Cory Doctorow on Musk and the AI bubble

The reverse centaur

In automation theory a "centaur" is a person assisted by a machine. Doctorow's book, _The Reverse Centaur's Guide to Life After AI_ (Verso), is named for the inversion: a human "conscripted into acting as an assistant to a machine". His examples are deliberately unglamorous. The warehouse worker urinating in a bottle to hit algorithmic fulfilment targets. The person sat in a self-driving truck on minimum wage to stop it crashing. The lawyer on reduced money checking Gemini's grasp of precedent.

The conjuring trick

Doctorow's line on capability is blunt: AI "cannot and will never render us obsolete". A machine that predicts the next likely word has been mistaken for a mind because, as he puts it, "in a state of nature, we don't encounter sentences that don't have sentence writers". He reserves particular contempt for the word "hallucination", which he says is useful only to describe what _we_ do when we impute a writer to word salad.

The numbers

- When he wrote the book it was a $700bn* bubble. At the time of the interview it was **$1.4tn**, and he expects *$2.4tn.

- Nine US tech companies account for 35% of the entire US stock market valuation.

- He cites Stein's Law (anything that cannot go on forever eventually stops) against Keynes (the market can stay irrational longer than you can stay solvent). Bubbles are easy to predict, timings are not.

Why capital keeps paying

The motive, Doctorow argues, is older than the loom. "The one thing a boss does not want is co-determination." AI is the promise of "products without product designers. Workplaces without workers, screenplays without screenwriters, movies without actors, hospitals without doctors and nurses." It does not have to work to attract investment, any more than face cream has to work.

He is equally hard on the critics, borrowing Lee Vinsel's term criti-hype, meaning critique that feeds off and feeds back into the hype. And he notes our collective amnesia. Crypto was going to replace global finance, the metaverse was going to replace tourism. "We have such poor object permanence."

Why it matters

The framing here is useful for anyone trying to think clearly about AI without picking a side in the doomer-versus-booster shouting match. Doctorow separates three things that usually get mashed together: what the technology can actually do, what it is being sold as, and who benefits from the gap. His argument about Grok and non-consensual imagery makes the point sharply. Ban the harm, yes, but "if we want to actually target the harm that Musk is doing, we have to make his investors think he can't make money."

Key takeaways

- The threat is financial and industrial, not existential. A $1.4tn bubble whose blast radius lands on everyone except the capital allocators.

- "Criti-hype" is a trap. Overstating AI's powers in order to attack it does the marketing for free.

- Automation's appeal to bosses is about ending co-determination, not about cost or quality.

- "You cannot make billions of dollars without hurting lots of people. And you can't hurt lots of people without, in some sense, believing that they're not really people."

Photograph by Jessica Pons for The Guardian.