AI panic distracts from the real bubble threat

AI bubble – Cory Doctorow’s new book argues that today’s AI doomism misses the core problem: a massive, fast-growing tech bubble built on promises to replace workers. In an edited conversation, Doctorow says the answer isn’t waiting for AI to “go away,” but forcing a diff

For a lot of Americans, the mood around artificial intelligence has curdled into something like déjà vu—only faster. Nostalgia that once took “at least a decade or more” to kick in now arrives after just a few years. when people look back to a time “before the president was drenching us in crappy AI memes.” Kids already use AI for far more than school. nearly every company on the S&P 500 is experimenting. and layoffs have come in waves as bosses swap workers for chatbots. Even the pope has weighed in, decrying what he called “new forms of slavery” that AI threatens.

The negativity is overwhelming, and a new Pew Research poll captures how lopsided it feels. Only 16% of Americans predict AI will be positive for society; 40% expect it to be negative, 31% say it will be equally good and bad, and 13% are unsure.

Cory Doctorow’s point—laid out in his new book. “The Reverse Centaur’s Guide to Life After AI”—is that the panic is aimed at the wrong target. AI, he argues, is a tool. And like rockets and medications, it can be used to harm or to heal. The constructive case may be harder to find right now. but the possibility is still there—especially if people stop treating AI as an inevitable fate and start demanding control over how it’s used.

Doctorow. a sci-fi author. tech activist. and journalist. frames the problem less as a breakthrough in machine intelligence and more as a manufactured economic moment. He says the AI bubble is “the biggest tech bubble we’ve ever had. ” and that it’s built on the promise that companies can replace labor—so the labor issue and the bubble can’t be separated.

His skepticism comes packaged with history. He says society keeps forgetting the mechanics behind bubbles because. politically. it treats the past like it doesn’t matter—“object permanence of like a two-month-old. ” he puts it. losing “games of peekaboo to our policymakers.” In his telling. the material reason bubbles form isn’t mysterious greed alone. It’s the problem firms face when they “saturate their markets” and stop growing—because investors revalue share prices drastically when the future earnings story changes. Shares are “a claim on the future earnings,” he says, and claims on more growth are worth more. When growth stalls. executives compensated in stock lose leverage with the market—and companies can’t conjure money the same way; “You don’t need Charlize Theron in a bathtub explaining this to you. ” he says. contrasting stock creation with the hard-to-get cash companies need if they aren’t growing.

Doctorow’s criticism of today’s AI story is also a criticism of how people talk about it. He argues that slogans like “AI will take our jobs” or the claim that “someday AI will be smarter than humans. ” or even the definition of AI as “artificial intelligence. ” have become stock frames that don’t match the reality of what companies are selling. He calls his approach an attempt to “be a better critic”—not just about labor. but about the economics underneath the hype.

In the book’s organizing metaphor. Doctorow borrows a term from automation theory: a “centaur” is “a person who is assisted by a machine.” He flips it into what he calls a reverse centaur—“a beast with a horse’s head and a human’s ass”—and argues that’s close to what happens when humans get “ridden by tech.” In his view. Uber. Amazon. and other systems powered by AI algorithms can turn workers into the ones being controlled. rather than assisted.

If the bubble is the central force, what happens when it finally pops becomes the question people can’t stop asking.

Doctorow says the numbers have only grown worse while the world argues about the narrative. When he wrote the book. he says. the AI bubble was “$700 billion.” Now it is “$1.4 trillion. ” and he says it doubled in a year as spending rises faster than before—making the “rate and the size… both increasing.” “The only thing worse than a $1.4 trillion bubble is a $2.8 trillion bubble. ” he adds.

He agrees that the burst will be “horrible. ” but he rejects the idea that the worst outcome is guaranteed to be more austerity and less relief. He says the instinctive response to bubbles in the last 25 years has been austerity. and that austerity drives normal working people into “the arms of fascists.” In his view. austerity creates “political instability” and “economic precarity and misery” that “casts a long shadow across multiple generations.”.

He offers a different warning: the danger may be that policymakers only know one move. If austerity is the reflex, then the social infrastructure that bubbles consume doesn’t come back the way headlines suggest.

The stakes are bigger than individual companies. he argues. because firms rely on collective expertise built across entire workforces and supply chains. That “process knowledge” is “held collectively across the whole workforce and its supply chain. ” and he says it “can’t just be conjured back into existence. ” a point he links to the long. slow work of building industrial capacity—naming examples like the CHIPS Act and Build Back Better. both of which relied on years of cultivation rather than quick fixes. If society wants to maintain the “invisible infrastructure” that he describes as the foundation of modern industry. he says. it can’t keep responding with cuts.

Doctorow’s politics land on one question: whether leaders will do anything other than austerity. He doesn’t say the answer is likely. but he points to a surprising development in American political rhetoric—one he describes as alarming and also “somewhat heartened.” He says Donald Trump is making arguments for “doing a bunch of things” that Doctorow himself didn’t manage to convince people to support. including replacing what he calls “useless Democratic leadership” with “political radicals.” Doctorow goes further. saying Trump’s stance includes calls to “stop using the American internet” and “stop using oil. ” and that Doctorow believes Trump is actually pushing people away from oil and away from American technology.

That shift, in Doctorow’s telling, is part of a wider tendency to convince people that their progressive leaders “suck” and should be replaced. He ties the appeal of those messages to the kind of investment capitalism that his book attacks.

The billionaire question—especially Elon Musk—hovers over Doctorow’s argument. Doctorow says Musk becoming “the world’s first trillionaire” matters, but he immediately undercuts the significance of the number. The “trillionaire” label. he says. belongs in “finger quotes. ” because it is “paper wealth.” He calls Musk’s situation the “poster child for buy. borrow. die. ” arguing that Musk’s liquidity has been limited.

He also pushes back on how quickly the public accepts victories out of speculative bets. He says it’s striking that the world made a trillionaire out of someone who “has only failed for the last five years.” Doctorow describes the pattern as an era of “exploding rocket ships” and the Cybertruck. alongside Twitter and Grok—“the world’s most advanced money furnace. ” he says.

That money furnace, for Doctorow, doesn’t only affect civilian life. He turns to military uses of AI when he’s asked about Defense Secretary Pete Hegseth’s claims about Anthropic. Doctorow says Hegseth is “picking fights” with Anthropic over the use of AI in the military. and that Hegseth argues the companies’ technology is so powerful it poses a national security risk.

Doctorow’s answer is blunt: in his view. for people like Hegseth—and for the customers Dario Amodei hopes to reach—AI becomes useful when the job doesn’t need to be done well. He cites AI targeting in Gaza during what he calls the “genocide. ” saying that if Israel claims the system is calculating the killing of up to 100 civilians to get one terrorist leader. Doctorow argues the deeper issue is that “they’re not sure if there’s a terrorist leader.” In that scenario. he says the result is targeting “100 civilians.” He describes the method as an “accountability sink. ” warning that systems can be used to pretend that someone didn’t choose violence.

He doesn’t stop at the Gaza example. He describes the “robot wars” as already here. referencing a “recent story of a Black Hawk helicopter getting downed by an Iranian drone. ” and pilots being “rescued by a drone boat.” He also points to a New Scientist report saying Ukraine tested fully automated drones that entered “Terminator mode” and killed Russian soldiers without any human input; Doctorow says the test was “two years ago” even though it was reported more recently.

That timeline matters to him. He worries about the technological feedback loop where errors—he calls them “hallucinations”—intersect with systems designed for speed. In his fear scenario. a system is asked to kill quickly “at such a cadence that no one could meaningfully intervene to review the kill decisions. ” and he says this could lead to many more deaths.

But Doctorow also pushes back on a common cinematic worry. He says the risk isn’t only that AI will be “in charge of the nukes.” He argues that the larger problem is the existence of nuclear weapons. and that even if AI isn’t the direct controller. the weapon systems may be coupled to something “like AI” that still behaves in dangerously unpredictable ways.

He compares the modern threat to “WarGames. ” the 1983 film. saying the movie’s fear wasn’t that a computer decided to end humanity on its own out of gameplay logic. He says the fear was the real-world equivalent: automated systems mistakenly interpreting something like a flock of seagulls as missiles and triggering retaliation.

Doctorow then shifts to what he believes is a workable alternative: not a blanket moratorium. and not the kind of rules that could backfire by creating new surveillance tools. He says he is open to international treaties and a moratorium only as an idea—but he doubts it because he sees backfire patterns in youth social media bans that. in his view. function as “a thinly-veiled mass surveillance effort.”.

His bottom line is a commitment to existing constraints. he says: “at a bare minimum. the Geneva Conventions.” He argues that accountability should still attach to people who order or enable war crimes even if AI is involved. “I care that you blew up the hospital. ” he says. adding that those responsible should face war crimes tribunals “doesn’t matter if you ask the AI to do it. or if you did it all on your own.” In the modern era. he notes. consequences might amount to “a podcast. ” and he calls for more severe penalties.

When asked about China and the United States. Doctorow says they are both major drivers of AI. but he argues the technologies are used differently. He describes American AI efforts as potentially building “a mass surveillance system like China has. ” and explains how that kind of system works: always watching creates compliance through fear. not performance.

He returns, again, to the motive that he thinks fuels the entire AI machine. When he asks what investors bet on with a $1.4 trillion bubble. he says he sees a world where capital allocators think other people “aren’t quite real.” He describes the psychic turbulence of bosses who feel in control but suspect the toy steering wheel they’re holding is connected to a system run by others. In that framework. AI becomes a way to “wire the toy steering wheel into the drive train of the car. ” giving executives a shortcut to getting rid of workers without guaranteeing that replacement actually works.

Doctorow’s approach to the moral debate is equally skeptical of blanket condemnation. Asked whether it’s immoral to use AI. he says opposition to “statistical inference” is “weird. ” and he argues that calling AI immoral per se elevates the technology to an extraordinary status—one that he says is “foundational to the investment bubble.”.

He also argues that scraping—essential to the way modern AI models are built—is not a special moral crisis. “Scraping is good,” he says, adding that websites and web technologies have coevolved with scrapers since early internet days. He draws an analogy to publishers resisting libraries lending ebooks: publishers complaining about libraries would miss the point. he says. because libraries predate paper.

That’s part of why he resists a culture war framed as “AI good” versus “AI bad.” In his view, the real questions are who uses AI and who gets harmed.

Language, too, becomes part of the fight. Doctorow’s previous book, “ens**ttification,” has already become a meme, and he’s asked what it feels like. His answer is that semantic drift is the point, not the problem. He says people try to get him “mad because of the semantic drift,” urging him to scold colloquial uses. He rejects it. saying language works differently: if the term is used more widely. the odds rise that someone will look it up. learn the technical. political. and economic analysis behind it. and then move from talk into action.

His hope is that some of those readers will eventually get involved with the Electronic Frontier Foundation, which he has worked for for “25 years,” and which he calls “the most important and most effective digital rights group in the world.”

Doctorow’s broader argument. in the end. is less about waiting for an AI bubble to pop and more about what people do while it’s still inflated. He says leaving the bubble to “fester” increases the likelihood of catastrophe—his phrase is “an above-the-knee amputation.” The work. he argues. is to “save what we can save” and decide what comes next before the momentum locks in who pays the price.

For all the doom talk, he insists, the choice isn’t between resignation and panic. It’s whether the country treats AI as a political and economic battlefield—one where accountability, labor protections, and international war rules matter—rather than a magical force that can’t be shaped.

AI bubble Cory Doctorow United States politics labor data centers mass surveillance war crimes Geneva Conventions Pete Hegseth Anthropic Dario Amodei Elon Musk Donald Trump