During assembly on Monday, November 27, Sam Altman '03 was interviewed by JBS Business Club leaders William Howlett '18 and Ian Sajjapong '18. Sam is president of Y Combinator, which funds early stage startups throughout the world; co-founder of OpenAI, a non-profit research company focused on building safe and accessible AGI (artificial general intelligence); and on the Stanford University faculty.
Andy Abbott, who was Sam's 11th and 12th grade principal at Burroughs, introduced Sam. Abbott said Sam was both an official and unofficial student leader. He was captain of the water polo team, editor of the yearbook and founder of the Gay-Straight Alliance (now Spectrum). As an accomplished computer science student, Sam helped rebuild the Burroughs website. But beyond this and more, "Sam was a voice at the school. He was frequently at this podium giving sound-offs. ... He did an awful lot to help us be better as a school." After Burroughs, Sam attended Stanford but left at age 19 to develop Loopt, one of the first apps that let smartphone users share their location with others. He became president of Y Combinator in 2014. Abbott said, "Sam never graduated from Stanford, but he now teaches there. ... He is one of the most important voices in the tech industry, and I would wager that every entrepreneur in America would give her or his right arm for five minutes with Sam Altman and we're going to get him for almost an hour."
Q&A Summary (not verbatim)
William: How do you go from running a small startup, Loopt, to running one of the largest, most important organizations in Silicon Valley, Y Combinator?
Sam: It didn't start that way. Y Combinator for a long time felt like this kind of hack that wasn't going to work, and that no one took very seriously [even though Loopt was the first company ever funded by Y Combinator]. The idea behind Y Combinator is to take people in their early 20s, late teens, give them money [even though they have] no business experience, and sort of hope for the best. Now it's accepted as a good idea, but 10 years ago, it certainly wasn't. I like helping early stage companies (nuclear fusion, supersonic planes, new cures for cancer, nanobots, build a base on the moon) ~ it's really fun. And I had become friends over many years with the guy who started Y Combinator. When he was ready to retire, I took over.
Ian: You have been head of Y Combinator for about three years. What goals have you had and what goals or plans do you have for the future?
Sam: About 30,000 people apply each year. We pick about 400 and invest $120,000 in each, which is enough money for them to survive for a year. And we help them with whatever they're doing. We do this twice a year. The first goal is to just do more of that. We're only at about 5 percent of the size we should be and how we can scale that up [is the goal]. There are all of these network effects ~ if you become one of our companies you have this full economy of thousands of other companies. So goal one is to scale it up. We now operate around the world, in about 50 to 60 countries this year. We want to fund any company that can have a meaningful impact on the world and be a $10+ billion company. We will fund companies in any market. Companies learn so much from each other, even if they're doing totally different things. Our goal as an organization is that we want to be the institution that creates the most innovation in the world. Our mission is to make the future really good for everybody. That's mostly about startups but it's also about advocating for certain policies. We are in the middle of this huge change in the world. We're at the very beginning of this automation revolution where basically everything is going to be done better by software. And trying to figure how that will go well is a big challenge for society. We want to be a voice for how that can be done well.
Ian: We have a lot of aspiring entrepreneurs in the audience. What qualities do you look for in choosing people/companies/ideas to fund?
Sam: I'll talk about a few of the non-obvious ones: The most important non-obvious characteristic we look for in entrepreneurs is determination. This is more important than anything else. Starting a startup can be really hard ~ things go constantly wrong. And most people make the very rational decision to give up. A good idea is also hard to evaluate. If you imagine a Venn diagram of startup ideas: two circles, one looks like a bad idea and the other is a good idea ~ you want the things in that very narrow intersection. Because the obvious good ideas get done by the big companies or a lot of other people. We also look for ideas that get better with scale [e.g., social media]. We look for founders who are good communicators. If you can't communicate well, you will not be able to hire, you will not be able to raise money, you won't be good at selling your product, you can't get the press to cover you. We look for people with original ideas. I think the hardest skill to develop, the thing that school trains you for the least, is having original thoughts. That trait (having truly original thought) is shared by almost all of our most successful entrepreneurs but the trait you see the least often in the general population. Finally, we look for companies that have a real sense of mission. It is relatively easy to get a company started. It is harder than ever to scale a company because there are so many. If you don't have a real mission, a real reason that people want to join you and do this crazy thing, you won't be able to get real talent or you certainly won't be able to keep them. Being able to explain why it matters is super important.
William: You co-founded OpenAI with Elon Musk to help develop a general artificial intelligence. Why did you found this organization and what do you hope to accomplish?
Sam: We are trying to build true Sci-Fi AGI ~ a computer program that is smarter than humans in every way. I think this is possibly the most important project in the world right now. This is going to change everything about everything in not obviously good ways, and I don't think people take it seriously enough. [Within 10 or 15 years], we will have computer programs that are much smarter than humans in every way, and, hopefully, they like us. But if not, it's pretty bad. It can be really good. It can be true utopia, where we have enough of everything, we eliminate poverty, we cure all disease. But that's not the default. We are really interested in how we get this built in a way that is good for humans ~ not in a way where it's humans versus machines. The rate of progress is amazing. We're building custom hardware that's getting 10 times faster every year. The thing that makes AI work is that the hardware is getting more and more powerful, and we've accelerated the curve. So it's going to come soon. We are a non-profit. We're only focused on trying to make this go well for humanity. I think we are the first species to design our own descendants, sort of like intentional evolution. And very soon [there will be] a very broad conversation about how we want that to go. But right now it's a very small set of people building it.
William: How do you think society should address the advent of automation and artificial intelligence and its effect on employment?
Sam: If aliens landed on earth and they were really smart, the first question would not be about jobs ~ it would be about survival. That's the question people don't talk about enough. I do expect jobs to change, but technology always changes jobs. That's what people forget. Very roughly, every 75 years, 50 percent of the jobs get eliminated by technology. And then people find new things to do. My general framework is all repetitive human work that does not require an emotional connection will be done by machines 10 to 15 to 20 years from now. But that's okay. There are plenty of new things to do. I do think we'll have to change the social contract, the economic model we have. But what people forget is when machines do all this work, the cost of business services goes to almost zero. So we'll be able to do something like a universal basic income, and people will be able to work on whatever they want.
Ian: You recently entered the political field through a progressive policy initiative, the United Slate. Can you tell us about your political vision?
Sam: We just want to get some candidates elected who understand technology. I think in general our leaders don't. If you try to go to D.C. to talk about AI policy, which I've done, it really feels like a fool's errand. It would be really great to get some people in who understand this important piece of the country, and how we make sure it works for everyone in a really good way. So it's really that simple. We're trying to identify candidates who have some new ideas and understand this, and then give them technology and money to get elected.
Ian: What are the most important policies you think need to be changed?
Sam: Fair economic growth. We need to figure out how to use technology to grow the economy and then have policies so that everyone benefits from that. In California, in particular, we have this wild cost-of-living problem ~ how do we address that? And, then, how do we make sure that the U.S. stays at the forefront of R&D?
William: It seems that technology can exacerbate economic inequality. How do you think you can use it to reduce it?
Sam: It definitely does. Technology creates wealth but tends to concentrate it. Technology is going to do its thing (and we probably can't and probably shouldn't stop that), [so what we think about is] how to build social policies on top of that to keep the world fair. Certainly, basic income is one idea people talk about.
William: How did Burroughs prepare you for what you've done? What was most valuable about your time here?
Sam: I have incredible gratitude for Burroughs ~ more than for my college. I still think about all of my Burroughs teachers and I can't remember the names of my college teachers. The world and college push you to hyper specialize. The thing that has worked well for me is knowing a little about a lot of things. The most interesting ideas happen at the intersection of two fields, or more fields. The really great investments we've made have been because someone was able to put a few things together. I would encourage all of you to do this ~ and this was the thing that was really great about Burroughs ~ I learned a little about a lot of things. I don't know how, but I think I also learned about new thinking here ~ that's really important as well. One of the things I learned at Burroughs that was really valuable was that I didn't have to stick on some traditional path. When I got to university, it was like this very single track. One thing I didn't learn that I wish I had was that my default belief, until I was 25, was that things just automatically get better. I realize now that the default is that things actually get worse. Things only go forward when people make it happen. [So the question is] how are you going to drive the collective humanity ~ it doesn't just happen.
William: What are the biggest lessons you've learned so far and what are your plans for the future?
Sam: I kind of like what I'm doing. I'd like to figure out how to ramp up my time more on OpenAI. But, I do generally believe that startups are the best way to get new ideas in the world.
Ian: What is your typical day like and what do you most enjoy ?
Sam: Most of my day is meeting with the companies we've funded and helping them with whatever their problem is ~ I probably meet 12 companies a day. The fun thing is that they all have very bad but different problems, so it never gets repetitive. The number of ways that a startup can have existential crises never stops amazing me. The second biggest thing I do is finding, recruiting, selecting new companies to fund. And at Open AI, it's technical problems mostly.