How AI Made Monomania Look Like Progress

A recent investigation accused Delve, a YC-backed startup run—as so many YC-backed startups now seem to be—by people barely past adolescence, of rapidly producing fake compliance certifications for startups eager to close enterprise deals. In a December TikTok interview, Delve's co-founder carried herself with the earnest authority of someone newly entrusted with it. She corrected her interviewer on the company's valuation. Not $200 million, but $300 million. She did so politely, almost fastidiously, as though restoring some small but essential truth.

When asked about the worst advice she had ever received, she answered without hesitation: "Sheer grit and hard work won't take you there…if your ladder is against the wrong tree, you're not gonna end up at the right destination." What mattered, then, was not effort but orientation: finding the right tree. Place the ladder correctly and you might approach what she called, without any visible irony, your "global maximum."

It was meant as advice, but it had the feel of doctrine. Somewhere, implied but never named, stands the Correct Tree, waiting to justify the climb. I'm sorry, Sweetheart, but there is always another tree—taller, straighter, and more promising. Forget your "global maximum."

I have spent the past year among people who call themselves founders, a term that feels at once inflated and strangely exact, like "visionary" or "thought leader." These are founders of companies, yes, but also of enclosed systems of belief in which the product appears not merely useful but inevitable. They speak of markets as though they were territories to be discovered, of users as populations to be managed, and of growth as something closer to fate than contingency. Beneath the language lies the same assumption: explanation can be deferred. The question of why something should exist will resolve itself later, preferably after Series C.

There is nothing especially new in this posture. AI has only sharpened it. Melville caught it in Ahab: the exquisite narrowing of the world until only one object remains, until the self becomes an instrument calibrated toward collision, whatever the cost. What was once called monomania now passes, in the age of AI, as focus, genius, and conviction.

Founders speak a version of this language, though they would resist the comparison. "If we don't build it, someone else will," they say, as though inevitability were a form of modesty. More often it functions as absolution. The question of whether something should exist dissolves into the assumption that it will, rendering the future a kind of alibi.

In Ahab's case, the whale ceases to be an animal at all. It becomes a surface onto which necessity, injury, and grievance are cast, until obsession begins to resemble purpose. The rest of the world—his crew, the dim memory of wife and child, even the plain fact of consequence—recedes. You see the same pattern in San Francisco startup culture. An idea begins modestly enough: automate this, connect that, remove some minor friction. Then it gathers funding, which is to say conviction. Then it gathers language, which is to say inevitability. It is no longer a tool but a solution, no longer elective but necessary. Questions of purpose, use, and proportion give way to the single question of scale.

The difficulty is no longer what can be built. Given sufficient capital and computation, nearly anything can now be made, including things that not long ago would have required entire teams of software engineers. The difficulty lies elsewhere: the capacity to build has outrun the capacity to ask, with any seriousness, what it is we are building for. Imagination has become, almost imperceptibly, its own justification, and the mere fact that something can be described begins to pass for proof that it ought to be made. What falls away, in the process, is any tolerance for uncertainty.

Melville offers, in passing, another figure for that uncertainty: Bulkington, who prefers the instability of the open sea to the false security of land. The sea offers no fixed trajectory and no guarantee of arrival, only exposure. It neither comforts nor pretends to. There is, perhaps, a kind of honesty in that—an honesty in remaining exposed, in resisting the seduction of false direction.

But that is precisely the condition this culture is designed to escape. The work begins to justify itself, and the language shifts to accommodate it. Small distortions—an exaggeration there, an omission here—come to seem incidental, then necessary. The end, undefined but assumed, begins to consume the means. One becomes, as Ahab names it, "madness maddened": aware enough to perceive the pattern, but not free enough to leave it.

There is, in this condition, a strange sympathy. Certainty exerts a pressure of its own, until hesitation begins to feel like failure. Even those who sense the danger find themselves unable to step away, carried along by the very motion they distrust, and by the money, or the promise of money, on which they have learned to depend.

And once that motion becomes collective, it acquires a flattering name. In contemporary terms, it is called progress: framed as unstoppable, and therefore good. Resistance becomes indistinguishable from irrelevance. These are consoling ideas because they remove responsibility. The ship appears to sail itself.

But it does not. What is built does not remain contained. It expands, integrates, and becomes part of the structure others must inhabit. A tool assembled quickly, if believed in strongly enough, becomes difficult to refuse.

At that point the deferred questions return in another form. What is this for—not in intention, but in effect? What does it permit, and what does it foreclose? To think teleologically is not to renounce ambition. It is to resist confusing compulsion with purpose. The two often feel identical from the inside. But they are not.

Artificial intelligence does not resolve this tension so much as erode the pause in which it might be examined. By making creation easier, it makes justification easier to neglect. Answers accelerate. Questions linger, shoved under the bed, kept carefully out of sight.

The TikTok interview with Delve's co-founder remains memorable precisely because it is not exceptional. Her protective confidence, her crude metaphor, her naïve assumption that a life can be oriented toward a single optimum: all of it is entirely ordinary. She is right, in a limited sense, that effort alone is insufficient. But she is mistaken, I think, in believing that success can be located so precisely. That mistake is common now, perhaps because it is flattering. It allows intention to stand in for judgment.

While intention is plentiful, foresight is rarer. The trouble is that good intentions so often feel like a sufficient defense, when they are really no defense at all. The only safeguard available is a form of vigilance that resembles hesitation: the capacity to stop and ask whether the path one follows is chosen, or merely inherited from the logic of motion. It is not an attractive practice, which is one reason it is so easily deferred. It does not scale or attract funding, and it feels inefficient, resistant to narrative. It is, nevertheless, the only interruption available.

Otherwise one wakes to find oneself high above the ground, surrounded by a forest without edge, each tree indistinguishable from the last, each ascent promising—again—that this time, finally, the climb will explain itself.