Mira Murati's deposition in Musk v. Altman has produced the most detailed public record yet of how Sam Altman was fired from OpenAI over Thanksgiving weekend 2023, including 78 text messages she exchanged with Altman over a 14-hour stretch and trial exhibits showing her role in feeding concerns to the board months before the ouster. The former CTO of OpenAI was named interim CEO on November 16, 2023, when four board members — Helen Toner, Ilya Sutskever, Adam D'Angelo, and Tasha McCauley — unanimously voted to terminate Altman. Within two days, she had been replaced as interim CEO by Emmett Shear, and by Monday, November 20, she was the first name on a letter from more than 750 OpenAI employees demanding Altman's return.
The exhibits, surfaced in the Elon Musk lawsuit against Altman and OpenAI, also include a document dated September 30, 2022 — a little over a year before the firing — in which Murati listed complaints about Altman's management style. She wrote that 'constant panic around our projects, people, goals, etc generates chaos and churn.' Toner testified this week that Murati and Sutskever had materially advanced the board's own concerns about a pattern of deceit and Altman's 'resistance' to oversight.
According to Toner's testimony, Murati funneled screenshots, text-message documentation, and allegations about Altman's earlier time at Y Combinator to Sutskever, who packaged them into a 52-page memo for the board. That memo, alongside Toner's own concerns, formed the evidentiary base for the November 16 termination vote. The board's public explanation — that Altman was 'not consistently candid in his communications with the board' — was deliberately vague and immediately spawned days of speculation on X.
Key facts
- 01Trial exhibits in Musk v. Altman show Mira Murati and Sam Altman exchanged 78 text messages over 14 hours during the November 2023 weekend.
- 02Four OpenAI board members signed a document terminating Altman on November 16, 2023, naming Murati interim CEO.
- 03Ilya Sutskever delivered a 52-page memo to the board built partly on documentation funneled to him by Murati.
- 04More than 750 OpenAI employees signed the letter threatening to defect to Microsoft, with Murati listed as the first signatory.
- 05A September 30, 2022 document shows Murati had complained to Altman that 'constant panic' around projects 'generates chaos and churn.'
The 78 texts between Murati and Altman, sent between Sunday evening and Monday morning, show Murati's position shifting in real time. Around 2:30am on Monday, Altman asked, 'can you indicate directionally good or bad? satya and others anxious.' Murati replied, 'Directionally very bad. Sam this is very bad.' She told him the board had 'walked me through all the reasons and the issues with you and why you can't be ceo.'
“Murati and Altman exchanged 78 text messages over a 14-hour stretch as she moved from interim CEO of OpenAI to first signatory on the 750-employee letter demanding his return.”— Jaeden Schafer
When Altman pressed on why the board had spent the weekend signaling it wanted him back, Murati answered, 'They want to have a new ceo in place tonight (not me.' She described Shear, the incoming interim CEO, as 'New guy is rando twitch guy,' and said she was 'hoping Satya can help undo this.' Altman, for his part, offered to walk away: 'i'm still willing to just walk away if that helps. if they are ramped up for crazy lawsuits against me then i'm not sure what.'
Between November 17 and 20, Murati also exchanged texts with Microsoft CEO Satya Nadella, who had publicly offered to hire every OpenAI employee into a new Microsoft subsidiary to be run by Altman and Greg Brockman. In one message, Murati told Nadella she was 'not putting [her] name on this,' referring to the board's Sunday statement that its decision was 'the only path to advance and defend the mission of OpenAI.' Early Monday, she texted Microsoft CTO Kevin Scott that the group was 'close to having the board resign.' Scott replied, 'For real this time?'
Murati then asked Nadella to issue a public statement of support, writing that 'It's very important that we don't lose researchers to Demis or Elon' — a reference to Demis Hassabis at Google DeepMind and Elon Musk's xAI. By that point, Sutskever had signed the employee petition demanding Altman's reinstatement, and the board's position was collapsing.
The most awkward exhibit is the employee letter itself. More than 750 staffers — out of a company that then numbered around 770 — signed it, and the document specifically chastised the board because 'within two days of your initial decision, you again replaced interim CEO Mira Murati against the best interests of the company.' Murati was the first signatory. Yet Toner testified that Murati had told the board she would only serve as interim CEO if the board could 'legitimize' the decision, and then declined to help legitimize it.
Toner's characterization of Murati's behavior in the days after the firing was blunt. 'She was waiting to see which way the wind would blow, and she didn't realize that she was the wind,' Toner said, calling Murati 'strikingly unsupportive' and 'remarkably passive.' Toner added that Murati 'seemed totally uninterested in telling her team that her conversations with us had been a significant factor in our decision to fire Sam.' In the texts, when Altman asked whether to send the employee letter to the board, Murati told him it 'wouldn't matter' and that the directors 'don't care if everyone quits,' only that they didn't want Altman's 'hand on agi.'
The countervailing read is that Murati was navigating an impossible position: she had raised real governance concerns, was handed a CEO seat she hadn't asked for, and watched the board fail to defend its own decision while Microsoft openly recruited her staff. The exhibits do not show her lying to either side so much as triangulating between a board that wouldn't make its case publicly and a workforce that had already chosen Altman. That is, in fairness, what most executives would do when the institution above them had already lost.
For OpenAI, the depositions arrive at a delicate moment. The company is fighting Musk's suit while running a $50B-plus annual compute bill and racing Anthropic and Google. The 2023 ouster has long been treated inside the industry as a governance failure that Altman survived; the trial record now reframes it as a governance failure in which the people best positioned to hold him accountable — Sutskever, Murati, Toner — either left, recanted, or were outmaneuvered. Whatever Musk's case ultimately wins or loses on the merits, the discovery process is doing the work the board's original blog post refused to do: putting the receipts on the record.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
