Cory Doctorow's AI bubble argument took a sharper turn Friday: in a new Pluralistic essay, the author and technology critic said the defining feature of the current AI boom is not the software, but the financing story around it.
Doctorow wrote that AI, treated as a technology, is ordinary: useful in some places, abusive in others, and comparable to other software add-ons. His target was the investment frenzy around it, which he described as an “extraordinary” economic pathology built on claims that AI systems will replace large numbers of workers while still producing acceptable work.
According to Doctorow, AI boosters have spent more than $1.4 trillion, most of it in the past year, chasing that premise. He argued that the money has helped push data center construction, mass scraping, AI features stuffed into existing products, and corporate efforts to replace employees with chatbots.
What did Cory Doctorow say about the AI bubble?
Doctorow argued that some AI investors may believe the technology will transform the economy, while others may only believe that other people will keep buying the story long enough for them to profit. That distinction is the spine of his essay: AI “solipsists” believe other people can be reduced or replaced, while AI “cynics” bet that managers and investors will act as if they can.
He connected the first group to what he has previously called “billionaire solipsism,” the idea that extreme wealth encourages people to treat workers, users, and victims as less real than themselves. In the essay, Doctorow cited Elon Musk’s use of “NPCs” for critics, Leona Helmsley’s “little people” tax remark, Jeff Bezos’s warehouse and delivery workforce, and Mark Zuckerberg’s interest in chatbot-mediated social platforms as examples in his broader critique.
The second group, in Doctorow’s telling, does not need to believe that AI can do the work it is being sold to do. It only needs to believe that executives can be persuaded to fire workers and buy AI systems anyway.
How does the Keynesian beauty contest fit in?
Doctorow invoked John Maynard Keynes’s “beauty contest” model of investing. The idea is that speculators may not choose the asset they personally think has the strongest fundamentals; they choose the asset they think other investors will reward, then try to sell before sentiment turns.
Applied to AI, Doctorow wrote, the speculative question becomes whether an AI salesperson can convince a boss that software can replace staff. If enough executives believe that pitch, valuations can rise even if the underlying systems perform poorly.
He also pointed to a Fortune report about SpaceX stock, index funds, passive investing, 401(k)s, Nasdaq 100 and Russell index inclusion to argue that retail retirement money can be pulled into assets ordinary investors did not actively choose.
What risks did Doctorow identify?
Doctorow warned that the costs of a collapsed AI boom would not be limited to investors. He argued that seized or cleared land for data centers could be left with unfinished facilities, that skilled workers displaced by chatbots may not be available when those systems fail or disappear, and that energy use tied to AI infrastructure will leave environmental damage behind.
He also claimed that seven AI companies account for 35% of the S&P 500 and warned that a crash in those stocks could invite austerity politics. That claim is part of his argument, not a market forecast verified in the essay.
The post ended with Doctorow’s central distinction intact: AI may be a normal technology, but the bubble around it is dangerous because financiers can profit from belief itself. The code is only one part of the machine. The sales pitch is doing plenty of work.
This story draws on original reporting from Pluralistic.
