The Geometry of Useful Men

He had encountered the man before.

Not in San Francisco. Not on Twitter. Not at a governance retreat in the Berkshires or an AI safety symposium in Berkeley.

In a novel.

The discovery arrived midway through a nineteenth-century book he had purchased less from curiosity than obligation. One eventually reads the classics when one spends enough time around people who mention them, if only reluctantly and always hoping for a cliffnotes. He had acquired it at a conference — one of those conferences where the bookseller’s table functions as a form of credentialing, where the act of purchase is observed, where the titles signal allegiance. He had not expected to be arrested by it.

But somewhere between the provincial climbers, the failed journalists, the minor bureaucrats, the salon intellectuals, and the men perpetually waiting for patronage to solidify into appointment, he experienced the unpleasant sensation of recognition. Not of a character. Of an architecture. Of the precise arrangement of hope and adjacency and deferred reward that he had understood, until that moment, as his own original condition.

The character was not identical.

The character wore worse clothes.

The character traveled by carriage rather than between airports, attended salons rather than Substacks, wrote letters rather than newsletters. His patrons kept him waiting in antechambers rather than in comment threads. The humiliations were slower and more legible.

But the structure was unmistakable.

Always adjacent to power. Never quite possessing it. The banker made the loans. The industrialist built the factory. The minister signed the decree. The intellectual attended the dinner. The intellectual explained why the dinner mattered. The intellectual went home afterward and wrote about what it portended.

He had met this man in Balzac. He had met him in Stendhal. He had met him in the back half of every nineteenth-century novel in which a young man of talent and insufficient capital arrives in the capital and discovers that talent without capital is not an asset but a service — a thing to be rented out to those who possess what talent cannot supply.

He closed the book and stared at the wall for some time.

The venture capitalist gets the carry. The founder gets the exit. The post-rationalist gets a $75,000 fellowship from a foundation the venture capitalist funds, an invitation to speak at a conference the founder sponsors, and a Substack with ten thousand subscribers.

The differences were technological. The similarities were social. The differences were, he now understood, irrelevant.

What disturbed him was not the resemblance itself but the chronology. For years he had imagined himself part of something historically novel. A new intellectual formation for a networked age. New institutions. New forms of coordination. New epistemologies. New ways of being in relation to knowledge and to power. The language had been genuinely new. The concepts had felt genuinely new. He had lived inside this newness with the conviction of a man who arrives at a discovery rather than a man who arrives at a destination others have reached before him.

And yet here, waiting for him in a novel written nearly two centuries earlier, stood his ancestor. Unchanged in any essential particular. Wearing the period costume of a different era but occupying the same position in the same room relative to the same door through which the same men with actual capital occasionally passed, nodded, and continued.

Not a revolutionary. Not a builder. Not even a parasite, which would at least have implied a certain ruthlessness, a certain clarity about the transaction. Something more ambiguous and, in its ambiguity, more Balzacian.

A court philosopher without a court.

A minor functionary of legitimacy.

A man whose livelihood depended on remaining close enough to power to interpret it, but never close enough to exercise it — for exercise would have required equity, authority, obligation, all the things that ecosystems do not issue and institutions once did.

Balzac understood this with the precision of a man who had lived it. His intellectuals are never simply failed. They are structurally positioned. They occupy a niche the system requires and does not reward. They provide the service of meaning — the narration of power back to itself in terms power finds flattering or at least interesting — and they receive in return the sensation of proximity, which they mistake, for as long as they can, for participation.

The banker had changed. The technology had changed. The language had changed. The velocity had changed. The position had not. The position, it turned out, was not historical. It was architectural.

What Balzac understood, and what every age rediscovers too late, is that societies produce surplus interpreters whenever the routes to power narrow. The frontier closes. The factories consolidate. The platforms scale. The term sheets go to fewer hands. Most of the talented do not become ministers. Most do not become magnates. Most do not become founders. They become, with varying degrees of self-awareness, explainers. Analysts of the very narrowing that has produced them. Theorists of the coordination failures that have left them precisely where they are.

The realization hurt not because it was harsh but because it arrived late. Late enough that the identity had hardened around it. Late enough that the self-conception had calcified into a life.

At twenty-five he had imagined himself entering a vanguard.

At thirty-five the vanguard had begun to feel, obscurely, like a waiting room.

At forty-five he had read a nineteenth-century novel and understood that the waiting room had always been the destination — that the vanguard was the name the waiting room gave itself during periods of sufficient funding, and that Balzac had described every subsequent iteration of it two hundred years before the venture capital arrived to make it briefly, expensively, almost convincingly real.

The tragedy was not that the profession existed.

Every civilization produces it. Every civilization requires it.

The tragedy was that nobody had told him — and that the ecosystem, which rewarded the narration of all other systems, had developed a single powerful blind spot: itself.

The book was called *The Geometry of Useful Men*. He had forgotten the title, or believed he had, until he found himself unable to verify this with any confidence.

The protagonist was named Vernet. Édouard Vernet. A young man of good education and modest means who had arrived in Paris from Lyon in 1831 with three letters of introduction, a facility for argument, and the unexamined conviction that these constituted a plan.

The scene he kept returning to was in the third part.

Vernet is at a dinner given by a banker named Castaing. The table is long. The candles are many. The other guests are men of consequence — a deputy, a newspaper proprietor, a man whose precise function nobody states but whose deference everyone calibrates carefully. Vernet has been placed at the table’s middle distance: close enough to the center to imply inclusion, far enough to preclude it.

Castaing asks him to explain something. The new theories of industrial coordination. The English model. What it might mean for French manufacturing. Vernet explains. He explains well. He has, as the narrator observes with Balzac’s customary surgical neutrality, the gift of making complexity feel like intimacy — of rendering the difficult not simple but legible, which is a different and more valuable thing in a room full of men who have not had time to read.

The table listens. Castaing nods.

“You should write this down,” Castaing says.

“I have written it down,” Vernet says. “I sent you the monograph in February.”

A pause. The candles do not flicker.

“You should write it somewhere people will read it,” Castaing says, and turns to the deputy.

He read this passage on a Tuesday in November, sitting in the kind of coffee shop that charges eight dollars for a cortado and offers, in exchange, the comfort of being surrounded by others doing the same work of waiting dressed as working.

His phone buzzed.

It was a message from a man named Garrett, who ran something called the Center for Epistemic Infrastructure. They had met at a conference in October. Garrett had seemed, at the conference, like a man with resources. He had the particular ease of the funded.

*Really enjoyed your piece on network epistemology,* the message read. *We’re putting together a small working group in January. Stipended. Would love to have you involved.*

He stared at the message for longer than it required.

*What’s the stipend?* he typed, deleted, retyped, deleted.

*What’s the focus of the working group?* he wrote instead.

The response came quickly. *Governance frameworks for distributed knowledge systems. We’re trying to build something serious.*

He thought about the monograph he had sent Garrett in September. Not a monograph exactly — a long essay, forty pages, on exactly this subject. He had received no response.

*Sounds interesting,* he wrote. *What’s the timeline?*

*January retreat, then ongoing. We’re still figuring out the shape. Very much a conversation at this stage.*

A conversation, he thought. At this stage.

He looked back at the book.

Vernet, after the dinner, walks home through streets that Balzac renders in the specific cold of a Paris November — the kind of cold that is less a temperature than a verdict. He is not unhappy. The dinner went well. Castaing noticed him. The deputy had asked a follow-up question. He had handled it with the ease of a man who knows his material.

He has not been offered anything. He will not be offered anything for some time. He will attend three more dinners, write two more monographs, be mentioned favorably in a letter he will never see, and eventually receive a minor appointment at an institution whose importance will prove, in retrospect, to have peaked six months before his arrival.

None of this is visible to him on the walk home.

What is visible to him is the lamplight on the wet stones and the feeling, which he does not examine, of having performed well in a room that was not yet ready to receive him.

He would be ready for it when it was ready for him.

He was always almost ready.

He typed: *Happy to talk. What’s the stipend structure?*

This time he did not delete it.

The response took four hours.

*We’re still working out the details on that side. But we’re really committed to making it worth people’s time.*

He put the phone face-down on the table.

Around him the coffee shop continued its business of hosting people between things. A woman to his left was on a call about a podcast. A man across the room was annotating what appeared to be a grant application. The music was low and chosen to suggest seriousness without demanding it.

He picked up the book.

Vernet, in the next chapter, writes a letter to his father in Lyon.

*I am making progress,* the letter begins. *The connections here are everything one was told they would be. It takes longer than one expects to find one’s position, but I am confident that the work speaks for itself, and that the right people are beginning to hear it.*

He had written almost this exact letter. Not to his father. To himself, in the notes app on his phone, in the early hours of various mornings, in the specific grammar of a man explaining to himself why the present situation is temporary.

The letter continues:

*Castaing has noticed me. I believe something will come of it.*

The narrator does not comment on this.

The narrator does not need to.

Garrett messaged again the following morning.

*Also — are you on the advisory board for the Threshold Institute? We’re trying to map the ecosystem a bit before January.*

He was not on the advisory board for the Threshold Institute. He had applied to be on the advisory board for the Threshold Institute eight months ago and received an automated response followed by silence.

*Not currently,* he wrote.

*Ah okay. We’re trying to get a sense of who’s in the conversation.*

He understood, with the clarity that arrives only after sufficient repetition, what was happening. He was being assessed not for what he knew but for where he stood. The working group was not looking for expertise. It was looking for adjacency. His value to Garrett was not his forty pages on distributed knowledge systems. His value was his position in a network — a network in which he occupied, he now understood, the middle distance.

Close enough to the center to imply inclusion.

Far enough to preclude it.

He opened the book to the dinner scene and read it again.

*You should write it somewhere people will read it,* Castaing says.

*I have written it down,* Vernet says. *I sent you the monograph in February.*

The candles do not flicker.

He closed the book.

He opened his laptop.

He began, for the third time that year, to write something serious about governance frameworks for distributed knowledge systems, in the hope that this time the right people would read it, and that reading it they would understand that he was not a man to be offered stipends and working groups and conversations that were still figuring out their shape.

He was a man who should be offered something real.

He wrote this for eleven minutes.

Then he opened his email and replied to Garrett.

*Would love to be part of the conversation,* he wrote.

*Let me know when you have more details on the January retreat.*

He sent it before he could read it back.

Outside, November continued its business.

The final section of *The Geometry of Useful Men* was shorter than he expected.

Balzac — or whoever he had decided wrote this book — did not linger. The prose contracted. The sentences, which in the dinner scenes had the expansive quality of a man who believes time is available, became clipped. Functional. The narrator had seen enough. He was moving toward the conclusion with the efficiency of someone who has made this trip before.

Vernet receives his appointment in the spring of 1837.

It is at something called The Society of Public Economy, which occupies three rooms on the second floor of a building in the seventh arrondissement that smells of paper and old ambition. The Institute had been founded in 1829 with serious money and serious intentions. By 1837 the serious money has moved elsewhere and the serious intentions have calcified into a specific set of questions that the Institute asks in a specific set of ways and publishes in a journal that seven hundred people receive and forty read.

Vernet’s title is Associate Fellow for Industrial Correspondence.

The salary is modest. The title is not nothing.

He is thirty-four.

His colleagues are men who arrived before him at the same destination by different routes — a failed newspaper editor, a former deputy’s aide, a man who had written a celebrated pamphlet in 1831 and had been, in some administrative sense, arriving ever since. They are not unhappy. They have achieved the plateau. They work. They publish. They attend the right dinners and are seated at the middle distance.

The narrator describes Vernet’s first week with a precision that is almost kind:

*He understood, settling into the chair that had been Moreau’s before Moreau’s retirement, that he had arrived at the place his trajectory had always been carrying him. This was not failure. Failure would have required a different destination. This was something more exact: the fulfillment of a logic he had not written and could not now rewrite. The Institute would continue. He would continue with it. Outside, Paris would continue its business of producing men like him in quantities the city could not absorb, and absorbing them anyway in the places cities keep such men — the institutes, the journals, the academies, the offices of deputies who need things explained. He sharpened his pen. He began to write. History, indifferent to his productivity, continued its preparations elsewhere.*

He read this paragraph three times.

Outside the coffee shop — a different one now, he had been moving between them for months in the way that people move between ecosystems when each one stops generating warmth — November had become December without asking permission.

His phone was on the table.

It had been producing, for the past seventy-two hours, a low-grade emergency of the kind that no longer felt like emergency because it had become the ambient texture of the professional situation.

The Threshold Institute had merged with something called the Meridian Foundation for Governance Research. The announcement used the word synergy once and the word ecosystem four times. The advisory board of the Threshold Institute — the one he had applied to join eight months ago and been quietly declined — had been dissolved and reconstituted with different names. He recognized two of them. Both were newer to the scene than he was. Both were unassociated with frameworks that had, in the past eighteen months, become liabilities.

Garrett had not messaged about the January retreat.

He had messaged about a postponement. The working group was being restructured. The foundation’s priorities were evolving. There would be a new conversation in the spring about what the new conversation would look like.

He had replied: *Sounds good. Keep me posted.*

He had not heard from Garrett since.

A tool had been released three weeks ago that produced, in the time it took to finish a cortado, the kind of synthesis that had taken him, at his best, the better part of a week. He had used it twice, experimentally, with the feeling of a man pressing on a bruise to confirm it still hurts. It was not as good as his best work. It was considerably better than his average work. It was, more relevantly, free, infinitely patient, and not waiting for a fellowship from a foundation whose priorities were evolving.

He picked up the book.

*History, indifferent to his productivity, continued its preparations elsewhere.*

The preparations. He had been thinking about the preparations.

Balzac’s world ended — not immediately, not in the novels themselves, but in the logic the novels contained — in the Commune, in the trenches, in the specific catastrophe that fin de siècle Europe was assembling from its surplus of educated men with nowhere to put their energy, its narrowing routes to power, its consolidating capital, its interpretive class expanding faster than the productive base that might have justified it.

The Vernets of 1837 did not see 1871 coming.

They saw dinners. They saw appointments. They saw the middle distance.

What they did not see — what the narrator saw, what Balzac saw writing from inside the thing he was describing — was that the system producing them was also producing the conditions of its own violent correction. That surplus interpretation was not a stable equilibrium. That the gap between the people who explained the world and the people who lived in it without explanation had a tensile limit. That legitimacy, endlessly manufactured and decreasingly tethered to anything outside itself, eventually encountered events that did not require interpretation because they were too large and too loud to be narrated into manageability.

The revolution. The war. The inflation that destroyed savings, credentials, positions — all the paper instruments of middle-distance existence — overnight.

He looked at his phone.

The feed was doing what the feed did, which was produce a continuous scroll of premium mediocre epistemology about events that were, increasingly, declining to be epistemically managed. Inflation numbers. Election results from countries that had recently stopped behaving predictably. Supply chain disruptions that the frameworks kept explaining and the supply chains kept ignoring. A war that the governance researchers had written about extensively before it happened and had failed to prevent, a failure that had not diminished their output on the subject, only increased it.

The discourse had become more sophisticated.

The events had become less legible.

He thought: there is a ratio here. There is a point at which the ratio inverts. There is a point at which the distance between the interpretive class and the physical world becomes not an intellectual problem but a political one — not a question of epistemology but a question of who absorbs the costs when the frameworks fail and the costs are real and large and distributed among people who were never in the conversation.

He knew what happened then.

Everyone knew what happened then.

The knowing was, itself, a form of helplessness.

He read the book’s final page.

Vernet in 1843, six years into the Institute, is at another dinner. Different house. Same geometry. He is older. The hunger has not gone but it has changed character — less the hunger of a man who expects to eat soon and more the hunger of a man who has reorganized his life around not acknowledging it.

Castaing is there. Castaing is always there.

Toward the end of the evening Castaing asks him about the new industrial theories. The English model again. Or rather, a new version of the English model, since the old version has been superseded. Vernet explains. He explains well. He has, the narrator notes, lost none of his facility. If anything it has improved — become smoother, more assured, the awkward edges worn away by repetition until what remains is a performance so accomplished it has forgotten it was ever uncertain.

Castaing nods.

“You should write this down,” Castaing says.

Vernet looks at him.

Six years. Three monographs. Forty-seven issues of the journal. Two hundred and thirty dinners at the middle distance.

“I have written it down,” Vernet says.

The candles do not flicker.

“Several times,” he adds, quietly, to no one.

Castaing has already turned to the deputy.

Outside, the narrator tells us, in the novel’s last sentence, in the flat declarative tone Balzac reserved for verdicts:

*The city continued its preparations. The men who understood it continued their explanations. Neither was listening to the other, and neither would have known what to do with the information if it had.*

He closed the book.

His phone buzzed.

It was a newsletter from someone he had met at a conference in 2021. The subject line was: *Toward a New Framework for Epistemic Resilience in Volatile Geopolitical Environments.*

He did not open it.

He sat for a while in the coffee shop that charged eight dollars for a cortado and hosted people between things.

Through the window, December continued its preparations.

He thought about Vernet at the dinner table.

He thought about the city that was not listening.

He thought about what the city was listening to instead — the older, louder signals that frameworks could describe but not contain, that coordinates could not neutralize, that discourse could sophisticate but not prevent.

The inflation that made the credentials worthless.

The war that made the governance models irrelevant.

The revolution that came not because nobody had explained the system but because the explanation had become the system, and the system had forgotten that outside its own self-referential coherence there existed a population that had never been in the conversation and was now, with the blunt instruments available to people who have never been in the conversation, proposing to end it.

He thought: we are producing Vernets at scale.

He thought: the city is making its preparations.

He thought: I have written this down.

He opened his laptop.

He began to write.

Outside, the century continued its business, indifferent to his productivity, moving toward whatever it was moving toward with the momentum of things that have been building longer than anyone in the coffee shop had been paying attention.

He wrote for a while.

Then he ordered another cortado.

Then he checked his email to see if Garrett had been in touch.

He had not.

The preparations continued.