The Most Important Founder You've Never Heard Of

Speakers: Sam Parr (host), Shaan Puri (host)

Shaan Puri: Sam, I think today we should talk about somebody who is one of the most important founders in the world, one of the most brilliant founders in the world that nobody talks about. I don’t even know how to say this guy’s name properly, and he is one of the most important tech founders in the world. His name is Demis Hassabis.

Sam Parr: Is he currently the guy who’s warning people? Is he on a podcast tour warning people?

Shaan Puri: No, he’s pro-AI. So, he’s not warning people.

Sam Parr: No, then I don’t know anything about him. Enlighten me.

The Most Important Founder You’ve Never Heard Of [00:00:45]

Shaan Puri: Okay. So, this guy is “Demis the Menace” is what I’m going to call this guy because this guy is an absolute animal. I watched this documentary called The Thinking Game. It’s on Prime Video if anybody wants to go watch it. I’d heard good things from some smart people. So, I thought, okay, let me check it out. Let me just first lay out my case for Demis as Billy of the Week because he’s kind of legendary.

I didn’t understand how much of a prodigy this guy was. This was a documentary that was pretty straightforward, but this could have easily been a movie like The Social Network. The Social Network basically covered the most transformative young brilliant founder from the 2004 to 2010 era, which is Zuck. It talks about Zuck in college and how he’s this kid and all the ups and downs he goes through trying to build this thing.

Demis is maybe that guy now. Him and Sam Altman. They’re both basically two guys who are creating the most important technology of all time.

Sam Parr: You think those are the two guys? Those are the guys.

Shaan Puri: Well, Elon would be the other, right? Elon’s the obvious other person that needs a movie and has a crazy life. But this guy I think is the most underrated, less talked about for who he is. He started this company called DeepMind. DeepMind got bought by Google and DeepMind is basically Google’s AI play. The DeepMind team, which was basically a research team that was building AI, is the reason that OpenAI exists. It is the reason that ChatGPT exists. It is the reason Elon is interested in AI.

It was very much because Elon met with Demis and basically Demis big-dogged him a little bit. He was like, “Oh yeah, I’m working on the most important thing ever.” And Elon, who’s building rockets and electric cars, he’s like, “I’m saving the planet. I’m going to space. That’s my portfolio.” And Demis said, “Well, what we’re building will be the most important invention humans will ever make. It will be the last invention. It’s artificial general intelligence.”

So, a computer that can think and learn better than humans. The reason why this is called the last invention is because once you invent an artificial general intelligence, it’s basically like its own little species. Computers that can think and learn will then do the thinking and learning and inventing at a far faster pace than we will. So they’ll invent all the new stuff after that. He has that conviction throughout the documentary.

Sam Parr: And he’s had it since he was a kid.

Demis Hassabis: The Chess Prodigy [00:03:15]

Shaan Puri: Okay. So here is the cool story. That’s the very basic setup, but here’s the story. He grows up. He’s got these hippie parents. His dad is a musician and they look very bohemian. He gets into chess and by the age of six, he is one of the best chess players in the world among all humans. First, he wins the under-eight championship when he’s only six in Europe. At one point, he’s ranked the second-best chess player in the world for his age. So he’s an elite chess player as a young kid.

He used his chess—he would go to—his parents would basically drive him to these chess tournaments. He would win as this—and he looks tiny even now, he looks baby-faced—he looked like such a little kid when he’s sitting there at these tables. He would basically go win prize money and he used the prize money to buy his first computer. Chess gets him a computer. When he gets a computer, he starts making games on the computer. He builds a chess game, builds other little games, and he starts a hacking club with friends at school. He’s basically like, “Wow, computers and chess, this is my life.”

Sam Parr: If you had to create a stereotype for a good movie character who’s a James Bond villain or a mega genius, this is how they all start. The stories all start.

Shaan Puri: And by the way, he tells the story of this incredible origin story. He goes, “My parents took me to this tournament of 300 best players in Europe and it was on a mountain in a church.” It shows the church and it shows 150 chess tables lined up. 300 players are going to be there and he’s only, I don’t know, eight years old or something at this point. He’s tiny and he’s playing against the Danish national champion. A 30-year-old man is playing against him.

He describes basically the chess tournament was no timer. This game with this 30-year-old dude goes for 10 plus hours and he’s just playing him. He’s like, “I’m pretty sure it’s a draw, but this guy’s not conceding that it’s a draw.” So he just has to keep playing and this guy’s just wearing out this little kid over hours and hours and hours. They only have five pieces on the board and it’s just a stalemate basically, but he won’t say it’s a stalemate.

He describes how at the end basically this guy tricks him a little bit and he ends up losing when what should have been a stalemate. He makes one wrong move at the end and the guy laughs at him, stands up and laughs and says, “You should—it should have been a stalemate. You should have just done this and it would have been a stalemate.” He rubs it in his face basically. He’s so upset at this tournament and this grown man humiliating him. He looks around and he’s just like, “What am I doing?” He’s like, “That was a horrible experience.”

He goes, “If you took the 300 people in this room, the brain power in this room that we’re just spending on this 10-hour tournament here, we could cure cancer.” He’s like, “Forget chess. I’m not going after chess anymore. I’m so done with chess after this bad experience. I’m going to go for computers. I’m going to try to figure out how to harness the brain power of humans and combine it with computers. How do I get computers that could think?” The documentary is called Thinking Game because they interviewed him when he was a six-year-old and a TV network was like, “So, why do you love chess so much?” And he goes, “It’s just a good thinking game.”

Sam Parr: What a fun hang.

Shaan Puri: What a fun hang. Can you imagine if your boy played with him?

From Video Games to Cambridge [00:07:15]

Shaan Puri: Okay, so listen to this from Boy Wonder. He gets into Cambridge, but he’s too young to go. So, he has to wait a year to go to Cambridge because he decides, “I’m going to Cambridge. I’m going to study AI.” He’s 14, 15 years old at this point. In his gap year, they need him to wait. He can’t go until he’s 17. So he says, “Okay, why don’t I try to get a job? I’ll work in the meantime. And I’m not going to do chess tournaments. I’m going to do something with computers.”

This company called Bullfrog, which made the most popular computer games at the time in Europe—they were the number one production company of games—they held a contest. It was also cool to see gaming was so new at the time. The CEO of the gaming company was like, “Dude, there was no recruiters. We couldn’t be like, ‘Hey, go get us the best game programmers.’ There were no game programmers. It wasn’t even a job yet.”

It just reminded me of what the frontiers always look like. It’s little signals of you’re in the right spot when there’s not even recruiters for the thing. There’s no agencies yet for the thing. There’s no name for the job. For context, by the way, he’s 50 years old now. He’s 49. So, we’re talking the late 80s.

Sam Parr: Right? Yeah. Long time ago.

The Peter Thiel Connection and DeepMind’s Origin [00:12:30]

Shaan Puri: Okay. So then he decides at some point that he’s going to start this company. Now nobody really believes in AI at the time. In fact, in the scientific community, AI was not a thing because it’s not science really. There’s no testable hypotheses that you could go do. You couldn’t go into a lab and do AI. The entrepreneurship community also didn’t really respect AI. It’s a sci-fi topic. No, there’s been no commercial companies doing this. So there’s nobody who believes in this.

Well, guess who believes when nobody believes? Guess who loves a good old contrarian bet? Thiel. Thiel backs—Thiel becomes the first backer of DeepMind.

Sam Parr: Are you kidding me?

Shaan Puri: No. So, how legendary is Peter Thiel? That he’s the origin…

Sam Parr: I think that people talk about this, but I don’t think it’s talked about enough. I think Tim Dillon’s a comedian where he was like, “Everyone thinks the President of the United States is most powerful, but there’s one person who’s never around. You can’t see him, but he truly runs everything. And that’s Peter Thiel.” He was saying that at Trump’s inauguration, it was JD Vance, who’s a Thiel guy. It was all the CEOs, Thiel guy, Thiel guy, Zuck, Thiel guy. Peter Thiel is the guy. I recently read a whole bunch of old quotes from him and everything he says is timeless and has been true so often.

Shaan Puri: He’s like a city, dude. He’s like a place that people are from. “Oh yeah, Zuck? He grew up in Thiel. Oh, Ethereum? You like that? Well, he grew up in Thiel. Oh yeah, Elon Musk? Yep. His first company merged with Peter Thiel’s company and Thiel was the CEO.”

Sam Parr: So a lot of times it starts with—was it Plato or Socrates? Socrates taught Plato, Plato taught Alexander the Great and also Aristotle. It’s sort of like there’s this one person that’s the lineage. It’s very strange.

Shaan Puri: So Thiel becomes the backer. The second significant backer was Elon Musk. Thiel tells Elon about this. Elon meets Demis. Demis says that big dog line of basically, “I’m working on the most important thing in the world.” Elon is like, “Wait a minute. What’s going on here?” He ends up funding this.

Teaching AI to Play Games [00:15:15]

Shaan Puri: Okay. So he gets a little bit of funding from some crazy believers. Now the part of the movie that I think is just incredible is showing them building this monster that is AI.

Sam Parr: When you say build, is there actual physical building as well?

Shaan Puri: No, so at the time it’s them on a whiteboard with really complicated math equations talking about, “Well, what if we took this technique from deep learning and we merged it with this technique over here about neural nets and what if we could get something new?” And that’s what they did. They combined these two different ways of learning. Don’t ask me what any of those words mean, but the thing they show is a little TV screen with the Atari game of Pong.

It’s so funny that this whole thing starts with Pong and it starts with games. You had the most brilliant people in the world staring at this Atari game trying to be like, “How can we teach the computer to play this game?” He talks about because he grew up on chess and he was super competitive and games were how he learned to think. He’s like, “Maybe games will be how the computer learns to think because games have rules, they have rewards, they have clear definition. You can see all the information and you could do it a bunch of times and get better and better and better. You can run a lot of simulations very quickly.”

The rate of learning—just like how he basically was like the way kids learn is games—maybe that’s the way we can build a childlike computer program to also learn.

Sam Parr: I think one of his breakthroughs was when they played the Asian game, right? Go.

Shaan Puri: Yeah. So before that it starts with Pong. I didn’t actually know this. They basically said, “Look, don’t tell it anything about Pong. Just tell it one thing: score go up is good.” At the beginning they show it and they’re all watching it. The computer hits it—the game hits it—and then their AI player doesn’t even move its paddle. It’s down one. Next time it moves its paddle the wrong way. Down two. Next time moves its paddle the wrong way. Almost recovers but misses it. Down three. And then it hits the ball once and they’re all like—but then it still loses the point.

It starts out terrible. By a hundred games, it’s competitive. By 200 games, it’s as good as the best humans at playing the game. And by 500 games, it’s never losing a point. And they’re like, “Okay, that was remarkable. Let’s carry on.” They had this first objective, which is let’s—without telling it—because again, the goal was he goes, “What is AGI? It can think and it can learn. So we can’t just tell it the rules. We can’t just tell it how to win. We can’t give it strategy and then it executes it. No, no, no. It has to figure it out itself. Like a kid learning how to walk and it stumbles and it starts to figure out, ‘Oh, if I put my center of mass here, that’s how I walk.’”

They wouldn’t tell it anything about the game except for whoever has a higher score at the end, that’s a good thing. Go for it computer. They had it learn like 50 games. The next one was Brickbreaker. If you ever played that game on Blackberry where it’s breaking bricks. Same thing. 100 games terrible. 200 games it’s pretty good. As good as most humans. 500 games it’s unstoppable. It figured out this strategy in Brickbreaker where you tunnel in through the sides and then the ball will just keep bouncing on the top and break all the bricks on its own without it having to hit you.

Next let’s do chess. They show it doing chess. One of the first “aha” moments was it started to invent its own strategy a little bit, like it’s got its own style. It’s got its own little attacking style. That’s pretty cool. It beat Stockfish, which is the best chess program out there. They’re like, “Well that’s good because Stockfish beats all the pros. If this beats Stockfish that means it’s the best at chess.”

Then they went to Go. Go, I didn’t entirely understand. It almost looks like Chinese checkers, but it sounds like it’s more complicated. They claim that it’s the most complicated game on Earth because it has the most permutations on how you could possibly win or lose.

Sam Parr: Right. There are more board configurations in Go than there are atoms in the universe. So you can’t just think it through. There’s too many combinations. So you have to be actually fluid that in any situation you’re in, you’re able to figure out the right move. People had always thought Go is too hard. No computers had ever beaten Go before.

Shaan Puri: So they start and they created this program called AlphaGo. AlphaGo basically—what they did, which was this is kind of nerdy but I liked hearing how they did it—they gave it 100,000 games from strong amateur players. They said, “Here’s 100,000 games. Learn from this past game.”

Sam Parr: So they gave him the play-by-play move-by-move thing and it learns all that.

Shaan Puri: And then it said, “Cool. Based on what you learn, now play yourself. See if you can get better.” And it played itself like a million times. Okay, so that’s kind of interesting. Maybe that’ll get a new result.

AlphaGo and the Infamous Move 37 [00:20:45]

Shaan Puri: They go to Korea for this test. They’re like, “We’re going to go play this guy, Lee Sedol.” Lee Sedol is a grandmaster Go player. He’s one of the best players of the past two decades. He’s the man. They show them getting off the plane and there’s hundreds of photographers taking pictures like “Today: Computer versus Man, Man versus Machine.” I didn’t actually see any of this when this was happening. I don’t know if you did either, but in this small corner of the earth…

Sam Parr: Dude, this story line is as old as John Henry. You remember John Henry who was the strongest man who was using the jackhammer through the mountain trying to race the new steam engine who can pile through stuff? He works so hard that his heart explodes and that’s the legend. That’s basically what happens except the guy’s mind exploded. This story line is perfect.

Shaan Puri: So they sit down and the game is going as usual. They have a line from Eric Schmidt. Eric Schmidt is from Google. He was the former CEO of Google and a super technical guy. Google had bought DeepMind at this point.

Sam Parr: Dude, I saw the price. One of the greatest deals of all time potentially.

Shaan Puri: They bought it for I think 400 million pounds. So it was 500-something million dollars. There’s a great line from Demis in this. I don’t know if you saw this part where they’re like his investors didn’t want to sell. He said this line that I really like—it was kind of a frame breaker for me. I don’t think most people when they listen to this line would even think twice about it. But for me, it was a little bit of a frame break.

He was basically in a frenzy. He’s like, “This is so important. There’s so much to do. My life is only so long. I want to see this happen.” He’s like, “We have so much to do. If we can just get this funding and be left alone to go do what we needed to do, then I might actually get to see this thing in my lifetime, and that’s what matters.” He’s like, “What’s a few billion dollars for five years extra of my life getting to work on this?” He goes, “I could sell for a few extra more billion and make billions of dollars. But let me ask you something. If you’re going to die, would you spend billions of dollars to live an additional five years? Of course you would.” And that’s what he said he was going to do.

Sam Parr: Here is such a good line. Someone changed my perspective on having children. Someone was like, “Do you think you’re going to love your kids when they’re born?” I was like, “Yeah.” He’s like, “Well, then why wouldn’t you have them sooner so you have an additional life with them, time with them?”

Shaan Puri: He has another line later that’s kind of like this. He was talking about what a new breakthrough that they were going to have and he’s like, “It’s going to be the most exciting thing ever. How will we get sleep? I won’t be able to sleep.” He was just that fired up 10 years into the mission. I just thought like when people talk about mission-driven, this is what they mean when the guy’s like, “There’s so much to do. I don’t know if it’ll happen in my lifetime. The most exciting thing in my life is if this happens while I’m still alive. I will do everything in my power to make this happen while I’m still alive.” I thought that that was just a next level of mission-driven excitement.

Passion, Resourcefulness, and Public Speaking [00:24:30]

Sam Parr: I want to read you a cool quote. Okay. So, I’m reading this book. It sounds silly, but hear me out. The Quick and Easy Way to Effective Speaking by Dale Carnegie.

Shaan Puri: Oh, very cool.

Sam Parr: Dale Carnegie famously wrote How to Win Friends and Influence People. He was actually more famous originally because he created the Dale Carnegie speaking program. They had locations all over the country and hundreds of thousands of people went through his programs, including Warren Buffett who says it was the most important class he ever took. He had the diploma from the speaking class on his wall next to his office, not his college diploma. He even taught—he was a Dale Carnegie instructor.

There’s this amazing quote. Basically Dale Carnegie, one of his premises is that public speaking—he calls it the “royal road to self-confidence.” He says if you want to be a more confident person, you should actually learn how to public speak because when you control the minds of many men, you control yourself. It makes you more confident. One of his axioms for how you get better is you have to envision the end goal.

He has this amazing quote from William James, who is like the godfather of modern psychology. William James says, “In almost any subject, your passion for the subject will save you. If you care enough for a result, you will most certainly attain it. If you wish to be good, you will be good. If you wish to be rich, you will be rich. If you wish to be learned, you will be learned. Only then you must really wish these things and wish them with exclusiveness and not wish 100 other incompatible things just as strongly.”

My point being is whatever you truly want, if you want it bad enough, your passion will carry you enough to acquire all the skills and have the determination to see it through the end. I wrote this down that this guy—you see it from the beginning and where he is now—this quote applies to him.

Shaan Puri: That’s great. Okay, little segue. Have you ever seen the Tony Robbins TED talk he gave?

Sam Parr: Yeah, probably. I’ve seen many of his talks.

Shaan Puri: Tony’s normal talks, like his seminar, is a 4-day, 12-hours-a-day on stage thing. So a TED talk is 18 minutes. He gets on stage, he’s like, “All right, I usually talk for 12 hours at a time. Let’s see what I can do in 18 minutes.” He gets to this point in the talk and he’s like, “What stops us from getting what we want?” And people are like, “I don’t have the time.” He’s like, “Yeah, all right. Time. I don’t have the money. I don’t have the skills. I don’t have the network.”

He writes all these resources that you lack down. Then one guy in the crowd goes, “I didn’t have the Supreme Court justices.” He looks through the darkness. He’s like, “Who said that?” And it was Al Gore. Al Gore, who had just lost the presidential election in Florida—there was a recount and he was two justices short or something like that. Everybody has a big laugh and then Tony says, “I don’t think that’s why you lost because I saw you yesterday on this TED stage talking about climate change.”

Gore is super passionate about climate change. He was one of the big advocates for climate change. Tony goes, “If you had talked like that in your presidential debates, you would have never needed the Supreme Court justices. You were on fire yesterday. I didn’t see that when you were debating Bush.” He basically says, “The only resource you need is resourcefulness.”

He goes, “Because look, if you’re just lit on fire to do something, you just ask yourself the following question: If I’m determined enough, if I’m charismatic enough, if I’m charming enough, if I’m playful enough, if I’m creative enough, if I am motivated enough, I’m persistent enough, can I not achieve anything I want? Can I not overcome all those things that I lacked?”

Of course. You didn’t have the resources, you didn’t have the money. Well, but if you’re determined and you’re charming and you’re persuasive, you’ll go get the funding. It’s like this master skill that’s underneath. I actually catch myself doing this all the time where I feel like I lack something. I’ll literally go say that almost like an affirmation: “Well, if I’m playful enough and I’m determined enough and I’m charismatic enough and I’m persuasive enough and I’m determined enough, can I not get this thing I want? Of course I can.” “Oh, they’re closed. I could probably get them to open.” “Oh, this guy said no. I could probably get him to say yes.” That’s a universal skill we all have if you remind yourself.

Sam Parr: That’s pretty badass. And I don’t even think you need to be charming-charming. This guy Demis was pretty black and white. But when I listen to him, I’m like, “You’re an unstoppable force. You care about this so much.”

The Sputnik Moment in China [00:29:45]

Shaan Puri: He is what Paul Graham calls a “fierce nerd.” I think that fierce nerd essay is actually hall of fame level for Paul Graham and you see it when you see somebody like Demis and how competitive he is with foosball and chess and then he’s also that way with trying to win the protein folding problem.

All right, back to the story. They’re sitting there with the best Korean Go player in the world, Lee Sedol. There’s this move, Move 37. I think if they write the book of humanity or the movie of humanity, Move 37 is the “uh-oh” moment. It’s like the moment in movies where in a rom-com the guy bumps into the girl, she drops her papers on the ground, then they pick them up, and they look each other in the eyes. It’s the spark. This is the spark of where AI really took off and it’s Move 37.

Basically, they’re playing Lee. The expectation is Lee Sedol will win because Go is so hard and he’s the best, but we’ll put up a good showing. We’ll be as good as the best players against Lee Sedol. In Move 37, the computer does something. Right away the announcers are like, “Oh my, oh, what is that?” Lee Sedol—you can literally—they show him sweating and thinking and he’s like, “What the hell just happened?”

They go, “I think we might have just seen an original move by AlphaGo.” Lee Sedol is just—he doesn’t know what to do. He’s really perplexed by this move. They go, “No human would have made that move.” It was the first time that it wasn’t just pattern matching—mimicking what a human would do or say, but less good than a human would say or do it, or maybe it’s a little bit faster because it’s a computer. It was the first time it was like, “That was novel. That was a creative breakthrough.” And it beats Lee Sedol. It’s like when in a horror movie the robot turns and says, “I’m in charge now.” This is that moment.

Sam Parr: Exactly. And I had never actually seen the clip and the way the movie shows it I think is wonderful. So then right afterwards, Eric Schmidt’s like, “Holy shit.” And he goes to Demis and he goes, “What’s next? Where does this end?” And he goes, “When we beat the Chinese guy.” I didn’t even know about this part. It’s like then there was a Chinese guy who was the actual number one ranked player in the world. They go to China to play this guy.

Now it’s like, what’s going to happen? This computer just beat Lee Sedol. Can it beat the Chinese guy? I just love that they even called him “the Chinese guy.” It was the most relatable thing that this absolute super genius with a 10,000 IQ said. I was like, “Oh, he’s just like me. He would just call him the Chinese guy.” So they go and they play. Had you ever heard about this?

Sam Parr: No. If you go on YouTube and you type in Move 37, there’s videos with hundreds of thousands and millions of views. It’s all retelling the story of Move 37 or there’s Magnus Carlsen talking about how Move 37 teaches you about XYZ. It’s become an acronym or an analogy for…

Shaan Puri: The 4-minute mile, right?

Sam Parr: Yeah, exactly. It’s exactly what it is. A 4-minute mile. It’s just a phrase that doesn’t even mean Move 37 anymore. It’s grown beyond that.

Shaan Puri: Totally. I see your little public speaking brain is picking up on all these magic tiny words.

Sam Parr: Thanks, Dale.

Shaan Puri: Thanks, Dale. Thanks to our guest today, Dale Cardigan.

Sam Parr: Thank you.

Shaan Puri: Okay, so then they play the Chinese guy. Now here’s the crazy thing about playing the Chinese guy. AlphaGo is whooping ass and it’s putting the pressure on the number one…

Sam Parr: Chinese guy just smoking cigs while he’s doing this because that’s why…

Shaan Puri: He’s actually a pretty young-looking guy. But the crazy thing is as he starts to put the pressure on the Chinese guy, they cut the feed in China.

Sam Parr: No way.

Shaan Puri: How badass is that? They cut the feed of the broadcast. They’re like, “No, we will not show. We will not lose face like this.” They call that in the movie. They’re like, “This is like the Sputnik moment where China was like, ‘Wake-up call. We’re getting into AI.’” This actually triggered the AI race for why China got so into it. And how they cut the feed was so dramatic. I thought that was incredible.

Sam Parr: That’s crazy. Okay, awesome.

AlphaFold: Solving the Protein Folding Problem [00:34:15]

Shaan Puri: Okay, so then they continue with the game. Let’s fast forward. They do Starcraft next. Starcraft’s interesting because both players are playing at the same time. So it’s not turn-by-turn like you go…

Sam Parr: Starcraft like a fighting game? I don’t know what it is.

Shaan Puri: Yeah, I think it’s called a MOBA or whatever. It’s basically a game where you have a map, you have a base, they have a base, you got to attack their base with characters, you got to move them around the map. There’s a fog of war. The whole map is not revealed. Both players are playing simultaneously. So now it’s even harder.

Sam Parr: You’re acting like you don’t know what you’re talking about. “Like I don’t play that.” But anyway, so there’s this main character. Here’s what he does.

Shaan Puri: I don’t play Starcraft, but I’ve been around enough dorks to know enough. All right. So it doesn’t actually beat the best Starcraft player in the world. That guy wins. Okay. But it was a good showing.

Anyways, then what’s the next stuff that really stood out to me? There’s this one last part about protein folding. Are you familiar with what they’ve been doing with this? All I know is that no one had ever solved it and basically within days or weeks or something like that, they solved something that took 50 years to get up to that point in progress.

Shaan Puri: Actually, it took years, which is cool. I didn’t actually realize this. Demis is basically talking about—they’re like, “All right, we did good in games,” but he’s basically like, “AI-assisted science is going to be the thing.” I don’t think this gets talked about very much nowadays. Like maybe AI could cure cancer. But this guy is seriously like, “No, AI should cure cancer.”

Sam Parr: It’s not clear how math or that type of… why do you need more data and more effort? It’s as if in order to cure cancer you’re just like, “Let’s throw these 50 drugs at them. Oh, that one kind of worked. Let’s soup up the drug and throw it 550 more times.” That’s sort of how in your head you think cure cancer, not like, “Can you math your way out of it?”

Prediction: The Core of Artificial Intelligence [00:37:00]

Shaan Puri: Correct. Now, what I’ve realized in watching this and hanging out with AI people is one of the most important things in the world is basically prediction. I remember I invested in this guy who was a self-driving car entrepreneur. He had worked at the Uber self-driving car team and he took me to this little garage and he had this golf cart that he had rigged in 2019 or 18. This was pre-pandemic.

Sam Parr: Before that, maybe. Yeah, I remember. No, that was probably the time. It was right before I started my fund. So, 2018, 2019. You’re right.

Shaan Puri: And he drove me around in a self-driving golf cart in a self-storage facility. It’s like you get a peak of the future. That was amazing. This is before Tesla had it and whatever. But it wasn’t perfect. You could only do it in a very controlled environment. But he basically said, “Look, everybody who’s working on this knows there’s these four or five steps of self-driving.”

I don’t remember all of them but one was vision—so you got to see the world. Then based on what you see, the next step is prediction. “Okay, I saw that that car was right there. Where will it be in 2 seconds? I need to predict where it’s going to be.” That’s the whole basis of self-driving: planning, prediction, and then an action step.

That kind of planning and prediction step is the key to how AI affects all these industries. ChatGPT is planning and prediction of what is the next token—or let’s just use word. What’s the next word that would probably go in this sentence? “Roses are…” I think it’s going to be “red” because I’ve seen “roses are red” so many times that my prediction score very confidently would say the next word is “red.”

Okay, great. How do self-driving cars work? Same thing. If I see a car there, my prediction is it’s going to be here in the next 1 second. So therefore, I need to do a new action. The same thing applies to science and curing all these diseases, which is you need to know what a protein structure looks like. Based on the shape of the protein structure, if you can predict the protein structure, then it’s not so hard to figure out what should you attach to it to either destroy that protein or soup it up and make it more strong. You know where to bind on the protein.

Okay, so I didn’t know about this thing, but it’s called CASP. CASP is this competition that has been going on for years and it’s basically the Olympics of protein folding. If you do a sequence—amino acids—you’re like, “Oh, it’s got this amino acid, this amino acid, this amino acid.” You get these 10 amino acids. Cool. You know what’s in it, but you don’t know what it looks like. You don’t know the structure of how it’s folded up into this little tiny knot. A very unique structure.

Sam Parr: When you say folding, figure out the shape of the knot.

Shaan Puri: The shape of it. And you need to know this shape in order to design a drug that’s going to do anything to it.

Sam Parr: Like who could kill it or grow it or shrink it?

Shaan Puri: Imagine I said, “Hey, you’re going to park this car at this address.” Cool. But if you don’t know what the garage looks like, you’re just going to smash into the house. You might know the location of it, but you don’t know where to park the car. So, how do you park the drug that’s going to attack this, that’s going to either kill it or enhance it? You need to know the shape of it.

The way they do it is one by one. They created this competition to be like, “Can anyone use computers to predict the protein folding?” because doing this manually is untenable. For years, if you look at the graph, it was like 20% or 30% prediction accuracy for a decade. Demis decides, “This is what we’re going to do. We’re going to throw our resources behind this.”

The first time they do it, they win the competition, but they’re like, “Great, we’re trying to go to the moon and we just have the tallest ladder. The ladder doesn’t get you to the moon.” They were actually incredibly disappointed. He’s like, “This was a bitter taste of we really tried. We won, but not by enough to even solve the protein. We’re here to solve the protein folding problem, not win the competition.”

To solve it, you need 90% plus accuracy. He describes this next year where they basically went back to the drawing board and tried to come up with new ideas. I thought it was a cool CEO moment. He was like, “I know when you need to come up with a creative idea, you can’t force it. Squeezing it doesn’t make creativity come out when you just push the team. ‘We need an answer now!’ That’s not going to get the best idea.”

Sam Parr: That’s interesting because that’s the opposite of what I would think. There have been people who would say constraints are the answer.

Shaan Puri: So they used constraints, but what they didn’t do was put everybody into fight-or-flight mode. When you’re in fight-or-flight, it’s kind of like why your best ideas come to you when you’re in the shower or when you’re relaxed or when you’re asleep or when you’re on a walk. The brain has two modes. One is executive mode where you’re doing tasks, and that’s good at doing tasks, but it’s not good at making new connections between existing fuzzy data.

Sam Parr: Dude, that’s so interesting because this is how Henry Ford—one of the things—basically there’s an engine block. It’s a block of metal and you put cylinders in there and that’s how a combustible engine works. But before that was one block, it used to be two blocks and it kept breaking. Imagine two blocks and screwed things together and it was holding them back from taking over the world.

Henry Ford got a team of four engineers out of a company of thousands of people and he brought them to a small office. He goes, “This is your guys’ workshop.” And they’re like, “What are we doing?” He’s like, “You see that big ass block of metal? Figure out how to put four holes in there and four pistons and make it work.” And they’re like, “Henry sir, that’s impossible.” He goes, “I’ll see you guys in a quarter.”

Apparently the story is that he went back like eight quarters in a row. So it was something like two years and then finally they got it. But it took two years. He did allow them: “This is four of you. This is your job. Just figure it out and let me know.”

Shaan Puri: Exactly. This is also how—if you read about Steve Jobs with the iPhone—he was like, “No keyboard on the phone.” And they were like, “But the Blackberry keyboard… you got to write emails.” He said, “No keyboard on the phone.” And they’re like, “But how would we… the accuracy of this? I mean screens today don’t…” He’s like, “No keyboard on the phone.” And so then they had to go invent multi-touch and figure it out. He gave them the constraint—you got to do it in these constraints—but then I give you the time to go explore and figure out which path might work.

Sam Parr: That’s interesting.

Shaan Puri: And they also did this with the game thing. The first one was train on 100,000 games. Then they created AlphaZero where they said, “Now try to make it win with no prior human knowledge.” Because he’s like, “If we’re ever going to do new novel things, you got to assume we were not going to have a database of 100,000 good humans at doing this to use.”

They created AlphaZero, which could win in chess and Go just by playing itself like 10 million times. It figured it out. Similarly here, they were like, “You got to go back to the drawing board.” He described, “First I’m going to give them the constraint. Second, I’m going to let them be creative and try to go to the drawing board, figure out multiple different possible ways this might work.”

He goes, “And then when they pick one, I know this is when it’s time to push. Because first we will get worse than we were before. Then after some time, we’ll pick an approach and we’ll get right back close to where we were before. And that’s when it’s time to push. I’ve seen it so many times before.” And we’ll explode through. I was like, “That’s pretty dope how he kind of had developed judgment on the scientific process and the creative process enough to know when do you push and when do you not push?”

Sam Parr: Dude, that’s so great. We’re learning all these techniques and I’m putting it all together. Hormozi had this cool thing that he said when I talked to him once. He was like, “Basically, I’ve noticed that when you start something new, the results go down 20% right off the bat.” So, if you’re training your sales team on something new, their conversion rate is actually going to drop from 50% down to 40%.

Eventually it will go up if you pick the right thing. The question is basically make sure you pick the right thing because if it’s going to go down 20%, that means you need it to double its improvement in order for it to be worth it. You pick the right thing, otherwise you’re just back to square one and you went down 20% for a quarter. Knowing that these J-curve progress things exist is important because the amateur would panic. The amateur would not go forward. That was actually my biggest learning this year running a company was expect new things to suck or bring everything down. Therefore, make your project selection perfect or high quality.

Isomorphic Labs and the Future of Medicine [00:46:45]

Shaan Puri: Right. And so they end up crushing the thing and they show how they did it. They end up getting 90% prediction accuracy and they basically solve the single protein folding question. Now there’s also multi-protein and variations and all these other harder things, but it’s pretty crazy that the line graph was like 20% or 30% and then went to 90% in one year when they went back to the drawing board and figured it out.

How ecstatic they were! They’re like, “Yo, this just changed the world. People don’t realize this yet but this just changed the world.” It reminded me of your inflections thing, which we should say—describe your inflections thing for entrepreneurs. I think this is one of the best axioms or principles you have on entrepreneurship.

Sam Parr: Yeah, basically—and I didn’t invent this. I think it was Mike Maples. I don’t remember exactly, but basically the idea is that in order for a lot of big breakthrough ideas to truly happen—not like small businesses that make tens of millions of dollars, but like culture-changing companies—you basically have to have inflections.

There’s a handful of inflections that matter. There could be regulatory inflections. During COVID, BetterHelp and all these telemedicine things existed because we changed the rules on who doctors can serve. It could be a cultural inflection, like the Me Too movement that changed a bunch of stuff. Or why does Uber exist? Well, there was a technology inflection. Everyone now had a cell phone that had GPS on their phone. Therefore, they could call an Uber wherever they were.

There’s about five or six different categories of inflections. Then you have to spot the inflections to know what’s actually worth going after because you need an inflection in order for a culture-changing company to exist.

Shaan Puri: Right. I think this AlphaFold stuff or figuring out the protein folding is a massive inflection. I didn’t really know what the businesses were around this but I Googled afterwards—I was talking to Grok and asking it about this. There’s some pretty cool companies.

First of all, Google has their own company they spun out from this: Isomorphic Labs. Basically Google has spun out this company that is trying to cure all disease. That’s the mission. No big deal. Their thought process is, “Well, with AI, we can from first principles change the way that drug development and discovery works.” Because if we can predict how the proteins fold, then we can have a way higher hit rate on the targets we design with the drugs.

Then we should be able to simulate if it’s going to work before we even get to clinical trials. We should be able to run hundreds of thousands of simulations to see how effective this is going to be to get the probability of success higher so that when we enter a trial, we have a way higher hit rate. This company, by the way, their first round of funding was $600 million as they spun it out of Google and DeepMind. Demis is the CEO, I think, of Isomorphic Labs. There’s a world where Google becomes the drug company that cures—right now they’re working on malaria and these different things. He’s the CEO of DeepMind as well.

Sam Parr: Yeah. Wow. The H1 on isomorphiclabs.com—the headline is “Solve All Disease.” “We’re entering a new era of drug discovery. One where the frontier of AI can unlock deeper insights, faster breakthroughs, and life-changing medicines.” If I was doing a Sarah’s List episode right now, Isomorphic Labs would be one of those where I’d be like, “Go be a PR person there. Go be a junior account manager there.”

Shaan Puri: Yeah. Does the cafeteria workers… yeah. You guys literally show up and you say, “Hey, I’m the best coffee bringer to your desk ever. Give me a job here. I will find ways to be useful every single day, whether it’s in any job you have. I need to be at this company because I can’t think of a more noble mission, but actually a shot of cracking it because there’s a new tech vector to go chase.”

Sam Parr: So, how does the documentary leave it?

Shaan Puri: The end is weird because they’re still in the beginning stages of what they’re doing. They end the documentary, but the AI stuff is just starting to work. They end it with after the AlphaFold thing, all these researchers and drug companies are like, “Hey, can we get access to this because if we know protein structures, this will be tremendously helpful.”

They’re like, “Oh, we should set up a service where you can request a protein and then we tell you how it’s folded.” And then Demis was like, “Can we just fold all the proteins?” And they’re like, “What?” And he’s like, “How long would it take to fold all proteins known in existence?” And they’re like, “We can do that in like 2 months.” He’s like, “Well, why don’t we do that? Let’s fold them and give it all away.”

He’s like, “Let’s just make it open for anybody. Let’s go run the computer, fold all the proteins, give it all away.” And so that’s what they did at the end. They folded 200 million plus—basically every known protein in existence—and they made it available. The end is basically researchers from around the world showing up on their Google Analytics like logging in. They’re like, “We have 100,000 concurrent users. We now have 3 million users.” That’s everyone from someone in Africa running a small lab to universities to Eli Lilly who are all using them to be smarter and better about how they do medicine.

Sam Parr: Dude, how are all the guys who work at Isomorphic Labs and DeepMind and Demis—how are they all not like Andrew Tate looking dudes, like the most tan, jacked dudes ever? Because if you can cure…

Shaan Puri: Taking peptides…

Sam Parr: If you can cure all disease, how are they not the hottest people on Earth?

Shaan Puri: I think the way you become that smart is you don’t care about stuff like that.

Sam Parr: No, you become that smart because you were bullied, but now you’re going to seek revenge. The issue with bullying going away is that none of these nerds are going to exist. It’s like, I know we’re getting close when Demis is 6’4” and has visible lats. Yeah. Why does he not look like Adonis? That’s my question.

The Gorilla Analogy and Human Desire [00:53:15]

Sam Parr: I don’t pay attention to the news too much but when Larry Ellison and Masayoshi Son and Trump and Altman did this thing where it was a hundred trillion dollars or some ludicrous number, it was under the premise of “this is going to cure disease.” Larry Ellison—I do know that Larry Ellison’s in his 80s I think, or close to it.

Shaan Puri: Looking good, Larry.

Sam Parr: Close to it, and he’s looking great and his wife is like a 30-year-old but for some reason they don’t look like that much of a different age, even though there is literally a 50-year difference. It was under the premise of, “We—Larry’s interested in solving death and therefore we must do that.” Whenever I hear that, I just think that’s just words that are meaningless. But now that I know a little bit more about the topic just from you now, is that actually a legitimate thing?

Shaan Puri: Well, I’m glad you’re asking me because as a pre-med student, I’m clearly an expert on this. So hard to know.

Sam Parr: Someone who took one class on this 15 years ago.

Shaan Puri: Yeah. That’s what you got a C in physics. Had to repeat in the summer. It’s like on Instagram when people—you see Instagram videos of people with their children and the kid’s on an iPad or screaming and someone’s like, “Well, as a mother I could never.” It’s like, dude, you mean as a human being? I don’t care. “As a mother” does not mean that you are right. Shots fired at the mom. Not special. As a father—brother, everyone’s a father. I don’t care.

There’s a line where they’re talking to one of the OGs of artificial intelligence and they were saying, “What are predictions?” and he goes, “It’s hard to predict what’s going to happen as we make this intelligence into superintelligence. It’s like asking a gorilla to explain Einstein’s theory of relativity.” When I heard that I go, “Oh yeah, we’re going to be the gorillas out of this thing.”

Clearly if you’re making intelligence smarter than any human, you’re creating the next race. It’s like to an animal, if they just saw a human at first, they’d be like, “Looks kind of skinny. They got a little funny little extra appendage on their hand. All right, cool. They walk upright. Oh, cool. But they’re pretty slow, actually.” And then you fast forward a thousand years and you see the Blue Angels flying above you. I don’t know. I have no sense of time. Speaking of gorillas, we are a few brain cells away from gorillas. But just what humans have done is incomprehensible to our closest animal relative. That’s what’s going to happen here, which I think is pretty crazy.

Sam Parr: I don’t know if you listen to Mark Manson. He’s the man. Basically, he did a podcast—I think it was his second most recent one. It’s a Q&A about finding your purpose, failing better, and the AI future. He tells a story about how he built an AI product recently. Someone asked him, “What do you think about the future of AI?” And he was like, “Well, I just built an AI product. And what I realized a few things. One, AI is amazing in that it’s better than 95% of people at certain things. But the vast majority of value created in the world is created by people who are 99.9% better than other people at human things. You still need these experts and AI can be great, but it’s not an expert.”

He also said there’s maximalists who think that AI is going to come really soon and take over the world and we’re all going to be worthless. And there’s other people who think that it might happen over many decades, but we’re probably going to be fine. He was like, “I tend to be in that category.” The reason being is that when a lot of people think about AI, they think that it’s just going to take all of our jobs and we’re not going to work anymore. But human desire is not fixed.

Often times people think desire is fixed, meaning once you hit a certain level of productivity, you will not do stuff. He’s like, “That’s just false.” For example, if we look at the Industrial Revolution, people said the same thing when certain stuff started happening. And then you look at the Victorian era where we started getting electricity, things like that. People made the same claim of “we’re not going to work again, we need universal income.” He’s like, “Humans just always want more.” Because of that, I don’t think that there’s ever going to be a point where we are useless. It’s just going to be different. I thought that was a really great perspective on it. That’s one of the first times I’ve heard a perspective on it other than maybe Dharmesh talking about it where I felt calmer. Desire is not fixed. We will evolve.

The Importance of Documenting the Journey [00:58:30]

Shaan Puri: Yeah, it’s interesting. I don’t know. Have you read this? There’s this book—I haven’t read it yet, but it’s called If Anyone Builds It, Everyone Dies.

Sam Parr: I don’t think I’m going to be reading that one.

Shaan Puri: Yeah. I mean, it’s a crazy documentary. One my meta takeaway is I love that they were filming this the whole time. I’m glad that smartphones and video and these video platforms are so popular now because imagine 10 years from now, I think we’re going to have 10 times the number of documentary behind-the-scenes building-it type of things.

Right now it feels like a fluke whenever this happens. For example, we did a podcast about the Kanye documentary and the craziest thing about the Kanye documentary is not about Kanye. It’s about the guy who just decided, “You know what? I’m going to just film this young guy in Chicago over a 10-year period because I think he’s got something here.” Which is one of the greatest calls ever. Before Kanye was Kanye, this guy started filming. I think we were lucky that that ever happened. That’s a lottery ticket level win for society that that guy just decided to film this religiously when there was no reason to believe that he should do that.

Conor McGregor did this on his way up and he’s like, “I’m going to be the best. Yeah, I’m a plumber now and there’s never been an Irish champion, but I’m going to do it.” He basically started filming a documentary at the beginning and because of that, you get this incredible look at what it was like on the come-up. It’s incredibly inspirational whenever this happens. I’m just glad that they did this and I hope more people do this. That was my big picture takeaway, really not even about DeepMind itself.

Sam Parr: I was going to say that the need for human craft goods—for example, you could buy anything you want but some people still want the handmade stuff from Italy and they want to know the story behind it. When you’re talking about the story about him playing the Chinese guy and the Korean guy, there’s still a human that is half of the story. It was necessary to the story. We are still drawn to stories and story in my head is sort of an analogy to where humans fit into this thing. We’re still drawn to these human elements of all of these stories which makes me believe we have to be part of this experience. We’re not going to be completely outsourced because the most interesting part is that this genius guy has called his shot this whole time and has been interested in this for years. That’s actually the most compelling part.

Shaan Puri: Yeah. Although, you don’t want to be relegated to, “Well, we’ll still make handmade goods.” That’s the 1%. 99% is the mass-manufactured things. You also don’t want it to be where, “Well, we’ll always be interested in human entertainment,” and everything else will be done by the AI.

Sam Parr: No, but I don’t mean that. But I mean like when you fly on a plane, you’re not like, “Can I get the one where the pilot’s doing all the work?” You’re like, “Okay, cool. This is run by a computer that’s way safer than a pilot. Great.” I’m glad there is a pilot, but the computer could fly. I think that I’m not—part of me is nervous, but I do like… well, a big part of me is actually nervous.

Shaan Puri: Most of me.

Sam Parr: It’s like when people say, “Well, some people say it’s not that big of a deal.” Most don’t, but some do. Some few people got their head in the sand. Most don’t, but some do. I just think that we still are going to play an important role and I’m not too worried. Although I am very worried.

Computational Biology: The Next Billion-Dollar Opportunity [01:03:15]

Shaan Puri: There’s also this funny thing that happened on the documentary point. Did you see this founder of Robinhood talking about his documentary?

Sam Parr: No. What’d he say?

Shaan Puri: I didn’t know this. Vlad from Robinhood, when Robinhood was getting started, he put up a Kickstarter saying, “Hey, I’m going to try to build this company that’s going to change the way the financial markets work. And if you guys fund $10,000, we’ll film it because how cool would it be to see Steve Jobs building Apple? How cool would it be to see these guys building Google? That’s what we’re going to do.” And then it didn’t hit its Kickstarter goal and they didn’t do it. Is that crazy? The Kickstarter project’s still up. You could see the trailer, by the way. I could see why it didn’t get funded. The trailer was garbage. But with the idea and him being like—he’s like, “Yeah, in hindsight, I was right. That would have been awesome. But we didn’t hit the 10K. We only got to $2,000 donated, so we didn’t do it.”

Sam Parr: So funny. Oh my god. I mean, did he actually compare himself to Steve Jobs or he saying like…

Shaan Puri: I don’t remember what he said in the Kickstarter, but I think he’s not comparing himself. He’s just trying to get you excited about why should you care about this company you’ve never heard of? And he’s like, “Well, imagine if the great companies had had this at the beginning. You would have wanted that, right?”

Sam Parr: Oh my god, that’s so funny.

Shaan Puri: He called his shot. Nobody cared.

Sam Parr: But think like—I’ve talked about this. I love doing home movies. I try to every day take a three-minute video of my family of us doing something and I have it on a secret YouTube channel.

Shaan Puri: What’s it called?

Sam Parr: Sam’s super top secret channel. Don’t look. But all the videos are unlisted. You couldn’t even see if you wanted to. But because a YouTube short can be three minutes now, I’m doing three because the problem with a lot of these videos that you take with your family or with your friends, they’re just 10 seconds and you’re like telling your friend like, “Wait, repeat. Hold on. Do it again.” Versus, “Here we are.” You remember when you’re a kid and your dad’s like, “Here we are. Christmas morning 2004. We’re doing this.” And those are the best. Those are the best videos. Thank God we have these phones because that’s really what a video should be is a three-minute to five-minute video of someone narrating saying what’s going on and no one’s performing. Versus now whenever I pull my phone out it’s always like, “Wait, hold on, tell me that joke again.”

Shaan Puri: How do you feel? Are you—so I definitely land on the optimistic side. I think he’s a badass and I loved hearing his story. I think it’s super mission-driven. I think it’s super cool that these guys believed 20 years before anybody else. I think it’s super cool they filmed the thing. I think the breakthroughs they’re doing—I like seeing use of AI that’s not all the same, which is like today the ChatGPT experience is so dominant that it feels like that’s what AI is.

It’s kind of like early internet—getting online and being like AIM or AOL News: “This is the internet.” And it’s like, “No dude, there’s so much more that’s going to come.” That’s how I felt when I heard more about the protein folding stuff and how they did the game stuff and how that’s going to apply to all these other domains. I walked away being like, “Oh man, if I was young and just trying to figure things out, if you’re high potential and you don’t know where to go, I got two words that’ll probably make you a billion dollars: Computational Biology.”

Just go there and just go play around. If you’re an entrepreneur, forget building a GPT wrapper, build an AlphaFold wrapper. I think you could build a multi-billion dollar company like what Cursor did. Cursor basically said they didn’t make the model. They were like, “Let’s take Claude, but we’ll just wrap this in a tool that programmers can use that will be very useful for programmers.”

Go do that for pharmaceutical companies. Go do that for research labs. Or you realize that with protein folding, people are going to want to test protein synthesis in the real world. That demand is going to go up 100x. Go create wet labs and just do the tests for the people who are using computers to come up with their hypothesis. That demand is going to go thousand-x. I was like, “Man, there’s so much opportunity for anybody who wants it.” I got two words for you. I thought it was going to be like “suck it” but it was “Computational Biology.” You led me down a path. I thought you were going one way, not the other way. That was great. This was a great episode. All right, that’s it. That’s the pod.

My First Million Wiki

Explorer