四人组纳帕谷之行

Speaker 1 (00:00:45): Everybody give it up for the dictator himself, Chamath Palihapitiya. The Rainman, yeah, definitely David Sacks. And your Sultan of Science, David Friedberg. Chamath Palihapitiya (00:00:57): I love how it's freezing, and there's one heat lamp. One heat lamp. We're all just I mean, these are literally, like, $80 for the night. Jason Calacanis (00:01:07): Yeah. They cost about $50 a night to rent, so I just thought $250 was enough. Yeah. That was fun. No. No. I mean, this is a super expensive place, especially since this is what happens. This is my life. Speaker 4 (00:01:18): Oh my gosh. Jason Calacanis (00:01:19): I come in. And they're like, J Cal. I'm like, yeah. It's like, Chamath's been here since 4:30. I'm like, yeah. That's fine. That's great. It's, you know, you got here early. And they're like, He's been ordering wine on your tab. I'm like, oh, no. So the event just flipped from being in the black speaking David Friedberg (00:01:35): in the red. Jason Calacanis (00:01:36): A lot of people had questions for us, so we thought we would do, like, a little Q&A, but I thought I would start with maybe asking the besties what has it been like, I think, having this podcast get so big and the scrutiny it's under, but the excitement and just how it's affected your life and maybe how you operate in the world? David Sacks (00:01:57): I think I may have told this story on the pod, but I'll tell it again because I do like the story. So I think I was visiting Chamath in Milan, I think it was either last summer or Speaker 1 (00:02:10): two years ago. And David Sacks (00:02:12): we were literally just walking down the street in Milan. And somebody looked at Chamath and stops him and it's like it's a tourist from Australia Jason Calacanis (00:02:23): Yeah. David Sacks (00:02:23): Who's a fan of the pod. Yeah. And he wants to take selfies. I mean, he definitely, you know, Chamath was definitely his favorite bestie, and it was like this excited conversation and that all happened, and then he, you know, moved on. And Chamath turned to me and said, oh, by the way, just some context here is that these two we're we're in like a big fight. And there's some question about whether the pod Chamath Palihapitiya (00:02:46): would yeah. David Sacks (00:02:47): Whether the pod would survive. And I think we were just talking about that as this guy came up to us. In any event, so he comes up, takes a selfie, and then leaves and Chamath turns to me and said, these 2 idiots better figure things out because I like being famous. Speaker 7 (00:03:05): I'd love to hear you guys. Talk more about in the current state of AI and particular software companies. I know Chamath, you got 80.90 going love to hear about the the lower cost of actually engineering software and building companies versus actually genai doing replacing human judgment, replacing judgment, self-driving the software itself. And those two dimensions feel like they're not exclusive to each other, but they are two different axes. Love to hear kind of current thoughts on the return potential from investment standpoint. David Friedberg (00:03:38): I think that Chamath Palihapitiya (00:03:41): the entire software stack is gonna get rebuilt. And I think it's gonna be a bunch of 10, and 20, and 30 person companies that do it. And I think the core artifact of a new company today are two decisions. The first decision is that you must demand that the people that join you use every available tool at their disposal, and you measure them against the counterfactual where that counterfactual is what your productivity was at a traditional company, and you need to be 100 to 200% more productive. So if an engineer is writing, you know, x amount of features or x amount of code, they need to be at a 2 or 3 x. And if you do that in an existing company, all I've seen is organ rejection. So that's the first decision. So the people that join you, you must expect that they do that, which means that you also have to deprive them of any help along the way. No administrative help. No HR help. No finance help. All of that stuff needs to be a workflow or a bot or some form of AI logic. So that's one critical decision. And I think the second is that the document that matters more than ever is that PRD that defines the v0 MVP of the product because I don't think you should be writing much code. I think what you really need to be doing is defining in excruciating detail what the feature set is, what the parameters are, what the guardrails of how features should behave in certain boundary conditions, fed into an LLM that then spits out the code that then compiles, and then you have a working product. And that is a 10, 20, 30 percent effort that can probably topple giants. It'll take a couple of years for us to get to that place, but in that world, there is no room for $500 million into a SaaS company. It doesn't make any sense. And the reason is the 10- and 20-person company can price that thing at just a fraction of what the incumbent charges because the incumbent has an OpEx load that's not correlated to features. It's correlated to the fact that you have fifty thousand people across fifty offices all over the world with all this ridiculous infrastructure that all of a sudden becomes obsolete. Chamath Palihapitiya (00:06:01): So I think there's two models. One is more of the lightweight approach. So like Sacks and Jake helped both seed a lot of stuff, get it off the ground, spin it up, and then raise money. Like he just did with Glue, which is like I'm gonna build something. Get it to a place where the rest of us can kind of pile on. But the other version is much more distributed, and I'll say a dirty word here, but I think it matters like it's like this Web3 model does actually apply here where there are some really interesting early adventures in crowdsourcing technical completion. I would encourage you to look at a project that I'm starting to focus a little bit on called Bittensor. And I was like, what is this thing? And it's basically a thing where you go and solve technical problems and produce features that people require, that need, and you get paid in this underlying currency. That it's actually distributed form of venture that actually makes sense. Like a bounty? Essentially. Yeah. We're like, you know, fifteen companies that need a replacement to Salesforce just say, here's the bounty, and here's the feature set, and it's a crowdsourced or a community-sourced PRD, and then some LLM spits out the code. Like, that is this crazy new world that we're going to do. So it's all about speed of execution, I think. Jason Calacanis (00:07:16): Sacks, do you have any thoughts on it and what you're seeing, you know, in the field right now? David Sacks (00:07:20): Yeah. I mean, I think in having conversations with enterprises and businesses, what they basically want is to be able to throw all of their data somehow into an LLM and be able to ask it questions. So they've seen ChatGPT, and they basically want to be able to do that, but with their own enterprise data. And it sounds easy, but it's actually, like, pretty hard to make that happen. One problem is that the LLMs don't have that big, what's called a context window, which is kind of like you could it's almost like active memory or something like that. And we're in the early days of it's like in the early days of computers where the, you know, you had to worry about your RAM. Was it like 8 megabytes, 16, then 32, then 64. Remember that? Jason Calacanis (00:08:04): You're gonna get some David Sacks (00:08:05): We had swapping out. Yeah. 3 1/2-inch disks or whatever. Eventually, just people stopped talking about RAM because it went away as a constraint. Yeah. But we're still in the days of you have to be selective about what you feed into the model because it just can only absorb so much information. So if you give it all of your enterprise data, how does it even know where to look? And so you've gotta figure out how to optimize it so that you can give the the AI model, the right chunks of data to give you the right answers. And that's actually, it's a complicated problem right now. And and, we're dealing with it at Glue. We're, but we just invested in a company. A startup called RAGI, which is basically RAG as a service. Does retrieve augmented generation as a as a service. It's to basically help solve this problem. So I think that that's kind of the stage that we're at is everyone kinda knows where they wanna get to and we're just dealing with some of the limitations so that the model can work you know, effectively. And, you know, what you see right now are like glimpses of, you know, genius or greatness. Like, if you feed the exact right chunks into the model, you'll get, like, answers that can blow you away. But then sometimes you'll just get an answer that kinda isn't that good. And, again, the reason is because the model didn't really know where to look for the answer. So, you know, right now, it's like, again, glimpses of something amazing, and we just have to kinda get to more predictability around that. And there's a lot of effort that's going into solving some of these sorts of issues. It's not just LLMs. It's also Yeah. Although it is at the LLM level, it's also the infrastructure around it. You know? Jason Calacanis (00:09:40): Yeah. And to put the two thoughts together, I think the people who embrace this are becoming bionic at work. I mean, we had many challenges in podcast production around things like transcripts or doing show notes and learning, hey, about this person who's a guest, they were on these three podcasts. Now we have producers at our company who will take those three podcast interviews, download the MP3, put them into a, you know, a chat GPT, and say, what are the key points in here? What are the key time stamps? And even what questions would you ask? Or what are relevant topics this person's talking about? And these data sets are disparate and they're not connected. So you can't say, well, what's David Sacks talking about if you're gonna have one of your podcasts on Twitter because there's no way to get that information. I could tell you what the topics are. It's pretty cool. Anyway, and those individuals are becoming so valuable in organizations, and it's allowing people to do so much work so fast and then unlocking things that were just not possible. And that to me is very obvious inside the startups we invest in because when they're resource constrained, they look for the fastest, cheapest way to do things. And if you watch those startups, they're the same ones who when other people were racking servers, they were using AWS. They were using cloud computing, and they figured it out first. And then they built businesses that you may have heard of, like Dropbox, YouTube. That whole wave was a group of people who just fundamentally said that the paradigm of cloud computing is gonna work, even though it's not perfect right now, even though it's slow or expensive or breaks, whatever. I'm just gonna go all in on that. And Dropbox made this critical error. I don't know if you remember that, Sacks, where they stored up their own data centers instead of, you know, they were just scared to have their data at, I guess, AWS or whatever it was. And so it's gonna be unbelievable how many businesses become viable with just three, four, five employees. Jason Calacanis (00:11:33): And that's what I'm seeing on the field. Friedberg, any thoughts on AI right now? David Friedberg (00:11:38): I just I think that there's a chance that SaaS gets kind of obviated. I've shared this in the past. I think that there's a chance that SaaS looks like this kind of temporary phenomenon that occurred between the ubiquity of the internet and the prevalence of AI for writing software. If you've worked with any enterprise, non-traditional enterprise, from a traditional industry, meaning they didn't grow up in software and they're not software native. They have IT departments, and those IT departments are really good at procurement of software. And that's really been their primary function, but they all try to hire software engineers and they generally suck. But they understand the needs of the business. And I think that if they can have software written for them, at basically a cost of zero to improve their workplace productivity, which was the original thing that was enabled by software in the first place, it starts to really change how folks are sourcing and using software, that it can actually be written for them in real time by software. And so I think that the tooling and the capabilities to enable the enterprise to write their own custom workplace productivity tools and their own custom workplace software, whether that's customer-facing or internal-facing, really becomes kind of the standard going forward. And you know, you see this, what I think Doug Leone called the greatest business model in human history, which is SaaS, getting kind of blown up. In that new landscape. So I think there's a chance that that happens, non-zero chance. And I've just seen this in a bunch of settings where folks are having software written for them by the software. And they kind of just really state clearly, here's the layout, here's what I want it to do. And with a couple of rounds of iteration, they can get it working pretty well, and they don't need to go pay some third party a per-seat license fee per year to use it. Speaker 4 (00:13:28): I have heard about your poker love through the podcast. I've been listening for a few years, and it was really fun to experience it with you last night. Could you comment a little bit on, like, how you think about your utility function personally? Is it networking, strategy training? Is it fun, like just comment a little bit on how you Jason Calacanis (00:13:46): all think. Why we play poker? There is networking that has occurred at it, but I think also, you know, you made a good comment about it, or maybe Bobby Baldwin said it. He never saw or it was, you know, a poker player, like, get old and lose their mental faculties. Like, they Chamath Palihapitiya (00:14:03): No. Bobby Baldwin ran the CityCenter in Aria for a long time. He plays in our game, and he made this comment. He's like, Have you ever seen a poker player get Alzheimer's or dementia? Jason Calacanis (00:14:14): Right. That was it. Chamath Palihapitiya (00:14:15): Saying that about Doyle Brunson, because when he died Jason Calacanis (00:14:18): sharp. Chamath Palihapitiya (00:14:18): Super sharp. Sharp as a tack. His body failed before his mind failed. Yeah. And it's true. It's like it's one of these unique games where you can really stay mentally sharp as you get old. Jason Calacanis (00:14:26): Yeah. Absolutely. Sacks, you still love it? Yeah. Definitely. David Sacks (00:14:32): Yeah. Look, I like it for the same reason that you guys like it. There was a phase early on where I would play in the World Series of Poker. I would go to that. And then I was just like, you know, I actually don't like sitting there for three days playing with strangers. It's actually not that fun to do that. And the thing that's fun is just playing with your friends. So now just playing fun games. David Friedberg (00:14:55): I first started playing poker. Did I ever tell you this? I worked at a pool hall in upstate New York. I was sixteen years old. I got paid $4.25 an hour to work at this pool hall. I cleaned toilets and scrubbed down the tables, and the guy sat at the pay phone all day was the bookie. For upstate New York, and he would take all of the bets on the pay phone and write them down. And one day, after he got to know me for a couple months, he said, come and play poker with us at our home game, and they played limit hold 'em, like 1-2 limit, 50¢-$1.12. And these fuckers took all my money. And I was sixteen going to college, like, fucking cold. I had to walk through the snow. I just couldn't afford boots or a jacket. It was really shitty. And they took all my money. Jason Calacanis (00:15:36): Yeah. David Friedberg (00:15:36): And I had to call my mom, and I lied Jason Calacanis (00:15:37): to her. And I'm like, mom, David Friedberg (00:15:38): I lost my jacket. Can you, like, loan me 150 bucks? Western Union it so I could buy a fucking jacket. So that summer, I bought all the books from the back of the poker magazines to learn how to play poker strategy, and it was all Hold'em at that time. And then my freshman year of college at Cal, I made $10 that summer playing limit Hold'em at the Oaks Card Club, playing tournaments in 6-12. And after that, I learned so much about investing in life and, like, perseverance through ups and downs because you sit there, you grind it out as long as you have a positive EV decision, you know, you made the right decision. Over time, the money will come to you. You don't need to, like, win every fucking hand. And to learn that at that age, I think, was really influential to me in entrepreneurship and investing and in decision making later on in life. It was very important to me. And so when I sold my company in 2013, I was introduced to Chamath and invited to come in and play in the home game. And for me, it was like very kind of nostalgic because I hadn't played really much since 2001, 2002. What's the most Jason Calacanis (00:16:39): impactful advice you've been given and how has it shaped your careers? Early on, I got you a good piece of advice. One was from Mike Savino who's here somewhere. He's speaking tomorrow. There it is. And he said, the piece of advice his dad gave him was look to the left and come in an hour before that guy and then look to the right and come in and stay an hour later than that guy. Just basic hard work stuff, and I had gotten that from my mom as well who worked three or four jobs to put us through school. And my dad who worked really hard. And so the hard work ethic, I think, was the key to a lot of my success because I just decided I'm just gonna outwork everybody because I was coming not from Harvard or Stanford, but from Bay Ridge, Brooklyn, and, you know, going to night school at Fordham. So I had to use, so there were no connections. There was no network. I just had to literally kill what I ate. And then after I sold my first company, I was at TED at the Billionaires' Dinner with my book agent, John Brockman. And I still had a chip on my shoulder, but I had sold my company, and he said, Hey, schmuck. You made it. You don't need to stop fighting with everybody. And I was like, okay. I'll stop fighting with everybody. I still Chamath Palihapitiya (00:17:52): You took that advice? Jason Calacanis (00:17:53): I took it sometimes. I put the coin. When did you do that? David Friedberg (00:18:00): Yeah, Jason Calacanis (00:18:00): I just got it here. You just right now. But it was actually because I had to fight for everything, I think I just kept fighting to try to get to the next, you know, level. And, you know, at a certain point, I decided I would be like super magnanimous and just be helpful in elder statesman-like and it that's why I'm not perfect, but, you know what? Well, I mean, when you don't see me working, you don't see me working with startups, you know, like, I am tireless in my ability with patience from them and supportive. And I just Chamath Palihapitiya (00:18:31): and I Jason Calacanis (00:18:31): and I try to do that with my friends as well. It's just, you know, what's the point of being successful if you can't support the people around you who you love and try to pay for it? So those are just David Friedberg (00:18:38): You are, like, the Chamath Palihapitiya (00:18:40): the most reliable wingman of all of us. Jason Calacanis (00:18:42): Oh, thank you. You know, valuing your friends and your family is, like, just such critically important advice because at the end of the day, all you have is your memories, and you make them with your friends and your family. And so I am on a mission to make great memories with my friends and family as many as possible. And actually, this is part of it. Like, you all being here and just that's why I always tell you, like, make a couple friends. It's like, it's hard to make friends, especially as you get older, and people kinda tighten up their circles. Like, make a couple of new friends, you know, and go do things together. Anyway, some philosophy. I don't know. David Friedberg (00:19:17): Advice that you got? David Sacks (00:19:19): The best advice, career advice I got was from Peter Thiel, who advised me not to go to law school. And, unfortunately, I didn't listen to him. I went anyway. But it ended up not mattering because I ended up joining PayPal after I graduated from law school. So I wish I had a better story. The best career advice I've actually heard for Silicon Valley was the advice that Eric Schmidt gave Sheryl Sandberg when she joined Google. Which is when you get invited to take a seat on a rocket ship, don't ask which seat. Yeah. Just get on board. I actually think that is perennially great advice for anyone in Silicon Valley. You know, there aren't that many of these rocket ships, so if you get a chance to be on one, you should just take it and worry about the titles and all that David Friedberg (00:20:03): kind of stuff later. Chamath Palihapitiya (00:20:05): Yeah. The details later. David Friedberg (00:20:05): I think, like, the biggest piece of advice that I still struggle to take every day, every time I have, though, it's had profound impact on my life is to focus. I think the bigger a portfolio you develop, the less alpha there is. You try and minimize beta, but you take all the alpha away if you do that. In 2009, I focused my company on the agriculture market, which was crazy. We were in, like, seven verticals and doing all sorts of stuff. When I focused on, again, I went deep and I'd, like, went on a long hike in Iceland, and I came up with this crazy idea for the product for the app market and came back. We launched it, and we did, like, $30 million in sales that year. And that made a huge impact in my life and the trajectory of the business changed. And every time I focused and avoided distractions, it's made a huge impact. I think my decision to become CEO again back in November was a really important one for me. And I realized I was on lots of boards and doing lots of investing and thinking about lots of things, but now I can really dig deep. And I think that everyone assumes that there's this power law in the world that you are a passive participant in. You're either gonna catch one of the power law winners or you're gonna lose it. Therefore, you've got to portfolio your way to a power law to catch a power law. I don't give a shit about that because I think my job in life is to make the power law and to make that outcome. And I feel like that has made a huge impact on how I think about life, and it's really every time I've tried to act on it, it's actually paid back significantly. So focus is the biggest piece of advice I've gotten that I think matters. David Sacks (00:21:36): So the version of that that I tell founders all the time is always be all in on your best idea. Because founders, sometimes I have a whole different idea. Scientists are doing multiple startups. So I'm just thinking about pivoting, but they're still hanging on to the old idea. And always give them permission to pivot. I'm like, don't worry about what you said you're gonna do whatever two quarters ago when you raised this money, do whatever you think the best idea is right now, like, let's go all in on that. Do not hedge your bets. I think that's just really important for founders. It goes back to what you're saying. It's just so hard to execute any idea that you can't be hedged as a founder. David Friedberg (00:22:16): Seven Samurai. Yeah. Akira Kurosawa did such an amazing job representing this ethos, I think. Like, you strike and you strike perfectly, you strike with full force and with all your energy and everything you have, your whole character, everything, and you can accomplish incredible things. Chamath Palihapitiya (00:22:33): You have to put one foot in front of the other every day, and you have to focus on tangible progress. And where that fails is when most people, and I do it a lot, and I've tried to get better as I've gotten older, is when I get comparative and I compare myself to the other person, the other company, the other funding round. There's so many reasons for you to feel like you're less than something else. And the reality is that has nothing to do with you. You're not in control of that, but it's so hard. And then if I don't take that medicine, I become insecure, and then I make mistakes that are entirely avoidable. So it's just tangible progress, the things that I can control. That's probably the most useful piece of advice that I try to remind myself of every day. And then I kind of have these two jobs. One is I'll incubate companies from time to time if I get intellectually curious enough. But the other part is just as an investor. And the best piece of advice I got as an investor is you are a big wave rider. And the swells that become ginormous are not visible when you paddle into them. And so you gotta commit and you gotta go, and you are going to have less than a 100% hit rate, and you will get massively washed out, and you just gotta get to the surface and swim back out. And as an investor, that's been helpful because I've had some huge wins, but I've also had some huge losses and some total embarrassments and flameouts. And it's the totality of that that allows me to swim back out. So those are two different things, but the entrepreneurial one and the life one is more valuable because it's so easy to get distracted and feel insecure because of what somebody else is doing. And it always ends up screwing me up. David Friedberg (00:24:29): I think Jason Calacanis (00:24:29): that's a good note to end on. Give it up for my besties. And, we're gonna take a little walk over here, for a little surprise if I could have my four besties follow me over here. And walk over here because it's a very special day or a special time of the year. It's the Sultan of Science's birthday again. So come on around, and we got him a beautiful cake and then a tiny little vegan cake. And we're all gonna sing the Sultan. Happy birthday. Three, two, one. Birthday. Birthday to you. We're gonna be prepared.