In the latest discovery phase of the Musk v. OpenAI lawsuit, OpenAI co-founder and former Chief Scientist Ilya Sutskever sat for nearly ten hours of deposition and submitted a 52-page memorandum outlining the core basis behind the board’s November 2023 decision to remove Sam Altman as CEO. According to multiple U.S. tech and legal media outlets, these filings now represent the most detailed and authoritative public reconstruction of the dramatic OpenAI power struggle that stunned the global AI industry two years ago.
Related articles
The Core Allegation: A “Persistent Pattern of Lying and Manipulation”
In the deposition, Ilya alleges that Altman repeatedly demonstrated what he calls a “persistent pattern of lying” — including pitting executives against each other, withholding key information, and intentionally generating confusion inside the organization, which ultimately prevented the board from effectively carrying out its fiduciary oversight duties. According to Ilya, the ultimate reason behind the extreme board action was simple: the board could no longer trust the CEO on matters of critical governance.
This narrative aligns with the previously cited “lack of candor” wording — but Ilya’s sworn testimony elevates the claim from external speculation into formal legal evidence. The Information describes the testimony and memo as reflecting a complete breakdown of “credibility and fiduciary trust.” The Washington Post had previously reported long-standing senior employee complaints about Altman “creating chaos and making oversight impossible” — and Ilya’s deposition now directly reinforces that pattern.
Shock Disclosure: Board Considered Merging with Anthropic and Having Dario Amodei Become CEO
The most explosive new disclosure from the deposition is that at the peak of the crisis, the OpenAI board seriously discussed the possibility of merging with Anthropic — and even considered Anthropic co-founder Dario Amodei as the potential CEO of the combined entity.
This indicates the board was not looking for a temporary patch — it was evaluating an extreme, full reset of OpenAI’s leadership and strategic trajectory.
This revelation is particularly consequential because Anthropic was founded by former OpenAI researchers who left over philosophical and governance disagreements — and the two companies later emerged as the two most important rivals in global frontier AI. Ilya’s testimony marks the first direct confirmation that such a merger + CEO replacement scenario was seriously explored.
The “Brockman Memo” Emerges as a New Focal Point
Another key element surfaced in the case: a second critical document known as the “Brockman memo.” A new federal court discovery order now requires Ilya to produce this memo and appear for a second supplemental deposition. The judge ruled that the document is relevant to the case — and that the plaintiff has the right to probe Ilya’s potential financial stake in OpenAI to assess whether his testimony could be influenced by economic incentives.
Legal outlet Law360 reports that there is ongoing procedural friction over Ilya’s compliance — including a brief threat of contempt motions — but the legal direction of travel is clear: the court is accelerating toward obtaining more internal records and testimony.
If the Brockman memo and Ilya’s 52-page memo converge on material factual points — and can be further validated against contemporaneous emails, meeting notes, and third-party testimony from investors and senior executives — the historical narrative of OpenAI governance between 2019 and 2023 could be fundamentally reframed.
The Most Important Crisis in AI History Was Not About Safety Philosophies — It Was a Governance Collapse
As more internal documents, emails and audio records enter the public record through forced discovery, the unresolved gaps within the 2023 OpenAI drama are likely to be filled in the weeks ahead.
While the global AI landscape — and OpenAI itself — have both undergone seismic changes over the past two years, Ilya’s deposition recenters the core truth:
The defining crisis was not a debate about alignment trajectories — it was a breakdown of fiduciary trust.
As the Musk v. OpenAI case moves deeper into evidence discovery, these materials are rapidly becoming primary historical source material for future analysis of AI governance, regulatory frameworks, and institutional design at the frontier of AGI development.
⸻
Timeline Recap
• Nov 17–21, 2023: Board fires Altman → internal revolt → Microsoft intervenes → Altman returns → board is restructured
• 2024–2025: Musk lawsuit proceeds; OpenAI repeatedly responds publicly; discovery enters substantive phase
• Oct 2025 – Now: Disputes intensify over Ilya’s additional testimony and memo production; court orders forced production of the “Brockman memo”
更多精彩内容，关注钛媒体微信号 （ID：taimeiti），或者下载钛媒体 App
