What Altman Will Do to Make OpenAI Live Up to Its Name Again; Why You Should Care About Stability’s Implosion

In classic OpenAI fashion, big drama dropped on Friday. Jan Leike, who formerly co-led OpenAI’s superalignment team with Ilya Sutskever—the co-founder and chief scientist who temporarily ousted CEO Sam Altman in November—explained on X why he left the company. Altman & Co. failed to give superalignment, which aimed to be a check on OpenAI researchers developing the AI, enough resources and say over the company’s work, he said. Disagreements about it pushed Leike to a “breaking point.” 

Altman didn’t take the bait. Ever the politician, he thanked Leike for his contributions and said OpenAI needed to do more work on safety, without elaborating. (OpenAI’s research team has absorbed superalignment, as we first reported on Wednesday.)

The debates over AI safety at OpenAI first burst into public view the day Sutkever fired Altman (see here), but Leike’s statements aren’t the kind that Altman wants aired in public. Government officials are sure to raise the issue with Altman the next chance they get, and rivals will try to hire OpenAI talent that may be sympathetic to Leike’s view (if any are left).