The Man Who Builds for the Decade Ahead | Founder of Google X, Waymo, and Udacity

28:02We're going to make the same mistakes. But if a self-driving car makes the mistake, it corrects the mistake. And all other cars will never make the same mistake again, ever, including all the unborn cars. So you have this incredible multiplier in the digital world that humans don't have, which is they can build our knowledge so much faster. So it's just a question of time when these curves intersect. And I always believed in the early 2020s would be the time when the car would be safer than humans. And now we'll see it. We have cars on the streets that are safer than human drivers. Do you think Waymo is going to be the only player?

28:37No. Do you think Tesla will come soon? There's a lot of activity in China. There will be mostly overlooked here between Baidu and Pony and Diddy and others. It is very, very impressive right now where there's Waymo equivalents driving across certain places in China. And yes, eventually I think even in the United States there's competition now. Tesla has made the choice to really focus on cameras only. As opposed to? Cameras plus LiDAR plus radar. And that's the choice they've made. As a result, I think they're having a slightly harder time getting the same reliability levels, but they're going to get there.

29:13Elon's very outspoken about that. That's good for him. He's very outspoken about the idea that Waymo's wrong and he is right about choosing just cameras. Because that's how humans drive. And why can't cars drive that way if humans drive that way? Look, I mean, I was taking the position that we should build technology that works. There were people at DARPA at some point that put a humanoid robot into a car that physically moved its artificial legs to push the gas pedal and the brake pedal. To me, that's not the correct solution. I think we should just interface with the car. In terms of sensors, the nice thing about lighters and radars is they give you direct range.

29:48So if something comes across you've never seen before, like this famed plastic bag, you can tell how far away it is, irrespective of what it is. And that gives you an extra layoff a curtain of safety that I think is important today. So, so far it looks like the only car that's truly driven 100 million miles without a driver inside has been Waymo. Yeah. And so you think, do you think he'll eventually add LiDAR? I think eventually the thesis is correct. People can drive with eyes only. So at some point machines will drive with eyes only. It's just going to take a little longer. That's fine. And his point is once we get to that point, it's going to be cheaper for us to do this than Waymo.

30:29Look, if you look at the cost of LiDAR, it's come down tremendously. You find world-class automotive LiDARs, long range for roughly 200 bucks in China. So if you argue about 200 bucks, yeah, it's real money. If you build cars, 200 bucks is real money. And Elon told me this, like a car, he told me at some point, is a box of stuff where no component costs more than$3. And to make a car cheaper is really, really hard. And I believe that's true. so adding 200 bucks or maybe 40 bucks for a radar is real money but if i can add it today and have a service today and drive safely would i do it probably yes especially if i can build a right sharing company like waymo where the car is being utilized for many many people i think if you look at the balance sheet of a waymo the lighter cost will be relatively small when armatized over many miles.

31:21So you think it's just a relatively moot point over the coming five years or so? It won't matter. Look, my personal belief is at this point you want to make the car as safe as humanly possible and you really can't risk running over a person. So as a result you do whatever it takes to make this thing safe and our attitude when I ran it was really to say look let's double down on safety over cost and that's not entirely correct. In the end, it's going to be a cost game. It's no question. But safety was more important than cost. So we were happy to use, back in the day, lighters that cost multiple hundreds of thousands of dollars.

32:02Are there Waymos in New York City now? I think they're trying to get there from what I understand. That seems insane. Why? I mean, like, when was the last time you were in New York? Look, here's the thing about New York. The cars drive, if you're in a taxi in New York, you're like inches from the car next to you. But it turns out a self-driving car can calculate the distance much, much, much better than people. Because it has these amazing lighters that just sit there and tell you up in millimeter accuracy how far this other thing is away. When it comes to precision, what would you win? Okay, but part of the driving demeanor of New York City is an aggressiveness.

32:45Correct. that defines how you move around in New York. Same in San Francisco, maybe not quite as extreme. Maybe not as extreme, but that might be true. But it's not as extreme. Like in New York, you know, you kind of got to run some red lights to get anywhere. Even if you're like turning left or whatever, it's just kind of how it is. People are walking around everywhere in a way that it's not the same in San Francisco. So we learned this the hard way early on in Google Chauffeur, which is on highways. If you stick in San Francisco to the posted highway speed, you become an obstacle. In many, many cases, people exceed the speed by 10 to 15 miles per hour.

33:23If you roll with the traffic, you're generally safer than if you slow everybody else down and force them to drive around you, okay? We also learned at four-way intersections early on, if you want to be polite and wait until it's your turn, you're never gonna go. You have to, when it's your turn, you have to inch forward aggressively into the intersection to signal to other drivers, now it's my turn. I'm going to take it, stay away. That's what's happening. You're constantly communicating with other traffic. Same is true for lane changes and so on. When you look at Waymo, over the years it has become much more quote-unquote aggressive to make it safer, right?

33:58So if you're overly careful, if you drive overly carefully, you don't add to public safety. Just slow everything down. So if you look at self-driving cars today, and many people remark this to me, I'm obviously not on the team anymore, but I believe that the Waymo car has become significantly more assertive. For sure. Probably got sick of getting cut off. Yeah. No, that's a good thing. I think traffic works in a certain way and you have to adjust to the way traffic works. When you go to Manhattan, you have to adjust more, obviously. If you go to Bangalore, you have to adjust because of different traffic rules.

34:32You go to Boston. So you don't think, it doesn't matter the city, is your point? I think you can adjust to any kind of traffic rules. Yeah. You've come full circle from there's no way this thing's going to work. Most extreme is go to India where the size of the vehicle determines the right of way. If you have the big bus, you go first. If you have the small vehicle, you go last. Yes, all that can be learned. Fascinating. Do you think you're, how old are your kids? 17, four and two. 17, four and two. Will your four-year-old own a car? My 17-year-old hasn't learned it yet. I don't know. He or she takes.

35:09He takes Uber or Waymo. Like there's no interest in driving your car. And if your son, if your 17 year old is taking Uber versus Waymo, as a parent, do you get more nervous because it's a human driving? No, no, no. I think Uber and Lyft are very, very safe. Professional drivers are much safer than casual drivers. I got more nervous when he goes with his drunk friends after a party. Do you feel like these advances in these language models is going to change, is going to meaningfully progress any of the behavior of these self-driving cars? Oh, it's already happening. If you look at the way data is being interpreted, machine learning has played a massive role for the last decade already, very, very deeply.

35:57Really understanding what's inside a camera image, what is the object. When it predicts something, looking at the 3D point cloud is one thing, but if you understand this is a dog, this is a human, this is a child, and maybe you understand even more, like these are two people who look at each other, then you're much better able to predict what's going to happen next. And so how do you think, from assuming the rate of progress maintains relatively the same, how do you think the behavior of these cars is going to improve, practically speaking, call it in the next five years? Well, I'm looking forward to the time when most cars are self-driving because then they can massively change everything.

36:37For example, highway lanes are six feet wide and cars are by a large part, three to four feet wide. And the space between cars is gargantous because human drivers are not good at reacting. If you had a precision driver on a highway, you can make the lanes much smaller and increase the capacity. Yeah. A full highway is 92 % empty. If you take a full highway and look at a picture from the helicopter and count the number of pixels. Is that true? Yes. of traffic a full highway that is like slowed down because it's so congested it's 92 % empty and you can do the test just go to Google Maps find a congested highway and count the number of pixels that have no car from above it's roughly 90 % give and take and so when Larry came to you and was like let's do flying cars is that how it happened?

37:32yeah he came and said let's do flying cars? that was another funny story good But Larry, I just admire Larry. He's so out of the box creative about things. Talk about someone who got rich and didn't take their foot off the gas pedal. That's true. And knowing Larry really well, he doesn't concern himself with how to spend his money in any meaningful way, to be honest. He really concerns himself how to make the world a better place to the present day. But back then, 2010-ish, he invited me for dinner and we had dinner again. We geeked out together. We love geeking out together. He's obviously much smarter than me, so I learned from him.

38:08He had a spreadsheet and said, look, Sebastian, I talked to Elon, I worked it out. I think batteries are good enough for us to just fly. And he had put together the very basics, like lift over drag, like the basic mechanics that govern energy and power consumption in flight, both for the vertical takeoff phase and the horizontal flight phase. And he also said, look, Sebastian, if you use batteries, you can use electric motors. Electric motors are quiet, whereas gas motors are noisy. And one of the reasons why you don't like helicopters is because they're so incredibly noisy. They're illegal for most part in like Napa and in San Francisco for private flight because they're noisy.

38:47And they also have a single point of failure. There's a bolt in the helicopter that keeps the blades together called the Jesus bolt because if it breaks, you'll die. Whereas if you build an electric vehicle, you could have like 10 motors, 20 motors, who cares? Massive redundancy. And I was sitting there thinking, okay, I learned my lesson. I'm not gonna go to Larry and say, it can't be done. I'm not that stupid. Let's make the same mistake twice again. So I said, okay, that's interesting. So I helped him start what ended up to be become Kitty Hawk and find the right initial CEO, a wonderful professor from Stanford and build the very first prototypes.

39:22And going back to your 92 % vacancy on the freeway. It's in my mind now, you're thinking about it. Well, I guess, were you solving the wrong problem by going to the air? No. If the ground was so inefficient or that could become so much more efficient? I'll give you any speed where the air is the right solution for everything in transportation. 60 % of the landmills in Los Angeles is dedicated to cars, more than half. It's the driveways, garages, freeways, all this stuff together. If you could invent a flying car that could take off a new rooftop, you have twice the real estate in Los Angeles. A bit different from San Francisco, but the density in Los Angeles is particularly bad for cars.

40:06We have built vehicles that can fly roughly a third of the energy of a Tesla per mile, a passenger mile, and twice the speed. And because they're going to go straight line as opposed to have to stop at intersections, they go roughly 10 times as fast. So imagine you come from San Francisco to the airport in like four minutes reliably. Now, people always say, how about a congestion on the ground? Isn't there congestion in the air? But the highway on the ground is basically one-dimensional, whereas the space in the air is three-dimensional. It's much, much, much, much bigger. If you don't like the traffic coming at you, you go 20, 30 feet higher and gone is the traffic.

40:43You can't do this on the ground. So, you can de-conflict by going higher and lower. And the sky is effectively empty. So, if you took all the traffic in the Bay Area that's on the ground right now into the air, you'd still be hundreds of yards away from the next traffic because the air is so much bigger. That makes sense. Good. But give me this look. Did it work? So we got into the prototype stage and we eventually sold the biggest part off to Boeing. We were able to document, we had like 27 ,000 test flights, able to truly document A, safety. The stuff was really safe. We never had an incident.

41:21We were able to document energy efficiency, quietness. We knew that if we flew a thousand feet above you, our noise level would be the same as in the library, like 38 decibels, which is like nothing. We were able to show the end-to-end thing that really worked in theory, but we weren't able to build a business. Did you get in any of the flying cars? Yes, I did. That's super fun. Where did you fly from? We flew over Lake Las Vegas. We had a phase where we tested it over water in case something goes wrong. You wouldn't immediately die. You just take a bath. But that was, yeah, it was a few years ago.

41:57It was super fun. The other thing that was really cool is, I mean, one version of it is autonomous. Like you just fly these things, fly them itself, which I think is a correct solution. But we also had a version that you can fly yourself. And it turns out flight today for a regular pilot is very complicated to fly an aircraft. But to fly like a drone is super easy. It takes you 10 minutes to learn how to fly a drone. So we built a plane where we trained 50 pilots on a single day to fly it. And through all these 50 people, random people, it took about one hour of training. And after one hour, you were a proficient pilot.

42:30How much money did Kitty Hawk raise? I can't talk about that. Maybe the question behind the question is, should it have just been done the same way that Waymo was off of the Google balance sheet so that you'd have more time to figure out the commercial model? Well, so Larry and I talked about it. And at the time, it felt like self-driving car was a stretch for Google. Flying cars felt like too much of a stretch. What year was that? 2010, about the same time. About the same time as? As a chauffeur. When Waymo was getting, Waymo, when self-driving was getting off the ground at Google. Yeah. No pun intended.

43:06And, but you were still at, you were still building the self-driving at that point? Correct, yeah. I, at the time, had many jobs. I was a professor at Stanford. I was head of Google X. I was running chauffeur with Chris Urmson. I was running an education company. And then you were finding the CEO? Yes. I often love to start things. If you want to criticize Sebastian, I'm probably better at zero to one person than one to 100 person. So for me, the moment when you invent something completely novel that's never been there before, it's never the most fun. And do you think that what enabled Waymo to ultimately go from interesting tech to what it is today was the longevity that it had within Google, basically?

43:50I'm sure they could have found outside funding. I mean, they're worth a lot. Now? Yeah. Yeah. I think the moment when it really flipped over to something that people took seriously was 2017 when Uber got to the race. You might remember Uwe Haggert, a good chunk of the Carnegie Mellon faculty and with the big fanfare, built his own team. And that's when people started realizing this stuff was real. They recruited a bunch of people from your team too, right? Only one person, Anthony. Okay. I won't go there because that's a touchy subject, I imagine. I was his mentor. No. Yeah, let's not go there.

44:30Sad. Look, I mean, overall, I'm actually quite happy. I think the technology is really having an impact right now. It's becoming real and it's unstoppable. And there'll be a time when we will find irresponsible in cities that humans drive. There'll be a time in New York, Manhattan, we decide it's a city for self-driving cars because it's so much safer. So you're saying, hey, maybe with long enough that the wound is healed. I generally believe that Silicon Valley is amazing because information does fairly freely from company to company. It's pretty common you work at Apple and then work at Google and work at Facebook.

45:07And in doing so, you might not take exact trade secrets, but you can take all your knowledge with you. And if we didn't have this incredible permeability in Silicon Valley, people would stay in their jobs. Like you go to the finance industry in Manhattan, it's different. If you work for a big hedge fund, you will not work for another hedge fund the next year. Or they'll sue you. Whereas in California, that's not possible. but in California we can actually change these jobs very frequently. And it's a big part of why Silicon Valley is so great. Because there's a common knowledge. Non-competes are not that enforceable in California, and therefore you allow great talent to cross-pollinate into other areas.

45:43Yeah. And that's a good thing. I mean, you look at the photographic suite on Android and iOS. The Android when it came off Google X, they know it really well, Mark Levoie built it. Like low-life photography and all that stuff. Today, iOS and Android are about the same, right? And it's great because all the smartphone users get to enjoy the best technology. When you were doing the professorship, the find a CEO for Kitty Hawk, the building self-driving, what am I missing? Udacity. Lots of other stuff. How did you manage the day? Look, I always believed that a way to organize an entity like Google Access to set very clear your goals, what you want to accomplish, find the best people in the world, and get out of the way.

46:31I never believed that we should micromanage. I never believed that we should be fuzzy about stuff. So let's take the self-driving car project. We decided that we want to build a self-driving car that could drive every street in California. It sounded crazy at the time, but Larry persuaded and convinced me it's the right objective. And credit to Larry and Sergey for doing this. And then we hired people like Chris Urmson and Dimitri Dolgov and other great people to build this with us. And in that process, I more and more got out of the way because it is a phenomenally great job. And they had clarity where they're going and ownership.

47:04That's the beauty of Silicon Valley. When you have clarity where you're going with mountain you're climbing and you have ownership, people are unbelievably amazing and they build amazing stuff. So they built a car that could within a year and a half drive a thousand reference miles that Larry and So I get handpicked in California as the hardest streets to drive, including Tiburon and down to Market Street and Lombard Street and Highway 1 to Los Angeles and around Lake Tahoe. And the car was just flawless doing it a year in, a year and a half in. So my life has always been to find and build amazing teams.

47:38And then when the teams work really well, which is a lot of hard work, basically get out of their way and be there when they need you, but don't micromanage. And in that way, you can multiply yourself. You can build multiple things. So Google X, we had a computational photography team. We had a blimp team that would put balloons into the stratosphere. We had a team building drones. We had a team, Google Brain, maybe the most impactful that really invented how to do indexing and YouTube and later became Gemini. We had a self-driving car team. We had teams that failed. We had a Google Glass team, which I loved, but Google Glass would end up being a failure.

48:17Maybe too early, I don't know. We had an architecture team that didn't work out. And all these things, what I really enjoyed and really learned from Larry and Sergey and from Eric Schmidt is it's really important to have an objective that is meaningful, that is far out there. Larry had long taught me that building something incremental is much harder than something radically new. He would say, look, building a pizza restaurant is harder than building a self-driving car. And it's true. Making a pizza restaurant, no one's going to give you credit. It's super hard. You have to put your retirement savings on the line and probably fail.

48:47whereas building a self-driving car, you get all the money in the world. It's easy in comparison. And if you build something radically different, it's often much easier than something incremental because you free your brain. You build something entirely new. So as long as we pick the right objectives, and Larry and Sergey really taught me greatly how to do this, we could build teams that could move the world and change the world. Were you managing all those projects at Google X? Me and I got asked through Tanner at some point to help me, and he was amazing. What about your family? Look, I have my second family now.

49:24I'm very happy. I spent a lot of time with them. I think my first wife was incredibly great in supporting me. I really, to the present day, are full of appreciation. There were many, many moments where all of a sudden I had to fly somewhere and all the plans changed and she carried the burden. Absolutely, yes. Meaning, expects you for dinner. Sorry, I got to go to... Yeah. Carnegie Mellon. But look, I mean, when you ask - Do you feel guilty about that? No guilt, never ever. Guilt and fear are just the wrong emotions in life, period. There's like one lesson to take away from this. Try a day without guilt and without fear and just try it out for one day and see how you feel and then take it for the rest of your life.

50:06No guilt, no fear. You can absolutely make mistakes and you can look back and say, I made a mistake. I would say about 80 % of my work is either mistakes or cleaning up past mistakes. And the way I come to this 80 % number is if you ask any CEO, now give them, like say, go to Brian Chesky, someone really amazing or Larry Page and say, give them what you know today. How much faster could you rebuild your company? Give them what you know today. Universally, they give you an estimate that's five times as fast. So it took them five years, it's going to be one year. It took them 25 years, it's going to be five years.

50:40And it's true. They could rebuild the same thing in a fifth of the time, which means four fifths, like four out of five days from Monday to Thursday, are wasted time, at least at the surface, where you did the wrong thing. Now, they're not wasted because you learned something interesting and you have to do experimentation to find the right answer. But my most recent company, we just now have an app we just launched, which is phenomenally successful so far. It's the fifth app we tried on the same topic. And four of them we had to throw away. so in in terms of efficiency it's it's actually more important to think about how to be efficient with time than to work like crazy if you can avoid if you can crank up your productivity from 20 to 40 in silicon valley um you're still 60 incorrect as opposed to 80 incorrect which is about the same number you doubled your productivity so for me it's more about can you be super wicked get smart about what to do and what not to do.

51:36And that's hard. I'm not good at this, but I feel this is more important than the workaholic side. No guilt. No, no. That's like guilt is what your grandmother, your Jewish grandmother invents to make you accountable to what they want you to do. So let me give you the counterpoint and just respond to it. Sure. A lot of people would say, well, my kids are growing up. I missed their first steps. I didn't hear their first words. I didn't get to see a lot of their youth. By the time they're 18 years old, they've already spent 98 % of the time that they'll ever spend with you or something. You're probably gone for a lot of those times.

52:17Good, here's what you should do. You made a mistake. You're looking back and said, if I could turn back the time, I would do it differently. If you have more children, don't make that mistake again. Learn from it. You're allowed to make a mistake, but only once. If you make it twice, you don't learn. But why guilt? Why does guilt help in that situation? Why does guilt feel better? My God, why should you punish yourself and pound yourself with guilt? But do you think you can choose to experience the emotion of guilt? Do you think that's a deliberate choice that people make when they decide? Do you think they decide, I want to feel bad about this?

52:49I think you can. You can change your life to be feeling this guilty. Absolutely, yes. Maybe not overnight, but once you realize it's just fundamentally a miserable force in life. like I have to find people say I'm so happy to feel guilty and it doesn't really change anything to the better like why don't you like if we all make mistakes every day I just told you 80 % of my work is mistakes like I make mistakes all the time um if I felt guilty about it and fearful I would I would stop sticking up my neck I would stop being innovative stop trying something new I would become a creature of habit like most people and just do the same thing over again but it feels safe.

53:27So that kind of feeling to me takes away from the beauty of human nature and what we can accomplish. Super interesting. This app that you built, when I go on the website to see what you're working on, we launched it this week. The website is shopongold.com. I'll tell you what the ambition is. It's serious. Yeah, so we have the nice thing. We just, we had it on the marketplace for a month and a half with people that just came and didn't pay random testers. And our seven day retention is now 95%. Our one day retention is roughly 60%. The insight is the following. When you go to your phone right now, I have about 15 different apps that relate to financial services.

54:12I have about eight to 10 different apps that relate to communications, like messaging apps, my various medical apps and so on, emails, SMS. I have about 25 travel apps. And when I navigate my phone, I find myself navigating the world's org chart. And I'm just one person. I don't need like 15 different messaging apps, Signal, and WhatsApp, and all that stuff, if I want to talk to the same people. Why isn't it just one messaging app? So for me, what Argentic AI would do is, it'll unify the world to give you one access point, irrespective of the world's org chart. And we're testing this in retail, in online purchasing, in shopping, specifically fashion shopping.

54:56So in fashion shopping, we have now under contract roughly 300 different vendors that provide things to buy, like skirts and dresses and all that kind of stuff. And we go out with an Argentine AI every night and find the perfect things for you. And we give you an hour every day called the golden hour where you can look at this stuff and buy it, typically 70%, 80 % off, 90 % off, chosen just for you. and when you like something, you can click it with a single button and we do all the work for you. We go to this website, we register you, whatever, and put in your address and your credit card number and all that stuff and this stuff comes to your house.

55:33You have no clue where it's coming from, which means you can now command 200 stores, 300 stores with a single app as opposed to having 200 apps on your phone or whatever, 200 websites, a single app. And then by us having the genetic error finding the perfect thing for you, we get roughly 86 ,000 new deals every night, which is a price drop or a sale, which happens to be exactly one per second. You as a human being, wouldn't even be able to monitor all this. You wouldn't even find it. We find this stuff for you and we serve it to you every single day fresh. Was this the original idea? No, the original idea was much more abstract.

56:12My God, we made every mistake on another planet. It was so bad that my entire founding team has left by now. Come on. Yes. I'm the last one standing. Yes. I'm the last one standing. Yes. I hired a whole bunch of people straight out of college. And after roughly a year, they were so, oh my God, it's not working. So we tried, my God, we tried, in the beginning, we tried in large language model. We felt like GDP is so amazing. We're going to build a chat TPD for shopping. And you could talk to it and say, look, I'm going to go to Hawaii and I want to have a dress for this kind of weather. And it would come something with you.

56:45It turns out that's blah. People don't want to talk about stuff. They want to see stuff. So you built a site where you can see lots of stuff. And that ended up to be just about a clone of Google Shopping. So when we were done with it, it was really, really good. People said, meh, it's like Google Shopping. So it took many iterations. But what we always believed is that a genetic guy is amazing. We should really be able to do all the hard work for you you don't want to do. So search the stuff, buy the stuff for you. We believe the world is dynamic. and we now know with the many, many sites we scrape every night that there's roughly a million new things every night popping up on the internet for shopping.

57:22That was great. We also believe that the industry itself is inefficient. If you're in the retail industry, you might have large margins. If you buy something, say, at H &M for 100 bucks, it might only be 10 bucks to make it. But one of the three things is never sold. So one of the three things either is being destroyed or being shipped to Africa, what have you. So the industry also has a problem, which is it wants to match its stuff to its customers. We figured if you build a really good AI that matches things to people, we can make everybody happy. We can make the industry happy, make customers happy.

57:55And that's where we are today. We have incredible retention right now. Your founding team quit? Yeah. And then what'd you do? I hired a new team. My God. That's easy. You didn't want to close down shop? Look, I mean, one of the things I learned in my life is that grit is really important to success. If you give up at the first intersection, when first speed bump, you're never going to make it. All the good entrepreneurs have grit. So you had co-founders or a founding team? I had a founding team of fresh out of college. Why did you decide to go fresh out of college? At the time, I just had a great connection to places like Princeton and MIT and Harvard on the East Coast and I found a lot of young kids wanted to be on the West Coast.

58:40It was just when the AI stuff took off like crazy. I'd raised some money from some good friends and I decided, why not? Give it a try. And they were amazing people. They just didn't have the grit to stick around. And I don't blame them for that. Look, innovation is hard. for the longest time when things are not working, you ought to work. You build something really hard that's not working and you throw it away. That's not an easy thing to do. Did you rehire another batch of younger folks. Yeah, they're amazing. Why don't you hire people that, you obviously know people that have been doing this for a long time in industry.

59:16You could call anybody. Look, I can tell you, in my life, all the great stuff from self-driving cars were done by people who weren't experts. And I give you the funny episode of me being an expert and how stupid the others. But all the people on the self-driving car team at Google, none of them were automotive engineers. They were all never built a company before. They were all software engineers. And every time we hired an automotive person, we ended up firing them because they were so set in their way of thinking. They couldn't understand things would be different. I'm sure if you go to, I don't know, Tesla and ask about battery technology, they have a similar experience.

59:56The fresh mind sometimes brings so much amazing energy and novelty to it that you really want to have solving a hard problem. and how does this fit into your framework of like doing only do things that change the world well if we can change with agentic ai the way people access the internet and build agentic ai does everything for you don't want to do we're massively changing the world i believe we'll be building the lead to the next cell phone operating system um which i believe will not be full of lots and lots and lots and lots of apps and i know i'm not the only one believing this It'll free our time.

1:00:35I mean, you can see it in many, many places where we waste time. Like, let's take a tax return. I believe you're going to build an app, not me, but someone will, where with a single button click, your entire tax return is being done automatically. So you don't do it to taxes anymore, which is something people really hate doing. There's so many places where we have to do stuff we don't like doing. Shopping is one example, right? So shopping is a lot of work. Young people spend an average of six hours a week shopping. It might also be fun work, but it's more fun when you get to see stuff you actually really like.

1:01:08And it's a good starting point. It's not regulated. It's not like the FAA is authority over it. It's not like self-driving flying cars. Yeah, flying cars or medical stuff and so on. Are you building it here? Yeah, in the Presidio. Oh, cool. How big is the team now? It's about 15 people. All engineers? Mostly engineers. We have some very good business people too. Does that feel like the right team size for a company that continues to search for, or maybe most recently, maybe finding product market fit? Yeah. So we had an inflection point, but now we're just hiring our first recruiter. To me, 10 people is plenty to search for product market fit, but it's not good enough to build a company.

1:01:44So now we've run - 10 is plenty though, isn't it? Yeah. Especially right now with these tools. Yeah. It's amazing. 10 is plenty. 10 is plenty. Yes. And the tools have become amazing. but now I mean as we go into a growth phase where we have to hack growth and get influencers on board and go into TikTok and all these places we want to be more than 10 people does this feel like are you the CEO yes are you going to remain the CEO of course you are yes well in some cases you've hired a CEO yeah with very mixed success yeah sometimes I hired CEOs and then two or three years later we let them go it happens I'm of course never going to talk negatively about people I've worked with.

1:02:23But yeah, and look, maybe I should leave at some point. But at this point, I'm very convinced that having me do this is not just fun. It's also good for the company. Are you writing code? Yeah, on a daily basis. Oh my God. I had a big review session where the entire company reviewed my code this afternoon. They're doing it this afternoon? Yeah. Let's see how that goes. Yeah. I mean, I do write code. I love data. what model are you guys using on the back end uh view slot clip um we play with gemini and others um and i'm doing my own neural network a small one right now that i train myself you train it yourself yeah i think it's not that hard i don't know how you have time for all this it's super fun look life is amazing you're also in good shape too thank you how do you do all this i have no time to eat

1:03:19no I spend I spend a lot of time with my kids when you say guilt and kids I love the time when they're like two three four years old it's such an amazing time and they're very labor intense but I spend every weekend every week with my kids I usually drive them to school I pick them up often I my my four-year-old son is trying to learn how to play chess so you're playing chess every day that's awesome yeah how much are you working um there's a lot of there's a lot of um hard work porn in silicon valley right now i i run two companies right now i also run a social media company and i um people are sometimes confused and say sebastian how does a half-time ceo work out for us and it's fair to ask the question it's a good question to ask but the reality is me full-time is probably not as good as me part-time because then I'll be more of a pain in the ass for my people.

1:04:14The right balance really is to get a team together that feels extremely empowered, where the direction is clear, where ownership is clear, where people can do their thing. And I don't always succeed in this. And I mean, I sit there and ask myself, on a scale of one or 10, how well does my company run? And I say my shopping company is probably running at seven right now. It's not a 10. There's stuff that I would like to improve. And it's typically me. can improve it. But when it's a 10 or a nine, it's unbelievably amazing how productive people become and how much they love it. But you have to make an environment where people feel their work is being appreciated and they clearly own what they own, that other people don't take away their ownership.

1:04:51It's really important. And it's meaningful. There's predictability in their life. You can't just thrash them every day with new ideas. You have to have a plan and a method for having people participate in decision-making and really make joint plans. So a lot of stuff I learned about people. It's actually interesting. I'm an engineer at heart, right? And I used to engineer machines. And you wake up in the night and you tell your computer to calculate the square root of two and it does it in the middle of the night. Never complaints, right? And if it doesn't do the right thing, it's your fault, not the computer's fault.

1:05:21It's very binary. And at some point I realized that's not good enough. I want to program organizations. I want to figure out how to make an organization. And people are so different. Like people, you can't just tell them what to do. It just doesn't work. People have to feel they want to do it. So you have to kind of evoke a passion in the people. They have to own it. If you just tell them, they're not going to do it. They do it maybe for a week. And no amount of money can change this. You can give them a really big check, they're still not going to do it. So how do you make an organization where people feel ownership and feel clarity?

1:05:51That's really the engineering that I believe I've mastered in my life. And it's super fun to do. Do you think people like to be engineered in that way though? We all are being engineered every day. like we talked about money in the beginning. It's not engineering to get people to say you want to get rich. But I do believe people want to have a, I actually believe people want to have a fulfilled life and people feel that they can make the world a better place for other people and that time is meaningfully spent. They're generally happy. They're generally happy people. And if you can create this environment where that's the case, then they're generally happy people.

1:06:26I appreciate you doing this. Thank you. It's a real thrill. I'm sure you're doing the same thing. What? make people feel good about themselves, I can tell. You're good at it. You're very good at it. I think when people know that you care, I think that's more empowering than a lot of things. Yeah, I agree. And there's a lot of ways to show that you care. I agree. And some of those levers might be financial. Some of those levers might be autonomy. me. Some of those levers might be title. Some of those levers might be just expressing that you care. And some of those just might be that you're interested in that person.

1:07:13Yeah. And you have to be honest about it. You can't fake it. It has to be sincere. Correct. It has to be sincere. Yeah. And I think in order to do that, it can't feel like work to you. It has to feel like play. Correct. And so, I think if you're in environments where you can genuinely, it's kind of second nature for you to do those things, it shows, I mean, people feel that. Yeah, and I can tell you have the same skillset and the same experience, it was amazing. It's rare, people often, especially young people often think they should just be pushing really hard and work themselves to death and scream at people.

1:07:50And there's certainly instances where screaming has made a positive difference. But to me, screaming, first of all, I don't like screaming. But it's also, it's not very sustainable. You can do this once or twice, but at some point, people are just very upset, and then they leave. I think in today's environment, with the younger generation, especially those that are kind of working for you, just doesn't work. Just doesn't work. That's just not how they operate. And that's great. It's just not how they operate. That's great. That makes Silicon Valley so special. Such a beautiful environment. If I was in politics or I was in banking, it would be so different.

1:08:25But that's something we have in Silicon Valley that's really amazing, that people care. And here's a good thing about Silicon Valley. When I go to the East Coast or to Germany, where I'm from, I always get the feeling that the people there have like a zero-sum picture in their mind. Like if I get promoted, it's because you didn't get promoted. If I make money, then I'm stealing it from the poor people, which is a very German attitude. And here in Silicon Valley, we do something amazing and everybody benefits. everybody gets the benefits of it. Maybe some get more and others get less, but by and large, everybody benefits.

1:08:56So we can create value out of nothing, whatever your value is. And that is such a beautiful world to be in that you can just do this. You can go into a company and say, let's build a company together and we all benefit. You know, everybody believes it. And it's true. Like it's historically always been true. And that's just an amazing inflection point in history. If we lived 500 years ago in farming, then I promise you the way the world was the moment you got born was the same the moment you died. You couldn't really create value in any meaningful way. Even the railroad barons had to work really, really hard.

1:09:28Today, a team of 50 people can build WhatsApp and have a worldwide impact that's worth billions of dollars. Do you think that's why, let's just say in Germany, innovation is so lacking relative to somewhere like here? Germany is sad for me. And I spent a lot of time there with politicians and I tried to really encourage them. um germany does have i believe a zero-sum perspective on things and obviously i have a very thorny history um which i acknowledge but the culture is is such that if you stick out your neck and try to do something crazy that people play racquetball with you they hit you over the head um the place where this works better than anywhere else is silicon valley like you know with my friends we only talk about ideas what to build we never talk about wealth or anything like this um the culture is just not permissive for this yet in germany it's getting better there's no founders but even the financing in germany is really really harsh on people it's very hard to to raise money and not be micromanaged um there's a beautiful book by sebastian malavi about the power law in silicon valley where he lays out that as an investor in silicon valley if you lose your money you lose say a million bucks but if you win you make a billion dollars So there could be 100x, 1 ,000x when you win compared to what you lose.

1:10:46And that asymmetry suggests you should just invest like crazy, which is what we've done in Silicon Valley. And it's true. When you take these companies where you invest a million dollars and you make$100 million back, you can easily invest in 50 other companies and lose all your money and still be in the positive. Great investors like Planet Programme have recognized this long ago and used it as a way to really promote the world and make the world a better place. And it's really important. And that DNA does not exist yet in many places of the world. People have not seen that yet. And there's many reasons for it.

1:11:19In the end of the day, it's culture. It's not just one entity. It's not just the investors or politicians or journalists or entrepreneurs. It's the entirety is in a different national equilibrium as we are in Silicon Valley. As John Doerr always says, you can only lose one time your money. Correct. John saw this long before. I was about to say I was born, but I'm not that young. That's right. Thank you. John is amazing. Thanks for having me. Are you hiring for your new gig? What's it called? So the company is called Sage AI Labs. The website is called shopongold.com. And yes, we're hiring 100%.

1:11:52For what? For anything. Okay. But mostly engineering. In Presidio? In the Presidio. We're going to move probably to the south of the market here. Oh, really? Yeah. Stay in the Presidio. No, it's hard to reach for most people. If coming from the South Bay. Yeah. I think here we have the Caltrain and so on. We're going to move at some point. All right. And when you hear the word grit, I ask every guest, what do you think of? Grit is the absolute, single, most important thing to have in life. If people ask me, how do you raise kids today? What are you gonna recommend? I would say curiosity and grit are the things that matter the most.

1:12:26If you are the person, and I've been guilty of this myself, it jumps too fast, you're not gonna succeed being a great entrepreneur. Look, every problem can be solved. Every problem that people walk on is what they walk on. Every mountain can be moved. And sometimes it's easy and sometimes it's harder. You always look at successful companies in awe and say, oh my God, these guys were so lucky and they're so great. And you forget the many, many, many failed companies. But if you give up to early, no matter what it is, then history will punish you. Thank you. I appreciate you. That's it for now. If you liked the episode, please leave us a review or go back into the archives where we've done more than 200 episodes with some fantastic folks.

1:13:10This podcast is a Kleiner Perkins production, and I'm Juven. Thanks for listening.