"How to Think About AI": Cory Doctorow on Big Tech, Understanding AI, Labor Automation & More

### TL;DR Cory Doctorow argues that the real driver of the current AI bubble is not technological progress, but capital's attempt to replace well-paid workers with low-quality algorithms, suppress labor costs, and extract profits. Workers need collective bargaining power—like Hollywood screenwriters—rather than more copyright to protect themselves. ## The Key to Understanding AI: Who Does It Serve? As a science fiction writer, Doctorow emphasizes that when thinking about AI, we must ask: "Who does this tool serve? Who does it work for?" rather than simply describing the technology itself. The current AI industry has invested $1.4 trillion, but annual revenue is only about $50 billion (and Microsoft's $10 billion to OpenAI is not real revenue). The unit economics are terrible—every additional customer or use makes AI companies lose more money. This is the exact opposite of the early internet's "improving marginal profit per user," forcing capital to recoup its investment only by mass-firing high-paid workers, replacing them with low-quality algorithms, and forcing society to accept inferior products. ## The Radiology Case: Augmentation or Replacement? Using radiology as an example, Doctorow points out that AI can indeed help detect tumors that humans often miss. But the real goal of today's AI companies is: replace 10 radiologists earning $3 million each with a $1 million/year chatbot, saving $2.7 million, leaving one doctor to "grade the AI's homework" and become a black hole of liability. Medical workers (radiologists) have market power due to supply shortages, so they can decide how to use AI tools (e.g., as assistance rather than replacement). However, Doctorow warns that capital-driven automation inevitably forces workers to accept worse conditions and produce lower-quality products. ## "Inverse Centaur" and Capital-Driven Automation Doctorow introduces the concepts of "centaur" (machine-assisted human) and "inverse centaur" (human driven by machines). Under capital-driven automation, workers are consumed by machines to the point of exhaustion—for example, in Amazon warehouses, high automation leads to injury rates three times higher than other warehouses because machines force workers to work at extreme speeds. In contrast, labor-driven automation usually improves product quality. Doctorow emphasizes that workers are best qualified to decide how to use tools, but investors in the $1.4 trillion bubble believe they can fire workers and squeeze those who remain. ## Key Differences Between the AI Bubble and the Internet Bubble Although the internet bubble left behind giants like Google and Amazon, the AI bubble differs fundamentally: * **Unit economics**: Every new user increased profitability for the internet; every new customer or use increases losses for AI. * **Public discourse**: In the 1990s, business articles discussed "how CEOs should handle workers demanding internet access"; today, the content is "how CEOs can convince workers to use AI, or be fired." This shows AI is imposed on workers, not driven by their demands. * **Growth illusion**: Today's big tech companies (like Google, Facebook) already hold over 90% market share, and stagnant growth leads to market revaluation. Executives must fabricate a "growth story," and AI provides this narrative, even if actual benefits are poor. ## The Power of Local Resistance: Monterey Park Bans Data Centers Doctorow highly praises voters in Monterey Park, California, for passing a ballot measure permanently banning new data centers near residential areas. He argues that such local political actions (e.g., taking control of minor public offices, engaging in zoning decisions) can really change material conditions. He cites the success of the "Moms for Liberty" organization (right-wing) to show that even "the dumbest people you know" can change rules through organizing. He calls on the progressive movement to abandon "voting with your wallet" (influencing politics through shopping) and turn to real local organizing. ## Hollywood Writers' Strike: Sectoral Bargaining Is Key The 2023 Hollywood writers' strike was the only successful workplace-level fight against AI. Doctorow points out that its core was not demanding more copyright (which is useless against monopoly studios) but relying on **sectoral bargaining** (the right to negotiate simultaneously with all employers in an industry). This was an exemption for Hollywood unions under the 1947 Taft-Hartley Act, allowing writers to collectively resist AI replacement. Doctorow believes creative workers face a crossroads: either side with their bosses and demand more copyright, or demand sectoral bargaining, thereby standing in solidarity with all American workers. ## Conclusion: From "Shopping Politics" to "Organizing Politics" Doctorow calls on people to escape the illusion of "voting with your wallet" and turn toward real collective action and grassroots organizing. He argues that even without deep technical knowledge, one can judge that "overall, I stand with all workers, not the bosses." --- **Source:** “How to Think About AI”: Cory Doctorow on Big Tech, Understanding AI, Labor Automation & More – switchflip (https://www.youtube.com/watch?v=OBUzl_IaWIw)