A lot has happened since OpenAI CEO Sam Altman was fired from his own company all the way back in 2023, but he’s now given his most detailed account yet of what really went down in the corporate coup which had captivated the tech world.
Sam Altman has spoken at length about his experience of being fired from OpenAI, and his subsequent return. Speaking on a podcast with Y Combinator’s Jessica Livingston, he said he was stunned when he first heard he was being ousted, and provided minute details how he navigated his return over the next few frenetic days.
“I don’t know what to say about it other than it was like totally crazy and surreal and felt like a dream,” Altman said about when he first learnt — on 17th November 2023 — that he was being fired as OpenAI CEO. At that point, he was in a hotel in Las Vegas with his partner Oliver. “I was caught completely off guard and I really did in that first moment think I was having some crazy vivid dream that I was about to wake up from,” he added.
Altman said the news had been broken to him by Ilya Sutskever, who was the Chief Scientist of OpenAI at the time. “I’d got a text message from Ilya (Sutskever) the night before saying, can you talk tomorrow? I said, sure. A few minutes before (the scheduled time) he said, here’s a Google Meet link. I was like, okay, let me go find a place to sit because I was in a hotel, and I (thought) it was gonna be a call,” Altman remembered.
“And then (the conversation about his ouster) happened,” Altman says. “It was like very strange. I couldn’t get any questions answered. And then they just put out a blog announcement, and turned off my email and stuff. So I’m out of town, my computer just got deactivated and I have a phone. I kind of didn’t think about what was gonna happen next, I don’t know if they did either, but it went crazy,” Altman remembers.
Altman says he immdiately got on non-stop calls to figure out what had happened, and to decide the best course of action. “Then the feeling of that afternoon was just like this kind of fog of war,” Altman says. “Like I had no idea what was happening. The team didn’t know what was happening. They said they were trying to get answers. I was trying to get answers. And it was like nonstop calls. It started as like, Sam must have done something horrible. Like this happens if you kill someone. I was just on call after call. Microsoft’s trying to figure out what’s happening. I’m talking to a couple of board members, talking to the whole executive team. We’re trying to stabilize things. (It was) very crazy,” he says.
Amidst the frenetic discussions, Altman remembered a moment he’d shared with his partner, Oliver. “So I’m on like these nonstop calls. Ollie has packed up our hotel room. The first time I had a minute where I was not on a phone call is as we were like getting on the plane. It was this very dramatic thing. It was in Las Vegas. It was super windy. We were like standing right outside the plane. (Oliver) been overhearing some of the calls and I had written him some notes. It was the first time we like really had a minute to talk. And as we’re about to walk up the steps, he puts his hands on my shoulders, and he’s like, Sam, I’m with you 100 percent no matter what, but what the f*** did you do?” Altman laughs. “And I was like, gosh, no! And he’s like, okay, I’m good. And we, like, got on the plane,” he says.
“There were a lot of personal power issues (over the ouster),” Altman says. “There were some legitimate AI safety issues that I really disagree with. I’m very proud of our safety track record. But I think other people have a different opinion on what we’ve done,” he says.
Altman conceded that had he been fired only over AI safety issues, he would’ve given credit to the board, but he felt the reasons ran deeper. “I will say if you believe that AGI safety is an existential issue and that the company or me was doing it wrong or otherwise unfit, then I admire the courage of convictions to say this is the logical conclusion rather than I think what most people do, which is say like, oh, AGI safety is this existential issue and it’s coming really soon. And what are we meant for dinner?” Altman says. Altman seemed to be implying that those in the board weren’t as focused on AI safety, because if they were, they wouldn’t have treated it as just another reason to fire him but an existential risk for the company and all of humanity.
“Anyway, by the time I had got back to San Francisco that evening, we understood a little bit more about it. There was this kind of dramatic thing where I had tried to get one of the other board members off in a way that I think I didn’t handle well, and there were concerns about safety and who was going to be in charge of stuff and whatever,” Altman says. He says that at this point, he was prepared to leave OpenAI, and was already thinking of next steps. “I was just like, whatever. I love OpenAI. I wish them the best. What I want to do is work on AGI research. I can figure out some other way to do that. I can do it at Microsoft. I can start a new company,” Altman says.
Altman says that others in the company were also trying to find details over why he was dismissed. “The executive team had been trying to talk to the board. Didn’t get any answers,” he says. He says that some people in OpenAI began taking things into their own hands. “People started like quitting the company. So a bunch of us met up that night and more of us met the next day. I had a call with two of the board members the next morning where they said we’d like to talk about you coming back,” he says.
“And I’m not proud of this, but my first reaction — in my head, my first reaction — was ‘absolutely not’. And then, because I was just like hurt and mad and whatever, you know, excited, and then my second thing was — and this is something I’m not proud of — I was like, I will come back if all of you resign right now,” he says.
Altman says that asking his board members to resign in return for him coming back didn’t help matters. “I think that was not a constructive thing. I think had I handled that differently, and said we can negotiate some version of the board where I feel like I can have a chance to actually function, (my return would’ve been sealed) that morning,” he says.
“But I responded with emotion. And then it kind of like kicked off this weekend of like crazy nonstop negotiation, planning, counter planning, whatever. I was mostly talking with (board member and Quora CEO Adam D’Angelo) who I think acted in good faith throughout it, but he was only one vote of several on the board,” Altman says.
“It seemed all weekend like I was going to come back and we were going to figure out a new board and it seemed like things were on a good trajectory.,” Altman says. But the drama wasn’t quite over. “And then Sunday night rolled around, and to my total shock, they’re like, we’re gonna appoint Emmett (Shear) as CEO,” he says.
“That was the angriest I had been the whole time,” Altman remembers. “Because I was like — the level of deception here — (is crazy). I was trying to keep things together. I was trying to like tell the company, hey, this is heading to a good place. Don’t go. Because all of our competitors, like these vultures, were circling, trying to recruit our employees,’ he says. Altman says that OpenAI CTO Mira Murati put her foot down over the appointment of Emmett Shear as CEO. “Mira Murati had told the board, because there were no good answers, she was gonna just resign and not do this. And they said, well we can’t have Mira do this, so we gotta pick somebody else,” he says.
“And at this point, I’m just like, what the f***? I’m ready to go to bed. Like enough is enough. So I announced that I am going to join Microsoft, and work on an AI research project there,” Altman says. “I stay up for like a couple of hours talking to people and I go to bed. And I hadn’t slept much in the last few nights, and I lay in bed and I’m very tired and still angry and full of adrenaline and I can’t sleep. So I take a sleeping pill, and I go to bed probably like three in the morning, something like that,” Altman recalled.
But things changed when he got up the next morning. “And then I’m rustled awake. And this time it really feels like a dream because I’m in a sleeping pill haze. Ollie’s like, you need to check your phone. And Ilya had said, I regret my participation in this, saying didn’t want to harm OpenAI,” Altman says. Ilya had posted on X on 20th November, saying “I deeply regret my participation in the board’s actions. I never intended to harm OpenAI. I love everything we’ve built together and I will do everything I can to reunite the company.”
“He didn’t say we should undo it, but something like that. And 700 — like 95 percent or something — of the OpenAI employees had signed this letter while I was asleep saying this is a horrible decision and we’re gonna resign. We’re gonna resign if Sam does not come back,” he says.
This pressure made the newly appointed CEO Emmett Shear also resign, and the board too relented. “Emmett had also told the board that was going to resign too. So, finally after some time — I cannot believe how long the whole thing went — we finally finished everything, and announced that we had sort of high level agreement, and that I was gonna come back, and the company got put back together. That part was all a crazy thing to live through but honestly you can kind of do a lot for four or five days on a kind of crazy adrenaline rush,” Altman says.
But the entire process took its toll on Altman. “The couple of months after (it happened) were just miserable. I was just exhausted. I was like in this stupor. I felt I would walk down the hallways and people would just avert their eyes to me. I felt like, you know, a patient with like some really bad diagnosis because they didn’t know what to say. They just felt bad,” he says.
The following few months were eventful as well. OpenAI continued releasing models and products without missing a beat. After disappearing from the public eye, Ilya Sutskever announced that he was leaving OpenAI, and launching his own AI lab named SSI, or Safe Superintelligence. Not long after, OpenAI CTO Mira Murati also left and started her own company, Thinking Machines, and took several OpenAI employees with her. OpenAI is still chugging along — it’s the most valuable AI company in the world, but its lead over its competitors has been seen to be diminishing in recent months. And Sam Altman continues to be its CEO, in spite of the narrow — and dramatic — brush he’d had with being ousted from the company he’d founded.