Sam Altman’s 2023 OpenAI Ouster: Early Reports and What Followed

SUMMARY

In 30 seconds

In this November 18, 2023 video, Matthew Berman follows the first day after OpenAI’s board removed Sam Altman as CEO and removed Greg Brockman from the board. He reviews posts by Altman and Brockman, reporting by Kara Swisher and reports of researcher resignations, while speculating about the board’s motives and what might happen next. Altman subsequently returned as CEO and Brockman as president. OpenAI’s March 2024 summary of an external review reported a breakdown of trust between the former board and leadership, and said the decision did not arise from concerns about product safety or the pace of development; that summary did not publish the full review.

1First reactionsAltman publicly thanked colleagues after the November 17 announcement, while Berman emphasized uncertainty about the board’s motives.
2Mission and governanceBerman relays comments attributed to Ilya Sutskever and speculates about safety and commercialization. The later review summary did not attribute the decision to product safety or development pace.
3Brockman’s accountBrockman described separate calls in which Altman was told he was fired and Brockman was informed he would leave the board.
4Reports of departuresEarly reports named Jakub Pachocki, Aleksander Mądry and Szymon Sidor as researchers who had resigned; later roles differed.

Takeaways

On November 17, 2023, OpenAI’s board removed Altman as CEO and said he had not been consistently candid in communications with it.
Brockman said he and Altman were surprised, described how they received the news, and resigned after being removed from the board.
Early claims about the reasons for the firing and who supported it were incomplete and should not be treated as established fact.
Altman returned as CEO later that month; OpenAI’s 2024 review summary attributed the episode to a breakdown of trust and said it did not arise from product safety or development pace concerns.
Reports of research staff resigning on the first day describe a fast-moving moment, not necessarily their eventual status.

TopicsOpenAI Leadership Company Culture AI Safety

FULL TEXT

Read the edited text by chapter

Show transcript

Transcript only

Altman’s removal and initial reactions

Berman responds to the board’s November 17, 2023 announcement and Altman’s post. He offers an early interpretation while acknowledging that much is unknown.

On November 18, 2023, Matthew Berman reports on the first day after OpenAI’s board removed Sam Altman as CEO. He says the available information is incomplete and reads Altman’s post thanking his colleagues and saying he would share more later. Berman, a former founder, expresses sympathy and regards the decision as a striking upheaval.

SLIDE 1 / 3Altman’s initial public responseHis post thanked colleagues and said he would share more about his next steps later.

Berman characterizes the decision as a betrayal and speculates that Ilya Sutskever and Mira Murati had split from Altman and Greg Brockman. That was his interpretation of early reporting, not an established account of Murati’s role: she was appointed interim CEO, was not a voting board member, and later supported Altman’s return. His estimate of how much time the board spent on OpenAI is also not substantiated in the video.

Berman says contemporary reporting suggested a tension between rapid commercialization and caution. He explains that at the time OpenAI had a nonprofit governing entity and a related for-profit business, giving the board a role unlike a conventional startup board. His account of the individuals’ motives and financial incentives is speculative and should not be read as a finding about why they acted.

Governance and reported reasons

Berman reads reporting and remarks attributed to Ilya Sutskever. His theory about safety and pace is not established by those early reports.

Berman cites Kara Swisher’s reporting and a remark attributed to Sutskever: he rejected the characterization of the decision as a coup and said the board was acting for the nonprofit mission. Berman also recounts reports that Altman, Brockman and Microsoft were informed shortly before the public announcement. Those timing details are reported accounts from a rapidly developing story.

Swisher reported that board members viewed Altman as manipulative and headstrong; Berman says the specifics and evidence behind their concerns remained unclear. The board’s [November 17 statement](https://openai.com/blog/openai-announces-leadership-transition/) said Altman was not consistently candid in communications with the board. Berman interprets that as lying, but the public wording did not identify a particular falsehood, so his interpretation should not be treated as a proven finding.

Brockman’s timeline

Brockman describes how he and Altman learned of the board’s decision. Berman discusses their response and possible next steps.

Berman reads Brockman’s post saying he and Altman were shocked and still trying to understand what happened. Brockman described Altman joining a video call with the board, without Brockman present, where Sutskever told him he was being fired. Berman predicts the pair might soon start a new company; at that moment it was a prediction rather than an announcement.

SLIDE 2 / 3Brockman’s account of the meetingBrockman’s post describes how he and Altman learned of the board’s decision.

According to Brockman’s account, Sutskever subsequently arranged a separate call and told him he would be removed from the board while retaining his role at the company. OpenAI announced the change around the same time. Brockman said most of management learned shortly afterward, while Murati had learned the previous night; this is his account, not proof of her involvement in the board decision.

Berman highlights messages of support for Altman and says many founders appreciated his past help. Altman posted that the public response felt like reading his own eulogy while alive, and urged people to tell friends they appreciate them. Berman interprets another post about his shares as a joke and imagines a rival venture; neither establishes Altman’s future plans or the value of any holdings.

Reactions, departures and the outcome

Berman reviews Altman’s social posts and reports of researcher departures. Later, Altman and Brockman returned and an external review addressed the episode.

Berman cites a report naming research director Jakub Pachocki, AI-risk lead Aleksander Mądry and researcher Szymon Sidor among people who had resigned during the first day of the upheaval. He expects other employees might follow. These were early reports, not necessarily permanent departures: OpenAI subsequently [announced Pachocki as its chief scientist in May 2024](https://openai.com/index/jakub-pachocki-announced-as-chief-scientist/).

SLIDE 3 / 3Early reported researcher departuresThe post records reports from the first day; it does not establish each person’s later status.

The situation changed quickly after this video. OpenAI reached an agreement in principle for Altman’s return on November 21 and later reinstated Altman as CEO and Brockman as president. In March 2024, [OpenAI summarized an external WilmerHale review](https://openai.com/index/review-completed-altman-brockman-to-continue-to-lead-openai/) as finding a breakdown of trust between the prior board and leadership; it said the decision did not arise from product safety, security or the pace of development. The company released a summary rather than the full report, so this is the review’s reported conclusion rather than a complete public account of every dispute.