In Sam Altman We Trust?

0:00Hey, everybody. Katie and Lauren and I are taking a week off from taping a new episode, but we wanted to share one of our favorite episodes of Uncanny Valley this week. It's all about OpenAI's Sam Altman. We trace his origin story from his Midwest roots to his tumultuous rise at OpenAI. And of course, we question whether he is the person we should trust to create our AI future. We hope you enjoy it. We have been a misunderstood and badly mocked org for a long time. Like when we started and said we were going to work on AGI, like people thought we were batshit insane. Sam Altman is the CEO and one of the founders of OpenAI, the generative AI company that launched ChatGPT about two years ago and essentially ushered in a new era of artificial intelligence.

0:50This is Wired's Uncanny Valley, a show about the people, power and influence of Silicon Valley. So today on the show, we're doing a deep dive on Sam Altman, from his Midwest roots to his early startup days to his time as a venture capitalist and his rise and fall and rise again at OpenAI. We're going to look at it all while asking, is this the man we should trust to guide our explorations into artificial intelligence? And do we even have a choice? I'm Michael Kalori, Director of Consumer Tech and Culture here at Wired. And I'm Lauren Good. I'm a senior writer at Wired. And I'm Zoe Schiffer, Wired's Director of Business and Industry.

1:36Okay, so I want to start today by going back one year into the past, November 2023, to an event that we refer to as the blip. The blip. We don't just refer to it as the blip. That is actually the internal phrase that is used at OpenAI to describe some of the most chaotic three to four days in that company's history. The company OpenAI, one of the top players in artificial intelligence, thrown into disarray. Being one of the most spectacular corporate fallout. The news on Wall Street today involves the stunning developments in the world of artificial intelligence. It really started on November 17th, this Friday afternoon, when Sam Altman, the CEO of the company, gets what he says is the most surprising, shocking and difficult news of his professional career.

2:29The shock dismissal of former boss Sam Altman. His firing sent shockwaves through Silicon Valley. The board at OpenAI, which at the time is a nonprofit, has lost confidence in him, it says. And despite the fact that the company is, by all measures, doing incredibly well, he's out. He's no longer going to lead the company. So he's effectively fired from the company that he co-founded. Yeah, and that immediately sets off a chain reaction of events. So his co-founder and president of the company, Greg Brockman, resigns in solidarity. Microsoft CEO Satya Nadella says that Sam Altman's actually going to join Microsoft and lead an advanced AI research team there.

3:09And then we see almost the entire employee base at OpenAI sign a letter saying that, wait, wait, wait, if Sam leaves, we're leaving too. Some 500 of these 700-odd employees. Threatening to quit over the board's abrupt firing of OpenAI's popular CEO, Sam Altman. And eventually there's this back and forth tense negotiation between Sam Altman and the board of directors. And eventually the board then installs Mayor Maradi, the CTO, as the interim CEO. CEO. And then shortly after that, Sam is able to reach an agreement with the board and he returns as CEO. And the board looks instantly different with Brett Taylor and Larry Summers joining, Adam D 'Angelo staying, and the rest of the board leaving.

3:53So all of this happened over a weekend in the first couple of days of the following week. It ruined a lot of our weekends as technology journalists. I'm sure it ruined a lot of weekends for everybody in the generative AI industry. But for a lot of people who don't follow this stuff. It was the first time that they heard Sam Altman, first time maybe they heard about OpenAI. Why was this important? Yeah, this was such an interesting moment because it kind of surprised me, Lauren. I'm curious if it surprised you that this story broke through on such a national scale. Like people went from not really knowing who Sam Altman was to being shocked and disturbed that he was fired from his own company.

4:29So I think at this moment, everyone was hearing about generative AI and how it was going to change our lives. Sam was like now officially sort of the face of this. And it took this chaos, this Silicon Valley drama to kind of bring that to the forefront. Because then in trying to understand what was happening with the mutiny, you know, you got a sense of like, oh, there are these people in these different camps of AI. Some who believe in artificial general intelligence that eventually it's going to totally take over our lives. Some who fall in the camp of accelerationists who believe that AI should just scale as quickly as possible, you know, unfettered AI.

5:06Others who are a little bit more cautious in their approach and believe that there should be safety measures and guardrails put around AI. So, like, all of these things came into sharp focus through this one very long chaotic weekend. So, we're going to talk a lot about Sam on this episode. And I want to make sure that we get a sense of who he is as a person. So, how do we know him? How do we understand him as a person? Like, what's his vibe? Well, I think Lauren might be the only one who's actually met him. Is that right? Yeah, I have met him a few times. And my first interaction with Sam goes back about a decade.

5:39So he's around 29 years old. And he is the president of Y Combinator, which is this very well-known startup incubator here in Silicon Valley. The idea is that all these startups pitch their ideas. And they get like a very small amount of seed funding. But they also get a lot of coaching and mentorship. And the person who is running YC is really like this kind of very glorified camp counselor for Silicon Valley. And that was Sam at the time. I remember talking to him briefly at one of their YC demo days in Mountain View. He exuded a lot of energy. He's clearly a smart guy. You know, he comes across as friendly and open.

6:14And people who know him really well will say he's really one of the most ambitious people who they know. But at first glance, you wouldn't necessarily think that he's the person, like fast forward 10 years later, meeting with prime ministers and heads of state around the world to talk to them about his grand vision of artificial intelligence and someone who's really positioning himself to be this power monger in artificial intelligence. Yeah, I mean, I think one of the interesting things about Sam is just that he seems like such an enigma. Like people have a really hard time, myself included, like putting a finger on like, what is he all about?

6:47Should we trust this guy? I think there are other CEOs and executives in the Valley who have such like brash personas, Elon Musk, Mark Andreessen, that you kind of instantly feel one way or another about them. You either really like them or you're really turned off by them. And Sam's somewhere in the middle. He seems kind of like a little more quiet, a little more thoughtful, a little more nerdy. But there's this sense of, like Lauren said, like an appetite for power that makes you pause and think like, wait, what is this guy all about? What's in it for him? He also wears a lot of Henley shirts.

7:18Now, I know this is not a fashion show. People who listen to our very first episode might be like, are they just going to talk about hoodies every episode? But he wears a lot of Henleys and jeans and nice sneakers, except for when he's meeting with heads of state where he suits up appropriately. And Sam, we have notes on the outfits if you want them, so call us. Well, Zoe, to your point, Paul Graham, who was at one time the head of Y Combinator, the incubator that Lauren was just talking about, described Sam as extremely good at becoming powerful. Ooh. So it seems like he's somebody who can sort of read the situation and read the room and figure out where to go next before other people can figure it out.

7:57So a lot of people point to parallels between Sam Altman and Steve Jobs. I feel like Steve Jobs was somebody who had a vision for the future and was able to communicate why it was important. And he had a consumer product that got people very, very excited. Sam Altman is also somebody who has a vision for the future, is able to communicate why it's important to the rest of us, and has a product that everybody's excited about in ChatGPT. I think that's sort of where the parallel ends. I think just the question of, is Sam Altman the Steve Jobs of this era a really good one to ask? And you're right in that they were both effectively the salespeople for products that other people made.

8:39They're, in a sense, marketers. Jobs marketed the smartphone, which literally changed the world. Sam has helped productize AI, this new form of generative AI, through ChatGPT. So that's a similarity. They're both or were both extremely ambitious, enigmatic, as Zoe said. reportedly impatient. They have some employees who would follow them to the end of the earth and others who have gone running scared. They've both been ousted from companies they were running and then returned, although Sam's absence was infinitely shorter than Steve Jobs' return to Apple. I think there are some key differences, though.

9:14One is that we have the benefit of hindsight with Jobs. We know what he accomplished, and we have yet to see if Altman is really going to be that defining character of AI for the next 10 or 20 or 30 years. The second thing is that while Jobs had some very real personality quirks and complexes by all accounts, I tend to see him as being put into this messianic position by his acolytes, you know, whereas Sam seems to really, really want to elevate himself to that position. And there are some people who are still really skeptical about that. That's really interesting. I mean, just to double click, as the tech people say, on the salesperson analogy.

9:51That can sound a little reductive as well, but I actually think it's a really important point because large language models have existed. AI has existed to some extent. But if people, regular users, don't know how to use those products or interact with them, do they really break out and change the world in the way that people like Sam moment think they will, I would say no. And so, you know, his role in launching ChatGPT, which is, I think, by all accounts, not like an incredibly useful tool, but points to like future uses of AI and what they could be and how it could kind of interact with people's everyday lives.

10:27Like, that is a really important change and an important impact that he's had. Yeah. And, you know, Lauren touched on this point too. Like, we don't know what the impact is going to be. We don't have the benefit of hindsight here. And we don't know whether these promises about how artificial intelligence is going to transform all of our lives, how those are going to play out, right? And there are a lot of people who are very skeptical about AI, particularly artists and creative people, but also people who work in surveillance and security in the military have a fear and a healthy skepticism around AI.

11:03And we are looking at this figurehead as the person who is going to lead us into the future where this becomes the defining technology of that future. So the question becomes, can we trust this person? And I feel like Sam's feedback to that would be like, no, you shouldn't trust me and you shouldn't have to. Like he kind of has said in interviews before that he's trying to set up the company where he doesn't have quite as much power. He doesn't have total control in that way. And he has definitely hinted that he thinks a lot of the decisions governing AI should be democratic. I think it's just, again, this question of like, do we trust those statements or do we trust more what he's doing, which is consolidating power and leading a for-profit company, what used to be a nonprofit?

11:49Yeah, we should underscore that, that he's always encouraging healthy debate. He encourages open dialogue about what the limits of AI should be. But that still doesn't seem to be satisfying the people who are skeptical of it. Yeah, I mean, I think it's also important to make the distinction between people being skeptical of it and people who are fearful of it. too, right? There are people who are skeptical about the tech itself. There are people who are skeptical that Sam Altman may be the figurehead that we need here. And then there's fear of the tech from some people who actually do believe in its potential to revolutionize the world.

12:23And the fear is that that may not happen necessarily in a good way. Like maybe someone will use AI for bioterrorism. Maybe someone uses AI to launch nukes. Maybe the AI itself becomes so all-knowing and powerful that it somehow, you know, becomes violent against humans. Those are some real fears that research scientists and policymakers have. So the questions about Sam's trustworthiness aren't only from, let's call it a capital perspective, you know, what's going on in terms of his open AI nonprofit or a for-profit, or can we trust this man with billions of dollars in funding? It's literally, can we trust him with, in a sense, our lives?

13:00Right. It's like the degree to which we should be worried about Sam Altman is kind of the degree to which you personally believe that artificial general intelligence is a real concern or that AI could significantly change the world in a way that could be potentially quite catastrophic. Mm-hmm. Altman was quoted in a New York magazine profile about him as saying, AI is not a clean story of only benefit. Stuff is going to be lost here, right, as it develops. It's super relatable and natural for people to have loss aversion. They don't want to hear stories in which they are the casualty. So we have a self-appointed figurehead, whether we like it or not.

13:39And we have to make decisions about whether or not this is somebody we can trust. But before we get there, we should talk about how Sam got to where he is today. So we'll do that right after a break.

13:57This week on the political scene from The New Yorker, Trump's rupture in the world order. Europe caught between two adversarial great powers. That's basically dialing back the clock to not only pre-World War II, but really it's a pre-20th century view of the world. And I would say it's a world of permanent insecurity that we're looking at. Join me, Evan Osnos, and my colleagues Jane Mayer and Susan Glasser every Friday on The Political Scene, available wherever you get your podcasts.

14:35Welcome back to Uncanny Valley.

14:41What do we know about Sam Altman, the man, before he was Sam Altman, the tech founder? Well, he's the oldest of four children. He comes from this Midwestern Jewish family. He grew up in St. Louis. And his family, by all accounts, he had a pretty good childhood, I believe. They hung out together. They played games. I think his brother has said that Sam always needed to win those games. and in high school he comes out as gay which was pretty unusual at the time he said that his high school was actually quite intolerant toward gay people and at one point he gets up on stage at his school according to this New York Magazine profile and gives a speech about the importance of having an open society and my takeaway from that was that he was willing from an early age to be different from other people and to kind of reject the dominant ideology there was also this really funny section in the profile where they said that he was kind of like a boy genius and at age three was like fixing the family's vcr as the mom of a three-year-old i was like what the fuck i was resetting the clock on the family vcr by 11 and i just want to say that i'm also a boy genius it's funny with profiles of some of these founders because i do feel like there's this like hagiographic tendency to be like and of course they were a child genius too no one's ever just mid in childhood and then ends up being a billionaire.

16:05There's always something special. Right. Well, there was something special about Sam, that's for sure. Yeah. So he starts at Stanford in 2003. This is kind of the time period when other young, ambitious people are starting companies like Facebook and LinkedIn. And so, I think if you were smart and ambitious, like Sam clearly was, you weren't going to go into the traditional kind of law doctor type fields, you were going to try and start a company, which is what he does. Yeah, he starts this company called Looped. He's a sophomore at Stanford. He teams up with his then boyfriend and they create what's basically like an early four square.

16:43And this is when Sam gets involved with Y Combinator for the first time. He and his co-founder receive a$6 ,000 investment. They're accepted into a summer founders program at Y Combinator. And this is where they get to spend a few months incubating their app with mentorship. They're with a bunch of other nerds. There's this great detail about how Sam worked so much during that time period that he got scurvy. Oh, my God. That also feels like mythologizing. It really does. So fast forward a few years. It's 2012. Loopt has raised about$30 million in venture capital. And the company announces that it's going to be acquired by another company for around$43 million.

17:24And this sounds like a lot of money to people who don't like startups and sell their companies. But by Silicon Valley standards, this is not really deemed a success. Sam is comfortable. He, you know, is able to travel the world a bit, find himself, think about what he wants to do next. But there's still a lot of ambition there, and we're still yet to see, like, the real Sam Altman. Does money seem like it's a motivating force for him, or what is he after during this time period? Yeah, that's a really good question. And this, I think, is maybe emblematic of his personality, right? Like, because he doesn't just stop there.

17:57He's still ruminating on a lot of things. He's still thinking a lot about technology. And then in 2014, he was tapped by Paul Graham to be president of Y Combinator. So he's immersed himself and a bunch of other technologists with all these new ideas. Around 2015, as he's thinking about all this stuff, that's really when we have the seeds of OpenAI. So let's talk about OpenAI. Who are the co-founders? What do the early days of the company look like and what is its mission when it takes off? Originally, OpenAI is founded as this kind of group of researchers who are coming together to explore artificial general intelligence.

18:33A group of co-founders, including Sam and Elon Musk, set up OpenAI as a nonprofit. They don't really have a conception of having like a commercial component to it or a consumer-facing app involved. It really does feel like a research org at this point. But Elon Musk, in classic Elon Musk form, wants to wrest more control away from the other co-founders. He tries to take it over repeatedly and kind of floats the idea that Tesla could acquire OpenAI and the rest of his founders kind of reject this idea, allegedly. And so ultimately, Musk walks away and Sam Altman gets control of the company. And I think it's important to also make the distinction that there are a lot of companies working on artificial intelligence at this time, about eight or nine years ago.

19:16and open AI saw themselves as the good guys in the group. That's the open. Yeah. Like basically if you're building artificial intelligence tools, you know, they may be harnessed by military. They may be harnessed by bad actors. They may get to the point where they become dangerous, like some of the dangers that Lauren was talking about earlier. And they saw themselves as the people who are going to do it the right way. People are going to do it so that AI can benefit society and not harm society. They wanted to make their tools available for free to as many people as possible. And they really wanted to make sure that this wasn't just like a closed box that ended up just making a few people a bunch of money while everybody else sat on the outside.

20:02So that was kind of how they conceived of being good. It was like the democratization factor more than the like trust and safety factor. Is that fair? Like, were they sitting around being like, we're really going to study the harms, all of the potential misuses? Or was it more like, we're going to put this in people's hands and see what they do with it? It's a good question. I mean, I think they saw themselves as quote-unquote value aligned. Yes, that's a term that they all used. Yeah, yeah. I mean, in 2012, there was something called AlexNet that was this convolutional neural network. It was able to identify and classify images in a way that hadn't been done before by AI.

20:38It kind of blew people's minds. I mean, you can hear NVIDIA CEO Jensen Huang talk about how that, like, convinced him to change the company's direction towards AI. So a big moment. And then in 2017, there were some Google researchers that worked on this paper that was – it's known as the attention paper now that basically defined the modern-day transformers that make up the T in chat GPT. And so you're right, Mike. There were these groups and companies that were – they were hopping on the train. And opening AI from the beginning was like, we definitely want to be on that train, but we think that we're more value-aligned than the others.

21:11Something that they realized very quickly is that if you're building artificial intelligence models, you need massive amounts of computing power, and they did not have the money to buy that power. So they turned to daddy Microsoft. Well, they try and raise as a nonprofit. And Sam says it's just not successful. And so they're forced to take on a different model. And the nonprofit starts to kind of morph into this other thing that has a for-profit subsidiary. And OpenAI starts to look kind of weird and Frankenstein-y even in its early years. Yeah. So by the early 2020s, Sam had left Y Combinator.

21:51OpenAI was his full-time thing. They had created this for-profit arm. And then that enabled them to go to Microsoft, Daddy Warbucks, and raise, I think, a billion dollars to start. And then what is Sam's own journey during this time? Is he investing? Is he just running his company? He's meditating. He is very into meditation. I mean, in classic founder, venture capitalist kind of fashion, he's investing in a bunch of different companies. He's poured$375 million into Helion Energy, which is like a speculative nuclear fusion company. He's put$180 million into retro biosciences, which is looking at longevity and how people can live longer.

22:38And he's raised$115 million for WorldCoin, which, Lauren, you went to one of their events recently, right? Right. Yeah. So WorldCoin is a pretty fascinating company. And I think in some ways is also emblematic of Sam's approach, ambition, personality, because what they're creating is, well, there's an app, but there's also an orb, a physical orb, like a ball that you're supposed to gaze into and it captures your irises. And then it transforms that into an identity token that puts it on the blockchain. And the idea that Sam has expressed is that, look, in the not so distant future, there's going to be so much fakeness out there in the world because of AI, and people are going to be really easily able to spoof your identity thanks to AI, thanks to the AI I am building.

23:21Therefore, I have a solution for you. And his solution is this world coin now just called world product. So here's the man who is accelerating the development of AI and saying, here are the potential perils, but also, I have the solution, folks. I'm pro anything that lets me not remember my many, many passwords. So you can scan my iris, Sam. But okay, what else was he investing in? During this period, he's also, you know, getting rich. He's buying fancy cars. He's raising those fancy cars. He gets married. He says he wants to have kids soon. He purchases a$27 million house in San Francisco. And then he's pouring a ton of energy into OpenAI and specifically launching ChatGBT, which is going to be the kind of commercial face of what was previously a nonprofit.

24:10Yeah, and that's really, that's a watershed moment. It's the end of 2022, and all of a sudden, people have a user interface. It's not just some LLM that people don't fully understand or comprehend that's working behind the scenes. It's something that they can go to on their laptop or on their phone and type in and then have this kind of search experience that feels very conversational and different from the search experience that we've known and understood for the past 20 years. And Sam is the face of this. The product events that OpenAI starts hosting intermittently start to feel a little bit like Apple events in the way that people and people, I mean, us tech journalists are covering it.

24:50And then the next year in 2023, before the blip, you know, Sam goes on this kind of world tour. He's meeting with prime ministers and heads of state. And what he's doing is he's calling for a very specific kind of regulatory agency for AI. He has decided that if AI continues to grow in its influence and power, that it's going to be regulated at some point potentially. And that he wants to be a part of that conversation. And not just a part of that conversation, he wants to sort of have control over what that framework looks like. Right. I mean, you kind of pointed to this earlier, but there is this ongoing debate about artificial general intelligence.

25:27and the idea that AI will become sentient at some point and maybe even escape the box and turn against us humans. And Sam recently said that this isn't actually his top concern, which I think is a little concerning, but he said something smart, I thought, which was that there's a lot of harm that can be done without reaching AGI, like misinformation and political misuses of AI. They don't really need artificial intelligence to be that intelligent in order to be very, very destructive. Yeah, and job loss. We should also mention AI's impact on labor because a lot of companies are eager to cut costs and they buy into an AI system that automates some of the work they used to rely on humans to do.