I strongly recommend reading the full letter from Mikita and his colleagues on their firing from OpenAI. They are raising vital issues that deserve everyones attention and the fact that they have been left guessing about the reasons for their simultaneous termination while working on essential safety work is not good. But also, between this letter and the recent White House Super Intelligence accords, I am realizing that in addition to the issues they raise there is a perhaps equally pressing need to improve the signatures of people who work on developing frontier models. Jasmine and Marks' are decent, but Dario, Greg, Mikita and Tomek should step it up if they plan on signing more important public documents in the future. (To be clear this is a joke and my handwriting is similarly unfit for public scrutiny)