Inside the Deposition That Showed How OpenAI Nearly Destroyed Itself

A recent deposition by OpenAI co-founder Ilya Sutskever reveals critical insights into the company's tumultuous governance and internal conflicts. In a nearly 10-hour testimony during the Musk v. Altman lawsuit, Sutskever disclosed a 52-page case against CEO Sam Altman, built predominantly on unverified claims from CTO Mira Murati. He detailed his year-long contemplation of firing Altman, based on perceptions of board dynamics rather than performance. During a crisis that unfolded in November 2023, discussions about merging OpenAI with competitor Anthropic almost led to the company's collapse. Board member Helen Toner expressed controversial beliefs that permitting OpenAI's collapse could align with its safety mission, showcasing ideological divides within leadership. Sutskever's reliance on single-source allegations and his underestimation of employee loyalty contributed to miscalculations in decision-making, as over 700 of 770 employees demanded Altman's return following his abrupt firing. This deposition uncovers a cacophony of poor governance decisions and a misread organizational culture at OpenAI.

Source 🔗