Inside the Deposition That Showed How OpenAI Nearly Destroyed Itself

In a recent deposition, OpenAI co-founder Ilya Sutskever revealed a tumultuous period within the organization, marked by panic and unverified claims. This 10-hour testimony was part of the Musk v. Altman lawsuit, which questions OpenAI's shift from a nonprofit focus to a profit-driven entity aligned with Microsoft. Sutskever disclosed a 52-page memo detailing accusations against CEO Sam Altman, which he prepared with information primarily from CTO Mira Murati without cross-verifying with other executives. He had contemplated removing Altman for over a year, awaiting favorable board dynamics. During Altman's firing, discussions of merging OpenAI with competitor Anthropic surfaced, a move that Sutskever opposed. The ideological divide within the board was stark, with some members suggesting that destroying OpenAI might align with its mission to ensure AI safety. This crisis highlighted governance failures, a rushed decision-making process, and miscalculations concerning employee loyalty, leading to widespread resistance against Altman's removal, all of which jeopardized the company’s stability.

Source 🔗