ref/sam-altman crisis --nov-2023
The November 2023 Board Crisis
Five days that shook the AI world — reported factually, with what is known, what is disputed, and what remains unknown.
How to read this page. The full reasons for Altman’s removal have never been publicly established. This page reports the documented record — statements, the independent review, and the participants’ accounts — and flags where accounts differ. Speculation is labeled as speculation.
Friday, 17 November 2023 — the removal
OpenAI’s nonprofit board announced that Sam Altman would no longer be CEO, effective immediately, stating that he “was not consistently candid in his communications with the board, hindering its ability to exercise its responsibilities.” The board also removed Greg Brockman as chairman. Mira Murati was named interim CEO.
The board at the time: Altman himself, Brockman, chief scientist Ilya Sutskever, and independent directors Adam D’Angelo (Quora CEO), Tasha McCauley, and Helen Toner (Georgetown AI-governance researcher). The vote to remove Altman was carried by Sutskever, D’Angelo, McCauley, and Toner.
The announcement gave no specific examples of the alleged lack of candor. Reporting in the following days surfaced several candidate explanations — tensions over commercialization speed, Altman’s side projects, a paper Toner co-authored critical of OpenAI’s safety practices, alleged undermining of board members — but none was confirmed by the board as the reason.
Saturday–Sunday, 18–19 November — the revolt
- Brockman resigned in solidarity with Altman.
- Investors — reportedly led by Microsoft and major venture firms — pressured the board to reinstate Altman.
- Hundreds of OpenAI employees signed a letter threatening to resign and join Microsoft unless Altman was reinstated and the board resigned. The letter was eventually signed by a large majority of staff (reported ~700 of ~770).
- Microsoft CEO Satya Nadella announced that Altman and Brockman would join Microsoft to lead a new AI research group.
- The board named Emmett Shear (former Twitch CEO) as interim CEO, replacing Murati.
Monday–Wednesday, 20–22 November — the return
Negotiations continued through the weekend. On 22 November, OpenAI announced Altman would return as CEO, Brockman would return as president, and the board would be reconstituted: Bret Taylor (chair), Larry Summers, and Adam D’Angelo retained. Sutskever, McCauley, and Toner departed the board. The company also announced an independent review of the episode.
The aftermath: what the review found
In March 2024, OpenAI released a summary of the independent review conducted by the law firm WilmerHale. Its conclusion, as summarized by the company: the board’s decision was the consequence of a breakdown in the relationship and loss of trust between the prior board and Altman — and not a response to concerns about product safety, the security of AI systems, OpenAI’s finances, or its statements to investors or the public.
Altman was reinstated to the board. The episode ended with Altman arguably more powerful than before: a new, more sympathetic board and a public demonstration of staff loyalty.
The perspectives
The board’s defenders (notably Toner, in later public writing) argue the nonprofit board was doing its chartered job — overseeing a CEO it had lost confidence in — and that the charter explicitly prioritized the mission over any individual, including Altman.
Altman’s supporters (including Paul Graham and much of the staff) argue the removal was sprung without due process or stated cause, destabilized the leading AI lab over an interpersonal breakdown, and proved the governance structure unfit.
Outside observers split: some saw a legitimate governance mechanism working as designed; others saw a small, unelected board nearly destroying enormous value on opaque grounds. The episode became the central exhibit in debates over who should govern AGI development.
What remains unknown
The specific communications the board found “not consistently candid” have never been publicly detailed. Participants have given partial, sometimes conflicting accounts. This site does not fill those gaps.
Source notes
- Board statement (17 Nov 2023): OpenAI announcement, quoted verbatim above.
- Timeline of 17–22 Nov: contemporary reporting (The Verge, NYT, Bloomberg, FT, WSJ).
- Staff letter (~700 of ~770 signatures): reported figures, contemporary press.
- WilmerHale review summary (Mar 2024): OpenAI announcement; conclusion quoted as summarized by the company.
- Toner’s account: her public writing/interviews, 2024 — attributed as her perspective.