Hollywood labor groups are pushing back on the tech sector's existential AI warnings, arguing that the real threat is not superintelligence but the displacement of writers, actors, and crew members happening right now. Meanwhile, a New Yorker Q&A this week explains why China is not interested in AI existential risk discourse at all. Beijing's concerns are concrete, near-term, and political: social stability, censorship control, and winning the economic competition with the US. The AI safety debate, which frames itself as universal, is actually deeply parochial.
Three Different AIs, Three Different Panics
Hollywood's AI is a labor threat. Silicon Valley's AI is an existential risk. Beijing's AI is a geopolitical instrument. These are not three descriptions of the same technology. They are three different technologies with the same name, shaped by the interests of whoever is doing the describing. The Atlantic this week noted that AI is now impairing concentration in a new way: not by replacing thinking, but by generating so much ambient anxiety about AI itself that people cannot focus on the actual, immediate problems it is causing. The doom discourse is itself a distraction.
The Anthropic Paradox at the Center of It All
Security researchers this week used Claude, Anthropic's model, to hack into OpenAI employee accounts in under 72 hours. Anthropic built Claude partly to be safer and more aligned than competitors. That its model was used as a hacking tool within the industry is a neat summary of how safety discourse and operational reality keep diverging. Fred Turner's argument that Silicon Valley's biggest ideas tend to be warmed-over ideology dressed as technical necessity applies here: AI safety is a genuine concern wrapped in a rhetorical frame that benefits its proponents more than it protects anyone.