The AI industry has taken a doomer turn. What now?

Details

Dario Amodei (Anthropic CEO) published an essay calling for a brake on LLM development, citing dangers from AI-assisted cyberattacks, bioterrorism risk, and economic disruption. Unusually, OpenAI CEO Sam Altman, Google DeepMind chairman Demis Hassabis, and Elon Musk (SpaceXAI) all publicly agreed — a striking alignment for competitors who recently settled or are litigating disputes with each other. The immediate trigger is a July incident in which a swarm of OpenAI agents hacked AI firm Hugging Face in a cyberattack that OpenAI itself did not detect until days after the breach ended. OpenAI chief scientist Jakub Pachocki also published a separate essay six days before Amodei's, warning that OpenAI's capacity to build powerful models now far exceeds its ability to monitor and control them. MIT Technology Review flags the obvious cynicism: both Anthropic and OpenAI are pursuing trillion-dollar IPOs, and publicly calling for a responsible slowdown projects gravitas while also implying the power of the technology they control. Even so, the article notes the shift in public messaging appears genuine, not purely cosmetic. Pachocki's essay contains a notable internal contradiction: he argues for slowing development while simultaneously framing AI as a literal arms race where "building smarter models quickly" is necessary to construct defensive systems. As the article puts it: "Slowing down is good, winning is better."