Good companies and good-looking companies

May 25, 2026

Du mußt dein Leben ändern

In 2014, I went to CERN’s Winter School on Supergravity, Strings, and Gauge Theory at the Meyrin site outside Geneva. Meyrin is in the middle of nowhere. In February, it is windy and bleak. The campus looks like a Soviet commune stuck in the eighties — low concrete buildings, corrugated metal cladding, peeling signage, a canteen, bicycles. Inside, the corridors run for kilometers under fluorescent light. The offices are small and often shared, two desks pushed against each other under the window; the hallways have chalkboards of their own. The chalk stays up for weeks.

No Glass Cathedral. No suits. Nothing like the metropolitan business school I would return to in Hyde Park: glass, light, money, polish, ambition, and the quiet choreography by which everyone learns to look like someone worth investing in. Meyrin is its opposite, and the closest thing I have seen to the communist ideal as it was originally imagined — work as play, conducted in common, hierarchy present but strangely beside the point. You do not know who is tenure-track and who is not. You do not know who is famous and who is there only for the term. You do not know who will be remembered and who will vanish back into the machinery of science. It does not appear to matter. The problem matters. The detector matters. The equation matters. The argument at lunch matters. The Higgs boson had been discovered there two years before. The Large Hadron Collider runs underneath. Internet was invented in one of those buildings, to help the scientists share their papers more easily. The achievements are enormous; the place that produced them is practically a monastery, stripped of ornament and almost embarrassed by display.

At CERN I saw work stripped of almost everything that usually confuses it. There was little glamour. There was little money on display. There was no obvious social choreography of importance. The place did not flatter the people inside it. It did not need to. The work itself supplied the dignity.

The people who work there are, in a sense, in self-exile. They have removed themselves from the markets that would pay them better and the cities that would entertain them more. Walking through the corridors at Meyrin, I thought of Perelman in St. Petersburg, living with his mother in an apartment on the outskirts of the city, having declined prestigious American appointments and then refused the Clay Millennium Prize — a million dollars for proving the Poincaré conjecture — after already refusing the Fields Medal in 2006 because he did not want to be displayed like an animal in a zoo. I thought of Ramanujan in Madras a century earlier, working out his theorems in the margins of borrowed books while clerking at the Port Trust, until Hardy brought him to Cambridge. I thought of Yitang Zhang, who proved the bounded-gaps theorem on prime numbers at fifty-eight, after years on the margins of academic mathematics. I thought of Grothendieck, who would die later that year in a Pyrenean village he had not left in two decades, having walked out of the Institut des Hautes Études Scientifiques in 1970 over military funding, burned much of his own archive from the world, and spent the rest of his life writing thousands of pages of mathematical, mystical, and philosophical work he later asked the world not to publish,In 1991, Grothendieck burned twenty-five thousand pages of his own manuscripts in the garden of his then-partner’s house and moved to a hamlet in the Pyrenees foothills called Lasserre. Jean Malgoire, a former student, managed to rescue about twenty thousand pages before they went into the fire; the rest were lost. He lived in Lasserre for the next twenty-three years on dandelion soup and his own dreams. In January 2010 he wrote a letter to another former student, Luc Illusie, asking the libraries still holding his books to remove these books from those libraries. as if even recognition after disappearance had become another form of theft.

In each case, the work had become almost separable from the world that was supposed to recognize it. Recognition was a corruption, or an irrelevance, or a contingency that arrived late or not at all. The discipline made room for them when it had no choice. They did not need the discipline back.

Andri Pol, Inside CERN.

I returned to Chicago that spring and the university looked different. The question hovering over my work was not whether I could do it but whether I belonged to the discipline that would judge it. I was a statistician working in quantum field theory and string theory. Interdisciplinary work was permitted but discouraged. The rule, spoken or unspoken, was that crossing fields was a privilege reserved for the tenured. First you served the discipline. Then, if you survived, you were allowed to leave it.

Epigraph of my doctoral dissertation, The University of Chicago, 2018.

A question began forming in my notebooks that I would later put in my dissertation: are the Renaissance painters heroes and gods, or are they manual workers?I probably got this from Sloterdijk? The question was not really about painters. It was about the meaning of work itself. Is work made noble by the institution that recognizes it, or by the thing it answers to? Is the scholar dignified because the university confers dignity on him, or because the act of inquiry already contains its own form of sovereignty? Was Perelman free because mathematics had crowned him, or because he had found something in mathematics that made the crown irrelevant?

The honest answer, for me, was that I was not Perelman. I was not brilliant enough, or pure enough, or indifferent enough for the pages alone to sustain me. Mathematics in and of itself would not have been enough. The work alone, with no audience but the work, would not have carried me through the years a serious life requires.

But there were other considerations too, less noble and more practical, though not therefore less real. By the time I was leaving academia, machine learning was already rising at the edges of everything I had been trained to value. The older university had justified itself as a place where knowledge was stored, transmitted, and authorized. But that function was already beginning to leak out of the institution. Who needed the lecture hall when YouTube could teach you the lecture? Who needed the seminar room when the best explanations were becoming searchable, replayable, and free? And now, with LLMs, even that question seems quaint. The university no longer has a monopoly on access to knowledge. Increasingly, it does not even have a monopoly on explanation.

Statistics, too, was changing under my feet. The discipline in which I had been trained still carried the prestige of inference, asymptotics, elegant parametric models, and the controlled humility of estimation. But the world was moving toward prediction, scale, computation, and engineering. The center of gravity was shifting from the statistician’s model to the machine’s performance. The old professional identity — the statistician as the person who specified the right model, proved its properties, interpreted its coefficients, and disciplined uncertainty into form — was losing authority. Not everywhere, and not all at once. But enough that I could feel the future receding from the thing I had been trained to become.

So leaving academia was not only a failure of purity. It was also a recognition that the institution I was leaving might not be the institution in which the next life of the mind would be fought. The hedge funds beginning to recruit my classmates were glass cathedrals of their own: monetized, confident, lit from above. But they would see me. They were also, in their vulgar way, closer to the new dispensation: computation, prediction, scale, feedback, judgment under uncertainty. I had not left China, learned the language of American meritocracy, and spent a decade inside its institutions in order to disappear into a noble obscurity. For every Chinese student enrolled at a private American university, there are thousands at least as able, and many more able, who never came close to the same door. They were not less gifted. They were less fortunate. Whatever I owed the work, I also owed something to the accident that had carried me to it.

To choose obscurity after such an improbable passage would not have felt like purity. It would have felt like ingratitude disguised as purity. I left academia after graduation and never looked back.The figures I have been describing embody a particular Western romance: the genius who withdraws, who refuses the world and is purified by the refusal. I did not grow up inside that romance. The tradition I inherited took a different view of what learning is for. 学而优则仕 — when one excels in study, one enters service. The scholar does not study in order to go on studying; he studies in order to act, to take office, to apply what he has learned in the world. Mencius makes the application an obligation: 达则兼济天下 — in adversity, perfect yourself alone; in success, better the world. To rise — 达 — is to acquire a position from which one is bound to benefit others; to rise and benefit only oneself is to fail the duty that rising imposes. And where the Western recluse answers the question of withdrawal by going to the mountains, the Chinese tradition answers it in reverse: 大隐隐于市 — the lesser hermit hides in the wilderness, the greater hermit hides in the marketplace. The mountains are the easy discipline. To keep one’s cultivation intact inside the noise and the money of the world is the harder one, and the higher. The cloister was never our ideal. The scholar who entered the world without losing himself was.

For years I told myself this was a story about leaving one kind of work for another. It was not. It was a story about discovering that work has two meanings, and that modern life is largely organized to confuse them. There is work as economic activity: the thing one is paid to do, measured by output, title, compensation, and rank. And there is work as vocation: the thing through which one’s powers are disciplined, enlarged, and finally judged. The first makes a living. The second makes a life. The tragedy is not that these two forms of work are different. The tragedy is that they sometimes resemble each other closely enough for a person to mistake one for the other for years.

The Chicago motto is crescat scientia; vita excolatu — let knowledge grow, and so let human life be enriched. The motto was chosen in 1894, and it rests on a confidence that the growth of knowledge enriches life. That confidence was reasonable when knowledge was hard to come by, when the act of growing it was itself a human achievement that conferred dignity on the grower. To know more than others, to discover what had not been known, to add something durable to the common store: these were not merely technical accomplishments. They were ways of becoming a certain kind of person.

In 2026 the premise has cracked. LLMs are the most extraordinary librarians the world has ever seen. If something has been written down, they have probably seen it. Knowledge is cheap now. No one wants to watch Usain Bolt race a motorcycle. We are now being asked to accept the same humiliation of the mind.

But perhaps it is not a humiliation. Perhaps it is a clarification. If knowledge itself is no longer scarce, then knowledge can no longer be the center of human dignity. The Chicago motto asked, in 1894, that knowledge grow so that life could be enriched. In the age of the librarians, the inverse question matters more: if knowledge is no longer worth growing for its own sake, can life still be enriched? The answer, if there is one, depends on what you do with the knowledge, and on whom you become while doing it.

This is why the question of work returns with more force, not less. The machine can summarize the paper. It can draft the memo. It can retrieve the precedent, imitate the style, generate the code, compose the email, and offer the plausible answer. What it cannot do is stand behind the answer. It cannot want the thing to be true. It cannot be corrupted by recognition or purified by obscurity. It cannot leave the cloister for the market and then ask whether it has betrayed itself. It cannot confuse a salary for a vocation, because it has neither hunger nor vocation. It does not work. It performs.

Human work begins where performance is insufficient. It begins where the person doing the work is changed by the doing of it. This is why the old examples still matter: Perelman, Ramanujan, Zhang, Grothendieck. They are not examples because they were geniuses, though they were. They are examples because their lives reveal the extremity of the question. What would you do if the world did not reward the thing you were made to do? What would you do if recognition came too late, or came wrongly, or came in a form that falsified the work itself? What part of the work would remain?

I do not ask these questions from a position of purity. I left. I chose the market. I chose visibility. I chose the glass cathedral over the concrete commune. And because I chose it, I do not have the luxury of romanticizing the cloister. Poverty is not dignity. Obscurity is not proof of depth. Institutions, for all their cowardice and vanity, can protect real work as well as distort it. Markets, for all their vulgarity, can reveal truths that monasteries hide. The question is not whether the cloister is better than the market. The question is whether either can be inhabited without surrendering the thing that made work meaningful in the first place.

What I have been looking for, in the years since I left Hyde Park, has a name. It is Hesse’s Siddhartha, not the Buddha — the Brahmin son who left his father for the ascetics, then left the ascetics for Kamaswami’s city, then left the city in despair for the river, and only at the end of his life understood that none of those leavings had been a betrayal. The cloister and the market were not opposites. They were stages. The river was what contained them both.

That is the meaning of work I want to recover. Not work as employment, not work as status, not work as productivity, not even work as knowledge production. Work as the river: the activity in which the divided parts of a life are gathered, disciplined, and made answerable to something beyond appetite and applause. Work is where the self stops merely expressing itself and begins to be formed. It is where freedom ceases to mean the absence of constraint and begins to mean obedience to the right constraint.

At CERN, beneath the concrete and the peeling signage, I saw one version of that obedience. In Chicago, under the glass and the institutional light, I saw another. In the market, I would see still another. The mistake was believing that one place possessed the meaning of work and the others had lost it. The harder truth is that work is not guaranteed by any setting. The monastery can become a career ladder. The market can become a discipline. The university can become a theater. The hedge fund can become a school. Everything depends on whether the work remains answerable to something real.

And so the question is not, finally, what kind of work should one do? The question is: what is your work making of you?

I can think. I can wait. I can fast.

We say we work to make a living. We say it because it is partly true, and because the partial truth lets us avoid the harder version of the question. The harder version is why the activity of work — the spending of one’s hours, attention, intelligence, body, and pride inside it — is bound up with what we take ourselves to be. A person can live, in principle, on very little. Most of what we do at work, and most of what we say about it, is not about staying alive. It is about something else. Status, perhaps. Usefulness. Recognition. Escape. Obedience. Mastery. The wish to become necessary. The wish to be seen doing something difficult. The wish to see oneself in the world and not merely pass through it.

Hegel, in the Phenomenology of Spirit, gave what is still the deepest answer. Two self-consciousnesses meet; one becomes master, the other slave. The master is recognized but recognizes nothing. He consumes without producing. The slave works on the world, transforms it, and through that transformation comes to recognize himself in what his labor has made. The slave, not the master, is the one who arrives at consciousness. Labor, for Hegel, is the path by which the human being becomes aware of himself as a self: by remaking the world, the worker finds his own image in it.

Marx took this and turned it against capitalism. Under capitalist relations, the slave does not receive the dialectical reward. The labor that should have produced self-recognition is alienated — sold as a commodity, its products taken away, its meaning emptied. The worker’s own power returns to him as something foreign: a wage, a commodity, a system, a boss, a market price. He does not see himself in his labor. He sees only what his labor can be exchanged for.

Lukács sharpened the diagnosis into reification. Under capitalism, human activity confronts the human being as a thing. The living motion of work becomes object, commodity, calculation. Time becomes billable. Talent becomes human capital. Attention becomes productivity. Education becomes credential. Personality becomes brand. The Hegelian dialectic, which should have ended in freedom, is short-circuited by the form of value.

Deleuze and Guattari pushed the diagnosis to its widest form. Capitalism, they argued, is the first social formation that has no inherent code. Older societies organized themselves around something untranslatable — a god, a king, a kinship, a craft — that could not be exchanged for anything else. Capitalism dissolves every such code and reterritorializes everything on the single axiom of money. Nothing is sacred in capitalism because everything is, in principle, priceable. This is not a moral complaint about consumerism. It is a structural observation. Inside the axiomatic of capital, there is no place from which work could be loved on its own terms, because the form of value has been extended to cover every register in which work could be valued.

He is opportunistic without seeming hungry. This is part of the style. Neediness would lower his price. He is chill. He does not overwork unless overworking produces visible return. He does not underwork in any crude or punishable way. He works exactly hard enough to remain optional, liked, plausible, and difficult to accuse. He understands effort as capital allocation. The question is never what the work requires. The question is what level of effort buys the greatest personal upside for the least personal exposure.

Remote work suits him, though not for the reasons he gives. He is not especially contemplative, independent, or capable of deep self-direction. Distance widens the gap between contribution and appearance. At home, he can manage surfaces more efficiently. He can be present without being implicated, responsive without being available, involved without being seen too closely. The screen is his natural habitat because the screen converts work into signals: green dots, quick replies, calendar blocks, polished snippets, performative busyness, the visible tokens of contribution detached from the slow embarrassment of actually being watched working.

He parades work because work, for him, is most useful when it can be seen. He works on planes, not because the work must be done there, but because the plane supplies the image: the overburdened man in transit, too necessary to be still. He sends Slack messages late at night, after having drifted through the day with remarkable ease. The timestamp becomes evidence of devotion. The content matters less than the hour. He has discovered that exhaustion, properly displayed, can substitute for contribution.

This gives him a moral language with which to accuse others. A colleague takes a trip, protects a weekend, refuses the late-night call, and he calls it selfish. The accusation is not that the work has suffered in any precise way. The accusation is that someone has broken the theater of sacrifice. He speaks often of fiduciary duty, compliance risk, optics, responsibility, the seriousness of the matter. These phrases are not constraints on his conduct. They are weapons for disciplining the conduct of others. He is usually the first to violate the obligation he has just named, and the first to explain why his violation is not really a violation at all.

No conflict survives contact with his self-interest. He rationalizes it away, reframes it, forgets it, converts it into an exception, or discovers that the true ethical risk lies elsewhere. Optics matter intensely when they can be used against someone else; they become vulgar and beneath consideration when applied to him. He does not care how he is judged by his peers because, in his private hierarchy, they are not really his peers. They are observers, obstacles, future subordinates, or people too small-minded to understand the scale at which he operates.

In public, he speaks the language of burden. He is losing sleep. The matter keeps him up at night. The stress is giving him a receding hairline. He performs depletion as proof of seriousness. But the exhaustion is curiously selective. It appears in the story about the work more reliably than in the work itself. What he wants credited is not sacrifice but the appearance of having sacrificed; not responsibility but the visible strain of seeming responsible.

This is the figure I have been calling the value-extractor. He is not stupid. He is not, in any sense the law would recognize, unethical. He is the fully formed subject of late capitalism: the worker who has internalized reification and the axiomatic together so completely that he treats his own time as a commodity and optimizes the return on it without remainder. Every meeting is a transaction. Every relationship is a network position. Every year of his life is a portfolio decision.

He may even know this about himself. He may have read the books that diagnose him. He may describe the system with fluency, irony, and contempt. But the cynical distance is not a refuge from the ideology. It is the form the ideology now takes. Žižek’s analysis turns on exactly this point. Older ideologies required belief: the subject was supposed to think the official story was true. Contemporary capitalism requires only behavior. You do not need to believe the mission statement; you need only attend the off-site, repeat the framing, optimize the metric. The cynic who knows it is all theater is more reliable than the believer, because the cynic has already absorbed the inconsistency and turned it into a brand. The person who says, of course it is all about the money, I am not naive, is not seeing through the system. He is the system, articulating itself fluently in the first person.

The deeper problem with the value-extractor, though, is not that he is alienated from his labor in the Marxist sense. The deeper problem is that he has misunderstood what work is for. He thinks the work is an instrument and the return is the point. He has the equation backwards. The return is the byproduct, sometimes large and sometimes meager, of work being well done. The work itself is the point. The value-extractor cannot see this because the categories through which he sees have been emptied of any non-exchangeable content. He has eyes for what can be optimized and no eyes for anything else.

This is most visible in his relation to what Polanyi called the tacit dimension. Polanyi observed that we can know more than we can tell — that all serious work rests on embodied, unarticulated, judgment-laden skill that resists capture in rules and KPIs. The chess master, the radiologist, the welder, the trader who can feel a market shifting before he can explain it: each knows in the bones of his practice what no document records. The workshop is precisely the form of work that transmits this tacit dimension, through apprenticeship and through co-presence at the bench. The value-extractor cannot see the tacit dimension because the tacit cannot be optimized for. He sees only the documented surface — the deck, the dashboard, the deliverable. He confuses the documented surface with the work, and confuses optimizing the documented surface with doing the work. He is, in the strictest sense, missing the point.

What he does instead is optics. Perception, for him, is reality, because in his world perception is what gets priced. He cultivates appearances of competence, ascends through the management of impressions, takes credit for the tacit work of others while contributing none of it himself. He pontificates, chest-pounds, cries foul, discovers injuries useful to his position, and converts the common work into a theater of advantage. He makes his own development impossible because growth requires submission to the discipline of the work itself, and he has lost contact with that discipline. He makes the development of others impossible because apprenticeship requires trust, and he has converted trust into exposure. He does not believe one can see him. He believes everyone is doing what he is doing and is simply less skilled at hiding it.

Work has become so abstract to him that he has nothing left to reason with except relations. He no longer touches the thing. He is the architect of a skyscraper who has stopped going to the site. He dreams in floor counts and skyline silhouettes and does not know whether the concrete on the twentieth floor has cured. That is beneath him. That is for the people who work with their hands. What he forgets is that the skyscraper actually depends on whether the concrete has cured, and that no polish on the rendering will compensate for failure at the slab. When the building falls, it falls on the people who were doing the work. He will already be on his next project, presenting himself through the credentials of his last employer rather than the record of what he himself did there, because the record will not bear the weight he needs it to bear.

He is quick to act, and this looks, at first, like decisiveness. It is not. He acts quickly because he lives episodically, in flashes of opportunity, and because planning would commit him to a structure he intends to escape. Planning is against his method. So is the public sharing of plans: a public plan ties him to it, and his entire mode depends on not being tied. What he does instead is hoard information and release it selectively. Secrecy is his most reliable instrument because secrecy preserves optionality. To share is to surrender future moves.

This is also why he lives from crisis to crisis. Crisis is the mode in which everyone else’s planning has been shattered and his improvisation reads as leadership. He gravitates toward it, generates it when it is not at hand — by withholding information until the moment of maximum pressure, by promising and forgetting, by stacking deadlines against one another so that the chaos requires him to step in. The older he becomes, the more he leans on this. He thinks it is leadership. It is only the substitute he has left, because he has spent decades opting out of the kind of work that builds an actual mental model of what he is supposedly running.

His relationships are positions that open and close. He claims loyalties he does not cultivate. You are my guy is a sentence he produces frequently and inexpensively, addressed equally to those he intends to use now and those he intends to use later. When he is younger the corresponding sentence is I will follow you, addressed to whoever currently holds the keys. He does not follow. He does not lead. He attaches and detaches. The relationship is the unit of his work, but the relationship has no interior — no shared project, no accumulated trust, no inherited debt. Acquaintances eventually notice and stop returning his calls. He does not mind, not really. The acquaintances are not the target market anymore. The target market is the new people, who have not yet been on the receiving end of the closing of a position.

What he holds against the world he holds quietly. He does not confront the casual injury, the offhand remark, the honest correction. He buries it, and the burial becomes its own form of fuel. I will prove them wrong is the motor running underneath his ambition. The people he is proving wrong are mostly people who never had the chance to know him in the first place — coworkers in passing, acquaintances who said something honest at the wrong moment, observers who paid a cost they did not realize they were paying. The injury is asymmetric. They were doing the small work of saying what they saw. He is nursing decades of wounds whose authors do not remember inflicting them.

Even ordinary interactions are not easy for him. The transaction is always running. His mind is racing — what does he want from me, what can I get from him — and the other person can feel the calculation in the room even when he cannot name it. He will tell you Sam hates him. Sam does not know him. Sam crossed his line of sight once, gave a neutral expression, and went into the file tagged hostile, because in the absence of any actual relationship the calculation has nothing to read but the worst possible interpretation of silence. The grudge does not need a real injury. It needs only an other.

This is what comes of not understanding how relationships actually work. Real relationships move through a cycle of time, test, injury, and repair. The friendship that has not yet survived a quarrel is not yet a friendship; the colleague who has not yet disappointed you, and whom you have not disappointed, and with whom you have not yet found your way back, is not yet a colleague. He does not have access to this cycle. He has two categories: people who are useful and people who are enemies. The third category — the people you trust enough to be hurt by, and whom you trust enough to forgive — is missing from his vocabulary. So he does not trust. He uses. He interprets the absence of trust as clarity, when in fact it is the absence of a faculty.

He even takes pride in his choice of associates. He likes working with people who are visibly second-rate, and he calls this his superpower: he can make do, he says, with the material at hand. What he does not say, because he does not know it, is that he has selected the second-rate material on purpose. Competent people see through him. Incompetent people do not. The associates whose lack of credentials embarrasses him to others are the same associates whose lack of credentials protects him from being seen. Their incompetence is not a constraint he has heroically worked around. It is a moat.

What he does not see, and what he gets burned by repeatedly without learning, is that the pattern keeps producing the same result. Each time, he is briefly successful, then briefly tolerated, then quietly avoided. He attributes this to the malice of others. He becomes defensive. He stops communicating about the work and starts communicating about himself: corrections, vindications, war stories from prior firms, the credentials of his ex-employers held up against the suspicions of the present one. He suspects the colleagues he has not yet replaced of laughing at him behind his back and scheming against him. Often they are. Not because they are conspiring, but because they have all noticed the same thing he refuses to notice, and have agreed, without coordinating, that the only thing they want from him is for him either to do the work or to remove himself from the room.

This is one reason the administered workplace is more fragile than it looks.See §3 All the uncaring intricate rented world, Cognitive Resignation American workers are often docile in the first instance, especially across title lines. They will do what they are told, or appear to do it, because the organization has trained them to translate rank into temporary obedience. But obedience is not cooperation, and work is not a single encounter. Work is a repeated game. The first command gets compliance. The second gets clarification. The third gets public pushback, often in the language of process, risk, fairness, or scope. After that come the quieter forms of refusal: delay, omission, literalism, non-response, the email not answered, the task completed in the thinnest possible sense. Eventually the manager feels as if he is punching air. Nothing is openly disobeyed; nothing quite moves. A small organization can survive for a while on direct force, charisma, fear, or the exhaustion of an extractor-leader who personally pushes every loose object back into motion. But scale makes dictatorship impractical. No one has the energy to dictate at sufficient resolution. At a certain size, the company runs only on what cannot be ordered minute by minute: trust, standards, memory, pride, shame, loyalty, and the tacit desire of people to see the thing done well. Where those are missing, hierarchy still exists, but it increasingly speaks into a void.

The value-extractor also changes the population around him. Extractors attract other extractors, not because they trust one another but because each recognizes in the other a familiar grammar: tactical warmth, optional loyalty, flexible memory, surface competence, and the language of principle used as leverage. Soon his world is filled with people who behave as he behaves, and this feels to him like confirmation. See, he thinks, everyone is like this. But this is not anthropology. It is adverse selection. The better people leave first because they are disappointed. They expected work, loyalty, standards, or friendship, and found only use. The other extractors remain only while the spread is attractive. When a higher-value field appears elsewhere, they detach with the same fluency with which they attached. The world he has built therefore confirms him only by emptying itself of anyone who could have disproved him.

Bauman traced the conditions that produced this figure. In liquid modernity, the older producer society — where identity was made by what one made, where a craft, firm, vocation, town, or institution could hold a person long enough to shape him — gives way to a consumer society in which the self is a brand, the job a temporary engagement, and durable commitment a failure of flexibility. The value-extractor is what becomes of a person whose long horizons have been dissolved. He is not a villain. He is a survival strategy.

Byung-Chul Han names the final form of this subject more sharply than anyone else writing now. We have moved past the disciplinary society of Foucault, in which the subject is constrained by external authority, into the achievement society, in which the subject exploits himself. The neoliberal worker is his own master and his own slave. The factory boss has been internalized, and the resulting subject is faster, more productive, more adaptive, and more thoroughly worked than any disciplinary regime could have produced. He is also exhausted in a new way: not tired from labor but depleted by the impossibility of ever being done. Burnout is not the worker breaking under an external demand. It is the worker breaking against an internal demand that has no ceiling. The command is no longer simply work. It is become more employable, more efficient, more legible, more networked, more resilient, more optimized, more yourself. Under such conditions, even self-expression becomes labor. Even rest becomes recovery for future output. Even freedom becomes a technique of management.

So we have the diagnosis. The question is what the alternative is.

It is not the cloister. The cloister is the lesser hermit’s choice — admirable but incomplete, refusing the harder discipline of staying inside the world. Nor is the alternative the market as we usually defend it: the cheerful brutality by which every human power is translated into exchange value and every life is invited to become an investment thesis. The tradition keeps trying to name a third thing. Hegel called it labor as the path to self-recognition. Marx called it unalienated labor. Arendt called it work: the making of a durable world, something more stable than consumption and more tangible than action. None of these is the value-extractor’s life, and none is the hermit’s. They name work in which the worker is present, the world is engaged, and the activity itself produces both a self and a contribution.

That third thing has a name in the older tradition. It is the workshop.

The workshop is neither sacred nor merely commodified. It is not enlightenment, and it is not return on capital. It is not the monastery, where the world is renounced, and not the market, where the world is priced. It is the doing of the thing in the presence of others doing the same thing, under the discipline of the thing itself. The worker is neither dissolved into the institution nor sealed off in private purity. He is formed by contact with the world and by contact with the others who are forming themselves the same way.

The workshop is a community, and the community is an ecosystem. It can absorb some number of value-extractors, because every real workshop now operates inside the late-capitalist environment that produces them. But the ratio matters. A workshop can survive a few; it cannot survive many. Once they reach a certain concentration, the tacit dimension stops being transmitted. Apprenticeship becomes credentialing. Senior craftsmen stop teaching. The work becomes the rhetoric of work. The institution goes vacant. The building remains, the titles remain, the meetings remain, the slides are still produced. But the workshop is gone. What remains is an architectural shell with the vocabulary of work and none of its inner life.

Jockeys over horses

The exchange that SBF had started to build, FTX, was Goldilocks-perfect. There was no concerted effort to skirt the law, no Zuckerbergian diktat demanding that things be broken… SBF himself seemed to be bred for the role of crypto exchange founder and CEO. Not only had he been a top trader at a top firm — and, thus, the ideal customer — but both his parents were lawyers. ‘And, so, he is committed to making the right chess moves for FTX to eventually be able to legally do everything they want to do in the U.S.,’ Bailhe says — ‘not by asking forgiveness, but by asking permission. […]

SBF is a Peter Singer–inspired utilitarian in a sea of Robert Nozick–inspired libertarians. He’s an ethical maximalist in an industry that’s overwhelmingly populated with ethical minimalists. I’m a Nozick man myself, but I know who I’d rather trust my money with: SBF, hands-down. And if he does end up saving the world as a side effect of being my banker, all the better. […]

But within the inner circle at Apartment V — a community of family and friends united by a philosophy that’s almost Pythagorean in its rule — there’s no unit of account. Love is the currency. Love is infinite. And infinity is a problem… infinity is also the solution, because it provides a shield from the spartan logic of utilitarianism. There’s no way to do an expected value calculation when one of the terms is infinite. The very incalculability of the love that exists in Apartment V makes it the one refuge from the whip that drives him. […]

After my interview with SBF, I was convinced: I was talking to a future trillionaire. Whatever mojo he worked on the partners at Sequoia — who fell for him after one Zoom — had worked on me, too. For me, it was simply a gut feeling… I don’t know how I know, I just do. SBF is a winner.Adam Fisher, Sequoia, Sam Bankman-Fried Has a Savior Complex—And Maybe You Should Too.

I was charmed by his nerdy affect as SBF, unkempt in a T-shirt and bushy hair, as he twirled a fidget spinner and rattled off tidbits about everything from M&A strategy to the macroeconomy to the importance of trust in business deals. It was all bullshit, of course, and I didn’t see through it… The answer is that, like any good con man, SBF told us a story we wanted to hear and were eager to believe. In hindsight, there were red flags aplenty—FTX using cutesy numbers like 69 and 420 in its funding announcements, the lack of a board or any other oversight, a CEO who played League of Legends instead of working—but we ignored them.

Sure, a handful of people saw through him, and they will come to receive the same accolades as those who foresaw and bet against the subprime mortgage boom in 2007. And in the coming days, many others will join Elon Musk and Mark Cuban in claiming they knew SBF was a fraud all along—even though they didn’t say a peep until last week.

The lesson here for me and others is the old adage “trust but verify.” I should have pushed SBF and his company harder to show me documents proving FTX was what he said it was, and not assumed that the likes of Sequoia would never back a house of cards. Finally, in business as in life, beware of messiah figures. This caution extends to the crypto community that, time and time again, rushes to build cults around leaders rather than to stay true to blockchain’s core ideal of decentralization.Jeff John Roberts, How SBF Fooled Everyone—Including Me.

There are good companies, and there are companies that look good. The two don’t always overlap. A good company can be ugly: slow to explain itself, awkward in a pitch, indifferent to the rituals of decks, demo days, founder mythology, and market maps. A good-looking company can be superficial: fluent in the precise idiom that unlocks meetings, memoranda, and term sheets, but with little behind the fluency.

This distinction matters because venture capital is not ordinary finance. It is finance conducted in the absence of most of the facts that later make a company legible. The public-market investor can inspect revenue history, margins, audited statements, comparable companies, and years of market behavior. The venture investor mostly confronts fragments: a founder, a prototype, a market claim, a few customers, a theory of adoption, and a story about a future that has not yet become measurable. This is why venture financing is surrounded by elaborate contracts, governance rights, liquidation preferences, board control, and staged rounds.

Where facts are scarce, signs become powerful. The founder biography, the enemy, the TAM slide, the crisp insight, the named secret, the demo that appears to compress the future into five minutes — these are not decorative features of venture culture. They are substitutes for evidence. This is why looking good has become a discipline of its own. The early-stage company must produce not only a product but a theory of itself persuasive enough to travel through a partnership meeting.

Venture capitalists themselves have long tried to simplify this into doctrine. Don Valentine’s Sequoia emphasized the size and force of the market — the problem being solved, not merely the polish of the person solving it. Marc Andreessen, borrowing from Andy Rachleff, gave the industry one of its most durable tests: product-market fit, the moment when a good market meets a product that can satisfy it. In Andreessen’s formulation, you can feel the absence of it: customers are not getting enough value, word of mouth is weak, usage is not growing, sales cycles drag, and deals fail to close. You can also feel its arrival: the market begins to pull the product out of the company faster than the company can push it. Paul Graham’s simpler command — make something people want — points toward the same discipline: the company must eventually be judged by desire outside the room, not eloquence inside it.

But before the market speaks, the room speaks. And the room has a grammar.

Peter Thiel codified one piece of this grammar in his book Zero to One: the contrarian truth, the secret the world has not yet noticed, the monopoly disguised as a startup. Andreessen Horowitz institutionalized another: the founder as protagonist, the firm as patron, software as the medium through which history is now written. The grammar works. It attracts attention, organizes belief, and moves capital. It can help a real company become legible before its evidence has fully arrived. But by itself, it is indifferent to whether anything beneath the language is true.

The failures make this easiest to see. Theranos looked impeccable: the board of statesmen, the black turtleneck, the aura of inevitability, the pitch that opened doors at the highest levels of American capital. But the machine on the bench did not work. The SEC later alleged that Theranos misled investors about its portable blood analyzer, which could perform only a small number of tests while the company relied heavily on modified or standard commercial analyzers. FTX looked like maturity arriving in a feral asset class: effective altruism, regulatory fluency, Washington tours, and a CEO who seemed to have read the right books. The SEC later charged Sam Bankman-Fried with defrauding FTX investors and concealing the diversion of customer funds to Alameda Research. Neither was a failure of looking good. Both were triumphs of it. What collapsed was the distance between the surface and the thing the surface claimed to represent.

So the question is not how a company learns to look good. That game has been mapped, taught, and absorbed. Most ambitious founders learn it whether they intend to or not. The harder question is the reverse: under what conditions does a good-looking company become good? When does the grammar stop functioning as costume and begin to discipline the reality beneath it?

One answer is that reputation has to become a resource rather than a mask. David Hsu’s study of competing VC offers found that entrepreneurs were much more likely to accept offers from high-reputation venture capitalists, even when those investors received equity at a 10 to 14 percent discount. This suggests that elite venture capital is not merely money; it is certification. A famous firm on the cap table can make a company easier to hire for, sell to, raise for, and believe in. Hochberg, Ljungqvist, and Lu similarly found that better-networked VC firms had better fund performance, and that their portfolio companies were more likely to survive to later financing and exit. In the best case, the company’s good-lookingness becomes infrastructure. The signal attracts the people, customers, partners, and investors who help turn the signal into fact.

But certification is double-edged. It can accelerate reality, or it can replace it. A prestigious investor can help a young company recruit the executive it could not otherwise hire; it can also make outsiders stop asking whether the product works. A board seat can impose discipline; it can also become theatrical oversight. A brand-name round can buy time to learn; it can also buy time to continue a fiction.

The difference is whether the surface creates obligations. Good-looking companies become good when the story they tell begins to make demands on the company telling it. The enemy in the deck has to become an actual enemy in the market. The secret has to become an operational advantage. The monopoly slide has to resolve into distribution, switching costs, proprietary data, regulatory position, network effects, or some other stubborn fact that survives contact with competition. The founder biography has to stop being symbolic capital and become temperament under pressure.

The research on venture capital’s “value add” is useful here. Hellmann and Puri found that venture-backed startups professionalize faster: they adopt human-resource policies, stock-option plans, marketing leadership, and other internal structures associated with mature firms. Bernstein, Giroud, and Townsend found evidence that VC monitoring can increase innovation and the likelihood of successful exit, using the introduction of new airline routes as a shock that reduced investors’ travel time to portfolio companies. In other words, venture capital can sometimes do more than select companies that already look promising. It can help install the internal machinery that makes promise harder to fake.

Still, the founder myth needs qualification. VCs say they invest in people, and early investors often behave as if the founder is the best proxy for the company. But Kaplan, Sensoy, and Strömberg complicate the familiar “jockey versus horse” debate. Studying venture-backed companies from early business plan to IPO and beyond, they found that business lines remained remarkably stable while management turnover was substantial. Their conclusion: at the margin, investors should place more weight on the business — the horse — than on the management team — the jockey. That result cuts against the most theatrical version of founder culture. A company becomes real not when the founder’s myth survives unchanged, but when the business can survive revision of the myth.

This is why time is so clarifying. In the beginning, almost everything can be performed. Urgency can be performed. Insight can be performed. Technical depth can be borrowed from the one impressive engineer in the room. Customer love can be simulated by curated anecdotes. Even traction can be arranged to look cleaner than it is. But duration is hostile to mere appearance. Customers renew or they do not. The product compounds or it does not. The best employees stay or quietly update their LinkedIn profiles. The market either begins to pull the company forward, or the company must keep dragging its own story behind it.

The recent history of venture financing has made this both easier and more dangerous. Ewens, Nanda, and Rhodes-Kropf argue that falling startup costs — especially after cloud computing reduced the cost of early experimentation — changed the venture model. Investors could fund more initial experiments with less capital, abandon more after the first round, and reserve deeper commitment for the few that generated positive information. Nanda and Rhodes-Kropf also find that hot venture markets fund riskier and more innovative startups: more of them fail, but those that succeed can be more valuable and more inventive. Capital abundance, then, does not merely fund worse companies. It funds more experiments. But it also creates more opportunities for companies to confuse experimental possibility with demonstrated truth.

A good-looking company becomes good when it allows itself to be corrected by the world. That is the threshold. As long as the narrative governs every fact, the company remains primarily an artifact of persuasion. But when the facts begin to push back — when customers use the product in ways the deck did not anticipate, when the best market turns out not to be the one the founder announced, when the true moat is less glamorous than the slide, when the board insists on evidence, when the company abandons a beautiful claim because the business has discovered something truer — then the company has begun to live.

The surface is not the enemy. A company needs language before it has proof; otherwise no one would join, invest, buy, or believe early enough for the proof to arrive. The danger is when language becomes a substitute for contact with reality. The discipline is to make every polished sentence answerable to something outside itself.

A real company is not one whose story was right from the beginning. It is one whose story can be revised by customers, constrained by governance, strengthened by reputation, disciplined by capital, and humbled by time. A good-looking company becomes good when the performance stops being a costume and becomes a promise the company is forced to keep.