Cory Doctorow on bursting the AI bubble: roots and risks

Cory Doctorow has published a book-length critique and argues that massive spending, persistent losses and new "reverse centaur" jobs have inflated a fragile AI bubble that could collapse painfully. He defines the problem, lays out the scale in dollar terms and warns of the social effects of machines that turn people into peripherals.

What does Doctorow mean by "reverse centaur"?

A reverse centaur is a machine head on a human body: a person reduced to a squishy appendage serving an uncaring automated system. Doctorow gives the concrete example of an Amazon delivery driver who sits in a van surrounded by AI cameras that monitor gaze and speech, score behavior and report workers for missing quotas. The driver remains necessary because the van cannot complete the last-yard task, but the system uses and exhausts the human while the AI takes operational control.

Doctorow contrasts that with being a centaur, where a human is usefully augmented by a machine. He allows there are many AI tools that can be centaur-like, but his central claim is that the industry is funding and designing tools in ways that create reverse centaurs instead.

How big and fragile is the AI bubble, by the numbers?

Doctorow cites global capital expenditure as evidence of scale and fragility: CapEx globally rose from $700 billion to $1.4 trillion. He points to high-profile, loss-making projects as symptoms: Meta "wasted $60 billion" on the metaverse, spent $150 billion on AI in the last three years and says it will spend another $150 billion this year. He notes the sector is "turning over $50 billion a year" and must replace all of its assets every 24 to 30 months.

He also highlights financial concentration, saying seven AI companies currently account for more than a third of the stock market and that they "endlessly pass around the same $100 billion IOU." On unit economics he is blunt: "AI is the money-losingest thing our species has ever done. We have never lost as much money as we’ve lost on AI," and he argues every AI customer and every generation of models currently loses more money for the companies building them.

Why it matters

Those dollar figures imply a structural risk: the material basis for the AI narrative is massive, ongoing capital flows that subsidize loss-making models and sprawling data centers. Doctorow warns that when the investment mania stops, many models will disappear because it will not be economical to keep the data centers running. The result would be not only a financial contraction but also social harm: workers forced into reverse-centaur roles, surveillance and automated enforcement in workplaces, and the political power of firms seeking endless growth despite poor unit economics.

Doctorow ties this financial picture to ideology. He says AI appeals to leaders because it promises systems that reduce dependence on other people, and that desire combines with capital markets demanding growth to produce technologies designed to replace or discipline workers rather than empower them.

What to watch

Watch capital flows and data-center economics: whether global CapEx contracts from the $1.4 trillion level Doctorow cites, and whether major firms follow through on planned AI spending (he calls out an additional $150 billion planned this year by one firm). Also watch whether companies begin to shut models or cut data-center spending; Doctorow predicts many models will disappear if those costs become uneconomical. Finally, monitor workplace rollouts: increased surveillance and quota-driven AI monitoring are concrete signs of the reverse-centaur dynamic he describes.

Doctorow presents the critique as both economic and moral: the bubble is held up by narratives and subsidies as much as by technical progress, and its deflation would expose both balance-sheet losses and the human costs of designs that make people peripherals to automated systems.