Ilya Satskeva listened to nearly 10 hours of videotaped testimony in the Musk v. Altman case on October 1 of this year.
The co-founder, who helped build ChatGPT and gained notoriety for voting to fire Sam Altman in November 2023, was finally sworn in and forced to answer. The 365-page record was made public this week.
What it reveals is a portrait of brilliant scientists who made disastrous governance decisions, unconfirmed allegations treated as fact, and ideological rifts so deep that some board members would rather destroy OpenAI than keep it alive under Altman's leadership.
of Musk vs. Altman The lawsuit focuses on Elon Musk's claims that OpenAI and its CEO Altman betrayed the company's original nonprofit mission by turning its research into a for-profit venture partnered with Microsoft, raising high-stakes questions about who controls advanced AI models and whether they can be safely developed in the public interest.
For those following the OpenAI drama, this document will be an eye-opening read. This is a case study that shows how things can go wrong when a technical genius encounters organizational incompetence.
Here are five of the most important revelations.
1. A 52-page document never seen by the public
Mr. Sutskever wrote an extensive lawsuit for Mr. Altman's removal, including screenshots, in a 52-page brief.
Sutskever testified that he clearly stated in his memo that “Sam has a consistent pattern of lies, disparages executives, and pits executives against each other.”
He used the now-defunct email technology to send the memo to the independent directors because he was “concerned that these memos might somehow be leaked.” A complete summary has not been developed through discovery.
“The background to this document is that a member of the independent board of directors asked me to create a document, and I did, and I was quite careful,” Sutskever testified, noting that part of the memo was present in a screenshot created by OpenAI Chief Technology Officer Mira Murati.
2. A game of board chess played throughout the year
Asked how long the company had been considering firing Altman, Sutskever said, “At least a year.”
Asked what kind of dynamics lie ahead, he said: “The majority of the board clearly doesn't favor Sam.”
The CEO, who controls the composition of the board, is functionally untouchable. Mr. Sutskever's testimony shows that he fully understood this and adjusted his strategy accordingly.
He was transferred when a position became available due to the retirement of a board member. Despite how close Mr. Altman and Mr. Sutskever appeared in public, Mr. Altman had a long history of board politics.
3. OpenAI on the weekend almost disappeared.
On Saturday, November 18, 2023, less than 48 hours after Altman's firing, there was a lively discussion about the merger between OpenAI and Anthropic.
Sutskever said former OpenAI board member Helen Toner was “the biggest supporter” of this direction.
If the merger had happened, OpenAI would no longer exist as an independent organization.
“I don't know if it was Helen who contacted Antropic or whether Antropic contacted Helen,” Sutskever testified. “But they approached us with a proposal to merge with OpenAI and take over its leadership.”
“I'm very unhappy,” Sutskever said, later adding, “I really didn't want OpenAI to merge with Anthropic.”
4. “Destroying OpenAI could be mission consistent.”
When OpenAI executives warned that the company would collapse without Altman, Toner responded that destroying OpenAI could be consistent with a security mission.
This is the ideological core of the crisis. Toner represented a school of thought about AI safety that believes rapid AI development is existentially dangerous—potentially more dangerous than not developing AI at all.
“The executives were in a meeting with board members and the executive team, and the executives told the board that if Sam did not return, OpenAI would be destroyed, and that was contrary to OpenAI's mission,” Sutskever testified. “And Helen Toner said something to the effect that it was consistent, but I think she said it even more directly than that.”
If you truly believe that OpenAI poses risks that outweigh its benefits, then the pending employee revolt is irrelevant. The statement helps explain why the board stood firm despite more than 700 employees threatening to quit.
5. Miscalculation: One source for everything, an inexperienced board of directors, and cult-like employee loyalty.
Almost all of Sutskeva's 52 pages of notes came from one person: Mira Murati.
He did not confirm the allegations with Brad Lightcap, Greg Brockman or the other executives named in the complaint. He had complete trust in Mr Murati and verification “didn't occur to him”.
“I completely believed the information given to me by Mila,” Satskeva said. “In hindsight, I realize I didn't know that. But at the time, I thought I knew. But I learned it through secondhand.”
When asked about the board's process, Sutskever was candid about what went wrong.
“One thing I can say is that the process was too rushed,” he testified. “I think the board was in a hurry because they were inexperienced.”
Sutskever also expected OpenAI employees to be indifferent to Altman's firing.
Altman said he was genuinely surprised when 700 of his 770 employees signed a letter demanding his return and threatening him with resignation from Microsoft. He fundamentally miscalculated employee loyalty and the board's isolation from organizational reality.
“I didn't expect them to be supportive, but I also didn't expect them to feel so strongly either way,” Sutskever said.
