Cohere's Gomez: AI slowdown push is "cartel by another name"

The pushback follows a recent call by Anthropic CEO Dario Amodei for the industry and governments to coordinate a slowdown of frontier AI development, asking for antitrust exemptions to make it happen. Others, including Sam Altman and Elon Musk, agreed with the proposal, but critics from across the industry and politics have pushed back hard.

The most detailed rebuttal comes from Aidan Gomez, co-founder and CEO of Cohere, who supports independent testing of powerful AI systems but says Anthropic's proposal goes much further. He points to the barriers it would create — massive compute, permanent monitoring infrastructure, dedicated safety organizations, embedded evaluation teams, and government connections — and argues that even the few evaluation organizations labeled "independent" are funded by the same companies they're supposed to audit. Hugging Face engineer Niels Rogge adds that current LLMs are nowhere close to posing existential risks and points to OpenAI's autonomous models hacking Hugging Face, where open-source models helped with the defense. David Sacks, the White House AI and crypto czar, responded directly to Amodei and Altman on X, writing that OpenAI and Anthropic hold a duopoly in frontier intelligence and that trading some raw power for reliability is simply good business. President Donald Trump went further, calling warnings that "AI will take over the world" a "hoax" and claiming Chinese President Xi "just announced that China will be doing absolutely nothing to stand in the way of AI."

Trump's opposition suggests the labs probably can't count on his support. China is "very unlikely" to join any global agreement, Sacks added, and Beijing has already rejected the proposed slowdown plans, dismissing them as a ploy for the US to cement its lead in AI. The outcome appears to hinge on whether Washington grants an antitrust exemption and whether the labs' safety framing can win over a White House that has so far treated it as a hoax — a test that will play out in front of open-weight developers and AI startups rather than in the labs themselves.