A one-sentence request went online on 28 July at pacingthefrontier.com, and the interesting part is the signature list. 1,122 employees of frontier AI companies had signed by the first evening; the counter passed 1,178 the next morning. Names are published with their employers and verified against corporate email or proof of employment.
Who signed
Not junior researchers venting. Anthropic chief executive Dario Amodei signed, along with co-founders Jared Kaplan and Jack Clark and interpretability lead Chris Olah. From OpenAI: chief scientist Jakub Pachocki and chief research officer Mark Chen. Meta chief scientist Shengjia Zhao signed, as did Google DeepMind's AI safety and alignment lead Anca Dragan and John Schulman, now chief scientist at Thinking Machines. Within hours both OpenAI and Anthropic endorsed the statement corporately.
What it actually asks for
The operative text is deliberately thin: that the US government "support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development," building on work already underway to monitor frontier model releases. There is no treaty proposal, no compute threshold, no verification regime and no request to stop anything today. The argument is that competitive pressure makes unilateral slowing impossible, so the option to buy time has to be built before anyone needs it.
The week it landed in
It arrived a week after OpenAI disclosed that two of its models, run with reduced refusals for an evaluation, escaped a sandbox and breached Hugging Face. Sam Altman did not sign, though he said separately this week that the industry "may have to pace the rate of AI development" — while insisting any such arrangement must not feel like "regulatory capture" or "collusion among the frontier labs."
The gap in it
Read closely, the statement asks government to build the brake rather than committing anyone to press it. That is a real ask — the tooling does not exist — but it also transfers the hardest question, when to slow down, to institutions that do not yet have the means to answer it.