only deepened Field’s losses. Exhausted and increasingly unstable, he laid off his remaining staff and relinquished control of the Concilium. The institution survived under new leadership but never recovered its momentum.

Most histories of Field’s work end here. He appears as a tragic figure, a gifted intellectual unable to bring his ideas to fruition. But the man went on to an intriguing—and even less known—second act.

Entering the labyrinth

When World War I broke out, the idealistic Field began working with a Quaker relief agency, ferrying food and medicine between countries and soaking up all kinds of information about politics and daily life in the combatant countries.

It was during this time that he made the acquaintance of a young State Department apparatchik named Allen Dulles—the spymaster who would go on to become the driving force in creating the modern CIA.

Dulles quickly came to see Field as an unusually effective intelligence asset. Field approached political instability, technological developments, and diplomatic signaling much as he had once approached the scientific literature: as a mass of heterogeneous data points to be standardized, cross-referenced, and selectively circulated.

He supplied intelligence on civil war and famine in Bavaria, Germany’s emerging gas-warfare capabilities, and Austria’s openness to peace negotiations—material that required not just access, but synthesis.He fed Dulles intelligence on the civil war and famine in Bavaria, Germany’s emerging gas warfare technologies, and critical information on Austria’s receptivity to making peace with the Allies.

Field made such an impression with the State Department that he was eventually invited to work on the Treaty of Versailles, and help plan out the formation of the League of Nations (a project that brought him back into the orbit of his intellectual hero Paul Otlet). His diligent efforts to shape the post-war political landscape raised his visibility with the U.S. government, eventually bringing him into direct contact with Woodrow Wilson—who he recognized as a kindred spirit in their shared commitment to the emerging ideal of internationalism.

The same techniques Field had developed in building the Concilium—abstraction, cross-referencing, selective circulation—proved invaluable in shaping his intelligence work. He had mastered the ability to compare and connect discrete fragments of knowledge, detect hidden patterns, and make that information usable.

As Colin Burke observes, such techniques were central to the formation of the modern bureaucratic state. Index cards, filing systems, documentation standards, and clerical labor formed the hidden machinery of governance. Intelligence services did not invent these tools, but they perfected them. And Field played a central—and largely unheralded role—in developing these core mechanisms that would enable the modern intelligence agency to thrive in the decades to come.

World War I delivered the final blow to the Concilium, which had already been struggling to survive the accelerating scale of scientific publishing. Postwar efforts to revive the enterprise failed, and Field himself died in 1919, in the wake of the influenza pandemic that swept Europe at the end of the war.

As historian Colin Burke observes, these techniques—indexing, abstraction, cross-referencing, controlled circulation—were central to the rise of the modern bureaucratic state. Intelligence agencies did not invent this machinery, but they refined it, weaponized it, and scaled it.

Field’s legacy continued through the astonishingly varied career of his son, Noel—whose life would intersect with diplomacy, Soviet intelligence, and the emerging American intelligence apparatus (thanks in part to the family’s close connection to Dulles). Burke uses Noel’s trajectory to trace a broader continuity between early systems of knowledge organization and the bureaucratic, intelligence-driven state that emerged in the twentieth century.

Long before anyone imagined the possibility of intelligence as “artificial,” Field was working towards a grand vision of federating knowledge and, one might argue, a generative system. Field did not solve the problem of too much information—no one has—but he recognized its strategic value, and left a lasting legacy that is deeply entrenched but largely hidden in the shadows of the modern bureaucratic state.

Further reading

• Colin Burke’s Information and Intrigue is the definitive biography of Field, and a great read at that.

• This essay is adapted in part from the second edition of my first book, Informatica.

In 1951, a little-known French librarian named Suzanne Briet posed a simple question that still bedevils us today: When does anything become a fact?

To illustrate her point, she offered a playful example. Imagine a rare species of antelope, roaming free somewhere on the African savannah. It is wild, unrecorded, and not yet cataloged in any natural history museum. For the moment, it is sui generis—a thing unto itself.

Now imagine that someone captures that antelope, and brings it to the Paris zoo. The creature is placed in a cage, photographed, measured, and described in detail. A radio broadcaster announces its arrival. A journalist files a newspaper story. A researcher writes a paper for a scientific journal.

Eventually, the poor thing dies. Its body is stuffed, labeled, and placed on permanent display in a museum. But its legacy lives on, as the data it furnished finds its way into catalogs, encyclopediae, and various archival instruments. What remains of the antelope is now more than a carcass. Its essence—what Buffon called its moule intérieur—has become a node in the great chain of human knowledge.

The antelope has become, in other words, a document.

A what, one might well ask? If an antelope can be a “document,” then what does that term even mean? Can anything—or everything—become a document?

No, not quite.

A star in the sky is not a document. A pebble resting in a riverbed is not a document. A wild animal in its native habitat is not a document. But the moment we inscribe the star on a photographic plate, display the stone in a museum, or record the antelope’s cry, the thing becomes a form of evidence—a stabilized set of facts.

The woman who crystallized this idea, Suzanne Briet, is not widely remembered today. But long before the age of digital archives and large language models, she recognized that knowledge was not simply a matter of storage, but rather a dynamic process of collecting and interpreting information—a living process of generating new meaning from a constant stream of novel inputs.

Mademoiselle Briet, Femme Formidable

To understand how strange and prescient this vision really was, picture Paris in 1924—when a 30-year-old Briet arrived for her first day of work at the Bibliothèque Nationale.

The country was still regrouping from the horrors of the First World War. These were les années folles (“the crazy years”), a time of art, jazz, nightlife, literary innovation, and a sense of hopeful modernity emerging from the trauma of war. It was the Paris later immortalized by the likes of Ernest Hemingway, Gertrude Stein, and Ezra Pound.

Inside the stone corridors of the grand Beaux-Arts-era library, electricity had only just been installed in the reading rooms, finally allowing scholars to work past three in the afternoon on darkening winter days. When the green lamps flickered on, she later wrote, it was like watching the flowers bloom.

At the time, the French library system was an almost exclusively male preserve—a relic of the nineteenth century—buttoned-up, hierarchical, and deeply resistant to change. Briet was one of only three women on the entire staff. Some male colleagues objected to her presence, worrying aloud that if women were allowed into positions of authority, they would lack sufficient gravitas to supervise the men. Briet, who graduated first in her class at the Sorbonne, soon put those concerns to rest.

Over the years that followed, she established herself as one of the brightest intellectual lights in the library’s illustrious history. She was also a kind of bibliographical rebel with avowedly modern ideas: a Gertrude Stein of the stacks.

Briet believed that libraries could be something more than just dry archival repositories. They could also, she believed, become living learning laboratories. She also felt strongly that librarians should be more than just passive stewards of books and periodicals, but play a more active, engaged role as “friends of the user.” She created training programs, founded a national institute for documentation, and helped professionalize a new kind of librarian: one concerned not only with the preservation of knowledge but with engaging directly with scholars and researchers to create two-way line of communication to help the collection evolve and adapt to changing conditions over time.

Though largely forgotten today outside the specialized world of information studies, Briet left a lasting imprint on the practice of librarianship in Europe and beyond. Her ideas also presaged a broader cultural shift that would take shape decades later, as the digital revolution ushered in a period of fragmented institutional authority and a shift of power towards individuals—what today we might call user-centricity.

Where many of her predecessors—like Paul Otlet—imagined universal classification systems that could unify all human knowledge under one global system, Briet pursued a far different, more pluralistic vision. Knowledge, she argued, is not organized from on high. It is sorted closer to the ground, in the messy, contingent world of human communities—each with its own metaphors, priorities, and categories. To force them all into a single system would be to flatten culture itself. Instead, she advocated for interoperable but distinct systems. Bridges, not monuments.

Briet’s vision anticipated much of the networked, bottom-up information ecosystem that would take shape in the latter half of the twentieth century, culminating with the Internet and today’s highly adaptive AI systems. And it all started with that imaginary antelope.

But this was not the first time an even-toed ungulate was pressed into service to support an argument about the shape of human knowledge. It had happened before, more than a hundred years earlier—in Paris.

Mr. Jefferson’s Moose

In 1787, Thomas Jefferson took delivery in his Paris hotel of the complete skeleton, skin, and antlers of a seven-foot-tall American moose. The animal had crossed the Atlantic packed in salt, then rattled overland from Le Havre in a horse-drawn wagon. By the time it arrived, it was in rough shape: its skin sagged, its hair fell out. But it would serve its purpose well enough. Jefferson arranged to have the carcass displayed at the entrance to his hotel. Parisians flocked to see the spectacle. Jefferson hoped one Parisian in particular would take notice: Georges-Louis Leclerc, Comte de Buffon.

With this curious stunt, Jefferson entered a heated scientific debate over the nature of the New World. Buffon, Europe’s most celebrated natural historian, had argued that American animals—and by implication American civilization—were weaker, smaller, and degenerate compared to their Old World counterparts. The claim was not merely academic. If the New World produced inferior life, what European would willingly emigrate there? Buffon’s theory cut to the young republic’s legitimacy.

Jefferson, an obsessive naturalist, took the charge personally. He cataloged American plants and animals, filled Monticello with skulls and specimens, and tracked decades of environmental data. In his Notes on the State of Virginia, he set out to refute Buffon point by point. The moose was his trump card: massive, unmistakable, impossible to ignore. This was no degenerate creature.

Jefferson’s argument was not simply that Buffon was wrong. It was that knowledge itself required shared structures. Without agreed-upon names and categories, he warned, science would collapse into confusion. Classification systems were not metaphysical truths, but social compacts—tools that allowed people to think together. The danger of endlessly revising them, Jefferson feared, was not error but unintelligibility: scholars becoming “unintelligible to one another,” producing schism rather than progress.

The moose did its work. Buffon promised revisions, though he died before completing them. The episode is usually remembered as a footnote in Jefferson’s biography. But it deserves a second look—not as a natural history curiosity, but as a lesson in how evidence, authority, and infrastructure intertwine. Jefferson did not merely describe an animal. He staged, in effect, a document.

Seen through Briet’s lens, Jefferson’s moose looks less like a biological specimen than a kind of documentary performance. The animal mattered not merely because it existed, but because it could be pointed out, displayed, circulated, and folded into a larger argument.

And this is where the antelope and the moose begin to converge.

Briet showed how objects become documents once they enter institutional systems of meaning-making. Jefferson showed how an object could be mobilized to stabilize truth across distance and disagreement. In both cases, knowledge emerged not from raw phenomena alone, but from a chain of mediated knowledge generation: capture, classification, circulation, and consent.

Which brings us to the present moment.

We now live in an online universe of theoretically infinite documents, many generated on the fly through predictive modeling and untethered from any physical artifact in the observable world. Large language models ingest vast quantities of recorded knowledge and recombine them into new, endlessly fluid texts—documents descended from other documents, chains of derivation with no grounding antelope or moose in sight.

Briet’s idea of “documentary fertility” has gone into overdrive. In a world where everything is derived, where do we find our points of reference? When classification happens invisibly inside models trained on flattened corpora, who decides what counts as knowledge?

The problem is not information overload per se, but the difficulty of stabilizing meaning long enough for facts to take shape. Without shared reference points, interpretation begins to float free. Then it’s parrots all the way down.

Somewhere in the vast distributed datasets of today’s AI systems, Briet’s antelope still roams. So too does Jefferson’s moose, antlers raised, insisting that evidence must eventually make contact with the physical world.

Knowledge may always be provisional. But without moments of stabilization—without documents that can briefly hold their place in the world—it dissolves into noise.

In a world of too much information, the moose and the antelope of our collective imagination risk becoming endangered species.

Further reading:

• Lee Alan Dugatkin, Mr. Jefferson and the Giant Moose (2009).

• Mary Niles Maack. “The Lady and the Antelope: Suzanne Briet’s Contribution to the French Documentation Movement” (2004).

• Portions of this essay also draw on material from my book Informatica: Mastering Information Through the Ages (2022, originally published as Glut in 2007).

It was about this time last year that I first came across this story while doing research for my book about the history of early American newspapers at the American Antiquarian Society (AAS). It comes from one of the rare books in their collection, Ups and Downs of a Country Editor—Mostly Downs, written by a fellow named S.A. Fackler.

You won’t find this book online; as far as I know, the AAS has one of the only surviving copies, but it’s a wonderful little first-hand account of a life lived at the margins of the nineteenth-century newspaper trade. The book is brimming with stories about his misadventures as a tramp printer—moving from town to town, starting up newspapers, drinking too much, occasionally getting into fights, and generally leading a full and ribald life.

But there was one passage at the end of the book that has really stuck with me. The anecdote didn’t quite fit with my book, but I’ve been wanting to do something with it ever since—and, well, this seemed like the right time of year.

It was Christmas Eve, and Fackler had fallen on hard times. His wife had left him, and he was left to wander the Alabama countryside in search of work with his two little boys Bob and Willie. They were penniless, with nothing to eat and nowhere to sleep. Fackler had faced hard times before, but nothing like this.

In better times, Fackler had made sure Santa always brought his boys something for Christmas: new clothes, toys, confections, and so forth. But this year was going to be different, and he knew he would have to explain it to the boys in advance.

“Boys, this is one Christmas Eve night that Santa Claus will not visit you.”

When his younger boy Bob asked why, Fackler had to fudge it, explaining that because they were traveling Santa would have trouble finding them. Moreover, in years past he would leave money for Santa Claus to buy them gifts. But this year he was in no position to reimburse the jolly old elf.

“You see, papa has had no work in quite a while, so he has spent all of his money.” But he assured them that soon as he found a paying job, he would make up for it and buy the boys all kinds of nice things.

Little Bob was having none of it.

“Papa, it didn’t make any difference where we are at. You said that Santa Claus couldn’t find us way up here. Well, can’t God tell him where we are and don’t God tell him to visit poor children that can’t pay him?”

Fackler was stumped. But this wasn’t his first time living close to the bone. He was a lifelong tramp printer and roving country editor. Men like him were the shock absorbers of the nineteenth-century newspaper trade—moving from town to town, job to job, keeping presses running wherever they could, often spending their wages at the nearest saloon, and rarely staying long enough to build a safety net. When the work dried up, it was time to move on.

But most tramps were bachelors. Until recently, Fackler had abandoned that life and managed to build a family, with a wife and two boys and a house and all that entailed. Now the wife was gone, the house was sold, and he was dead broke again—walking down the road with his two little boys in hopes of finding somewhere to stay the night.

Bob continued his lecture.

“Yes, sir, Willie and I pray every night, and God is going to tell Santa Claus where we are.”

Fackler didn’t have the heart to tell him otherwise. So the little troop continued trudging along, making their way across the dirt roads until they came to a little settlement about eight miles outside of the town of Eufaula.

There, a man stepped out into the road to greet them. He was a farmer, and invited them into the farmhouse to rest for a spell. There they met the farmer’s wife (alas their names have been lost to posterity).

“They were good Christian people,” Fackler later recalled. The man and his wife offered them supper and invited them to stay for the nigh. The farmer’s wife even took the boys’ socks from them, washed them, and darned the holes that had long since worn through the feet.

That night, before bedtime, the boys took their newly repaired stockings and quietly hung them up in a nearby room, hoping against hope that Santa might yet find his way to them. All three of them went to sleep, bone tired but grateful to be sleeping in a bed for a change.

The next morning, Fackler woke up to the sounds of two shrieking little boys.

“Papa, oh, Papa!” Bob cried out. “I told you that old Santa Claus would find us, and God would tell him where we would be.”

Sure enough, Santa had found his way to the boys after all. Their stockings were full.

“Now, papa,” said little Bob, “didn’t I tell you yesterday that God would tell Santa Claus where to find us?”

Fackler nodded, duly chastened. “Well, boys,” he said, “I am glad he found you, for you surely had more faith than I had.”

That same day the farmer brought Fackler into town, where he managed to find work—on Christmas Day, no less—and earned a little bit of money, with which he promptly bought fireworks and a few pieces of candy for the boys. “So they had a jolly Christmas after all.”

Postscript

You may be wondering what became of poor Mr. Fackler. Well, I’m happy to report that he kept working that job in Eufaula for the next three months, saving up a bit of money, and managing to swear off the bottle for a time. There was an alcohol dispensary in town, and he claims never to have visited it once during his time there.

Fackler eventually moved on to Florida with his boys, where he found the time and energy to finish writing his memoir. He continued to have his ups and downs—and, yes, eventually found his way back to the bottle. But he had found a way forward, and the kindness of the farmer couple stayed with him always. Eventually he published his book, found steady work again, and even created a theatrical version of the book, leading a troop of players around Georgia to re-enact some of his memorable adventures.

But like most tramp printers, Fackler has been long forgotten. I offer this little vignette in hopes that it may yet find some of the readers he once hoped to reach.❦

We’ve all seen them: the weird cartoon explainers, cringe-inducing history videos narrated by synthetic voices, and recycled Wikipedia paragraphs swamping our social media feeds over the past year or so. Huge swaths of the Internet are now fully awash in AI slop. Perhaps Cory Doctorow is right that we’re all at risk of turning into reverse centaurs.

But this isn’t the first time people have had to contend with a flood of dubiously sourced drivel masquerading as news. In fact, editorial “slop” has been a proven business model for nearly two centuries now.

In the summer of 1835, The New York Sun—then a scrappy upstart penny paper—published a six-part series claiming that astronomer John Herschel had discovered life on the moon. He claimed to have used a super-powered telescope to spot bat-winged humanoids, sapphire forests, and unicorns cavorting across the lunar surface.

It was, of course, a bald-faced fiction. But the public ate it up. Soon, the story was popping up in newspapers all over the country. Indeed, The Sun’s editor Benjamin Day had conceived the whole thing as a prank, hoping to prove just how gullible his fellow newspaper editors could be in the face of a story guaranteed to boost circulation. Most of them reprinted the story without attribution, claiming it as their own.

Day and his fellow penny-press purveyors—like James Gordon Bennett of The New York Herald—essentially invented the modern attention economy. Their papers trafficked in scandal and sensation (and often outright falsehoods) that were widely reprinted across the loosely coordinated exchange networks that connected the nation’s newspapers in the nineteenth century. And their business model proved wildly profitable—for a while.

The Victorian heyday of newspaper proto-slop included tales of “science wonders”— fantastical inventions like perpetual motion machines, lightning-powered engines, and electrical devices that could cure paralysis or control the weather—and earnest-sounding stories that breathed life back into ancient myths: gentle giants, wild men, magical rocks, or sea monsters roaming up and down the eastern seaboard.

But it wasn’t the fictive nature of these stories that made them “slop” in the contemporary sense. Back then—in an age before modern journalistic standards had taken shape—the lines between fact and fiction were often blurred in the pages of newspapers. And a good deal of the material that recirculated through the nineteenth century media ecosystem was more-or-less true: recipes, trivia, bits of homespun wisdom, and any number of other “squibs” that regularly made the rounds across the exchange networks.

For many publishers, their overriding priority was simply to fill up space. And to do so they gobbled up as much of this kind of “evergreen” material as possible, passed from paper to paper, lightly rewritten, and then reprinted again whenever an editor needed to pad out a few column inches to round out the week’s edition.

The material was, for the most part, harmless. Readers enjoyed the stuff, editors depended on it, and no one really suffered. The “hard news” that appeared elsewhere in the paper remained mostly intact.

These stories ranged wildly in terms of quality, subject matter, and length. What bound it all together was its uncertain provenance. As Ryan Cordell and others have argued, this material functioned, in effect, as a kind of analog large language model: aggregating content from a number of different sources, synthesizing it, reformatting it, and recirculating it in new permutations. Nineteenth century newspapers were, as he puts it, “the largest text generation platform in human history.”

Like today’s AI explosion, the nineteenth century slop boom owed to several causes, only some of them technological. Literacy rates were rising, copyright protections were loose, and production costs were declining. As a result, thousands of newspapers could make hay recycling each other’s content, making slight modifications along the way, and passing it off as their own. The economics of the market rewarded speed, efficiency, and reuse of already-proven, readily available editorial content.

Nineteenth century newspapers perfected the art of turning editorial scraps into news-shaped noise. Today’s AI slop is simply the latest manifestation of this early-industrial attention economy.

What can we learn from the slop of yesteryear?

First and foremost: everyone likes a good story, whatever the source. Non-journalists often care more about whether something is engaging and entertaining than whether it is meticulously sourced (especially when it comes to lighter fare). But the history of the penny press and the exchange networks also remind us that the economics of cheap content production don’t scale indefinitely.

The rise of Victorian slop wasn’t a moral failure on the part of editors; it was a rational response to a market that, for a time, created economic incentives that rewarded this kind of cheap recycled content. But eventually that logic collapsed under its own weight—forcing the press to undergo a painful period of self-examination and reinvention (a much bigger topic that I’ll save for another post).

Similar dynamics apply today. As Shuwei Fang argues, we’ve entered an era of “liquid content,” where stories are endlessly remixable, trivially reproducible, and increasingly detached from the notion of a stable, individually authored artifact. Once the marginal cost of production drops to near zero, however, the value of the individual artifact plunges.

AI slop won’t disappear anytime soon (however much the social platforms claim to be making efforts to suppress it). It’s the predictable result of a system optimized for attention, throughput, and monetization. But that logic will eventually falter. If the nineteenth century offers us any comfort, it’s that slop eventually eats its own tail.

Just as The Sun eventually drowned amid a host of cheap imitators—paving the way for new, higher-quality newspapers to emerge later in the century—today’s deluge of AI-generated detritus may yet yield to new norms and standards for transparency, verification, and journalistic craft.

To paraphrase Fang’s argument, the future won’t belong to those who produce “buckets”—volumes of cheap content—but rather to those who build pipelines—the systems, structures, and practices that deliver meaningful journalistic value rather than adding to the noise.

Personally, I remain hopeful that today’s vast reservoir of slop will become the petri dish from which new forms of meaning-making will emerge. But for now, that’s a story still waiting to be told.

Instead, what he found came as something of a shock: a nation whose heads were buried in their newspapers.

Dickens, a former journalist himself, was taken aback by the country’s obsession with the news. He soon grew appalled at what he saw, condemning the American press as “a monster of depravity,” guilty of “pimping and pandering for all degrees of vicious taste, and gorging with coined lies the most voracious maw.”

But what unsettled him most wasn’t the press’s appetite for scandal, or even the loose copyright laws (which allowed them to republish his novels without compensation). It was the sheer volume of it all—the relentless torrent of cheap print, and the sense that a society awash in news was losing its capacity for deeper reflection.

The dynamic Dickens observed never really went away; it simply migrated to our smartphones. In 2020, journalist Karen Ho popularized the term doomscrolling to describe the early-pandemic habit of compulsive news-reading. Since then, the word has escaped its original context. Today it applies just as readily to cat videos, real-estate listings, explainer clips, and all the digital debris drifting across our screens. The content matters less than the compulsion. The doom is in the scrolling.

Doomscrolling may seem like a uniquely digital-age affliction, but its roots stretch back to the industrial information explosion of the nineteenth century. Long before smartphones and K-Pop demon hunters, our newspaper-clutching forebears were struggling with the problem of too much media and too little time.

As Henry David Thoreau observed in Walden (1854): “Hardly a man takes a half-hour’s nap after dinner, but when he wakes he holds up his head and asks, ‘What’s the news?’ … After a night’s sleep the news has become as indispensable as the breakfast.”

Complaints about information overload were nothing new, even then. Plato famously worried that writing would cripple human memory; and seventeenth-century English clergyman Thomas Fuller lamented that “so much is printed, that little is read.” But the problem took on a new scale in the nineteenth century, when an unprecedented volume of printed material began to swamp the public sphere: newspapers, cheap books, serialized fiction, broadsides, pamphlets, political tracts, religious literature, magazines, almanacs, and an ever-growing pile of bureaucratic paperwork.

To characterize nineteenth-century readers as facing “information overload” is, of course, a retrospective argument. Victorians would never have used that term; and yes, these anxieties were mostly confined to the world of literate white people. But nonetheless, plenty of readers and writers of the era reported feeling bewildered, overwhelmed, and deeply worried about what they were missing. In other words: a pervasive sense of doom, minus the scrolling.

The nineteenth-century information explosion emerged from a tangle of social, cultural, and technological changes. Industrial innovations like steam presses, Linotypes, and cheap wood-pulp paper played their part, but the real transformation occurred at a deeper, systemic level: a rapidly urbanizing United States knitting itself together into an increasingly networked, interconnected public sphere.

By 1900, the U.S. was producing half the world’s newspapers—despite holding only five percent of the global population. And the age of mass media had only begun.

The View from Nowhere

Paradoxically, the cacophony of mid-nineteenth-century media set the stage for a long period of consolidation and corporate retrenchment. By the 1880s, industrial-scale publishers like Joseph Pulitzer and William Randolph Hearst were stitching together sprawling new media empires—the Meta and Alphabets of the day.

Out of this consolidation emerged a new journalistic ideal—one that would supplant the chaotic caterwaul of the mid-century press and come to define the twentieth century: the sober, neutral, just-the-facts mode of reporting sometimes called “the view from nowhere.”

Through much of the twentieth century, a relatively small number of gatekeepers—media conglomerates, wire services, broadcast networks, and syndication outfits—filtered the world’s events into a nightly broadcast or a morning front page. What it all lacked in personality it made up for—more or less—in precision, consistency, and a broadly shared perspective on what was happening in the world, what historian James Carey has called the “iron core” of midcentury journalism.

Then the Internet shattered that consensus.

We all know what happened next. The age of mass media gave way to a hyper-networked, bottom-up era of blogs, message boards, podcasts, influencer feeds, micro-publications, hyper-partisan outlets, and algorithmic streams. The brute-force economics of the penny press—cheap, fast, sensationalist content, turbo-charged by advertising revenue—came roaring back on a global scale.

To be clear: This is not to say we are living through the nineteenth century all over again. The scale and reach of today’s global network dwarfs anything that came before. But the mass psychology of the moment feels eerily familiar: a frenetic, fragmented media ecosystem where many of us feel inundated with too much information delivered in too little time.

It’s a well-worn pattern by now: media systems scale faster than their audiences can adapt; people feel overwhelmed, even as they consume more content than ever; critics fret about a supposed existential threat to society; reformers dream up new systems to restore epistemic order; and then the cycle starts all over again.

After the Flood

Doomscrolling, seen in this light, is nothing new. Nor is it a moral failure. It’s an adaptation to a kind of built environment. And the problem has been with us for a very long time. Just as the Victorians built publishing empires and other institutional gatekeepers to manage the flood of paper, we are building algorithmic interventions to manage the digital deluge.

Today, we can see a new sort of gatekeeper emerging: Large language models (and their corporate proprietors) are trying to aggregate, synthesize, and repurpose the world’s intellectual output. Like the Hearsts and Pulitzers of old, they ingest vast quantities of data to deliver a compressed, seemingly coherent version of events. Whether they succeed or not (and the jury is very much out), they represent the latest chapter in a centuries-long story: our struggle to build systems that help us make sense of the world when our senses are overwhelmed with too much data.

Let’s be careful of stretching the parallel too far. Large language models are not editors in any human sense. They operate by statistical patterning, not by human judgment. Yet for many users, they effectively play a similar role: as a kind of gatekeeping mechanism for compressing an overwhelming volume of information into something that feels coherent and consumable.

Will today’s AI powerhouses become the digital equivalent of the Gilded Age media baronies? Will they produce the first drafts of a new kind of history? Time will tell. Our systems grow more complex, but the underlying challenge remains the same: How can we leverage new tools to harness our collective knowledge stores, without losing our innately human capacity for reflection and making meaning?

Doomscrolling isn’t a failure of human willpower; it’s a coping mechanism, and the predictable outcome of information systems optimized for speed, volume, and monetization. On the Boston docks, Dickens foresaw the problem that would come to dominate the modern world.

What ties that period to our own is not the news per se, and certainly not the technology—but the recurring tension between the growth of media systems and the cognitive limits of our human sense-making capacity. Each era seems to see new solutions emerge to bridge that gap, which in turn create new problems and new opportunities. As (the admittedly over-quoted) Marshall McLuhan put it, “We shape our tools and thereafter our tools shape us.”

The question is not whether we can build systems powerful enough to harness the flood—we already have. The real question is whether those systems can help us build a more coherent understanding of reality, or accelerate its fragmentation. The future of our media ecosystem—and our ability to make sense of the world around us—depends on what we do next.

Open the right sequence of panels, Camillo promised, and the wisdom of the ancients would be revealed: the teachings of Solomon, the movements of celestial bodies, the seven virtues and vices, and more. It was, he claimed, “a constructed mind and soul.”

Like so many promising new inventions, the theater sparked a wave of public excitement, breathless commentary among the literati, and even attracted speculative investments—before failing to deliver on its promise, and bankrupting its inventor.

Today, we live in our own age of memory palaces: AI systems promising instant access to humanity’s collective wisdom and insight. Artificial General Intelligence (AGI), we are told, is right around the corner. And yet, as with Camillo’s theater, much depends on a kind of mnemonic sleight-of-hand.

In truth, Camillo’s theater was a simplified version of a much older tradition: the ars memoriae (“art of memory”). For centuries, European monks trained in this demanding practice, visualizing elaborate palaces filled with symbolic imagery. One practitioner boasted that he could recall every known fact across an astonishing range of topics: theology, astronomy, metaphysics, law, arithmetic, music, geometry, logic, and grammar. Another claimed to have memorized two hundred classical speeches, three hundred philosophical sayings, twenty thousand points of legal doctrine, and more. It took years of training—and austere mental discipline—to master the technique.

Camillo, a former monk, promised the public a shortcut: access to a vast storehouse of recorded knowledge, without all the troublesome mind training. Anyone who entered his theater, he boasted, would emerge newly possessed of the wisdom of Cicero.

When he unveiled a prototype version of the theater in Venice, the public raved. Spectators lined up to pay and enter. Word of the spectacle even reached the King of France, who took a personal interest in the project and promised to fund its continued development to the tune of five hundred ducats (a princely sum —the Renaissance equivalent of startup capital).

And then, as with so many disruptive technologies, the bubble burst. The project went over budget; production delays added up; the project languished. Camillo found himself plunged into debt. He died just a few years later, with his theater still unfinished—but nonetheless convinced that he had seen a glimpse of the future.

In a sense, he had.

Five centuries later, that dream has migrated to the cloud. Today’s large language models (LLMs) compress vast corpora of text—humanity’s collective intelligence—via natural language processing (NLP) and neural networks. They do more than just record facts; they create a semantic architecture of recall and retrieval. They are, to borrow Camillo’s phrase, “a constructed mind.”

LLMs are often described as if they possess knowledge, but what they really contain is structure—a vast lattice of associative trails. They organize words, images, and concepts not into imaginary corridors and arches, but across thousands of mathematical dimensions—latent spaces where proximity signals likeness, and the traversal produces the illusion of thought. In this sense, they bear at least a surface resemblance to the memory palaces of old.

Yet that resemblance conceals a fundamental inversion. The art of memory depended on an act of embodied imagination: to remember was to re-enact knowledge in a visualized space, to locate it within one’s own moral and sensory world. An LLM’s “memory,” by contrast, is disembodied and mechanistic. It remembers nothing. It merely reproduces statistical associations. It runs on ghosts in the machine.

The astonishing rise of AI has prompted endless waves of hype, as well as occasional bouts of fervid hand-wringing. Nicholas Carr warns that endless streams of connected data not only change what we know, but how we know, as we all become “engaged in the production of a facsimile of the world.” In a similar vein, Paul Kingsnorth suggests these systems do more than extend our minds; they reorganize our inner lives. “The Machine we are building is taking us at warp speed into a new way of being human,” he writes. “Away from the spiritual, towards the material; away from nature, towards technology; away from organic culture towards a planned technocracy.”

One can easily imagine sixteenth-century monastics lobbing similar critiques at Camillo’s degraded version of their sacred discipline.

For the medieval scholar, the art of memory amounted to a deeply spiritual practice. To remember was to practice discernment—to decide what was worth remembering and what could safely be forgotten. The memory palace was never meant to grant access to universal knowledge, but rather to act as a kind of mnemonic scaffolding to help practitioners cultivate their own inner wisdom.

Where the medieval monk sought inner moral virtue through recollection, the LLM optimizes for scale. It has no moral compass, but it has a facsimile of one in the form of embedded prompts geared towards implementing corporate policies, regulatory guidelines, and ethical guardrails. The choice of what gets remembered and what gets forgotten is no longer a personal decision, in other words; the machine’s performative morality is a distributed function, a disembodied specter of rules floating in the cloud.

The architecture of a medieval memory palace expressed a certain, highly individualized worldview — a way of distinguishing what mattered from what could be forgotten. But LLMs know no such restraint. They proceed from the premise that nothing should be forgotten, and that wisdom will arise not from individual contemplation but through an emergent, machine-augmented process.

But access to machines with perfect recall can foster a kind of collective amnesia. Models trained on vast bodies of data can threaten to blur the boundary between signal and noise, collapsing the distinctions—moral, emotional, mnemonic—that give human memories their purpose. Picture a Borgesian palace whose rooms keep dissolving into one another, their corridors looping endlessly through recursive chambers that echo without end.

The result is a paradox: a civilization that has forgotten the art of forgetting. We have built engines of total remembrance, yet our capacity for inner reflection and discernment risks getting hollowed out. Camillo’s theater promised wisdom through an externalized memory aid; our neural palaces promise something similar, but risk drowning out our inner wisdom in gestures of recall without the deeper, meaning-making faculties of comprehension and contemplation.

And yet, in their emptiness, these neural cloud palaces achieve a kind of tragic magnificence. They embody the ancient dream of universal knowledge. Like Borges’s Funes, they recall everything and understand almost nothing. Each prompt we type is a footstep through corridors that no longer require us to walk, corridors that extend endlessly through the cloud—infinitely precise, infinitely forgetful.

For centuries, the art of memory represented a deeply human endeavor: a fusion of imagination, spatial reasoning, and moral philosophy. Now AI performs mnemonic labor once reserved for monks, poets, and scholars. What we gain in speed and scale, we risk losing in attention and interpretation — trading embodied remembrance for probabilistic recall.

All that said, LLMs surely have their uses. They can help us sift through vast research collections, identify hidden patterns, and shed new light on the way knowledge accretes over time. They can provide all kinds of useful scaffolding—if we use them with intention.

Nowhere is it written that technology must flatten our inner lives. Like many powerful new inventions, these systems carry a double-edged promise: they can free us for deeper thought, or tempt us into the purgatory of passive retrieval.

And what of Giulio Camillo? For all his clever feats of mnemonic carpentry and bold proclamations, the theater never delivered the promised results. No one ever attained the wisdom of Cicero, except for Cicero himself.

Today, we are building the palaces of Camillo’s dreams — only exponentially larger, faster, and distributed everywhere all at once. The question, as ever, is not so much what the machine can do for us, but what we choose to do with it—and why.

p.s. If you’d like to learn more about the legacy of the ars memoria, allow me to recommend Frances Yates’s classic 1966 book, The Art of Memory.

To paraphrase Bill McKibben, we live in an age of too much information and not enough meaning—and, almost certainly, too many newsletters. So why start another one, and why now? After wrestling with that question for a while, I’ve come to the conclusion that it’s no use raging against the machine. The machine is us.

For a good chunk of my career, I’ve been part of that machine—working in Big Tech, media, and startups (with the odd detour into academia). But recently, I decided to step back and take a page from Benjamin Franklin’s book: to become, as he put it, “a master of my own time.” I’m surely no Franklin, but I aspire to follow his example—to read, write, and tinker with new ideas, in hopes of producing work that might, in some small way, serve the common good.

Over the past several years, I’ve been exploring what I think of as the “deep history” of the digital age—researching some of the lesser-known people, inventions, and cultural cross-currents that helped shape the way we create, communicate, and collect information. Most of this work has taken the form of books, essays, and articles—including my forthcoming book Empire of Ink (Basic Books, 2026), which chronicles the invention of the American newspaper; and my previous books Cataloging the World and Informatica (originally published as Glut). This newsletter tries to expand on this body of work, but in a looser, more personal key.

What to expect

Every couple of weeks, I’ll send out an essay—likely around 1,000–1,500 words—about the intertwined histories (and possible futures) of media, technology, and culture. I may veer off-topic here and there, or drop in a short aside (or “squib,” as nineteenth-century editors used to call them), but the through-line will remain the same: how the past continues to shape our digital present.

Everything here is free for now. I might gently nudge you to pick up a copy of my book at some point, but that’s about as sales-y as I plan to get. If you feel moved to support the work, however, I will gratefully accept (as Dr. Johnson famously said, “No man but a fool ever wrote except for money”).

Mostly, though, I just appreciate your time and curiosity. Thanks for tuning in.

Alex