Anthropic CEO Dario Amodei published a lengthy blog post on Saturday outlining new safety measures his company will take — bringing in third-party evaluators, for one — and calling on the industry as a whole to back a broader slowdown in frontier model development. OpenAI CEO Sam Altman quickly said he'd adopt Amodei's suggestion to give independent evaluators access "similar to that of employees." Elon Musk, who runs xAI, weighed in too: "Dario is right."
All three men have spent years warning about the risks of artificial intelligence, but in an industry this competitive, a genuinely coordinated slowdown has almost no precedent. Companies have long kept shipping new products to boost engagement, grab market share, and drive sales. Just how far the leading AI companies will go with new restrictions or safety checks remains an open question.