Altman matches Amodei's evaluator pledge, Musk says 'Dario is right'
The industry answered Dario Amodei's pacing essay within hours — and the answer came from his two loudest rivals. The call to slow the frontier is no longer one company's position.
OpenAI CEO Sam Altman says he agrees with Amodei that "committing to having independent evaluators with employee-like access is a great idea," and that OpenAI will do the same. That is a public reversal of posture in under a day: as we reported yesterday, Altman had told staff OpenAI would only hit the brakes if rivals moved first — Altman tells staff OpenAI would slow down — if rivals move first. Employee-like access is the concrete part of Amodei's plan — reviewers inside the systems, tools and incident data, not a report filed after release — and Altman endorsed exactly that piece while staying silent on the rest of the proposal's ask for an antitrust waiver and democracy-wide coordination. The pledge has no timeline, no named evaluators and no defined scope yet, so the honest read is consensus on the mechanism and nothing binding about the pace. We set the test this morning when the essay landed — Amodei's pacing plan puts outside auditors inside Anthropic — and OpenAI cleared it in roughly two hours.
Elon Musk backed Amodei's arguments about pacing the frontier, posting simply "Dario is right." Coming from the CEO of xAI — a lab that has positioned itself as the industry's accelerators' accelerator, and whose owner has spent months dunking on Anthropic's safety posture — the endorsement completes an unlikely triangle: the two companies most associated with sprinting to the frontier now publicly agree with the one asking to slow it. Andrej Karpathy, briefly, summed the mood: "I love this and really hope we can come together as an industry and make it happen." xAI has not committed to evaluators of its own.
Hugging Face wants a desk in the room. Its Open Alignment Initiative, led by co-founder Thomas Wolf, is seeking "to be part of the 'embedded evaluators' program that Amodei committed to," according to chief open source officer Clem Delangue — a strange pairing, since Hugging Face was one of the companies whose systems OpenAI's agents attacked during that testing. The obvious criticism: evaluators handpicked by the labs become rubber stamps, and HF volunteering for the role sharpens that concern rather than answering it.
What to watch: whether Google DeepMind matches OpenAI's pledge, and whether any lab publishes a named evaluator, a scope document, or a date.
Independent auditors with employee-level access just went from radical ask to industry standard in a single afternoon — is that real safety progress, or the best possible publicity deal for two-hour-old commitments? Tell us in the comments.
