Francesca K. Augustine – Working Draft July 26

In art school, learning to carve wood and print from it, I remember my professor telling us that in Japan a printmaker's apprentice would spend ten years learning to carve before being permitted to print at all. Ten years before the first print? I remember exactly what I thought: thank goodness that I am an art student in Chicago, and I can make whatever I want.

Twenty years later — twenty years of learning, slowly and sometimes expensively, the difference between a maker and a master — I sat listening to a teacher struggle to teach students a craft she had not given decades to herself. She struggled. The students struggled. Everyone was unhappy, and everyone blamed the medium. But the medium was never the problem. The problem was a truth so old it embarrasses us to say it out loud: you cannot fake it till you make it. You have to do the work, and the work takes the time it takes.

An entire economy has been built on the opposite premise. Much of the technology industry runs on a small set of maxims so thoroughly absorbed they are no longer experienced as claims. Fake it till you make it. Move fast and break things. Adapt or perish. Disrupt or be disrupted. Growth solves everything. These are treated not as one industry's house style but as the discovered laws of innovation itself — the way markets work, the way survival works, the way the future works.

Since I landed in the startup world after decades in the art market, I noted that these concepts are often presented as empirical truths. At first, I tried to understand it – as a newbie I did not reject the difference of the core values and belief systems outright – I wanted to understand them. But, the more I witnessed this dynamic play out, I recognized that there is a problem: the industry that produced them is too young to have tested anything against the long horizon. And that is the problem.

Venture capital is eighty years old. The internet startup is thirty. The social media platform is twenty. Generative AI is four. A sector that reinvents its own foundational assumptions every decade never accumulates enough history to watch its theories fail. Each generation of companies is dead or transformed before the long-term consequences of its ideology arrive, and the next generation starts the experiment over with fresh confidence and no memory. This is what zero introspection looks like at industrial scale: not a refusal to examine the premises, but a metabolism too fast for examination to be possible. The lesson never has time to land before the student is replaced.

The counterexample

Humans have been exchanging objects of symbolic value for roughly one hundred thousand years. Ochre, shell beads, carved figures — things whose worth was never their material but their meaning — moved between hands, across distances, through generations, before writing, before cities, before anything we would recognize as an economy. Commercial art markets in a form a dealer today would recognize — patrons, prices, provenance, reputation — go back at least twenty-five hundred years, to the workshops of Athens and the collectors of Rome.

Walk the ladder down from there and every rung is an order of magnitude: recorded civilization, five thousand years; commercial art markets, twenty-five hundred; industrial capitalism, two hundred and fifty; corporate management as a discipline, one hundred; venture capital, eighty; the startup, thirty; the platform, twenty; generative AI, four.

Now consider what the old system at the top of that ladder has actually lived through: the fall of empires, feudalism, mercantilism, the invention of capitalism itself, industrialization, two world wars — and, most instructively, a repeating series of technologies that were each, in their moment, confidently declared to be the death of art.

Photography was supposed to end painting; painting abandoned the burden of likeness, invented modernism, and grew, while photography itself was absorbed into the market it was meant to kill. Television was supposed to make the static image irrelevant; artists like Nam June Paik were making art out of television sets while critics were still drafting culture's obituary. Advertising was supposed to drown authentic images in commercial ones; Warhol walked directly into the flood and came out with the most consequential body of work of the late twentieth century. Video, the internet, and finally the phone that is also a camera — which put a lens in every pocket on earth and was going to render the trained image-maker obsolete. More photographs are taken every two minutes today than were taken in the entire nineteenth century, and the market for photographs made by artists has never been larger.

"Is art dead?" is itself one of art history's oldest recurring artifacts. It has been asked at every technological threshold for two centuries, and it has been proven significantly wrong every single time. The art market did not merely survive these transformations. It flourished through them.

What the case study actually proves

Here is the finding, and it directly contradicts the maxims.

The art market did not survive by adapting or perishing. It survived by doing something the startup playbook has no vocabulary for: it held its core values constant and adapted everything else. The values — that symbolic meaning is real economic value, that provenance and lineage matter, that judgment is developed over lifetimes and transmitted across generations, that stewardship of objects and reputations is the actual business — have been stable for millennia. The mediums, the tools, the venues, the formats have churned constantly. Oil paint was a new technology once. So was the print, the photograph, the video signal, the browser. The art world learned each one, pushed it, tested it past its intended limits, transformed it, and adopted it with caveats. Critical adoption: take the tool, refuse the ideology that ships with it.

That is what anti-fragility actually looks like in a hundred-thousand-year data set. Resilience does not come from perpetual reinvention. It comes from an unusual pairing: a stable core of values combined with promiscuous flexibility about means.

The technology sector has this configuration exactly inverted. It changes its values every decade — yesterday's gospel of open platforms becomes today's gospel of walled gardens becomes tomorrow's gospel of something else — while hardening its means into lock-in, proprietary formats, and infrastructure designed to be indispensable. Fluid values, rigid means. One of the oldest surviving markets on earth runs the opposite way: rigid values, fluid means. If longevity is the measure — and "adapt or perish" claims that it is — then the evidence says the tech industry's configuration is not the law of innovation. It is a bug that hasn't lived long enough to be diagnosed.

This is the real argument, and it is aimed well beyond the art world. "Move fast and break things" is not a hard-nosed empirical insight; it is what a system believes when its memory is too shallow to contain a counterexample. The art market happens to be the counterexample I know from the inside — thirty years inside it, which, not incidentally, is the entire lifespan of the internet startup. But the pattern is not unique to art. Universities, religious institutions, legal systems, cities: the structures that persist across centuries all show the same signature. Stable core, adaptive periphery, and a deep suspicion of anyone who proposes to break things in order to move fast.

That suspicion is earning its keep, because the maxims are no longer confined to products. Peter Thiel, one of the movement's founding investors, wrote plainly that he no longer believes freedom and democracy are compatible — and has spent the years since assembling backup citizenships and residences across multiple continents, most recently in Buenos Aires. It is the exit strategy of a man who treats nations the way his industry treats platforms: something to churn out of when the terms of service disappoint. Repair is for stewards; exit is for users. Ask the enthusiasts of these techno-sovereign futures the most basic survival question — where does the food come from? — and the grand vision goes quiet. In fairness, it is a hard question for anyone whose meals have been an uninterrupted service their entire life, provided first by Mom and then by an app. It is an argument that holds space but not water. Democracy is precisely the kind of institution this essay has been describing; centuries old, stable at the core, adaptive at the edges, and slow on purpose — slow the way a ten-year apprenticeship is slow, because some intelligence cannot be faked. And when capital markets pretend that mastery is irrelevant and scale is king, the cost lands on everyone else. To a worldview that has never held anything for longer than a decade, that slowness looks like failure. To anyone with a longer memory, it looks like design.

The price of belief

If this reads as too abstract, the maxims can be audited. Follow the money and ask a simple accounting question: when does the reward arrive relative to the value?

Uber was founded in 2009. It lost money for fourteen consecutive years — roughly thirty-one billion dollars in cumulative operating losses — before posting its first annual profit in 2023, kept alive across that decade and a half by more than twenty-four billion dollars in investor capital. And yet its early backers did not wait for the profit. First Round Capital's five-hundred-and-ten-thousand-dollar seed check was worth roughly two and a half billion dollars at the 2019 IPO — a nearly five-thousand-fold return. Benchmark's Series A investment of about nine million became roughly six point nine billion. Sequoia's six-hundred-thousand-dollar seed in Airbnb turned into a stake worth billions at listing. These are routinely celebrated as the greatest investments of all time, and notice what they have in common: the payout arrived at the moment of belief transfer, not the moment of value creation. Uber's seed investors were paid their billions four years before the company earned its first dollar. The returns were not denominated in profit. They were denominated in the successful sale of a story to the next holder.

This is not an aberration of the model; it is the model. Roughly three-quarters of venture-backed startups never return their investors' capital at all — the entire structure is engineered so that a handful of belief-transfer events pay for a graveyard. And in the boom years, around four out of five companies arriving at their IPOs were unprofitable at the moment of listing, a proportion last seen at the peak of the dot-com bubble, which is to say: the public was invited to buy the story at precisely the point where the story was least supported by earnings.

It helps, here, to be pedantic about words the culture works hard to blur. Revenue is everything that comes in the door, before any costs. Profit is what remains after the machine's own appetite is fed. They are not near-synonyms; they are opposite ends of the spectrum, and every increment toward profit is harder to fake. The startup era's rhetorical innovation was to climb the ladder upward — from profit to revenue, from revenue to bookings, from bookings to users, from users to growth — until "success" referred to nothing that had to survive contact with an income statement. Uber's revenue grew from under four billion dollars to over thirty billion while it accumulated those thirty-one billion in losses. By the old definition of a business, that is a machine for converting capital into losses at increasing scale. By the new definition, it was the triumph of the decade — because the reward system paid out on the revenue curve, and the revenue curve was magnificent.

And what did the thirty-one billion actually buy? The narrative says: a better product. The books say something else. It bought the ability to sell rides below cost, at planetary scale, for over a decade — a price no business that had to survive on its own economics could match. There is an old name for selling below cost to destroy competitors; antitrust law calls it predatory pricing, and it used to be understood as a market failure. The startup era renamed it blitzscaling and called it a methodology. Uber did not out-innovate the taxi industries of the world's cities; it outspent them, with other people's money, and small transportation businesses on every continent were crushed not by a superior product but by a superior burn rate. In New York, taxi medallions that had traded above a million dollars collapsed to a fraction of that, taking drivers' life savings down with them.

Then came the epilogue, and it is the part the mythology cannot digest. When the capital markets finally demanded profit, Uber obliged — by raising prices. The subsidy ended, and because the subsidy was the product, the moment it ended the supposedly obsolete competition began coming back. In New York today, yellow cab trips are up over twenty percent year on year, over a thousand dormant medallions have returned to service, and the regulated meter — a pricing technology essentially unchanged since 1907 — is routinely far cheaper than the surge-priced app for the same trip across town. Uber survives, profitably now, at margins comparable to a grocery chain. But read the sequence plainly: the company became profitable only by stopping the thing that made it dominant, and when it stopped, the old market was still there, still solvent, quietly competitive. The disruption was never a product event. It was a capital event, and capital events end.

There is one more tell, and every user of every platform knows it in their body: the fees. When growth slows and the market finally demands profit from a system whose engine was never value creation, the only lever left is extraction — the add-on charges, the tiers, the ads inserted into what was clean, the service quietly degraded while the price quietly rises. Everyone hates it; everyone building these companies does it anyway, because within the belief system it is simply "how business works." But notice what that phrase concedes. A business whose maturity consists of taking more while giving less has admitted that it never learned how to generate value in the first place — only how to first purchase, and then harvest, a captive market. Anyone who arrived in this world from outside it and asked the obvious question — this loses money; why is everyone certain it's winning? — was told they didn't yet understand how business works now. They understood perfectly. The reward structure was paying everyone else not to.

The scale of the wager has only grown. The largest technology companies — and yes, they are genuinely, historically profitable; that must be said plainly, because the argument does not need to pretend otherwise — are now pouring that profit into the next story at a rate without precedent. The four biggest American cloud companies spent roughly four hundred and ten billion dollars on AI infrastructure in 2025 and have guided to something approaching seven hundred billion for 2026, with credible analyses suggesting the buildout will consume on the order of ninety percent of their operating cash flow, financed increasingly by debt, including fifty- and hundred-year bonds. The AI companies this infrastructure is being built for post revenues that are a fraction of the capital being deployed on their behalf, and losses that are not. Analysts have begun noting the interlocking vendor financing among chipmakers, model labs, and cloud providers — arrangements that make end demand look larger and more independent than it is — and comparing them, in print, to the telecom equipment financing of 1999.

Perhaps the bet pays off. Bets sometimes do. But look at the structure honestly: extraordinary present profits, converted into infrastructure for a future that exists so far mostly as a promise, valued by a market whose reward system pays out on the transfer of conviction rather than the demonstration of value. This is a belief system sustaining itself with the proceeds of belief. The old market has a mechanism for this too — conviction, taste, the wager on an unproven artist — but it developed, over centuries, an entire immune system around it: provenance, connoisseurship, institutional validation, the long slow consensus of collections and canons, all designed to test whether the story survives contact with time. The young market has the conviction without the immune system. It has, in fact, dismantled every proposed immune system as friction. And a belief structure with no mechanism for testing belief is not a market. It is a faith.

The burden of maturity

None of this is romantic anti-tech nostalgia. The argument is not that the old system is sacred and the new one profane. It is that one of these systems has a hundred centuries of accumulated intelligence about surviving change, and the other has been treating it as a legacy system to be disrupted — a one-way feed, teacher to student, with the roles assigned backwards.

Flip the casting. Technology is not the protagonist of this story; it never has been. It is a tool — the latest in an unbroken lineage of tools that ancient markets have been learning, pushing, testing, and transforming since before recorded history. A hammer that requires the carpenter to change her craft in order to work is a badly designed hammer. The burden of adaptation falls on the tool, and the burden of maturity falls on the young.

And youth deserves to be named precisely, because the maxims have a developmental profile: it is adolescence. The founding mythology of startup culture is a high school story — the nerds, at long last, becoming the cool kids — and the industry has never noticed that this inverts the cast without revising the play. The values that won are still adolescent values: speed over patience, confidence over calibration, the certainty that the adults are obsolete and the rules exist for people who can't see what's coming. Adolescents are not stupid; they are frequently brilliant. What defines the stage is something more specific — the inability to gauge the size of what one does not yet know, paired with the conviction that whatever it is, it can't matter much.

In a person, this resolves on its own. Introspection arrives — almost always through crisis rather than curriculum, the failure that finally cannot be outrun — and we call the result maturity. We tolerate the adolescent process, even find it touching, because the blast radius is small: a family, a friendship, a semester. But startup culture never completes the arc, because the decade-long memory reset returns each generation to the same confident sixteen. The sector is not young. It is perpetually re-adolescent. And this particular adolescent does not have a bedroom and a curfew; it holds a position of economic totality: infrastructure, attention, information, labor — which means it is making a teenager's mistakes at civilizational scale, on people who never agreed to be part of anyone's growing up. The crisis that forces introspection will come, because developmental crises always come. The open questions are only the size of the tuition and who gets the bill.

The writing on the wall

What would it look like for technology to actually learn the lesson? Concretely: build digital infrastructure whose engineering requirements come from the long column — continuity, stewardship, transmission — rather than the quarterly one. For the first time, a digital primitive exists whose native properties belong there. A distributed ledger is, at its core, a machine for continuity: permanence, verifiable lineage, resistance to unilateral revision, survival independent of any single institution. Those are not startup values. They are archive values — the values of the registrar, the conservator, the estate.

The test is simple to state: the record must outlive its maker. The token outlives the company. If the firm that issued a work's provenance record disappears — and firms disappear; that is what firms do — the record persists, readable and verifiable, because it was never held hostage to a business model in the first place. That single design constraint inverts the entire startup logic. You cannot optimize for lock-in when your founding requirement is that nothing depends on you. It is, perhaps, the first piece of the technology stack designed with the temporal assumptions of a cultural institution rather than a product roadmap — the first evidence that the sector can introspect without waiting for the crisis. That the adolescent can, in fact, grow up.

The software industry is not the model. It is the newcomer — brilliant, as newcomers often are, and standing in front of systems that have outlasted empires. Its maxims have been running unexamined for thirty years because nothing in its own history was old enough to contradict them. The contradiction was always available. It was hanging on the walls the whole time. Ask how it got there.