Inside the Deposition That Showed How OpenAI Nearly Destroyed Itself

Ilya Sutskever's 10-hour testimony in the Musk v. Altman lawsuit reveals significant issues within OpenAI during a governance crisis. Sutskever presented a 52-page case against Sam Altman based on unverified claims from CTO Mira Murati. He indicated he had been considering removing Altman for at least a year, waiting for favorable board dynamics. When Altman was abruptly fired, discussions emerged about merging with competitor Anthropic, a move Sutskever opposed. Notably, Helen Toner, a board member, suggested that destroying OpenAI might align with its safety mission, indicating deep ideological divides. Furthermore, Sutskever acknowledged the board's inexperience and the critical miscalculations made in decision-making, such as relying heavily on a single source and undermining employee loyalty, resulting in 700 out of 770 employees demanding Altman's reinstatement. This situation underscores the challenges faced by groundbreaking organizations grappling with existential risks.

Source 🔗