- More details about Sam Altman's ousting at OpenAI have emerged.
- New reports suggest Altman may have, at times, been a manipulative leader.
In the days that followed Sam Altman's ousting from OpenAI on November 17, employees inside the company and several members of the broader tech community likened the move to a coup.
The narrative in the immediate aftermath of his firing was that broad ranks of OpenAI liked Altman and that his sudden dismissal was shocking — an erratic move by a board that was prioritizing ideology over the demands of its stakeholders and the wishes of its employees.
But over the past few weeks, new details have emerged that shed more light on the board's decision — which was ultimately reversed by a circuitous route — to fire Altman.
These new details suggest that Altman is a skilled corporate schemer who manipulated people and perceptions within OpenAI to maintain his own standing and that his tactics rubbed more than a few people at the organization the wrong way.
Altman vs. Toner
When OpenAI's board first announced Altman's dismissal on November 17, it didn't offer much of an explanation except that Altman had not been "consistently candid in his communications with the board."
But new reporting suggests that the board may have been referring to instances in which Altman played board members off one another — especially ones who disagreed with his aggressive approach to rolling out artificial-intelligence technology. From its inception, there has been tension at OpenAI over how carefully it should proceed, given the potential threat the technology poses to humanity.
Altman didn't always see eye to eye with the board member Helen Toner, for example.
In October, Toner, a researcher who works at a think tank based at Georgetown University, published a paper that not only praised OpenAI's rival Anthropic for delaying the release of its chatbot, Claude, but also criticized the "frantic corner-cutting" release of ChatGPT.
Altman called Toner about the paper and said it "could cause problems" with the Federal Trade Commission, which was already investigating OpenAI, The New York Times reported.
Toner offered to write an apology to OpenAI's board, but Altman later emailed OpenAI's executives himself and told them he had reproached Toner, the Times reported. "I did not feel we're on the same page on the damage of all this," the Times quoted from his email.
Their clash may have led Altman to sow tensions between Toner and another board member, Tasha McCauley.
Altman called other members of OpenAI's board and told them that McCauley — a tech entrepreneur and scientist at the RAND Corporation — wanted Toner off the board, people with knowledge of the discussions told the Times. McCauley later said this was "absolutely false" when board members asked her about the incident, the Times reported.
Altman vs. Sutskever
Altman and OpenAI's chief scientist (and former board member) Ilya Sutskever also had a strained relationship. Their differences were ideological to the core.
Sutskever was seen within OpenAI as an AI "visionary" with an academic approach that didn't necessarily go over well with an engineer such as Altman, those familiar with the situation previously told Business Insider. He worried that Altman was pushing OpenAI to develop technology too rapidly and wanted to take a more cautious approach.
Over time, Sutskever had also become frustrated with being "pushed out of decisions" about ChatGPT-5 and plans to scale the product and company, sources previously told BI.
Tensions came to a head in October when Altman promoted a researcher to a level equivalent to Sutskever, the Times reported. Sutskever saw it as a snub to his own standing in the company and, in protest, said to other board members that he might quit, which they saw as a demand that they choose between Sutskever or Altman, the Times reported.
Altman vs. everyone else
Some of the board's six members at the time found Altman to be disingenuous and a bit too calculating. Several of them had backgrounds in nonprofits or academia, and Altman's "move fast and break things" tech-executive approach didn't necessarily sit well with them, the New Yorker reported.
"They felt Sam had lied," a person familiar with the board's discussions told the New Yorker. They dreaded Altman's tactics so much that when they began talking about removing him, they were intent on ensuring it would be a surprise, The New Yorker reported. "It was clear that, as soon as Sam knew, he'd do anything he could to undermine the board," a person familiar with their discussions told the outlet.
In a meeting with OpenAI staff two nights after Altman was ousted, Sutskever said one explanation he had received from the board for Altman's dismissal was that Altman had given two board members two different opinions about a member within the organization, sources familiar with the meeting previously told Business Insider. The other explanation Sutskever offered was that Altman was said to have given the same project to two different people at the organization.
Altman himself has not denied he struggled with the board before his ousting. "It is clear that there were real misunderstandings between me and members of the board," he wrote on X a little less than two weeks after he was ousted.
The best interests of the company and the mission always come first. It is clear that there were real misunderstandings between me and members of the board. For my part, it is incredibly important to learn from this experience and apply those learnings as we move forward as a…
— Sam Altman (@sama) November 30, 2023
Altman hasn't publicly addressed the accusations that he was difficult to work with, but in an interview with Trevor Noah last week, he conceded that there was a need for more voices on the board concerned with AI safety. "I'm excited to have a second chance at getting all these things right. And we clearly got them wrong before," he told Noah.
That Altman's ouster was so brief suggested that he had the backing of the broader organization. It kicked off a wave of heart emojis on social media from OpenAI executives and a letter of support from staff threatening to quit if he wasn't reinstated. But not everyone at the company supported Altman.
A handful of senior leaders at OpenAI came to the board in the fall with grievances about Altman, the Washington Post reported. They suggested Altman could disrupt the workflow at OpenAI, and some — including those who manage large teams — said Altman would pit employees against one another in problematic ways, the Post reported.
The comments prompted the board to review Altman's conduct as CEO, and one employee told the board that Altman became hostile after the employee shared critical feedback with him. Altman then went on to undermine a member of that team, sources told the Post.
The fact Altman was so quickly reinstated as CEO suggests that none of these allegations were enough for the company's powerful backers, such as Microsoft. Sutskever, too, has expressed regret over his actions. Still, the word on the street is that Altman may have gotten the message that it's time to remake his image.