Last year, we featured a lengthy interview with tech journalist and science fiction author Cory Doctorow about his book, Enshittification: Why Everything Suddenly Got Worse and What To Do About It. Doctorow is back with a provocative new book that serves as a follow-up of sorts, focusing on AI and related issues: The Reverse Centaur’s Guide to Life After AI.

Doctorow does not enjoy talking about AI, but he is constantly asked to comment on it. “I made the tactical error of being sick of talking about AI,” Doctorow said. “So I wrote a book about why I think it’s a dumb thing to keep asking people to talk about, and now I have to talk about it.” Reverse Centaur is his attempt to “sort out the bullshit from the material reality.”

The Reverse Centaur Concept

In automation theory, a “centaur” describes a human augmented with a technology—like machine learning, or even just using autocomplete. A reverse centaur, Doctorow explains, “is a machine head on a human body, a person who is serving as a squishy meat appendage for an uncaring machine.” He gives the example of Amazon delivery drivers surrounded by AI cameras monitoring their every move, effectively a peripheral to the delivery van. Being a centaur is generally viewed as positive; few relish being a reverse centaur. Yet the AI industry seems intent on using tools to create more reverse centaurs.

It is one thing to incorporate AI into medicine to help radiologists process X-ray images and spot tumors they might miss. It is quite another to fire nine out of ten radiologists and let AI make diagnoses, with the remaining radiologist solely responsible for checking its work—and taking the blame for errors. This distinction lies at the heart of Doctorow’s critique.

The Bubble’s Scale

Doctorow is not anti-AI; he uses AI tools regularly. But he is alarmed at the hype, capital expenditures, unrealistic expectations, and potentially catastrophic economic consequences when the bubble pops. “The bubble doesn’t want cheap useful things,” he says. “It wants expensive ‘disruptive’ things: big foundational models that lose billions of dollars every year.”

Global capital expenditure on AI has grown from $700 billion to $1.4 trillion. Meta wasted $60 billion on the metaverse and $150 billion on AI in the last three years, planning another $150 billion this year. Seven AI companies account for more than a third of the stock market, passing around the same $100 billion IOU. “AI is the asbestos in the walls of our technological society, stuffed with wild abandon by a finance sector and tech monopolists run amok,” Doctorow says. “We will be excavating it for a generation or more.”

Why AI Appeals to Business Leaders

Doctorow argues that AI appeals to a fantasy of a world without people. This is especially attractive to corporate leaders haunted by the secret fear that if workers don’t show up, everything shuts down—but if bosses don’t show up, things run fine. “If that’s the case,” he says, “AI will let them wire the toy steering wheel directly into the drivetrain.” They can have an amazing idea without ego-shattering confrontations with experts who tell them they are idiots: “You just type some stuff to the chatbot, and it shits out your product.”

He contrasts this with earlier technological breakthroughs that workers actually demanded. The web became profitable because every new user made the web less unprofitable; every generation of web technology improved margins. “That’s the opposite of AI,” Doctorow contends. “Every AI customer loses money for the company, every use of AI by that customer loses money, and every generation of AI loses more money than the last. AI is the money-losingest thing our species has ever done. We have never lost as much money as we’ve lost on AI.”

Workers and Centaurs

Some workers find AI useful—they become centaurs who decide how technology assists them. Others become reverse centaurs, forced to produce more at the expense of quality and their own well-being, serving as “accountability sinks” when the AI screws up. This explains why some workers say they love AI while others hate it.

Doctorow insists he is not fundamentally anti-AI. He argues that scraping the web to train models is not inherently bad—indeed, it is socially beneficial. Rather than demanding new copyright laws that could be co-opted by big media companies, he advocates for new labor laws. The only workers who have beaten AI are Hollywood screenwriters and actors, because they are exempt from Taft-Hartley’s prohibition on sectoral bargaining—where all workers in a sector bargain with all employers. Extending sectoral bargaining to all workers would give them the power to shape how AI is deployed.

What Will Remain After the Burst

When the AI bubble bursts, Doctorow predicts a useful residue—much like after the dot-com bubble. “When the cryptocurrency bubble bursts, all that’s left are shitty monkey JPEGs and worse Austrian economics. But when AI bursts, you’ll be able to buy GPUs for pennies on the dollar. You’ll have your pick of applied statisticians who have interesting ideas but are stuck building what their bosses want. Open source models will be barely touched.”

He cites DeepSeek, a spin-out of a Chinese hedge fund given $6 million to play with open source models: they produced a model so good on commodity hardware that the market sold off $600 billion in 24 hours—the largest single-day decapitalization in history. Cheap hardware, skilled statisticians, and open source models make a better foundation than massive, capital-intensive foundations.

Doctorow also pushes back on “AI is coming for your job” messaging. “We have to distinguish between the AI doing your job and the AI being incapable of doing your job, but your boss is such a sucker that he fires you and replaces you with the AI anyway. There’s infinite evidence for the second one, very little for the first.” He points to Amazon’s “just walk out” stores that turned out to rely on people in India watching cameras—not AI. “The demos for AI have just turned out to be people in India pretending to be robots.”

Realistic Uses for AI

Doctorow himself uses AI in practical ways: Whisper to transcribe audio, local models to find typos in blog posts, or conversational interfaces to search media archives. “AI doesn’t have to be 100% accurate for that to be useful. It doesn’t have to be free from false positives. It can just be OK.” He also praises projects like the Human Rights Data Analysis Group (HRDAG), which uses AI to prioritize arrest reports for the Innocence Project of New Orleans, helping free wrongly convicted people. “That’s just good, and the proof is in the pudding.”

The key, Doctorow concludes, is not to be seduced by the hysteria of an AI jobpocalypse. The bubble will burst, and when it does, we must ensure the residue is used for socially beneficial ends—not for creating more reverse centaurs.

Source: Ars Technica News