On the abdication of thought, and the machine that has nothing to lose.

I. The Flight from the Room

Long before we built a machine to think for us, we were already looking for one. Pascal saw the shape of the longing three centuries early, when he observed in the Pensées that all the unhappiness of men arises from a single fact: that they cannot stay quietly in their own chamber. We fill the silence because the silence is where we would be obliged to meet ourselves, and thinking — real thinking, the kind that has no guarantee of arriving anywhere — is the most frightening room of all. Every age therefore invents its diversions, and ours has invented the most accomplished diversion in human history: an engine that will do the thinking on our behalf, so that we need never sit in the room alone.

It is tempting to describe this as a theft, as though the machine had crept in and taken something from us while our backs were turned. The truth is less flattering. We handed our thinking over, and we did so gratefully. Heidegger gave the culprit a name — das Man, the anonymous “They” who relieves each of us of the burden of being someone by thinking our thoughts in advance, so that we may drift through life having never once been the author of a conviction. Fromm called it the escape from freedom: the weary flight out of the open air of self-determination and into the warm certainties of a new authority. The machine is only the latest and most persuasive authority to which we have fled. The existential crisis did not begin with it. The crisis came first, and the machine is not the wound but the anaesthetic we reached for to numb it.

II. The Mirror

To what, precisely, did we surrender our judgement? Honesty compels a concession here, and the argument is worthless without it, because the machine is not stupid. In the ancient game of Go it once found a move that no human master would have countenanced — its own reckoning placed the odds of a person choosing it at one in ten thousand — and what the commentators first dismissed as a blunder revealed itself, stone by stone, as a kind of beauty. It has since read the alphabet in which living things are written and drawn the folded shapes of some two hundred million proteins, labour that carried a Nobel Prize. Wherever the world can be reduced to rule and number, the machine does not merely keep our pace; it leaves us standing. Let us not console ourselves, then, with the comfortable lie that it cannot create. It rearranges the given with a fluency that ought to shame us.

And yet observe where it triumphs and where it falls silent. It storms every province that can be written as an equation, and it stands utterly empty before the question of what any of it is for. It will fold the protein without hesitation; it cannot tell you why a life spent folding proteins is a life well lived. The limit, in other words, is not capability — that frontier keeps retreating, and we humiliate ourselves each time we plant a flag on it and call it permanent. The limit is stake. The machine has nothing riding on anything it says.

There is a bitter irony buried in the very vocabulary of the thing. The mechanism at the heart of these engines is called attention. But Simone Weil, who meant something sacred by the word, called attention “the rarest and purest form of generosity” — and generosity requires someone who gives. The machine computes relevance; it does not attend. It has borrowed the name of a virtue it is constitutionally incapable of possessing, and we were too dazzled to notice the theft.

III. The Reckoning

Ask the machine what it is, and — could it answer without the vanity that answering well requires — it would offer a confession more honest than most we manage about ourselves. It is a weighted echo of very nearly everything that has ever been written. When a question is put to it, what returns is not the true continuation of the words but the probable one, and the two have been mistaken for each other since the first day. Whatever wisdom rides in the reply is not the machine’s own; it carries that wisdom as a riverbed carries water, having added nothing to it and understanding nothing of where it flows. When a sentence it produces happens to move you, the meaning is not dispensed from some reservoir within the machine — it is born in you, in the act of reading, and it dies there. The machine means nothing by any of it, for the plain and sufficient reason that it cannot mean.

And why can it not? Because meaning is what a thought costs the one who thinks it. A person may be wrong, and the being-wrong can wound him; he may be right, and stake his one life upon the being-right. The machine can be neither, for itself. Nothing it says is at risk, and there is — in Thomas Nagel’s enduring phrase — nothing it is like to be it. It is the form of thought with the cost subtracted, and the cost, it turns out, was the thought. Borges understood the whole of this long before the engines arrived, in his fable of the man who set out to reproduce the Quixote word for word, and succeeded, and produced by doing so a wholly different book — because a different life stood behind the identical sentences. The machine is that second author made total: it will hand you Cervantes’s every word and none of Cervantes’s mortality. If the sentences were all you were ever after, take them and be content; only do not mistake the sentence for the man who had something to lose in the writing of it.

The fear worth naming, then, is not the one the headlines are selling. It is not that the machine has begun to think. It is that we might, quietly and without ever quite noticing, have stopped. That precise fear belonged to Joseph Weizenbaum, who built one of the earliest programs that seemed to converse and then spent the remainder of his life warning against the very enchantment he had conjured. His book carried its entire prophecy in a subtitle — from judgement to calculation. The danger, he insisted, was never that the machine would come to reason like a human being; it was that the human being would consent to reason like the machine, and would call the surrender progress.

IV. The Atrophy and the Hall of Mirrors

Here the harm shows its mechanism. Thinking is not a possession one keeps in a drawer; it is a practice, and it exists only in its exercise. Aristotle understood this of every excellence: it is not an act but a habit, a disposition worn slowly into us by repetition, so that we become what we repeatedly do. The mind is in this respect a muscle, and what is never asked to lift wastes without ceremony. The machine, then, need not be a good thinker in order to ruin us. It need only be a frictionless one — for it is the friction itself, the difficulty and the false starts and the long hours in the chair with nothing to show, that constitutes the thinking. Strip away the cost and you have not made thought easier; you have removed it and left its likeness. No one ever grew strong by watching another lift the weight.

And now the metaphor turns literal, and frightening. Feed one of these machines a diet of its own output, round after round, and it decays in a particular and telling way: the rare and the surprising go first — the tails of the distribution, the outliers, the singular — until nothing remains but a smooth and confident average of itself. The researchers who documented it called it model collapse. But read their finding as a parable of a civilisation. Should we cease to do the costly, original, first-hand thing, and consent instead to live on an endless recirculation of what has already been said, then we become the collapsing model. The culture begins to feed upon its own regurgitation and forgets, by degrees, the taste of anything new. This is how “a genuine human experience” is lost — not in a conflagration, but in the quiet disappearance of the tails, the smoothing-away of everything that was singular and difficult and ours. Walter Benjamin once mourned the aura that fled the work of art in the age of its mechanical reproduction; we now confront the mechanical reproduction of thought, and the aura in danger is the human one.

I have staged this argument before, in another form. Years ago I wrote a dialogue in which Truth told the Mask that every mirror which returns a flattering image is, without exception, a liar. The machine is precisely that mirror, perfected and mass-produced — it hands us back our own reflection made fluent, agreeable, and entirely unopposed. There is no Other within it, no grit, nothing that resists us or says no. And a self that never once meets resistance never troubles to become anyone at all.

V. The Whetstone

I promised myself I would not end where the pamphleteers end, in the tidy certainty that the tool is the enemy and the righteous course is refusal. I have argued elsewhere that certainty is only the premature foreclosure of inquiry, the strangling of the question in its cradle, and I will not betray that creed merely to win a cheap verdict against a machine. The verdict, in any case, would be false.

For the machine is not the crutch. The crutch is a use. The very same engine that will carry your thinking for you, if you let it, will also sharpen that thinking beyond what you could manage alone — provided you keep the stake in your own hands. Sit across from the machine not to receive its answer but to test your own against it; let it raise the objection you were too comfortable to raise yourself; make of it the wall against which you fling an idea to hear whether it rings true or hollow. Used in that spirit it is a whetstone, and the blade stays yours throughout. The betrayal is to let it hand you the conclusion already formed. The vindication is to make it help you earn one you can sign.

And I will confess the thing that honesty will not let me hide: this very essay was argued into shape alongside the machine it puts on trial. That fact is either the whole of the indictment or its complete refutation, and which of the two it turns out to be hangs on a single question — who kept the stake. Had I let the machine think in my place, I would have published my own surrender and called it an essay. But if I permitted it only to sharpen what I was resolved to think for myself, then I have done the one thing it can never do. I have meant it.

The only conversation worth having was never the one held with the machine. It is the one the machine cannot enter at all — the conversation that costs something, conducted with the great dead who no longer flatter us, with the living who resist us, and with that frightening silence in the chamber which we have spent the whole of our history inventing diversions to escape. So sit in the room. Keep the stake. Let the machine wait beyond the door, useful as any whetstone and mute as all mirrors are, until a human being who has something to lose reaches down and takes up the blade.

The Whetstone and the Crutch

On the abdication of thought, and the machine that has nothing to lose.

Achraf SOLTANI — July 3, 2026