Sam and Shaan explore the mindset of treating business as a sport, the Inner Game of Tennis framework, and a “letter to future self” technique for managing inevitable setbacks. They dig into Surge AI, the billion-dollar bootstrapped data-labeling company run by the mysterious Edwin Chen, and compare it to Scale AI. They round out the episode with a wide-ranging discussion on self-driving cars, AI in medicine, and using Gemini Live as an endlessly patient babysitter.
Speakers: Sam Parr (host), Shaan Puri (host)
Opening Riff: “Generative” Is the New High Agency [00:00:00]
Shaan: Dude, “manifest” is out. There’s a new word.
Sam: What?
Shaan: Generative.
Sam: Wait — is “high agency” out? Are we selling high agency?
Shaan: We’re selling high agency at the top right now. We’re spacking high agency. It’s gone. Taking that cash and plowing it into generative.
Week Recap: The Golf Hole-in-One Idea Goes Viral [00:00:30]
Shaan: All right. What did I miss? How was the week?
Sam: Week was good. What did we do? We had Chris Corner. That episode’s popped off — it’s over 100K on YouTube, so that’s going well. And dude, there were so many replies to one idea that was in that episode. I don’t know if you listened — the golfing one.
Shaan: The golfing one, dude. I got literally hundreds of replies from people who are like, “I could do this right here in my hometown.” People are sending PowerPoint decks. People are doing drive-bys, sending me videos of the spot where they think they could do it. They’re reaching out cold. It’s very intense.
Sam: How many people have replied to this? And what was the idea? Was it about betting on where you could hit it?
Shaan: No. Basically, on the way to the golf course in New Zealand — kind of on the side of the road — there’s a road driving by a body of water, and if you stop on the side of the road, there’s this thing called like the hole-in-one challenge. You buy a bucket of balls and try to hit a hole on a little golf float out about 100 yards in the water. If you get it, you win $10 grand. It’s just a fun thing to do with your buddies on the way to or from the golf course.
He was talking about napkin-mathing what he thinks it’s making based on available information. He was like, “I think this thing does like $300 to $500K in revenue.” And the costs are pretty marginal — it’s a person standing there with an iPad and a scuba diver who goes in once a week to fish out the balls. That’s it.
We basically said, “I think this idea could work at more places than just this random roadside thing in New Zealand. Let’s bring it to life. Who wants to do this?” A lot of people came out, so we’re going to make it an MFM project and see what we can do.
Sam: All the comments were like, “This is what I’ve been missing with MFM,” because we started with a lot of that kind of thing and then our interests grew and the content evolved to be a little different sometimes. And one critique is like, “Is this My First Billion?” because we talk about bigger ideas.
I was thinking — I’ve become acquaintances with Joe Lonsdale, you know, who’s worth billions. And I was with him recently when I got my Twitter check. Like, you know how you get Twitter money now? My Twitter was $1,000 last payment. The month before it was $600. And I was like, “Man, this is crazy. I just got paid $600 for tweeting.” He’s like, “Yeah, I got like $400.” And he was joking about how it feels just as exciting every once in a while to get a $400 thing as it does however much money he’s created in his lifetime.
And I was wondering — do you feel like, you just lit up when you talked about $300–$400K, and that may not really move the needle for you. But it’s kind of exciting, isn’t it?
Shaan: Yeah. Not because of the money. I think the idea itself is fun. Making it happen sounds like it’s going to be fun.
Business as a Sport [00:05:00]
Shaan: Actually, I was just watching an interview — the NBA Finals just ended, Game 7, the Thunder won. I watched it. They asked JD, “When you look back on this year, what are you going to remember? What were the high points?” And he said, “It’s weird, dude. When I think about this year, I remember me and Chad, we would go to the hotel room and do film sessions back when he was coming back from injury. Or these team dinners we were having.” He was like, “I couldn’t even tell you what happened in the last series. I don’t remember the recent games, but those inputs on the journey are so vivid to me.”
This is a really common thing — if you talk to pro players after their careers are done and you ask what they miss most, you expect them to say the big pressure moments, the championship games. And of course they love those. But what they always talk about is the team bus rides, the locker room, the camaraderie stuff along the way. The buildup is the stuff they miss most.
I think there’s that for entrepreneurship too. That’s a huge amount of the fun of it, and it’s what you get excited about. You need the number to sort of justify it — the number gives you air cover for why you’re acting like a little kid. The numbers help because, why are you taking this silly thing so seriously? But I think we would probably all do it without the numbers, or if the numbers were half as much.
Sam: Yeah. I’ve noticed the people who love MFM the most, and the guests you and I love the most, are folks who — I hung out with a friend of mine, and she was like, “I’m so good at going really high and going really low.” I was like, what does that mean? She’s like, “I can hang out with my homies from where I grew up and we just shoot the shit and be a little hood ratty, or I can go hang out with a billionaire and I love that too. I have so much joy doing that, and I can blend in with everyone.” And I think that’s what the podcast is — you like talking about these smaller things as well as the big things, and it’s the same type of person who loves both.
Shaan: Exactly. Also, do you think about business as a sport? Because that’s more and more become my mental model.
You meet a lot of people we know who have become successful but they’re still doing it. For many of them, I call it — they’ve already made the last dollar they’ll ever spend. Let’s say you make $30 million. At that point you’ve already earned the last dollar you’ll ever need to spend, especially once you take into account that $30 million could just sit in an interest-bearing account or the S&P 500 and it’ll double every seven years. So 30 becomes 60, 60 becomes 120, 120 becomes 240. That all happened over something like 25 years. So you don’t need to go earn the next dollar — but why do they anyway?
Part of it is I think it feels good to be good at something, and if you’re good at something it’s hard to stop doing it because the feedback loop of being good at something is strong. But I think in that same way, if you think about business not as a mechanism to make money, but as a sport you play — then it’s like, of course, just because you’re great at tennis and you won a tournament doesn’t mean you’ll stop playing tennis. Why would you? That’s your sport. You love to play the sport. You’ll basically play it till your body breaks down and doesn’t let you play anymore.
And it feels good to manifest. It feels good to have an idea and see it come into reality. It’s really fun flexing that muscle.
Sam: Manifest is out. There’s a new word.
Shaan: What?
Sam: Generative.
”Generative” as a Lens for People [00:10:00]
Shaan: Right, generative. I was on a podcast, and at the end I was like, “How was that? Give me the truth — from one podcaster to another, what was that like for you?” And he goes, “It was great because you’re extremely generative.” He goes, “It was also hard because you’re extremely generative.” I go, “What does that mean?” He goes, “I’ll give you one topic and you can almost bloom it — expand it into a story, a framework, a related idea, a simple example. You just generated all that content off the cuff.”
He said, “Biology is like that. Biology is extremely generative — you give him one thing and he’s able to take it from the origin of man to 100 years in the future and connect all those dots.”
I heard it once and I was like, okay, I don’t know if I just got insulted or complimented being called generative, but I’ll take it. And then James Courrier said the same thing. He’s like, “The reason we get along is because we’re both extremely generative. We like being around generative people. Why do we admire Elon? It’s not because he’s rich. It’s because he’s the most generative of all of us. He’s the least fearful, and that’s why he’s able to be more generative. He literally generates businesses — the Boring Company, Neuralink, SpaceX, Tesla. He’s generating kids. He’s generating ideas. He generated a president. He’s just doing so much.”
So I started using that little lens — looking at people and asking, how generative is this person? If you give them an inch, could they take a mile? What is their overall level of output? It’s not just businesses with someone like James Courrier — he started a church in San Francisco, created an incubator, a fund, a podcast. He’s just constantly creating things. And it’s not confined to one area — it’s his kids’ lives, his business life, everything.
I started to realize, yeah, I’m really attracted to that. I like people who are like that. And I want to be like that. Figuring out how to make that work is a fun challenge. So generative is the new word.
The Inner Game of Tennis [00:16:00]
Sam: Have you ever heard of this book called The Inner Game of Tennis?
Shaan: I’ve heard of it, but I’ve never read it. Is it good?
Sam: Yes. So The Inner Game of Tennis — I randomly discovered it at the airport. I was looking for a book to read on my Kindle, wanted something short, and I came across it. It’s written by a guy named Timothy Gallwey, and it’s one of those books that’s actually about life but uses tennis as the analogy.
The premise is that you have two selves. Self 1 is your critical person — when you’re playing tennis and you hit a bad shot, you go, “Why do I suck so much?” That’s Self 1. Self 2 is your animalistic self who doesn’t overthink. It learns by observing. The whole book is about how to be better by ignoring Self 1 and letting Self 2 do all the work. It gives you tips and tricks on how to listen to Self 2.
This sounds very woo-woo, and it is a little bit. But the book was written in the 1970s and Pete Carroll writes the foreword. Every new edition, they’re still releasing new editions where a whole who’s-who of leaders are writing about it. And I didn’t realize it, but after I started reading I was like, “Oh wait, Tim Ferriss talked about this as one of his favorite books of all time.”
It’s only about 150 pages. I read it in two days. It applies very much to business.
Shaan: This is incredibly fascinating. I love this type of book. It says, “The Inner Game of Tennis: The Classic Guide to Peak Performance. Introduction by Bill Gates, foreword by Pete Carroll.” That’s wild. So have you used any of this? Give me an example of how you’ve applied it.
Sam: A very simple example is lifting weights. When you lift weights, you’re like, “Okay, I have to lift this weight three times, and it’s the heaviest I’ve ever done, so I’m really scared.” You don’t listen to that at all. Instead, you just get under and go, “I’m going to let Self 2 do all the work. I’m going to trust Self 2. And if I fail, I will not be judgmental. I’m not going to say, ‘you suck.’ Instead, I will just objectively acknowledge — your knee moved in a strange way.”
So I’m just talking to myself like an objective machine, not an emotional person. “The ball was out. Okay. Noted. The ball was hit too hard.” And then you trust Self 2 to adjust. You do not judge it as “I hate this, I suck, this is bad.” It’s just — the ball was out.
Letter to Your Future Self [00:22:00]
Shaan: I did something similar to this — I didn’t even plan to talk about it, but it’s kind of in the same vein.
One thing I noticed is that anytime I go into a project, I have a lot of excitement and hope at the beginning. That’s obvious. And then the second obvious thing is that I’m going to hit some obstacles, walls, plateaus — something I don’t want to happen is for sure going to happen. I’ve never once experienced a project where I simply started, everything went as planned, and it had a happy ending. Like, that’s literally never happened. For me to expect that would honestly be a little foolish. Yet as soon as I hit those obstacles and those walls, I’m like, “Shit. I wish this didn’t happen. Why is this happening?” And I waste all this energy on something that was inevitable.
It’s like playing Mario and being like, “Oh my god, I can’t believe these Goombas are walking at me.” Dude, that’s the game. What did you think this was going to be?
So recently I wrote a simple letter to myself for like two or three months down the road. I’m just going to read you how I wrote it. I was like, “Hey, it’s me from the future. I’m writing this to you three months from now. First, congrats — you did so good, it turned out amazing. I’m really proud of you slash me.” And then I said, “This is a letter guiding you to some of the entirely predictable upcoming road bumps headed your way. Not only is it predictable that there will be road bumps, I can probably tell you right now what they’re going to be.”
Like, let’s say you’re trying to hire a head of sales. There are some entirely predictable road bumps: you’ll probably procrastinate starting because finding the perfect person feels hard. Then you’ll talk to some disappointing candidates. You may even find someone great, but the offer doesn’t work out — maybe they don’t take it or it’s not the right time in their life.
So you can basically tell yourself upfront: these four obstacles are probably coming. I’ve played this level before. And when they arrive, it takes the emotional edge off. You don’t feel betrayed. You don’t feel surprised. Like, yeah, I knew you were coming. Say hello to it — here you are, I thought I’d be seeing you soon. And I’d already thought about what I’d do to get around it before I’m in an emotional state.
“I’m probably going to meet a bunch of disappointing people in interviews. And it’ll feel in the moment like, God, am I ever going to find somebody great? But of course I will — I only need one, it’s a numbers game. I should probably expect to talk to 30 or 40 people and that 25 of them are going to be truly a waste of time. But that’s part of the process.” And you tell yourself that upfront. Then as it’s happening, you’re like, “Yeah, I already addressed this. I don’t need to react again because I already pre-reacted to the whole thing.”
Sam: Is this project you’re doing — is it big or small? Do you recommend doing this for a small thing or only a big thing?
Shaan: I don’t know. This is my first time actually doing the corny step of writing it out to myself. Like, “Dear Shaan…”
Sam: Yeah. And then it’s like, “PS, you’re pretty lame for writing this.”
Shaan: Exactly. I was like, all right, that’s three pages now. This was cool when it was a paragraph. But it was very helpful. I will do it again. It sort of blunts the pain, but the pain’s still there. It’s like when you get a shot at the doctor — if you’re hyperfixated on it and start hyperventilating, it’s a worse experience. If you look away, you might still feel a little prick, but you took the edge off. I think that’s what this has done for me.
Patron View: Nick Gray’s Personal Software Project [00:30:00]
Sam: All right, so we’re talking about big and small. Do you want me to tell you about a small thing and a big idea that to me are equally fascinating?
Shaan: Okay.
Sam: Go to patronview.com. I was with Nick Gray this weekend. I did this amazing vacation — my friend David owns a home in Utah and about eight of us plus our spouses and kids all went and hung out. It was amazing. Nick was there and I looked at his computer and said, “Nick, what are you doing?” He goes, “Let me tell you.”
So it’s called Patron View — patronview.com. Nick used to own Museum Hack, where you’d pay $100 and Nick or one of his tour guides would take you to the Met and give you a kind of guerrilla tour. That’s where he got really into museums, and he became buddies with the guys who do the fundraising. Being a business person, he was like, “It’s so fascinating that one person is donating a million dollars, $10 million, $20 million to these museums every year across tons of different institutions.”
So recently, with a mutual buddy Stacson Blake, they built this website. It’s pretty amazing. If you go to the Met or one of hundreds of other museums, every year they have to put out a PDF that explains who donated money and how much. He aggregated all of them — hundreds, maybe thousands — and used AI to upload them into a database.
If you’re fundraising for a museum, you’re going to be able to pay his service to find out who the whales are. And because of AI, he was able to build this — he told me — for $2,000.
The about page says: “We’re a research platform dedicated to documenting cultural philanthropy. Our research pulls from annual reports, 990 tax filings, institutional publications, official documents, and proprietary sources. This lets us present donor information that’s never before been displayed. We like to think of it as celebrating philanthropy and enabling development departments.”
Pretty cool, right?
Shaan: Awesome. It’s great. Nick — I’ve already really shouted him out on here a ton of times because he’s someone who’s made a big impact on me just seeing the way this guy rolls through life. He just does things for his own amusement. He does things on his terms, with high intentionality. He’s basically resisted the rat race. I think those are the people I admire the most — people who truly resist both money and status. If you think about the talented and successful people in your life, how many actually do that? It’s very, very few. I probably know two people like that: him and Jack Smith.
Sam: Yeah. You sort of watch their moves and learn from them. What’s funny about Nick is every two or three years, he likes to find a publicly traded company he loves and make a big bet on it. For the last four or five years, his bet has been Cloudflare. He loves it to the point where when he hosted an event, he specifically chose the Cloudflare event space. He’ll wear Cloudflare t-shirts. One time there was a 5K through Austin and he held up a sign that said “Cloudflare rules.”
He told me — at his birthday party, he had it at the Cloudflare office, and midway through he ran upstairs, got a product manager to come down and said, “Hey everybody, quick word from Jack from the marketing department. Tell us about the great things you got going on at Cloudflare.” And before Jack walked in, Nick said, “I need everyone to treat Jack from Cloudflare like a celebrity.” So when he walked in, everyone went, “Oh my god, is that Jack? Are you the VP of engineering at Cloudflare? He’s here!”
The stock is up 400% in the last 5 years, so he’s done pretty well. And if you click the about page on Patron View, he lists under Technology: “Patron View is built with modern web tech to ensure fast, reliable access to data.” He only put that in so he could list that he uses Cloudflare. I know that’s exactly how he thought.
The reason I’m bringing this up is: if you’re just starting to build a business or something, follow Patron View. Go there once a week and I would bet you’re going to see it evolve. It’s sort of like measuring your kid on the wall.
The Personal Software Moment [00:39:00]
Shaan: This is cool, and it fits into a genre I’d call personal software — or maybe social software.
Before the internet, the only people who made media were media companies. You got your media from the New York Times, newspapers, magazines, TV. When the internet came out and you got Facebook and Twitter and Instagram, social media became a thing and everybody became a little broadcaster. There was like a 1-billion-x increase in the amount of media created because everybody was doing it.
One clear thing I see happening now is that’s now happening with software. Software used to be something only software companies and software engineers could make. There are maybe 100 million professional software engineers in the world — 100 million out of 8 billion people. Now with Replit and Vercel’s v0 and all these different tools, it’s going to be like social media. I carry in my pocket a thing that can make software.
A guy like Nick, who before this probably couldn’t have taken his idea and made it into an app — he would have had to either learn to code or go hire expensive programmers — did most of this with AI. So this personal software category, which basically didn’t exist three years ago, is now going to have the same kind of 1-billion-x increase because anybody who’s got an idea can now make their idea.
Today it’s broken about three-quarters of the time and doesn’t quite work. But every 6 months that number goes down by 15%. Within two or three years, that’s going to be near zero. You’ll have an idea and you’ll make your app.
Sam: Everything I’ve been making on Replit and Lovable and Cursor — it’s basically just like a Figma replacement. Like I’m just drawing on paper.
Shaan: Yeah, it’s a sick mockup. Somebody called it a minimum viable promise. Instead of minimum viable product, it’s like you make a promise. You can see the promise of something. That’s what a lot of these tools are able to do today.
Edwin Chen and Surge AI [00:44:00]
Shaan: Have you heard of a guy named Edwin Chen?
Sam: Edwin Chen? I mean, there’s probably 6,000 of them on my Facebook feed.
Shaan: Edwin Chen might be on the chart of richest slash unknown slash youngest person in the world.
Sam: Is this the guy doing Surge?
Shaan: Yeah. Edwin Chen — around 2018, 2019, he worked at Facebook and was tasked with making some kind of Yelp-style product. What that meant was he had a list of 50,000 vendors and needed to figure out which were restaurants and which were grocery stores. He went and hired a firm to parse it out — manual work, offshore talent going through everything by hand. It took four to six months, so he basically just had to sit and wait.
He had this idea: he was going to make a better way to do data labeling. Data labeling is important because that’s what a lot of AI companies use. When a company like OpenAI wants to figure out if a certain reply is unethical — say, asking if it’s okay to hurt someone — they need really smart people, even philosophers or engineers, to go through all the potential answers and say, “I think this one fits what you’re going for.”
Edwin had the idea: I’m going to create a massive workforce of philosophers, engineers, Ivy League graduates who can go through and label all these answers as good or bad, so AI companies can use it. Kind of like their offshore talent. He started this in 2020. Now he has 100,000 people in the marketplace working for him as data labelers.
This company is completely unknown. If you go to Surge AI — I think it’s surge.ai — it’s a landing page with one paragraph. And it’s an amazing paragraph. Want to read it?
Sam: Yeah.
Shaan: “What made people like Hemingway, Coltrane, and von Neumann so extraordinary? Their life. The books they read, the stumbles they had — the reinforcement every time friends laughed at their jokes, and every time they didn’t. It’s the people they met, the places they explored, and every decision they made along the way. Data does for AI what life does for humans. It elevates the neural networks that know nothing about the world into intelligences capable of providing new art, sending rocket ships to Mars. Our mission is to shape AGI with the richness of human intelligence — curious, witty, imaginative, and unexpected brilliance. We wake up every day trying to produce the data that makes this possible.”
Sam: That’s amazing. Romantic.
Shaan: It’s romantic. This guy made a giant fleet of data labelers sound like the army from 300.
The business model: they have 100,000 trained annotators and software that shows tasks to their workforce. A company like OpenAI or Google pays Surge millions and millions of dollars, and Surge takes about 30 to 40% and gives the rest to the annotator. This company is only 5 years old, and it was leaked that they did $1 billion in revenue in the last 12 months. Edwin Chen is only 37 years old and he owns 100% of the company. They have not taken any outside funding.
Sam: Holy.
Shaan: Their biggest competitor is Scale AI, run by Alex Wang. Scale recently sold for something like 30 times revenue — they were doing $800–$900 million in revenue and just sold half the company to Facebook. I think it was for $28 to $30 billion. Which means Edwin, who’s 37 and has a five-year-old company, is presumably worth something like $30 billion.
And you cannot find him on Twitter. He has no blog. You can’t find photos of him. He used to have a blog, but you have to go to the web archive to find it because he took it down. And his customers say — “Edwin is not online. You can’t find him anywhere. And we like it that way.”
His business is very boring. The branding is basically nonexistent. It just does a very good job. Compared to Scale, where Alex Wang was just on Theo Von’s podcast, he was at the inauguration, he’s kind of the “it guy” right now — these guys are the exact opposite.
Sam: I did not know he bootstrapped the whole thing. I’d never heard of this company until Scale got bought.
Shaan: They were already winning before that. Surge was at $1 billion in revenue and Scale was at $750 million. And the reason Surge is winning is because they charge a premium. He’s like, “I wasn’t trying to get scale — I was trying to hire the best people, train them really well, and charge for it. I charged three times what Scale charges and the results have been better. People really like us because of it.”
Sam: This whole data labeling industry — I had no idea about this. I didn’t know people were behind the scenes making these decisions.
Shaan: This is one of the best picks-and-shovels businesses. Anytime there’s a gold rush, who makes the money? It’s the few people who find the gold, but the more reliable way is to sell picks and shovels to everybody rushing in. Scale and Surge were the best picks-and-shovels businesses, maybe besides Nvidia. They said, “Cool — everybody wants to make AGI, you’re all raising billions and billions of dollars, and all of you have the same problem. I’ll sell the data labeling service to all of you.”
And it’s so funny that now that Facebook is buying Scale, all that revenue has to find a new home. That’s the best news ever for this guy.
Handshake Pivots to Data Annotation [00:54:00]
Shaan: There’s another company called Handshake. Previously they were known as a job board or job network for recent college grads — basically helping 22-year-olds get jobs.
Sam: I remember using this. Like, “Oh, it’s interesting that nobody’s really built the one place to hire college interns.” They built a marketplace where you could post on a job board at your local college. It was kind of crappy though — very little liquidity in the market. But I remember thinking, someone should do this right.
Shaan: They noticed a few months ago that Surge and Scale were using their service to find data annotators. So they said, “We’re going to do that now.” In a very short amount of time they pivoted and that business is going to hit $100 million a year in the next couple months. What they did was they went and said, “You’re looking for a data annotation gig? We got you. Let’s get you trained and provide that service.” Now Handshake is building that business.
Sam: That’s so crazy. If you Google “Handshake data annotation” you can find the blog post they wrote announcing this.
Shaan: Right. And I don’t know how long this will last. In 7 to 10 years you may not need this anymore. Either they label enough data where then the model learns how to label data and you don’t need humans doing this — or they use RLHF alternatives, reinforcement learning without human feedback. Some people who are pure believers in AI think you won’t need the human feedback at a certain point. So this might be a get-while-the-getting’s-good type of business.
The Pandora Story: Data Labeling Has Been Around 25 Years [00:57:00]
Sam: Let me give you a potential counter to that. Tim Westergren founded Pandora. I think he started it around 2002. He told me this story when he spoke at one of our events. He said, “I raised $7 million and all of it went to hiring musicians — musicians who were teachers and didn’t make a lot of money.” For two years he had about 150 of them listening to music and gave them basically a scantron of all types of attributes a song could have.
If you’re listening to the Beatles, you’d fill out: “It sounds like it’s at 90 beats per minute, there’s guitar, it’s melodic, it’s lighthearted.” After two years of doing this, he put all the data into an algorithm he built. He told me he played a Beatles song and clicked next. The BeeGees came up. He was like, “The BeeGees and the Beatles? They’re not similar at all.” Then he kept clicking next. “Oh wait — they have the same melody. They make me feel similar.” And he was like, “It’s working.”
Originally his idea was to create a kiosk at Best Buy — you could say you’re interested in the Beatles and here are five other songs you could buy on CD while you’re there. Then the iPhone came out and he was like, “Oh my god, this is the exact way to apply this.”
This idea of data labeling has been around forever. I didn’t realize it when I was reading about Surge, but I was like, “Oh my god, this is exactly what Tim was explaining about how Pandora started.” So this has been around for 20 years.
Shaan: Yeah, that’s true. You say you don’t know if it’s going to be around, but it’s been around 20 years so far.
Waymo vs. Tesla: The Self-Driving Bet [01:03:00]
Sam: But it’s kind of like self-driving, which is coming out now. I’ve taken Waymo in San Francisco and Robotaxi in Austin. The Tesla Full Self-Driving just launched in Austin a couple days ago.
They took two different approaches. Waymo has this really expensive car — all-in cost is something like $150 to $300K with all the sensors. In addition to LIDAR, they hard-map the road. For years they drove around physically mapping roads. They can only launch in cities where they’ve mapped. Tesla took the other approach — cameras only, no LIDAR, and they’re not going to hard-map roads. They’re going to let people drive around and the car needs a brain smart enough to figure out a road even if it’s never been on that road before.
Elon was like, “LIDAR is not only something we’re not doing, it’s stupid and a dead-end path.” And everybody else was all in on LIDAR. “LIDAR makes it safer. You can’t do this without LIDAR.”
Elon’s point was: we humans drive with just eyes. We only have cameras. I don’t have LIDAR in my brain and I drive safely.
Shaan: What is LIDAR exactly?
Sam: It’s like you’re shooting some type of signal and it bounces back. You can see through things. A camera can’t see through an object. LIDAR can — it can sense that there’s another object behind something. The classic example: you’re making a turn, something’s obstructing your view, and there’s a little old grandma walking in the crosswalk you couldn’t see till you started turning. LIDAR would know there’s a moving object there.
For a long time there was a big debate. Some experts thought Elon was wrong. Some were like, “Elon is correct and we trust him.” Very smart people on both sides. And it was a very high-stakes debate because self-driving cars is one of the most valuable prizes there is.
I don’t think people really realize it. I think because people talked about it for so long, they got numb to it. This actually happened with AI too — people had been tracking machine learning and deep learning for a long time and didn’t realize when something had actually changed. And those people were almost late to the party because they mistakenly wrote it off: “Yeah, yeah, I’ve heard this before.” The same thing is happening with self-driving. And it’s going to be an extreme game changer — for society, for Tesla’s business.
Like, Tesla’s business is now: when you’re not driving your car — 95% of the time it just sits parked — you’re going to tap a button and say, “Go make me some money.” And like a dog, it’s going to go out there and start doing rides for people, earning you money passively.
Shaan: Dude, I think Morgan Stanley or Chase — one of the big banks last week — wrote a report on what the world’s going to look like with self-driving. And it was far more grand than just “you can play on your phone in the car.” It was like, the economy is going to look radically different because people are going to have so much more time. And there are something like 60,000 car deaths a year. What’s the world going to look like with more people?
I asked Grok last night: “What are the second-order effects of self-driving cars?” Here’s what it said.
Cities are going to look completely different. Right now, parking lots occupy 30% of all urban land in some cities. You’re not going to need parking lots because cars aren’t going to just sit parked — they’re going to be rolling around. You’ll need way fewer cars in a city. Just look around — how much space is dedicated just to parking? We’re going to look back at that the way we look at smoking in restaurants.
Sam: Yeah, exactly.
Shaan: The good version is that becomes green spaces and affordable housing. But who knows — maybe it gets co-opted for drone delivery staging areas.
Next: labor. Right now there are three and a half million truck drivers alone, plus all the Uber and taxi drivers. That job is just going to be gone. I don’t know what happens to that, but there we go.
Then the average person spends something like 90 minutes a day commuting. You’re awake 16 hours — you’re going to add roughly 10 to 13% more time to everybody’s day where they can now sleep, work, or play. You’re not going to have to think about driving, so the car becomes a new place for entrepreneurs to build experiences. Nobody builds car games right now. Car games are going to become a thing. People are going to relax, work, and build tools for that experience.
Insurance changes completely — Buffett’s big bets, Geico, all of it is based on human driving. If humans aren’t driving anymore, both the risk and the reward ratios change. And you’re insuring the software company rather than individuals. How does all that work?
And car ownership — today a car is both utility and a status symbol. If transportation is basically on tap, flowing like water — you push a button and in 30 seconds the car of your liking pulls up — owning a car is going to be like owning a horse. A small group of passionate people. You’re lucky enough to have the room or money, but that’s what it’s going to be.
Sam: It’s like horseback riding — it’s therapeutic. People like to brush a horse, pet a horse. That’s what driving is going to be. Male therapy. Just get behind the wheel and have control over something in your life.
Shaan: Or like punk. You’re going to feel the noise and smell the gas.
Sam: Yeah. It’s going to be a hobby. But I think it’s going to be a lot longer than people think — maybe 20 or 25 years, not 5 years.
Shaan: Are you sure? Why not 5 years? Waymo is doing 20% of all the rides in San Francisco now. That was zero not long ago.
Sam: Because a large percentage of Americans have to drive 60 miles one way to work, or they have to pull stuff or carry stuff. I think for the urban case, sure, you guys are probably already there. But there’s going to be like four sections of users. Young urbanites — yeah, you don’t need a car at all, you’re probably already doing Uber. And then on the far end is rural people who actually tow things. You say you can tow on an electric car, but go talk to someone in rural Texas who has to be driving stuff around all day. And then there’s everybody in between.
I think it’s going to take a minute for the full spectrum of people. When you say it’s not going to be 5 years, I mean user adoption, not whether it works. It’s going to work.
Shaan: Yeah. That guy towing stuff probably still has an AOL email address. It might be 40 years before that person.
Sam: Exactly. And I’m one of those — I’m romantic about my gas vehicle. I had an electric car and I got rid of it. In my head I acknowledge it’s better, I acknowledge it’s the future, but it sucks. It’s like our vegan friends: I get it, we shouldn’t kill creatures, but it tastes so good. But when you dip it in ranch, it’s fantastic.
Shaan: What’s crazy is in Austin and SF, people are actually paying more for Waymos.
Sam: Yeah. People want to not be around someone. That was unexpected. When I drive my BMW that has self-driving stuff, I feel way safer than if it were just me.
I’ve noticed about 20% of people — and it’s usually men. Every woman I’ve talked to hates self-driving and every man I’ve talked to likes it.
Shaan: I don’t have it, so I haven’t had the sample size to notice that. I’m curious if that’s common or if you’re just indexing on three people.
Sam: I mean, yeah, it’s like five friends. But the husbands use it and the wives are like, “Nope, I don’t mess with that.”
Shaan’s Elon Reply Goes Viral (With Mom) [01:20:00]
Sam: You want to do one more thing, or do you have something?
Shaan: I tweeted something out that Elon replied to over the weekend.
Sam: How did that make you feel? Did you clap and scream?
Shaan: No. I played it so cool. If you had seen me, you would have thought I was under the weather. What happened was I texted my wife and said, “Oh, Elon replied to me.” And then I forgot about it. Next day I didn’t even think about it.
Then my mom calls. She’s like, “Shaan, what did you say to Elon?” I’m like, “What?” She’s like, “What did you say?” My wife had put it up on her Instagram story. I was like, “Oh my god, I’m trying to play it cool over here and you made it lame city.” And then I got multiple phone calls from people. And I thought that was interesting — that was the only time my wife shared something from me, and it was when Elon replied to me.
The thing I said was: “Within a couple of years, not using AI while you’re doing your job will be the equivalent of coming to work without a computer. Like if someone just turned up and said, ‘No, I didn’t bring it today’ — you’d be like, ‘What the hell, dude? What’s the plan here?’” That’s how it’s going to be if you’re trying to do your job and not using AI constantly.
And Elon replied, “Sooner, probably.”
I started thinking about that — somebody else said, “Pretty soon being a doctor who’s not using AI as a co-pilot — like a radiologist who’s just trying to eyeball every MRI without running it through AI — that’ll be considered malpractice. Because you put the patient at risk by not including the second layer of AI diagnostics.”
The flip is going to go from “this doesn’t work, we don’t use it” all the way to “if you’re not using it, it’s considered malpractice — whether it’s corporate malpractice or medical malpractice.”
AI in Medicine: Fighting a Doctor with ChatGPT [01:24:00]
Sam: My doctor friend admitted to me the other day, “OpenAI is a better doctor than me.” He’s been a doctor for 10 years. For years, patients would come in and say, “Well, Google says this” or “WebMD says this.” Over the last 6 months, the only people who’ve done that are doing it with OpenAI. “Well, according to ChatGPT…” And he goes, “And they’re right. A lot of times the diagnosis is right.”
Shaan: Dude, I got in a fight with a doctor recently about this. Did I tell you this?
Sam: What happened?
Shaan: My mom had to have a surgery but she was on a trip. I’m calling in every time the doctor made rounds — she’d FaceTime me because she’s on the other side of the country. This one doctor comes in and says, “Yeah, your levels were fine.” And I had actually read the test through ChatGPT, and the levels were high for something. And she’s like, “Well, which level?” And I told her the term. And she’s like, “Yeah, that was high, but it depends on the exact number.” So I go, “What was the number?” She goes, “I’d have to check.” I’m like, “You’re the doctor.”
I asked her, “ChatGPT said if it’s above this number, you should consider doing this additional step. Do you agree we should do that step?” And she was basically getting pissed. She said, “Well, if you’re going to ask me questions, I’m going to need to go look at the number.” And I literally said, “Yeah, you are going to need to go look at the number.” What are we doing here? Why are you offended by me asking if you’ve seen the data from the test you just said to run, and then you’re coming back to discuss the results without looking at them?
Sam: I don’t really understand what’s happening. Well, I think what’s going to happen — have you been to a doctor with an AI scribe now? They have AI scribes, and I think the AI is going to step in and say, “Actually, he’s right.” Like, that’s what’s going to happen.
And if I were an entrepreneurial doctor, I would 100% start a new practice all centered around: we are AI first. We work with AI. You upload all your files, we have all the context. I think a lot of people like us have the sentiment where they’re like, “I trust a computer more than a human being, but I still want the human to put their stamp on it.”
And it’s not even just that the AI found the problem and the doctor didn’t. Sometimes it’s as simple as: the doctor came in, talked kind of fast, didn’t fully explain, and I still have more questions. So you go ask ChatGPT to explain it simpler or ask follow-up questions. Maybe you’re not as embarrassed to ask questions. You don’t feel like the person is in a rush to get out of there. So sometimes it’s not that the AI is smarter — it’s that it’s infinitely patient, an infinitely better communicator, and you don’t feel silly doing it. Those are other components of the doctor experience — bedside manner — that AI is better at.
Shaan: Yeah. I go to a concierge doctor now, and it’s not very expensive. But the reason I go there is most doctors have to see four patients an hour — 15 minutes each. I went to a doctor with an earache and he spent no time trying to figure it out. My concierge doctor, the average time is 45 minutes. So we can thoroughly walk through things. And if I can take all that information and then ping ChatGPT to further the conversation, that’s pretty brilliant.
I’m very eager to see what happens. People talk about AI being amazing for all kinds of stuff, but what they’re doing with medicine, drugs, cancer — that’s pretty astounding. I think that’s going to be the major breakthrough in the next couple of years.
Lazy Parenting with Gemini Live [01:32:00]
Sam: The other one — lazy-ass parenting. I’ll open up Gemini, and it has this camera mode.
Shaan: Why do you use different ones? You’ve said Grok and now Gemini, and we also refer to ChatGPT —
Sam: It’s like, you know, you go to different friends for different questions. You only ask me certain questions. Sometimes you go to Jack Smith, sometimes you go to Joe. You go to different people for different things. If you want something a little more real and objective, Grok is better. If you want code or creative writing, Claude is better. The catchall is ChatGPT. And Gemini has some advanced features.
Gemini has this thing where you just turn your camera on like FaceTime. I think it’s intended for like, “Show it your car and it tells you how to repair this.” But I pointed it at my kids and I’m like, “Hey, we’re playing charades. Guess what they’re doing.” And my kids get on the ground and start crawling and it’s like, “Hmm, seems to be a boy crawling. Maybe it’s a snake. Are you a worm?” And it tries to guess. They love it. I’m just chilling and letting them play with AI.
Another one I’ll do is: “Hey, I have a 5-year-old and a 4-year-old here and they want trivia questions. They like animals, they like Paw Patrol, they know a little about Pokémon but nothing too complicated. Ask them a bunch of questions. Cheer them on when they get it right. If they get it wrong, tell them the right answer. Keep track of the score. Here are their names. Go.”
That’s the prompt. And it plays trivia endlessly with my kids. They love it because it’s all audio — kids can do that, they don’t have to be on screens. And it’s like a tutor — an infinitely patient tutor. It’s not perfect — sometimes it starts and stops audio if you make any sound — but it’s already usable for us.
Shaan: I had no idea about Gemini Live.
Sam: Gemini is like that kid who comes back after summer break. Like, they’re kind of hot now, but you still have the old image of them. Their reputation hasn’t caught up to the reality yet. Gemini was basically out of the game — I was just doing ChatGPT and Grok. And then it changed. Like she got contacts, learned how to do her hair, watched a makeup tutorial, started rollerblading for surprisingly good cardio. And now Gemini can do things the other ones can’t do, but nobody’s really on it yet.
Shaan: Wait, so Gemini is hot now?
Sam: Gemini is hot now. Google’s hot.
Shaan: That’s hard for me to buy into.
Sam: Because you’re one of those jocks who’s stuck in seventh grade. You forgot what happened over seventh-grade summer. Yeah, you’re just stuck on ChatGPT.
Shaan: I don’t use Grok because I’m shocked when people say they use it. You go to twitter.com to use AI?
Sam: That’s the Twitter one? Yeah.
Shaan: Steph Smith just got a job at Grok with a Q. That’s unfortunate naming.
Sam: They should change their name.
Shaan: Yeah. And it’s AI — they’re making chips. They should change their name. Or at least the pronunciation. “Groke” or something.
Sam: Wait, what did I say?
Shaan: You’re saying “Grank.”
Sam: What is it?
Shaan: Grok. Like Crocs.
Sam: Like the shoes. Yeah. Like Crocs. Wait, and what’s the Twitter thing?
Shaan: That’s also — wait, no, that IS Grok. Oh, I thought they were different.
Sam: Nope. That guy’s a retired football player. Rob Gronkowski.
Overalls in Montana + Closing [01:41:00]
Shaan: Dude, I went to Montana to visit a friend last week and I wore overalls. I saw a photo of that and I just thought, holy shit, this guy’s got no limits. He’s just wearing overalls as standard wear.
Sam: It’s the best clothing. You can put your phone and wallet right there on the chest pocket. You’re holding kids, you have so many pockets, you have a vest right here. I love it. And my friend was like, “Oh, you got those? Did you think we’re all cowboys here?” She’s like, “You wore your overalls to Montana? Are you trying to make fun of us?” I was like, “What are you talking about? I’ve worn these for years.”
Shaan: By the way, I was very inspired by your Instagram post. You wrote something in the caption — you said, “From now on, I’m only taking photos that, if my kid looked at it 20 years from now, they’d be like, ‘My dad is pretty cool.’” That was great.
Sam: That’s because you have that photo of your father, right? When he was in his 30s and you’re a baby, and he’s doing something cool or wearing a cool shirt and you’re like, “Oh, wow. Dad was actually kind of sick.” You don’t see them like that anymore — they don’t have hair, they’re heavier, whatever. It lets you put a little respect on their name when you see it.
So I was smoking a cigar — which I never do — and someone had a camera. I used to hide it behind my back. And I’m like, “No, she’s going to be proud.” So I put it back in.
Shaan: Dude, you think smoking is going to be cool in 30 years? That’s going to be like you had a slave with you or something. It’s going to be crazy that you were just smoking with a baby on your shoulder.
Sam: Brother, have you seen the photo of those eight guys sitting on the beam off the Empire State Building during construction?
Shaan: That’s a great picture.
Sam: Those guys are crazy and dangerous, but they’re hard. That’s awesome. I will never be on a beam of the Empire State Building a thousand feet in the air. But at least I can smoke a cigar and look remotely cool. We should print this out. I want this framed.
Shaan: Three of them are wearing overalls very similar to the ones you were wearing.
Sam: What’s up? Same make and model. You just need one of those beret hats.
Shaan: And the courage to eat lunch a thousand feet above the ground. Even back then, co-workers were like, “Guys, what are you doing? There’s a cafeteria right here.”
Sam: Let’s just end this episode with that photo slowly fading away on our YouTube channel with some great music. Let’s leave people with a good feeling and a reminder of the men we once were.
Shaan: All right, that’s a wrap.