When the OpenAI board asked Sam Altman to return the day after he was fired, he initially felt defiant, hurt, and angry.
Sam Altman explains how to fire it and reset it in OpenAI
“It took me a few minutes to get over it and get over my ego and emotions to say, ‘Yes, of course I want to do it,’” he told me by phone on Wednesday. “Obviously I’ve really loved the company and I’ve poured my life’s energy into this for the last four and a half years full time, but actually longer than that for most of my time. And we’re making great progress on the mission that I care most about, which is the mission of safe and useful general artificial intelligence.
After a five-day attempted boardroom coup, Altman officially returned to the role of CEO of OpenAI on Wednesday. Microsoft, the company’s largest investor, plans to take a non-voting board seat as well.
During our interview, Altman repeatedly refused to answer the main question on everyone’s mind: exactly Why He was initially fired. OpenAI’s new board, led by Brett Taylor, will conduct an independent investigation into what happened. “I very much welcome that,” Altman told me.
Below is my full interview with OpenAI CEO Sam Altman and CTO Mira Moratti, which has been lightly edited for clarity:
Sam, I’d like to first address the elephant in the room, which is that we still don’t know exactly why you were fired in the first place. Why do you think you were fired?
Sam Altman: The Board will conduct an independent review here. I welcome that very much. I don’t have much to say right now but I’m looking forward to learning more.
Why do you think the board said it had lost confidence in you?
That would be a better question for them.
You He said on X Only now “it is clear that there was a real misunderstanding” between you and the board members. What are those misunderstandings?
I don’t feel ready to talk about it yet. I think it’s very important to let this review process run. I’m happy to talk about anything forward looking. I imagine there will be a time when I’ll be more than happy to talk about what happened here, but not now.
Can you tell me why you can’t talk about it now?
I just want to let this process go and not interfere with it.
You talked about Ilya Sutskever (Chief Scientist at OpenAI) In your memo (to employees). Can you tell me a little about why he changed his mind and decided to side with everyone?
Mira Moratti: We don’t know. You’ll have to ask Elijah that.
Sam, what was, in hindsight, the main driving force here that kept you coming back?
Altman: It was really interesting. Saturday morning, some board members called me and asked if I would be willing to talk about it. And my immediate reaction was kind of a challenge, and it was like, “Man, I’m hurt and angry, and I think this is bad.”
“It took a few minutes to let go and get over the ego and emotions.”
And then I immediately started thinking about it, obviously I really loved the company and I’ve poured my life force into this for the last four and a half years full time, but actually longer than that for most of my time. And we’re making great progress on the mission I care deeply about: safe and useful artificial general intelligence. But also the people here and all the partners who have bet on us, and Mira and the leadership team and all the people here who are doing a great job. It took me a few minutes to snap out of it and get over my ego and emotions to then say: “Yes, of course I want to do it.”
So the council asked you to come back?
Were you hesitant at first?
Not for long. There are a lot of emotions out there after it happened to me.
It was clear that the staff were with you. How big of a factor do you think that was?
We have certainly come through this with a stronger, more unified, focused and committed team. I thought we had a lot of conviction and focus before, and now I think we have a lot more method. So that’s the bright side of all this.
Throughout this whole thing, we have not lost a single employee or customer. Not only did they maintain the products even in the face of unmanageable growth, they also shipped new features. Research progress continued.
Do you want to return to the council?
This is going to sound like a PR topic: it’s not my focus right now. I have a mountain of difficult, important and urgent work. I want to be able to do my job well, but it’s not about (being) on the board or not. That’s not something I spend my time thinking about right now.
What did you do “Improve our management structure” He means? Will the structure of the nonprofit holding company change?
“I completely understand why people want an answer right now. But I also think it’s completely unreasonable to expect that.
That’s a better question for board members, but not now either. The honest answer is that they need time and we will support them through this until they break down and think about it. Clearly our management structure had a problem. The best way to fix this problem is that it will take some time. And I completely understand why people want an answer now. But I also think it’s completely unreasonable to expect that.
Why do you think this is unreasonable? I think people see a lot of ups and downs about what happened. It seems that it was disagreements, not violations, or anything like that.
Oh, just because designing a really good governance structure, especially for such an impactful technology, is not a one-week question. It will take a significant amount of time for people to think about this, to discuss, to get outside perspectives, to pressure-test. This just takes some time.
Will anything change about OpenAI’s approach to safety work as a result of the events that have just unfolded?
Intend: No, this has nothing to do with safety.
Reports about Q model breakthrough* Which you all did recently, what’s going on there?
Altman: There is no specific comment on this unfortunate leak. But what we’ve been saying — two weeks ago, what we’ve been saying today, what we’ve been saying a year ago, what we’ve been saying back in the day — is that we expect progress in this technology to continue rapidly, and also that we expect to continue to work hard to figure out how to make it safe and useful. That’s why we woke up every day before. This is why we will wake up every day in the future. And I think we’ve been extraordinarily consistent about that.
Without commenting on any specific thing or project or anything else, we believe that progress is research. You can always hit a wall, but we expect progress to remain significant. And we want to engage with the world on this and figure out how to make this as good as possible.
Final question: I’m sure you’re still thinking about all this. I know it’s very fresh. What lesson have you learned from this whole saga?
I guess I don’t yet have an eloquent and elegant sounding answer there. Obviously there’s a lot more, but I keep stumbling through it all. I mean there’s definitely a lot to say there, but I don’t think I’m up for it… all I’d like to get is a long, rambling answer at this point.
Well, we’ll save it again.
After we hang up, Altman calls moments later.
I’ve learned that the company can really function without me, and that’s a very beautiful thing. I’m very happy to be back, don’t get me wrong. But I come back without any pressure of, “Man, I have to do this, or else the company needs me or anything.” Selfishly, I feel good because I either chose great leaders or mentored them well. It’s so nice to feel like the company will be perfectly fine without me, and the team is ready and on track.