What We Know About Sam Altman's Ouster, The Morning After

0:00Welcome to Big Technology Podcast, special Saturday edition, special emergency edition, where we're going to try to break down what exactly happened at OpenAI, what is happening at OpenAI, where OpenAI and the rest of the AI field go from here. Because the firing of Sam Altman by the OpenAI board is one of the more remarkable, one of the more shocking, one of the more surprising tech stories that I've ever seen. And it's going to have plenty of implications, not just for OpenAI, but for Microsoft, for Amazon, for Google, and basically for the entire tech industry. That's why this is so important.

0:41And that's why we're going to spend a little bit of time here going through exactly what happened and what comes next. This is a solo episode, no guest. I'm just going to run through a little bit of my notes, a little bit of what I understand has happened here. And I thought it was just important to get something quick up on the feed to give you some perspective about this developing story. One note before we go ahead, this is obviously a developing story with a lot of fast-moving, quick-moving parts. So bear with me. We're going to continue to cover it on the Big Technology Podcast feed as we go on, as we learn more.

1:12I want to be very careful about the information that I share. But we are starting to learn what's coming out. And the first thing that I'll say is what happened. So we have a post on Twitter from Greg Brockman, who's formerly the president and the head of the board at OpenAI, the co-founder of that company, who shared a little bit of a play-by-play of how he and Sam Altman found out that both Sam was out and that he was going to be removed from the board. Okay, here's what he says on Twitter. He goes, last night, and this is from early this morning. Last night, Sam got a text from Ilya, and Ilya is Ilya Sutskever, who is the chief scientist and a co-founder of OpenAI.

1:51Sam got a text from Ilya asking to talk at noon Friday. Sam joined a Google Meet with the whole board except Greg, who was not there. Ilya told Sam that he was being fired, and the news was going out very soon. At 12.19 p.m., Greg got a text from Ilya asking for a quick call. At 12.23 p.m., Ilya sent a Google Meet link. Greg was told that he was being removed from the board, but was vital to the company and would retain his role, and that Sam had been fired. Around that time, OpenAI published a blog post. As far as we know, the management team was made aware of this shortly after, and other than Mira, who's the new CEO, Mira Mirati, who found out the night prior.

2:32The outpouring of support has been really nice. Thank you, but please don't spend any time being concerned. will be fine. Greater things. Okay, so basically what this looks like is a complete coup from the OpenAI board. And the board, of course, includes Sutskever, and then independent directors like core CEO Adam D 'Angelo, tech entrepreneur Tasha McCauley, Helen Toner of the Georgetown Center for Security and Emerging Technology. And that's basically it. So they decided that it was time for both Greg Brockman to be removed from the board and Sam Altman to go. Now, initially, it sounded like this was a conduct thing.

3:11I mean, to have the CEO removed last minute and, you know, basically announced blindsiding Microsoft, shared as the market was still open, totally unprecedented, at least not in any near memory. But we're starting to learn more, you know, in terms of what that conduct might have looked like. And initially it sort of said, okay, it seemed like it was maybe bad behavior and there might be some repercussions for the OpenAI board as we go forward in terms of the accusations that they made. But then it starts to come out that it might have just been a little bit of a difference of opinion in terms of the future and the direction of this company.

3:51So I want to bring your attention to a memo from the chief operating officer of OpenAI, Brad Lightcap, which Axios published today. So he says,

4:29This was a breakdown in communication between the board and Sam. Our position as a company remains extremely strong, and Microsoft remains fully committed to our partnership. Interim CEO Mir Maradi has our full support as CEO. We still share your concerns about how this process has been handled and are working to resolve the situation, and will provide updates as we are able. I'm sure you're all feeling confusion, sadness, and perhaps some fear. We are fully focused on handling this, pushing toward resolution and clarity and getting back to work. Our collective responsibility now is to our teammates, partners, users, customers, and broader world who shares our vision of broadly beneficial AGI or artificial general intelligence.

5:12Hang in there. We are behind you 1000%. Okay. So if it's not malfeasance, if it's not financial mismanagement, what is it? And the likeliest theory right now is safety. Simply that OpenAI was moving too fast, that it was pushing this technology, the cutting edge of this technology forward too quickly. And it got uncomfortable, both for Sutskever, who seems to be the person who spearheaded this move, but also for the rest of the board. Now, here's a reminder coming from Nick Thompson, who's the CEO of The Atlantic, that Sutskever was a protege of Jeff Hinton. who's now perhaps the most passionate critic of the risks of out-of-control AI.

5:55Jeff Hinton, of course, helped invent the modern genre of AI. He's part of this group called the Deep Learning Conspiracy, along with Jan LeCun and Joshua Benjo, who's a professor in Canada, basically helped make this current era of AI go forward with their deep learning model. And Hinton, who left Google, has come out passionately talking about the risks of this AI. And it's a good point to talk about a little bit about the corporate structure of OpenAI. It's very confusing. They effectively started as a nonprofit and they moved to a cap for profit. But they always kept the power in the hands of the board to decide whether or not to remove the CEO, whether things got, you know, if things got a little bit too hairy, shall we say, in the area of safety.

6:42And that seems to be exactly what they did. So where did these safety concerns come from? I think that we can go back to two moments. The first is just earlier this month in Developer Day inside OpenAI. So if you remember some of the coverage that we've done here about Developer Day, what OpenAI did was it made it possible for people to build their own chatbots based on its technology with no coding experience. They open it up to basically create your own GPT. And what it did is basically took this extremely powerful technology that the company had been very careful about building safeguards around and they opened it up to everyone.

7:23And I don't think it's a complete coincidence that this is happening right after that demo day because, you know, rather than having some of those safeguards, now everybody has it. And there's been some rumors, you know, going around talking about how this is a product of what happened at demo day. I think that is a potentially crucial moment where all of a sudden Sam Altman took the powerful technology that OpenAI had and democratized it almost too much for, potentially, for the liking of people within the company. Also, maybe something that also played a role is this copyright shield, although I doubt it, but basically the fact that OpenAI said, if you get sued for copyright, we're going to back you up legally.

8:01Maybe that played some factor. But there had also been this happening. This had also been happening in concert with some really unbelievable developments within the company. And I'm going to point you to this interview that Lorene Powell Jobs did just a couple of days before Altman's firing, just talking about the models that are underdeveloped. Right. So Lorene Powell Jobs says to Altman, what is the most remarkable surprise that you believe will have happened in your field or in your company in 2024? for. And Altman says, the model capability will have taken such a leap forward that no one expected.

8:38Jobs says, wait, say it again. And Altman says, the model capability, like what the systems can do, will have taken such a leap forward that no one expected that much progress. Here's Jobs. And why is that a remarkable thing? Why is it brilliant? And Altman says, well, it's just different to expectation. I think people have like in their mind how much better the model will be next year. And it's remarkable about how different it is. Okay, so clearly this technology had been advancing dramatically under Altman. And he had his foot on the gas pedal, taking funding, working to productize it as a company.

9:14And the company effectively had these nonprofit roots that were always going to be in tension with this model of saying, let's go ahead and push forward, you know, take funding from Microsoft, productize it, let anyone build their own GPT. And I'm not saying this is exactly what happened, but it sure points to the fact that the board structure and the ambition of Altman just came to a head. And that's why this firing is important, because it's not just the deposing of a leader. It's a reorientation of direction. Right. If this is actually the way that this is going to move, it's not going to be the same open AI anymore, because, you know, they may have been researching these incredible breakthroughs.

9:54But the way that they're going to move into the public's hands is going to be quite different. And the way that they're going to productize this is going to be quite different. It's almost like this cycle, like that OpenAI needed money to function, of course, because this stuff is extremely expensive to run. So it goes to Microsoft. Now, to justify this money, it builds products and it goes splashy, these big PR tours that Altman is going forward with. And then, you know, you use that money and then you research and then you have to go through another cycle and it just builds and builds and builds.

10:30And, you know, right now there is also this moment of competition in AI where you had open AI, but you had all these other companies like Google with Gemini and Amazon and Google funding Anthropic and maybe even Elon Musk's Grok coming in and being like, we can do the same thing too. And the pressure goes on and it builds. And there's a tension there between a company that's basically founded on the principle of developing AI safely and a company that wants to lead and be part of the game. And it's almost as if there's this go fast and go slow tension that a lot of people have been pointing out.

11:02And if indeed OpenAI has decided that it's choosing go slow, then it is a very, very different world for OpenAI going forward. And that is a serious implication to Microsoft, which we're going to cover in a second. But let me just quote Aaron Levy, the CEO of Box and a friend of this show. He says, this is not your standard startup leadership shakeup. Ten thousands of startups are building on OpenAI and have assumed a certain degree of technical velocity and commercial stability. this instantly changes the structure of the industry i mean levy puts it pretty much you know better than i can which is effectively saying in progress there are startups building on top of open ai's technology and this potentially if the go slow part of this company has won out this potentially changes everything about their roadmaps the way they're going to build and you know maybe gives rise to others but we're going to start to see this very very big change Now, one caveat here is that Mira Mirati, who is the now interim CEO, has been somebody that's been on board with this mission, right?

12:03She was the chief technology officer. She was involved in ChatGPT. She was involved in the Dolly. She was involved in the Microsoft deal. So putting her in charge, you know, and not Elias Eskever indicates that in some way, maybe they're continuing to go forward in a similar direction, although not exactly the same way. But it does seem right now, and of course, this is subject to change, that safety was a huge part of it. So where does this leave Microsoft? I mean, this is not a good thing for Microsoft. You know, Microsoft put almost all of its eggs in this OpenAI basket. Of course, it's developing a little bit on its own, but it will be impacted strategically here.

12:41And not surprisingly, the executives at Microsoft are not happy. This is coming from a Bloomberg story. It says executives at Microsoft Corp, the largest investor in OpenAI. By the way, they own 49 % of the company, still do. They were also taken by surprise, the story says. Microsoft CEO Satya Nadella was blindsided, quote unquote, by the news and was furious, according to someone with direct knowledge of his thinking. I mean, of course, Satya Nadella is going to be mad. He's going to be really mad. Selling open AI service through Azure has been a crucial component of Microsoft's strategic rise over the past couple of months.

13:17I mean, they're already in good shape, but they've really benefited from this AI. In fact, the AI narrative, AI and open AI in particular, has been central to their growth in 2023. It seems like Satya Nadella does not waste a moment to appear on stage or an opportunity to appear on stage with Sam Altman. They announced in their earnings call recently that they've gone from 11 ,000 to 18 ,000 clients of their open AI service within Azure in one quarter. So clearly this was where the momentum was for Microsoft. and I'm going to talk about it in a moment, but let me just preview it. If Sam Altman is not at OpenAI and decides to build a competing company, right?

13:59Just think about that. And OpenAI decides that the velocity gets a little bit slower, then it could potentially be a real strategic liability for Microsoft. Okay. Before we get there, let's just talk a little bit about the AI safety crowd. so this is from Louise Metzakis who's a friend of the program and is at Semaphore, she says if Altman was fired over AI safety concerns, that's way crazier in a lot of ways than a sex or financial scandal, it shows how deeply committed effective altruists are to the idea that it is essentially borrowed from science fiction, that AGI could kill all of humanity if true, I think this could completely radicalize people working on AI in one way or the other, either you say wow, this threat was so extreme that they got rid of Altman, or these people are so crazy in their beliefs, they got rid of Altman.

14:48And honestly, I think that AI doomers or AI, you know, whether AI can wipe us out is about to go on trial. I mean, it's really going to be a public, if this was the case that they actually decided to push Altman out because of safety concerns, we're about to see the public really get a hearing on these doomers. I mean, they could have potentially incinerated billions of dollars of value for these concerns. And the question is like, are they so prominent? Are they so present that it was worth doing this right now? You know, maybe they were a front for more interpersonal problems. Who knows? I mean, we have Sitzkever also tweeting that ego is the barrier to growth a couple months ago.

15:27Who knows? Maybe that was a subtweet. But we are going to start to see a deeper and more harsh interrogation of whether doomers, their belief that we're going to get killed by AI really stands up to scrutiny. And I want to say it's already begun because Eliezer Yudkowsky, who's like the high priest of AI doomerism, who talks about our potential to be killed by AI, had, I thought was a pretty innocuous tweet yesterday responding to the Altman news. He goes, Mir Marati reached out to me in 2022 for a one hour zoom call. Sam Altman never essayed any such content. Also, I don't think Marati has made any jokes about how funny it would be if the world ended, I'm tentatively 8.5 % more cheerful about open AI going forward.

16:10Now, of course, lots of ego, lots of me, me, me in this statement by Yudkowsky, but he was piled on by so many people after this tweet. And I think that's just anger coming out saying, we don't know whether the AI is going to kill us. I mean, and I've said this before, we seriously don't have any proof or any even rational arguments about how this will happen. And now you've basically kneecap the world's most popular startup because of it. So I do think the AI doom narrative is about to go on trial in a big way. And now let's get to what Sam Altman does next. So Altman is out, Greg Brockman pushed off the board, and then he resigned his position from within the company.

16:51And then there have been three more open AI, prominent open AI researchers that have left as well. You have Jakub Pachowski, the company's director of research. This is according to the information, Alexander Madri, the head of a team evaluating the potential risks from AI, and Simone Sidor, a seven-year researcher, they all resigned. Look, I do think that they all believe in researching this. They are in an amazing position to push the category forward. What happens next? It seems to me like they are going to start their own company, that they're going to continue to press forward. They're going to continue this mission.

17:27And by By the way, again, they weren't exactly building on proprietary technology. I mean, some of this, a lot of this, almost all of this is built on the transformer model that was researched within Google. You have a couple, you know, you take a couple million to train up a new model. You could potentially get right back in the game and maybe do it from scratch without any tech debt, any legacy risk. And that's clearly what a lot of people think is going to happen. People in the know. So here's Jim Phan from NVIDIA. He's their senior AI scientist writing on LinkedIn. And by the way, he was an intern at OpenAI.

18:00He says, zooming out, we will inevitably see the birth of a new heavyweight competitor or new heavyweight competitors. And that isn't necessarily a bad thing for the community. AI will be a bit more decentralized. New capital will pour in. Every party will act with more urgency. And new PhD grads will have more. We'll have at least one extra job offer to consider. So it certainly seems like we're about to see the formation of another AI startup, perhaps. Well, I would say likely with some of these open AI exiles. I don't think they've given up on the mission. And they're certainly going to have plenty of money thrown at them.

18:36I mean, you're starting to see lots of even joke letters coming in, you know, on Twitter of people who are talking about how they want to fund this company. But I mean, if you have Sam Altman, Greg Brockman, and just three of these researchers raise their hand and say, hey, we want to do this right, you know, potentially think about how much money they could raise. I mean, it's definitely in the billions. The question is what number of billions that's going to look like. And if that's the case, it really does put these these doomer strategies sort of to the test, right? Because if they really were so concerned about AI safety, and maybe you do what you can do, but effectively they're going to unleash one more startup, more startups are going to open basically a huge gaping hole in the AI world for more startups to come in.

19:22Maybe some that don't end up being so dedicated to safety. And the question is, what will happen to open AI next? small company, relatively given the amount of employees and other big tech companies, and losing this type of talent has to have an impact. We talked about it on yesterday's emergency podcast. There will be an impact there. There will be definitely an impact to Microsoft, and there'll be a great shuffling of AI competition. And I said on CNBC yesterday, but the AI Darwinism, I mean, it's about to go into another gear and an already intense competition is just going to be pushed up another notch.

20:00All right, before we go, I want to talk about a couple of other possibilities that might end up showing up. I don't think the end of the story has come out yet. I do think we're going to hear more. But I do have a couple of other thoughts of something that might have happened. This is, of course, total speculation. So take it for what it's worth. But I want to cover some of these other possibilities. One is that they could have been potentially running out of money. I know the board said there was nothing financial here. Let's see what happens. I mean, this stuff is super expensive to run. And on Big Technology Podcast here, Ranjan and I have talked about how the model is just vulnerable.

20:34And so that's potentially something that might have happened. Now, the other possibility is the breakthrough possibility, right? Which is that, you know, there's all these jokes about open AI achieving artificial general intelligence or human level intelligence. But there is a possibility that some massive breakthrough came out, which made the board say, OK, right now, we really need to deal with safety. and, you know, either you're in or you're out. Maybe there was dispute there. Maybe that was the communication breakdown. In which case, I'd be thrilled to see like what this breakthrough is. I mean, these comments with Altman and Lorene Powell jobs, you know, are something that I'm going to start paying way more attention to because we have to figure out what was brewing inside of there.

21:14You know, maybe it was something else like that other possibility. Who knows? Maybe there was some, you know, some actual behavior. You know, I don't think that's what it was, but you can never rule that out completely. and then maybe some combination of all of this. Okay, obviously a live wire story, lots breaking here. I'm going to get back to reporting, calling my sources. We'll try to have another podcast on the feed on Wednesday to go deeper into what happened. I don't think we'll do it before then. In the meantime, just stay tuned and check out Big Technology, the newsletter, sign up for premium if you can.

21:51Stay tuned to the podcast feed. But that is, again, what I think is going to happen. Two big takeaways here. A reshuffling of competition in the AI world is absolutely here. And again, for OpenAI, not just deposing of a leader, but a reorientation of direction for the company. That's exactly where it seems like we're heading, which is going to make this space, like Jim Phan says from NVIDIA, way more live, way more intense, way more competitive. And perhaps it will make it innovate even faster. All right, that will do it for me here. Thank you so much for watching if you're here on our live stream and thank you for listening.

22:29We'll see you next time on Big Technology Podcast.