Sam Altman (Co-founder & CEO of OpenAI; prev. President of Y Combinator)
Take a 10-minute break from whatever you're doing and read our conversation with Sam Altman. You'll be glad you did :)
|Michael Spiro||Sep 21, 2020||7|
We’re trying to help students & young professionals learn more about startups, tech, venture, and more. If you are interested in reading more interviews, just like this one, with founders, operators, and investors from today’s leading companies, feel free to subscribe below!
Interview Guest: Sam Altman (Twitter: @sama)
Current Role: Co-founder & CEO of OpenAI
What is OpenAI?: OpenAI is an AI research and deployment company working to ensure that artificial general intelligence benefits all of humanity.
Previously: Prior to joining OpenAI as CEO, Sam was the president of Y Combinator, the world’s leading startup accelerator, having helped launch companies such as Stripe, Airbnb, DoorDash, Twitch, Reddit, and Dropbox. Before joining Y Combinator, Sam founded Loopt (a location-based social networking app) while at Stanford in 2005. He later sold Loopt for $43.4M.
Quick Note: This interview was recorded via a Google Hangouts video between Sam and Michael (that’s me) in August.
(Huge thanks to Adam Kurkiewicz and Arun Sundaresan for helping make this one possible 😁)
Michael’s favorite quotes from the conversation:
On learning about startups: “Honestly, the best way to learn about a startup is just doing one. The first one may not work out, though sometimes it really does phenomenally well, but I think just getting going is super valuable.”
On OpenAI: “I certainly believe I will have much more of an effect with OpenAI than anything else I've ever done or will do in my career…make, develop, and deploy into the world super powerful general AI. We all think that this will be the most important technological development in a very long time, possibly in human history… but I will say I think in four years we'll have systems that are world-changing.”
On playing a long-term game: “Learning how to avoid the distraction of short term games, particularly games that even if you win, you still lose, is a very valuable skill.”
On getting good at sales: “But, the fun kind of sales is convincing someone of something that you deeply believe in. And when you think about it that way, at least when I think about it that way, that sounds cool and definitely something that is important to me to get good at.”
On knowing when something is working: “Most things don't work, but when something is truly legitimately working, it is unmistakable.”
THE FIRST FEW QUESTIONS I ASKED SAM ARE ABOUT HIS EXPERIENCES WITH Y COMBINATOR, LOOPT, AND OPENAI. WE ALSO DISCUSS THE BENEFITS OF STARTING A COMPANY OR JOINING A HIGH-GROWTH STARTUP EARLY IN YOUR CAREER.
Michael: As a first-time founder, how beneficial is Y Combinator?
Sam: I'm obviously biased here, but I think it's super valuable. One of the things that I think is important to understand about startups is that they are very much a pass / fail course. To a first approximation, you either succeed or you don't. You should be willing to do a lot of things, especially in the early days when nothing is working, to increase your chance of success.
You can clearly succeed without YC. We always tried to be open about that. I think anyone who told you that you had to do YC would be misleading you, but YC does try to make it a really great deal.
It is a lot of equity, but the company gets such a huge kickstart and peer group and access to a bunch of other networks that make it much more valuable. Clearly there have been many very valuable companies that started without YC, but we always tried to make it a really great deal.
Michael: You dropped out of school to pursue Loopt in a full-time manner as a 19-year-old. Can you touch on some of the key learnings there and what you learned from growing Loopt and ultimately selling it?
Sam: Man, there were so many! I think the most important one is to just jump in and do it. There's no pre-entrepreneurship class in school. Nothing that I learned in school other than how to program really helped me in any meaningful way, maybe also how to write actually.
I think general education is valuable, but there is no pre-startup like there is pre-med or something like that. I regretted dropping out of school because I was like, “I'm going to learn so much less.” I didn't get to learn the broad variety of stuff that I did at school, but my rate of learning for startup-specific knowledge was so high that I was really happy to just jump in.
I think there's all this bad advice out there about before you start a startup you should do this and study this and work at a company and get this kind of experience. Honestly, the best way to learn about a startup is just doing one. The first one may not work out, though sometimes it really does phenomenally well, but I think just getting going is super valuable.
Michael: In terms of joining a high growth startup and learning a lot there, versus actually starting a business yourself, do you think actually starting something and going through all that firsthand is the best way to learn versus kind of seeing it from inside a company but not as a founder?
Sam: Not necessarily. If you can have a choice between founding even a pretty good company and being one of the first 10 employees at Facebook or something like that, you should clearly pick the latter. So I think the most important thing is to do something that's really going to be successful or have a really high rate of learning and new opportunities, especially early in your career.
Starting a startup, you will learn something no matter what, but you will learn a lot more from success than failure. I think the sort of Silicon Valley failure point thing gets pretty weird. If you have a great opportunity to join a legitimate rocket ship company early, where you have very high conviction of the company and its founders, that is a wonderful opportunity. Those just don't come up often in life.
Michael: Great advice there. Onto OpenAI: talks of GPT-3 were everywhere a few weeks ago. I watched a bunch of videos of the API in action. Didn't get to use it myself...
Sam: We can fix that.
Michael: I’d love that! I've seen and heard a ton of great things. For our subscribers who don't know what GPT-3 is and maybe not even OpenAI, can you just give a super brief overview as a whole?
Sam: Sure. And by the way, to your previous question on the Impact Theory of careers, working at OpenAI is probably a better idea than starting a startup. I'm obviously biased there as well, but I genuinely believe it. I certainly believe I will have much more of an effect with OpenAI than anything else I've ever done or will do in my career. And so, I think the question of what you should do is just hard because these outlier things like OpenAI come along.
OpenAI is a hybrid of a research lab, a startup, and a policy think tank that's trying to make, develop, and deploy into the world super powerful general AI. We all think that this will be the most important technological development in a very long time, possibly in human history.
If you truly think about what it means to have computers that can think and learn like a human does, that revolution, call it the AI revolution, should be a bigger deal than The Agricultural Revolution, The Industrial Revolution, and “the computer revolution” put together. This is going to be a massive thing if we can do it. Maybe we can't or maybe someone else does sooner, but we're trying to figure out how to build those systems and then share the benefits of that and the governance over that with the world.
I think things are going to be so radically different. And, you know, in the long, long term, probably we create a new life form, some sort of descendent of humanity, and that goes off and explores the stars. And, I think we all deserve a say in that.
But sort of in the more prosaic short term, I think it is just going to sort of fundamentally upend the existing socio-economic contract that we need to think about doing this in a different way than tech companies traditionally do, or companies at all. I think we need to think about capitalism in a different way.
Michael: In terms of the vision for OpenAI in the next 5 or 10 years, because it's obviously a very long term game to really get to that next stage and ultimate vision. Where do you think the company will be in 5 or 10 years? What are some key objectives that you’re trying to hit?
Sam: Four years ago, we started with nothing and now we have GPT-3. If you think about the leap from nothing to GPT-3 and then doing that same thing again over the next four years, I think it could go pretty far.
Predictions about the future are always difficult, but I think we'll have systems that are, and I don't use this word lightly and I hate when Silicon Valley sort of calls everything world-changing, but I will say I think in four years we'll have systems that are world-changing.
IN EARLY 2019, SAM PUBLISHED A BLOG POST TITLED
HOW TO BE SUCCESSFUL
. IN IT, HE HIGHLIGHTS 13 THINGS TO HELP ACHIEVE OUTLIER SUCCESS. THIS POST HAS BEEN AN INCREDIBLE RESOURCE FOR ME, AND I WAS SUPER EXCITED TO DIVE FURTHER INTO A FEW TOPICS WITH SAM.
Michael: How to be Successful has been super influential for me so far. In that blog post, you discuss some of the benefits of compound growth and having a long term view. How can young people (students, recent grads, etc.) best maintain a long term view and focus while still focusing on short term goals such as getting good grades, landing internships and entry-level jobs, and so on? Do you have any frameworks to balance long-term focus with short-term success?
Sam: The framework I always use is to think about short term goals as stepping stones to longer-term ones. You don't want to get good grades just for the sake of getting good grades. Let's say what you really want to do is do AGI research, just so we pick something other than a company.
Then you could say, okay, that's what I want to do. There are many ways to do that. One path is to get really good grades, to get into a really good Ph.D. program, to do really good work, and then to get hired by a top lab or to get a professorship. That's perfectly valid. Another way is to sort of just say, I'm going to start doing the research now and publishing on the internet. And, that can work too if you're a good researcher.
So, what I always advise people on is, of course you have to balance long term goals with short term tactics, but it's the long term goals that matter and everything that you do in the short term should be done for the sake of where you’re trying to get long term — not for the sake of winning a game, and it's so hard not to try to win a game when a game is in front of you, really hard.
Learning how to avoid the distraction of short term games, particularly games that even if you win, you still lose, is a very valuable skill.
When I was like at YC, I'd mentor people all the time. That was my job. I don't have much time for it anymore, but the two or three people that I still really mentor, that are sort of young and at the beginning of their career, so much of what we talk about is how to pick the right long term goals and not get distracted by the short term peer pressure.
Michael: It's definitely difficult, especially when with short term tasks, you can kind of see the outcome and positive reward in a month or two months, whereas some of these long term goals and visions take years before actually reaping any of the rewards.
Sam: I think I'm really good at this. I'm 35 now and honestly, it's still hard for me. I know intellectually that I should be spending all of my time trying to make AGI and 95% of the time I'm good about ignoring other opportunities, but sometimes things comes along or friends of mine are doing something and it's so tempting. It's still so hard not to get distracted, even though I have this perfect setup not to be and I have no need to make more money or really do anything other than what I think is most impactful. It's still hard to live that in practice.
Michael: Onto another thing that you touch on in How to be Successful: getting good at sales. I hear a lot of students talk about getting into investment banking, consulting, or working at a big tech company out of school. I don't hear many of my classmates and friends really interested in sales, but I think there are obviously a ton of benefits of going and working in a sales role, especially at a fast-growing enterprise software company. What are some of the benefits of working in sales out of school? What are some of the key learnings that you can gain there versus in some of these other roles, whether that's finance, being a software engineer, and so on?
Sam: I didn't mean literally, like, you have to go get an enterprise sales job, although I think that's a perfectly good thing to do. What I meant more is, let's say you're going to be a startup founder: raising money, that's sales; hiring people and retaining them, that's sales; getting good press for a company, that's sales.
I think sales is often thought of as a dirty word and the actual activity of enterprise sales sounds unpleasant to me and something I'd be very bad at. But, the fun kind of sales is convincing someone of something that you deeply believe in. And when you think about it that way, at least when I think about it that way, that sounds cool and definitely something that is important to me to get good at.
Michael: And it's along the lines of storytelling in terms of, how can you get your story out in the best way that makes someone most compelled to buy your product, join your team, and so on.
Sam: You get to do this in any role, or almost any role, if you prioritize it.
FIGURING OUT WHAT PROJECTS TO WORK ON AND PUT RESOURCES BEHIND IS HARD. I ASKED SAM ABOUT HOW HE EVALUATES PROJECTS HE’S WORKING ON AND FIGURES OUT WHAT TO PURSUE FURTHER. SAM ALSO SHARES A GREAT VISUALIZATION OF WHAT PRODUCT-MARKET FIT LOOKS LIKE.
Michael: You've spoken before about trying a lot of things early on, seeing what works, and then putting as many resources as possible behind those things that do work. How do you figure out what to commit to? How do you evaluate what projects you're working on make sense to really pursue further and spend a lot of time on and put a lot of resources towards? I'm curious if you have any frameworks here.
Sam: Actually, that's always been pretty clear. That's not been the hard part. The hard part for me has been figuring out what things to work on first in that explore phase. But in that exploit phase, I do have to spend time and effort thinking about it and when I advise other people, I spend a lot of time trying to help them think through it.
Most things don't work, but when something is truly legitimately working, it is unmistakable. And so, my hard part has always been the discipline to cut off the other things. It's never hard to identify the thing that is truly working. You hear all of these different examples for what product-market fit means, but I think a visual that has always worked for me is: it [product-market fit] is when you switch from pushing a boulder up a hill to gravity pulling it down the other side of the hill. It's always clear when gravity is pulling the boulder down the hill.
Michael: I'm curious in terms of how long into a project it might take to realize whether it is worth pursuing.
Sam: I have never seen a project take more than a year or two to sort of figure out if it's working or not. Sometimes it fails later and sometimes it is only working a little bit and works more later on. And, sometimes it was working a lot and then later works less. But, projects do, at some point, have sort of an expiration date where they kind of fizzle out if they're not going under their own steam.
When we started OpenAI, it looked like any project that may or may not succeed. By two years in, it was very clear that it was going to be something quite good.
Michael: In terms of taking time to figure out what project(s) to pursue next and not rushing into something, do you have any advice for how students or recent grads can think about this? I know it's sometimes difficult to pursue a project or startup without having financial security and a good job on the resume to fall back on if it doesn’t work. Do you have any advice for how students can really try to figure out what projects do work in an efficient manner?
Sam: Look, all of these things come with risk trade-offs, and it's definitely risky to not jump into something right away and hack on projects for a little while. So that's a personal decision. I think in general, especially the beginning of your career, things are not as risky as they seem. If you're a good software engineer, you're going to find a good job, even if you go off and take a year. I think it's perfectly reasonable to go work at one of the big tech companies for a year and save up money and then go spend the time. The problem is, so many people get sucked into that lifestyle and they ratchet up their living expenses that they never do leave to work on something or start a company.
During college, I went and traveled around the world. In other parts of the world, you can live in hostels and eat food for extremely low money. That was great because I had not much money at all but was still able to travel. I would never be willing to go do that now. At the time, it didn't even seem uncomfortable or unpleasant to me, but there is no way I could do it now. Doing things like that before your kind of lifestyle style and personal brands adjust is very valuable.
There is always risk but you can make anything work and especially when you're young and don't have expensive tastes or a lot of life obligations, you really have much more freedom than you realize to say, I'm going to figure out a way to go hack on stuff for a while.
FAST FIRE QUESTIONS: ~60 second responses
Michael: Do you have a favorite book that you would recommend to our subscribers?
Sam: I could do that for like 100 books. I'll tell you the one that just came to mind, Siddhartha. If I was going to recommend a college-age student to go read one book about life that was both short and fun and a very high return on investment for reading it, that is what I would pick. I can never do these lists of favorites and pick one well.
Michael: This next one is another favorite question, so we'll see if we can narrow it down. Do you have any favorite restaurants in St. Louis?
Sam: I'm a big believer in favorite dishes, not favorite restaurants: like go to this place and get this thing. Sadly, my favorite thing to order in all of St. Louis was the mocha frappe at Coffee Cartel, which was this unbelievable dessert, but the place closed down. Something else I love is the falafel sandwich at Al-Tarboush on The Loop.
Michael: I'm anticipating playing a lot of poker given there isn't much else to do at school this semester. I know you played a lot of poker while in school. What attracts you to poker? Are there any business or psychological skills that you can learn from playing?
Sam: It's so much. I really try to get people to study the game. I think there's so much about business and human nature and risk and life and psychology and everything else to learn in there. So what attracts me to it besides it being really fun is just how much I think there is to learn.
There are all these concepts that people talk about. One that I never found a good way to learn or explain other than poker is the concept of implied odds. And I think in business, if you understand implied odds, like if you understand that if you take this action, it has these odds of success, but that that is conditioned on it happening, that is super valuable. It’s also a very hard skill to learn. I think about the implied odds playing out the whole game.
Michael: You somewhat touched on this earlier, but I have heard you discuss the risk of at the end of your career, and Jeff Bezos also talks about this, but kind of the risk at the end of your career of looking back and feeling that you missed opportunities or wasted things that you could have done. I'm wondering if you have frameworks for how young people can think about risk in that perspective and balance taking on risks with not doing things that are going to put them out of position to succeed and take on more risks in the future.
Sam: Definitely take the risks earlier. It's easy to take the risks earlier. It gets harder as life goes on. So the very beginning of your career is absolutely the best time to do it. If you have no reputation and no money and you take a risk and it fails, then you're still in a place of having no reputation and no money. It doesn't get easier over time.
I think I could say all of the standard things about like, no one ever regrets not taking the sort of standard path job of investment banker, but that actually is not that helpful. The thing to learn that I think is actionable here is learning to not fall into the peer pressure trap.
Michael: How can people, in particular students, not fall into that trap?
Sam: The thing that worked the best for me was finding a different peer group. All of the kids I went to school with were all focused on investment banking and consulting and so it made me want to do that, even though I actually didn't at all. I just wanted to win the game.
When I met, through Y Combinator, other people who were interested in startups and technology, then I wanted what they wanted, and that was an easy way to change my reference frame. I think the easiest way is to change who you hang out with. There are other things too, like all these ways to practice mental discipline that are somewhat helpful, but changing the peer group is probably the highest impact way to do it.
Michael: Sam, that was super enjoyable. I learned a lot. Thanks for coming on The Takeoff!
Sam: My pleasure. I enjoyed it. Thanks for having me!
Note from Michael:
This was an absolute blast to record. I was both excited and nervous at first, but this quickly became probably the most enjoyable conversation I’ve had for The Takeoff to date. Sam is wicked smart and has already achieved more than anyone can dream of (he’s only 35). But, what impresses me most is his desire to keep pushing forward, change the world for the better, and help people learn (he was president of YC and has countless blog posts designed to help people figure out their lives & careers) (
Thanks for reading our interview with Sam. I had a blast chatting with him and hope you learned a lot (I certainly did)!
You can find Sam on Twitter @sama.
I’m on Twitter @mspiro3 👋 (direct any thoughts / comments / questions to my DMs)
If you find The Takeoff valuable, share it with friends, or subscribe 👇 if you aren’t already (you’ll get an email every Monday with a new interview and every other Friday with some curated-content that you may find useful).