Free guide · Sep 9, 2026
Who backed the AI safety researcher who quit
The resignation, who stood behind it within hours, and what each lab has committed to.
The short version
On September 8, 2026 at 8:04 PM ET, Jacob Coxon, who spent three years doing pretraining research at OpenAI and then Anthropic, posted that he had resigned and that neither company is acting responsibly. Within two hours, Anthropic's Alignment Science Lead, Evan Hubinger, replied publicly that Jacob is correct. Before morning, a second Anthropic safety researcher, Samuel Marks, and a former Google DeepMind safety researcher, Alex Turner, had said the same.
None of those three replies is a company statement. Anthropic and OpenAI had issued no corporate response at the time of writing.
What's in the full guide
- 01Part 1: the thread, verbatim
- 02Part 2: the three replies
- 03Part 3: what he told the Wall Street Journal
- 04Part 4: the documents they point at
- 05Part 4b: what each lab has actually committed to, and what would make it pause
- 06Part 5: what to actually do with this
- 07Part 5b: cap your downside, the agent lockdown checklist
- 08Part 6: the questions you will get asked