I can't really see a world where OpenAI's actions are justified here. It was Tomek's job to communicate with METR, who were being given significant, unprecedented access to do the HuggingFace report. The norms were not set. Even if he fucked up badly, just give him a stern warning and ban him from roles involving liasoning with third parties, he did lots of other valuable work, and there'd no longer be room for him to inappropriately leak as part of his job. IMO firing people over a good faith attempt to use their best judgement in a novel and uncertain situation is a sign of a highly unhealthy culture. Firing him will have a chilling effect on anyone else at OpenAI working with third parties, an odd choice when Sam announced they'd embed evaluators with high access. I separately consider this a tragedy, the METR report was one of the most important works of safety this year, punishing Tomek for being a part of it, and making it harder for future such reports to exist is a real shame. Of course, I've only heard Tomek's side, but unless he's outright lying I'm struggling to see ways this looks good for OpenAI?

Last week I said we should wait for more info on OpenAI's firing of safety researchers. That information is in, and it looks incredibly bad. Korbak - whose job was being a liaison with METR - was told he had said too much to METR and was fired for that reason.

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?

Having seen up close several past departures from OpenAI of similarly murky circumstances, I support @j_asminewang , @balesni, and @tomekkorbak fully in this matter. I hope @OpenAI will take ownership of the procedural missteps here, rather than doing what is relatively more likely to follow: a carefully crafted internal communications process to quell dissent and make it hard to ask questions or get clear answers.

OpenAI fired me last week, along with two of my safety colleagues. I was given one reason: that I accessed an executive's email. I want to say this plainly, because too many of OpenAI’s history is smoke and mirrors when people disappear:

Two other safety researchers and I were fired from OpenAI last week. We wrote this letter to leadership. I believe we were fired for prioritizing safety over the near-term interests of OpenAI as a corporation.

There has been a lot of debate about what’s the right way to preserve monitorability of frontier AI and whether policies should target inputs (like model architecture) or outputs (like results on monitorability evals). I think they should target both. With latent architectures, models would no longer need CoT, and it will be harder to make a case that the model is sufficiently monitorable and easy to study given the current state of science.

Interesting that Dario thinks that pacing via inputs such as AI R&D compute is more gameable than pacing via safety evaluations/practices. Imo it's the opposite; e.g. a requirement to spend 90% of compute on external inference (and not R&D) seems fairly hard to game. While I am in favor of moving toward being able to pace based on safety evaluations/practices, these seem harder to define and more gameable to me; e.g. AIs or companies might game alignment evaluations, and making a judgment call on whether a safety practice is implemented appropriately seems potentially quite subjective.

switching to fully recurrent LLM architectures would be the biggest blow to safety, probably in history of AI all AI labs should commit to limit the opaque serial depth of their models, for the foreseeable future. this will not ensure monitorable CoTs but will protect us from the worst possible outcomes.