The abrupt removal of Sam Altman from OpenAI in November 2023 sent shockwaves through the tech world. After being briefly ousted by the company’s board, including his longtime collaborator and cofounder Ilya Sutskever, Altman was reinstated after a massive employee revolt. The initial explanation—Altman was “not consistently candid” with the board—left many questioning what exactly he’d been hiding, and what prompted Sutskever to turn against him. Now, new details have emerged from a legal deposition involving Sutskever, part of Elon Musk’s ongoing lawsuit against Altman and OpenAI, painting a picture of manipulation, conflicting information, and a growing lack of trust.
A Pattern of Manipulation and Conflicting Information
Sutskever’s nearly 10-hour deposition, punctuated by tense exchanges between attorneys, revealed a deep-seated dissatisfaction with Altman’s leadership style. According to Sutskever, Altman frequently pitted high-ranking executives against each other, offering conflicting information and tailoring his communications to satisfy different audiences. He expressed hesitation about relying solely on secondhand accounts later in the testimony, emphasizing the importance of firsthand knowledge.
OpenAI spokesperson Liz Bourgeois responded with a brief statement, stating that the events of 2023 were behind them and that the board’s independent review unanimously concluded that Altman and Greg Brockman were the right leaders for OpenAI.
The Making of a Memo
Sutskever and Altman co-founded OpenAI in 2015, with Sutskever initially securing a board seat and a role as chief scientist. However, by 2023, he had become increasingly concerned about Altman’s behavior. He testified that one or more of OpenAI’s independent board members had asked him, after discussions about executive concerns, to compile a collection of screenshots and documentation. He created a 52-page memo and sent it to board members Adam D’Angelo, Helen Toner, and Tasha McCauley.
When asked why he didn’t send it to Altman directly, Sutskever stated that he believed Altman would simply find a way to “make them disappear.” He also revealed that he’d been contemplating Altman’s removal for “at least a year.”
He also submitted a separate memo detailing concerns about OpenAI President Greg Brockman. Both memos were sent as disappearing emails, but “various lawyers” have copies.
“A Consistent Pattern of Lying”
The core of Sutskever’s concerns was Altman’s “consistent pattern of lying, undermining his execs, and pitting his execs against each other.” This sentiment appeared on the first page of the memo about Altman, read aloud during the deposition.
At OpenAI, Altman allegedly created division by offering conflicting information. Sutskever claimed Altman told both him and Jakub Pachocki, now OpenAI’s chief scientist, “conflicting things about the way the company would be run,” undermining Sutskever on multiple occasions.
Another instance involved former OpenAI research executive Dario Amodei (now CEO of Anthropic), who wanted to run all of OpenAI research and have Brockman fired. According to Sutskever, Altman didn’t take a firm position, allegedly playing both sides to his advantage.
Sutskever also highlighted claims from former OpenAI CTO Mira Murati, stating that Altman pitted her and Daniela Amodei (then an OpenAI research executive) against each other. The rivalry between the Amodeis and certain OpenAI executives is well-documented, culminating in Daniela and Dario leaving OpenAI to co-found Anthropic.
A Troubling Workplace and Growing Concerns
Murati provided Sutskever with screenshots and documentation of conversations, including a text message exchange between Altman and Brockman. She also revealed claims that Altman left his leadership role at Y Combinator due to “similar behaviors,” creating chaos and pitting people against each other. Sutskever considered this further evidence that Altman couldn’t be trusted.
Y Combinator did not respond to requests for comment.
The Evidence Mounts: Outside Validation
These accounts didn’t emerge in isolation. In May 2024, former board member Helen Toner publicly stated that Altman systematically concealed important information from the board, including his ownership of the OpenAI startup fund and “inaccurate” information about the company’s safety processes. She alleged that the board only found out about ChatGPT’s launch through Twitter.
Toner described years of difficulty in fulfilling the board’s role due to Altman’s withholding of information. “All four of us who fired him came to the conclusion that we just couldn’t believe things that Sam was telling us,” she stated.
She recalled executives telling the board, a month before Altman’s ouster, that they couldn’t trust him to lead OpenAI and feared he would not change. “They used the phrase ‘psychological abuse,’” Toner said.
A Brief Merger Proposal and Executive Departures
During the period of chaos following Altman’s removal, Anthropic reportedly proposed a merger with OpenAI, with Dario and Daniela Amodei participating in discussions. The merger didn’t move forward due to practical challenges.
Subsequently, OpenAI experienced a series of executive departures. Sutskever left to start Safe Superintelligence (SSI), followed by Murati, who announced her own AI startup, Thinking Machines Lab. Bob McGrew, chief research officer, and Barret Zoph, vice president of research, also departed.
The Financial Picture and Remaining Ties
Sutskever confirmed he hadn’t spoken to Altman in 10-12 months or to Brockman in 15 months. He believed OpenAI was paying his legal fees, though he wasn’t sure and hadn’t received a bill. He also stated he retained a financial interest in OpenAI, which had increased in value since his departure.
A Future of Questions
The lawsuit and depositions raise as many questions as they answer. The extensive financial stakes and legal fees involved suggest more depositions and details will emerge. Sutskever’s deposition provides a window into the tumultuous leadership changes at OpenAI and the psychology of those at its helm.
Near the end of his testimony, Sutskever was asked if he thought OpenAI employees would be happy about Altman’s ouster.
“I had not expected them to cheer,” he said. “But I [had] not expected them to feel strongly either way.”
Update, November 6th: Added comment from OpenAI
