A lesson I learned far too late in my time at OpenAI is that it's pretty easy for the company to fire you because they're afraid you'll raise some concern and pretty difficult for them to fire you once you've raised it.
By all accounts I've heard, the three weren't fired for whistleblowing. OK, what were they fired for? Was it just fear they were the kinds of people who might raise concerns when the time arose?
Well, that de-escalated quickly!
But the work's not done ...
Work at OpenAI, Anthropic, or X and want to be part of shaping the commitments about what pacing looks like and making sure execs keep the commitments they're laying out today? Join us at
And if you work at GDM, Meta, or another lab and want to be part of making sure your execs make the same commitments? Join us at