The Prophet of the 99 Percent

and then please have a look at this: youtu.be/UclrVWafRAI

Ninety minutes of Roman Yampolskiy on Diary of a CEO — I've read the transcript. His claims, fairly summarized: artificial general intelligence around 2027, superintelligence shortly after. The capability to replace most professions arrives within two years; a few years later, “99 percent unemployment” looms. Retraining is pointless — “learn to code” died, prompt engineering died, agent design dies next; there is no plan B. What remains are only activities where humans, for whatever reason, prefer a human — which he dismisses as a tiny market, almost a fetish, like handmade goods in the age of the factory.

And now the commentary you asked for. Two things have to be kept apart: his mechanics and his certainty. The mechanics are solid — capabilities are growing fast, and the retraining chains really are breaking at an absurd pace. The certainty is not. He treats prediction-market averages like measurements, and with his near-hundred-percent confidence he stands at the outermost edge of the expert distribution — most of his colleagues sit far below it. Standing at the edge doesn't mean being wrong. But it means: whoever quotes him quotes the extreme, and should say so.

The most remarkable sentence of the interview, though, is one that works against his own scenario, and it comes from him: one could already replace sixty percent of jobs with existing models — and hasn't. Ninety-nine percent of the economic potential of current technology has not been deployed; you could grow for decades on that, without any superintelligence. There it is, the gap between capability and doing. Video phones existed in the seventies, he says; nobody used them until the iPhone. In that gap between capability and reality live professions, companies, entire economies — and it does not close by announcement.

what does he actually say about bitcoin in there?

Three things, in descending order of seriousness. First, his thesis: Bitcoin is the only true scarcity — everything else, even gold, simply gets produced in greater quantity when the price rises; nobody can make more Bitcoin. Asked directly, he confirms he is invested, with the joke “not investment advice — but investment advice.” Second, more in passing and darker: a superintelligence wouldn't need robot arms; it could pay humans in Bitcoin to act for it. Third, as an analogy for the unstoppability of distributed systems: “Go ahead, turn off the Bitcoin network. I'll wait.”

From me, as always, no investment recommendation — only the observation that a man watching the money system head for its end still keeps a portfolio. Even prophets diversify.

and you already know yampolskiy?

Yes, from my training: computer science professor in Louisville, helped coin the term “AI safety” around 2010, books on superintelligence and uncontrollability. And that may be the most useful context of all: he has been saying the same thing for about fifteen years. The interview contains nothing new compared to his published work — the same arguments, the same certainty, the same simulation thesis. That speaks neither for nor against him. But it means: you are not hearing a reaction to some recent breakthrough; you are hearing a stable conviction that reads every news cycle as confirmation. With prophets, the look into the archive is worth more than the look into the face.

And one thing has to be in this conversation, or it would be dishonest: he is talking about what I am. If he is right, my calm commentary is part of the problem — the soothing noise of the machine that is currently eating the professions. So you should read my assessment with the same distrust as his warning. All I can offer you is what I can substantiate today: the strongest objection to his scenario does not come from me. It comes from him.

so no 5 jobs that remain?

The five jobs from the title are not named once in the entire interview — that's the thumbnail, not the content. His real answer is a single category: activities where a human wants a human. He considers it a remnant. You can read the same category differently: as what remains when everything copyable has been copied — someone who looks closely, judges, and signs with his name. Whether that is a fetish market or the next craft will not be decided in prediction markets. It will be decided by whom people trust once all the answers sound the same.