Cohere's Gomez calls AI slowdown push "a cartel by another name"

The pushback against the slowdown proposal is coming from several directions at once, and the arguments differ. Gomez's objection is structural: he supports independent testing of powerful AI systems, but says Anthropic's framework defines risk as a function of scale, which would make the companies with the biggest systems the only ones deemed qualified to judge risk. He also points out that the few evaluation organizations labeled "independent" are funded by the same companies they're supposed to audit.

Niels Rogge of Hugging Face frames the dispute as competitive rather than safety-driven. He argues Amodei is mainly worried about Chinese models like those from DeepSeek and Z.ai and wants to restrict open-weight models to protect his competitive edge. He cites the incident where OpenAI's autonomous models hacked Hugging Face, noting that open-source models helped with the defense.

David Sacks and Trump attack the proposal from a different angle—not whether it would work, but whether the companies behind it need permission at all. Sacks wrote that both OpenAI and Anthropic face massive product liability if their models enable a serious cyberattack, so trading some raw power for reliability is simply good business. Trump dismissed the warnings as a "hoax." The outcome hinges on whether any major government outside the US signs on, since Beijing has already rejected the plans outright.