Tesla Robotaxis Go Live, OpenAI Pulls Jony Ive Promotional Materials, Iran Updates | Yacine, Farzad Mesbahi, Sebastian Siemiatkowski, Jon Chu, Eric Romo & Thomas Mueller, Derek Thompson, Alex Robinson, Jesse Zhang

29:35We're going to get some emails in Discovery that are going to go on big tech history emails the X account and you know we're going to find some funny back and forths between Johnny Ive and Sam Olman. Speaking of lawsuits and emails and discovery, Elon's lawyers are apparently claiming that he doesn't have a computer as part of the lawsuit between OpenAI. Who doesn't have a computer? Elon. Elon doesn't have a computer. He doesn't use a computer. He doesn't even have a work computer. You posted this earlier today. I found this out because I posted earlier, if your CEO has a quote-unquote work computer and a personal computer, it's time to start polishing your resume.

30:14And so somebody responded to that with notes to basically claims that... How does Elon play Diablo? Well, he probably has a gaming rig. Oh, it's a rig. It's not a computer. No, but The Verge posted this. Your Honor, my client doesn't have a computer. He has a rig. He has a gaming rig. Oh, you want access to his rig? where he plays games on? You want to see if he's really that good at Diablo? There are people that would actually definitely want that for sure. The setup. Yeah. I don't know. Maybe the second greatest Diablo player of all time will subpoena that information. Be like, I want actual data.

30:53How many hours did he log? I need to know the real account information. So Wired is reporting this morning that the claim appeared in a court filing related to Elon Musk's ongoing lawsuit against Sam Altman and OpenAI. he said that the lawsuit also flags that he has posted about his laptop numerous times in the last year so people are trying to put him in the truth zone that is crazy so he doesn't have a computer he doesn't he doesn't have any apps installed on the computer he doesn't have linear installed on his computer i'm sure his team does his team must because how would they get anything done without a purpose-built tool for planning and building products like they need a system for modern software development how else would they streamline the issues projects and product roadmaps how else would they would they run such a big service with so few people on their team it would make no sense unless everyone is just okay i i i think we're gonna find the computer and when we find elon musk's computer we're gonna find there's gonna be one app on there linear for sure that's right linear.app an anonymous little birdie told me that uh he does have a massive a gaming pc with an ultra massive screen and indeed does not have a work computer lol so pretty base okay anyway uh wealthfront is filing to go public very exciting we've heard about this fintech platform before um they seems like the window is open john the window is wide open if you're a series d company series c company series b company maybe clearly gets out who knows i could see clearly getting out with chamath it's possible anything's possible when the window's open so go for it if you're a founder uh our advice to you just go public just take your company public and then go on a generational run compounding it 40 a year stock price for decades yeah a couple decades it's all it takes and then you'll be in the in the mag 7 yeah in many ways if you go public earlier those initial yeah years can be easier because it's such a small base.

33:01It's much easier to grow. The SEC requirements are so difficult. You could be short sellers could attack you. It's like that's just inspiration to grind harder. With putting up incredible quarter over quarter results. Blow them out. Blow out the shorts. Speaking of the public markets, Anthony Pompliano posted this morning. Today I'm announcing a$1 billion merger to create ProCap Financial, a Bitcoin native financial services firm. The company will be a publicly traded entity on nasdaq at the conclusion of the proposed business combination between my private company pro cap btc llc and columbus circle capital core corp one a publicly traded spec the ticker for the publicly traded entity right now is cccm and he has raised over 750 million dollars they will focus on acquiring bitcoin for their balance sheet while also developing products and services to produce revenue

33:59and profit from the Bitcoin on our balance sheet. Let's give it up. Some people would say that should be the first goal. Let's give it up for revenue and profit, everybody. We love it. Love to see it. Try to be the Warren Buffett of Bitcoin. Buy companies, invest in Bitcoin, funnel the Bitcoin back into buying more companies, produce more Bitcoin, and the flywheel continues. That's right. That's the goal. Well, whatever company he buys, got to put them on numeral numeral hq.com sales tax on autopilot spend less than five minutes per month on sales tax compliance um should we do some of the some of the downstream fallout from clueley yeah i just wanted to cover this post from will brown first uh zuck hired uh alex wang for a half to one percent equity he's basically a founding engineer thought that was good when when the when the 15 billion number was being thrown around just just a week ago everybody was freaking out but It's basically it's barely a point.

34:55Yeah, we I mean we identified that it's like through my point. It's like what would What would it take for Alex Wong to move? Mometa's market cap 1 % well, like I don't know a great llama 5 release that would probably do it right like there's there's a ton of things There's a lot of opportunity if you're just talking about that's being the stakes It's not that he has to overnight double the company in size He just has to he just has to you know implement a rational and effective AI strategy Ben Thompson was talking about this earlier today for a multi trillion dollar opportunity. Yeah, yeah, yeah And there's so many places where AI products can have an impact Ben Thompson was talking about this this idea that the researchers are really important But what Zuck has been on a terror about is hiring actually great managers when you look at Daniel gross Nat Friedman Alexander Wong like these guys are Everyone the criticism like they're not researchers But that's actually probably the benefit to the strategy because we are in the application war right now we are in the implementation of these AI tools and the capabilities of the models are clearly way way ahead of the of the actual foundation models like the like the actual science that's happened I was thinking about this just like For Instagram during the dolly moment.

36:12I was like I want like I wonder what would happen if you just took every single person on Instagram, their most recent photo, and Ghibli'd it, Studio Ghibli'd it, and just sent them to it. Send it to them, proactively. It would cost a ton of money to do this in the inference, but you do that, you send it to everyone, and you just say, would you like to post this? I bet you have the biggest day on Instagram ever, right? Because you're just creating all this viral fuel for the day. Or what does it take to distill that model down, bake it into just a filter, so when you upload a photo, but you can just select the Studio Ghibli thing and it runs efficiently.

36:47And that's almost, that's not really a research task. It's more of like product management. And I'm not even like a product manager at Instagram, right? But like, those types of like the volition that it'll take and the will and the incentive alignment that it'll take to actually roll out a product like that and really push the team to be like, no, we want to spend all this money rendering these Ghibli's because we think this is a calculated risk that will put us back in the AI image conversation. Or we want to, you know, spend all this money to distill down images, image generation, diffusion, to a point where it can just be a filter button on the Instagram app.

37:20Those are product management. Those are costs. It's almost like more of a CFO question than a research question at this point. We know that the Ghibli filter is doable, and it's great. And so how do you actually get it out into the world effectively? And that's more of a manager discussion. So he's basically a founding engineer, but I think he's going into more of a C-suite role. So we'll keep tracking it. Anyway, should we keep going into the AI application layer with the wrapper of all wrappers, clearly? The final wrapper, the final boss of wrappers. Yeah. We should get Tyler's take on this because the question that I think people are digging into is that clearly, the easy take was like, oh, this is like Theranos.

38:08And it's like, I don't think it's like Theranos at all. Like the product works like you you tried it and you would agree that it is a real product. Yes. Yeah, for sure. 100%. It's it does what it advertises. It does what it advertises. Now, there are some questions about like the actual long term value there. Or I think the risk that the the the bare case for clearly the risk is that the risk is that the opportunity does not align with the marketing strategy. So like the opportunity for helping students with their homework and this sort of like traditional nature of cheating aligns with the marketing strategy.

38:48I think they can dominate there. The enterprise opportunity of let me intake all the information from every call that you have internally, every sales call that you have. That's a very competitive space. That's a very competitive space that the labs will probably want to own because it's very valuable data. You can imagine OpenAI having a desktop application that's permanently running. They already do. They already have a, yeah, but they haven't adopted that pattern. Not to the same degree. And so, yeah, the risk is like if the pattern really works and people are like, this is what we want. We always want the screen capture and constant responses from LLMs.

39:29Like if it's a really big opportunity, you're going to face competition from the really big labs. If it's a smaller opportunity, it's hard to kind of scale up. So there's a lot of discussion. The other pushback was high level. You know, why are we funding these types of businesses that are loud and not, you know, truly sort of bending the world in any type of positive direction? And it has the cheating angle, which some people, I think that was very intentional. Yeah. The pushback I would have there is like it seems like it seems like there is enough venture capital right now. Like if you have a great idea in hard tech, manufacturing, et cetera, you will be funded.

40:10Yep. And it's you know, it's it's not like some great, you know, manufacturing American dynamism business is not getting 15 million dollars this week because clearly got 15 last week. And so it's not zero sum. Yeah. We had some good reactions. Kevin, a.k.a. Plant Daddy, was quoting a segment from our interview and said, Roy was basically saying, like, there's not enough viral content, which is actually an interesting take. His take at a high level is, like, if you make something, like, great, novel, engaging, it just will naturally go viral because there's actually a scarcity of, like, viral content.

40:53I think it's an interesting point, and I think that he's probably right to some degree. Kevin fired back at a high level and said, this is not correct. While it is true that when, oh, and sorry, the last thing, Roy was basically saying that consumption has just shifted so dramatically to TikTok, which is a new form factor, and so that's changed the dynamics. Kevin is an active creator, I believe, on YouTube, and I'm sure he's everywhere. so he was just firing back. He says, this is not correct. While it is true that when TikTok blew up in 2020, there was a brief moment where there wasn't enough short form content to satisfy consumption.

41:29That lasted 12 to 18 months. I remember it well because I was one of the first gardening accounts on TikTok in 2019. It was free back then. Literally any video popped off. Then millions of new creators came onto the scene coupled with 10x reduced friction in creating content due to TikTok's native editor, filter, and cap cut. You saw an explosion of short form videos so both the number of creators and number of videos dramatically increased consumption just increased even more because the videos were all five to 15 seconds long so one scrolling session equals 100 videos instead of 50 to 70 percent of one youtube video it's not true that for the next six to 12 month anything that you post has a potential to go viral will go viral that's only true when there's a paradigm shift and all platforms have to catch up the phase where ig meta released reels and YouTube released shorts to catch up to TikTok was truly a free attention grab era and thousands of creators made a fat bag during that era.

42:22From 2020 to 2022, I grew Epic Gardening from 300 ,000 to something like 8 million across all platforms. It was that free. That time isn't today though. He says, I think what you're seeing right now with Cluelee is someone applying a Mr. B style hyperscale influencer content model into something that VCs will salivate over B2C AI. Whether anything comes of it is another story, but it is somewhat comical to see the tech VC space go crazy for the viral attention grabbing when the type of attention they're grabbing as well as the scale of it is dwarfed by almost any top 1 % creator of which there are thousands of.

42:58Yeah, Christian Kyle was pointing out that clearly is getting millions of views on X, but only a few thousand views on YouTube. And so it's very, very platform specific and um and yeah but it's it he roy is really fantastic at applying that the the viral playbook that works for x on x and it's been it's been very successful i think they're successful on other platforms too um but it is interesting yeah and it's interesting slow ventures yeah created a fun megan light cap over there runs it we've had her on the show uh i believe right uh we had sam i don't know we actually had megan on but we should um basically saying hey there's to all these creators that are getting a massive amount of attention.

43:38If they have capital, they can monetize it. So some people are paying attention to this. It's still, like I've been surprised by how much the cheating buzzword has been really controversial. It seemed very clear to me that that was deliberately edgy and just kind of an edgelord statement and that it was pretty clear that the tool was very much just like embracing the new norm of co-pilots happening everywhere, and that cheating was deliberately used as a word to kind of spark controversy. But it really is acting as baggage on Roy. And so Sarah Guo over conviction said, we should celebrate hacking, not cheating.

44:28Cheating quietly pockets extra monopoly money, insider trading, plagiarism, inflated metrics. Hacking openly flips the game board, exposing rigged rules and inventing a better game for everyone. Wikipedia replaced Britannica. Airbnb monetized spare rooms. Linux openly challenged Windows. Both break the rules, but hacking doesn't compromise integrity. It reveals the old game wasn't worth playing. Yeah, it's interesting because he could have totally reframed the Clue Lee launches as hack everything, saying hack dating, hack your online interviews or whatever. And it probably would be much less viral, but much less controversial.

45:09Yeah. The other thing is like give yourself superpowers in any of these domains. Totally. Right? Yeah. I mean, the co-pilot thing is like the friendliest framing of all of this, right? It's like, yeah, it's just like something to help you out. Autocomplete, saying autocomplete for everything is not nearly as viral. And many people have said that over the last year. And I think this is what Roy is already battling with is he, I think generally has good intentions, right? He has his own intentions around. He clearly is excited about being famous. In many ways, he already is a niche tech celebrity.

45:43Yeah, over 100K followers. But at the same time, he does want the respect of insiders. You can tell that from some of his comments over the weekend and replies. So anyways, he's young. He's going to figure it out. Well, if he's going to try and sell this into enterprise at some point, he's got to get on Adio, customer relationship magic. Adio is the AI native CRM that builds, scales, and grows your company to the next level. Get on there, Roy. Lulu continued the conversation saying, one thing Cluey got really right is hiring only marketing people who have already built large followings on their own.

46:14Any marketing person that isn't bringing you an audience is just being subsidized by yours. I thought this was a good take. I don't know how, it's tricky because many people who have built large followings have their own businesses, but Roy's clearly created some sort of great relationship with them such that they can come in and earn a paycheck and at some point get paid very highly. Kind of unclear how much is cash and stock, but it's enough to get people to step back from a TikTok account that's 100 ,000 followers or something like that, but that type of person and understands the algorithm and actually can do marketing more effectively.

46:53And so I think there is an interesting lesson there in that the new marketing, like we transitioned from resumes to GitHub profiles for hiring programmers. We've seen this with folks, the interns that we've hired, where someone will come in and say, I built this this weekend, here's an example of what I can build, or here's an example of something I built in the past, an open source project. And the equivalent in the marketing world now is building some sort of following and actually understanding content creation at some level. And so, yeah, it's very bullish. Maybe marketing departments should be doing more proactive outreach to folks that are building large followings but not quite monetizing them just yet and saying, hey, you clearly understand Instagram, TikTok, whatever you're doing.

47:37You should come inside the organization and work here in our marketing team, which I think can make sense. Dylan Field had some alternative framing of the Clulay story. He said, Clulay isn't just good marketing. Roy hacked the algo feed with an antifragile mimetic virus. And the more energy you throw at it, the stronger it gets. Well, here we are throwing more energy at the mimetic virus, baby. That's what it's all about. We can move on from Clulay in a little bit. Teddy Blank says, Roy Lee disturbs people because by any measure, he should be a successful crypto slash Forex influencer, course sales hustler.

48:14But instead, he chose B2B SaaS, and it feels unnatural. Funny that he's saying B2B SaaS, because the previous post by Plant Daddy said B2C SaaS. And so no one even knows. They're doing both, I guess. Yeah, they'll sell to willing buyers at market prices. Also interesting, in the open source world, Ian has the story, while clearly spent that at$15 million on marketing videos. This random guy on Reddit built an open source the exact same project. Always interesting to see when these open source projects break through because I remember there was a competitor to Devon, OpenDevin, that was open sourced and everyone was like, Devon's cooked.

48:54But we saw the data and it seems like Devon's producing tons and tons of GitHub pull requests and has a very different go to market motion and that for a lot of enterprises that are doing and replatforming of their.NET application to Python, they might not want to use an open source product. They might want to have a team there that's actively improving the project on a daily basis. And so the open source did not kill Devin in that case. In fact, their more significant competition comes from OpenAI and another large enterprise scaled $100 billion company, not just some random developer. On the flip side, we saw DeepSeek and Manus come out with open source projects.

49:33Manus. But not quite these like, you know, I just open sourced it and made it free. Like DeepSeek and Manus, nobody would call those like true side projects. So it's unclear how much this will affect. At the same time, if you are hesitant about sending screen recordings to Gluly, maybe open source is the one that you actually want to run. You want to run this locally because of a privacy issue. But it will be interesting to see whether or not this is a factor in Cluley's rollout. I would be surprised if this is something that makes a difference to them. I would imagine it's much more about - Yeah, they're clearly marketing to people that don't know how to run open source software.

50:21Yeah, yeah, yeah. The thing that was interesting, Gary Tan shared, are people building brands ahead of the tech just to try to hold space right at the moment that the just right models hit. Somebody named Ethan Ding was highlighting that he was comparing Clue Lee and Icon, both of which have been very controversial. Ethan said, I'm surprised no one's connected Roy Lee's playbook to the Icon ad maker playbook a couple months back. Have an unusable product. Dominate Mindshare for potential buyers. Buy time for the AI to catch up, to bail you out. I don't, you know, Roy Lee's product is not unusable.

50:57We used it on the show Friday. It worked for the use case that they advertised the most. They clearly are dominating Mindshare. The question is, will AI catching up actually enable them to win in the enterprise or these categories? The issue is that the two categories that Cluely is going after being effectively like B2C, so like students, and B2B enterprise are two categories that ChatGPT is already dominant in. And so the question again is, are they going to be able to win head to head with OpenAI, even if OpenAI doesn't have as many viral TikToks being produced in-house? Yeah, he calls out the risk of churn here.

51:44I think that's very real. I think the bigger question is, clearly and very much is wrapper, but at the same time it's not trying to compete on the foundation model level. And so if the models get better, clearly should get better. And so you are very much riding the rocket or riding the wave of technological improvement, which seems like a good strategy. I guess it just keeps coming down to the question of, if this is truly the UI paradigm that people want to interact with ChatGPT through, you would expect the ChatGPT app to launch this. But, you know, who knows? Maybe it gets popular, and maybe there's other pockets of opportunity that come through.

52:29And there's a bunch of...

52:35In other news, Mark Gurman was reporting over the weekend. He said, Apple will need to leave its M &A comfort zone to succeed in AI. basically calling out that Apple has held various early conversations with different companies like perplexity and thinking machines around potential acquisition. People were pressing him saying why perplexity. Mark responded, which looks like it's an output from ChatGPT. I don't know. It's not. It's not. He actually wrote this. It's got so many MDashes. Look how many MDashes it has though john wait no no i think i actually maybe this is from i don't know i'm confused i can't tell anymore yeah this might be from chat gbt i thought so he says a proven consumer-ready product it fills a clear need a decently sized team and a reasonable valuation at around 14 billion perplexity wouldn't break the bank yeah and uh the timing too is interesting apple's long-running search deal with google is under threat from a u.s antitrust suit this iphone maker has to prepare for a post-Google search world, and acquiring perplexity would allow it to finally launch its own Apple-branded search engine.

53:43My question about perplexity going to Apple is what is perplexity building on top of that wouldn't be an issue in an Apple acquisition? Because I don't think that they trained a foundation model themselves, and I don't think they necessarily wrote a massive web scraper, so I feel like they might be sitting on top of a few APIs that might be renegotiated if all of a sudden Apple's the buyer on the other side. So Google search or Bing, I think Bing also had for a long time an API where you could search the web using Bing and resell that as a product. But if Apple's the buyer, it's going to be a different conversation with Microsoft, right?

54:31So they use a number of different, they have their sonar family of models. Those are perplexities in-house models. Then they have R1 1776, which was a deep seek. Then they have a bunch of different third-party models that they use, OpenAI, Anthropic, XAI. So they're routing to a bunch of different, but they say the default model is optimized for speed and web browsing, delivering fast, relevant answers. So perplexity has been solid. I use it once a day. I don't know if I. Oh, you do? At least once a day. I don't know if I'd use it if I wasn't running the show, but I get value from it day to day.

55:11Yep. I'm not sure if it'll happen. I mean, it certainly makes sense. I posted over the weekend that... Siri was an acquisition. And so even though Apple has this culture of let's never acquire anything or let's build everything in-house, everything needs the Apple polish, they do have a lineage of acquisition, although Siri was a much smaller acquisition, a much earlier stage. $200 million. And I think Siri had been around for like a couple of years. Kind of similar. And if you think about what a, I mean, it's 100 times expensive. $200 million in 2010. It's 100 times as expensive. It is a big deal, but still, it could happen.

56:06Well, in more important news, the All In podcast released their new tequila brand over this weekend in LA. There was quite a lot of pushback, but I thought the bottle was quite unique, and I would be down to sample it as well. Also, people were saying, oh, this is so out of character, but Elon Musk has Tesla tequila and Donald Trump has Trump tequila, apparently. And so, or liquor brands. I guess Trump has a liquor brand. And so this is a proven strategy. Anyway, we have our first guest of the show, our second guest technically, since we already had that call in, the CEO of Klarna. Welcome to the stream.

56:48How are you? What's going on? Welcome to the show. Haha, thanks for having me. I just realized the light in here is very, very odd, but I hope this is fun. I think you look great. You're good. You're good. Thank you so much. I imagine you had to stay up late for this, so we always appreciate it when someone can do that. Last man in the office, it looks like. Leading by example. Exactly. This is what your team said. I hope there's some more people here than me. Yeah, no, your team said this about you. They said that you're obsessed with artificial intelligence right now, obsessed with digging in the weeds, tinkering with stuff.

57:21probably goes back to the early history of the company. Maybe that'd be a good place to start. How did you actually, like, what was the inciting element or the inciting incident that led you to start the company? Wow. Well, so as you may know, I've been at this for longer than many others. So I'm at the 20th year of this venture. An overnight success. Overnight success. We love an overnight success. It's been easy. But I was 23 back then, so I'm not that old yet. 23. Did you ever have a real job? What? Did you ever have a real job or did you go straight into entrepreneurship? I had many jobs. I had many jobs.

58:00I started working when I was a 15 year old at Burger King, flipping burgers. That's great. Were you any good? What? Were you any good? Flipping burgers. Yeah. I think we actually had a competition for like making Whoppers really fast. No way. With quality. With quality. I think I scored pretty well.

58:24So 15, you're flipping burgers, 23, you're starting a massive tech company that will go on a 20-year run. What happened in between? Very shortly, well, first, for whatever reason, I always dreamt to be an entrepreneur, kind of an odd thing to do as a kid. My father really wanted me to be a doctor, but I was inspired by Richard Branson and Ingvar Kampra, the founder of IKEA and stuff. And I went to business school. Bad idea. I should have learned to code the first thing I did, but that was a bad idea. But anyways, and then, you know, after a few years of just like testing a lot of different jobs, I ended up.

58:59By very odd story. Going around the world without flying, coming back too late to continue my university studies. And as a consequence, ending up as a sales guy at a factoring firm that was doing like account receivables. Last thing in the world I ever thought I would work with. And at that point of time, I realized there was a lot of these very small newcomer e-commerce companies that were selling stuff online. And they needed a better payment solution because their customers didn't trust them, didn't trust that they would ship. So they preferred the idea of buy now, pay later, where you would first receive the product and then kind of pay for it once you had it.

59:37And that was pretty much the beginning of Klona. That makes sense. Yeah, it's awesome. Yeah, walk me through the more recent developments. How big is the company today? So we're about 100 million users worldwide. We do about$100 billion worth of payments volume annually. There's about half a million merchants, but a lot of them are really large. So there's like Nike, Macy's, Sephora, Walmart nowadays. And yeah, what else? on number of employees, about 3 ,000 people. Yeah. Wow. Yeah, what was the initial go-to-market? I seem to remember there were specific dynamics in Europe around cash on delivery being a unique kind of vector into growth.

1:00:27Can you walk me through kind of like what the early go-to-market was and how, and it seems like you almost like leapfrogged the technology curve a little bit. Well, the truth is, funny enough, even in the US, if you look at the beginning of credit bureaus, a lot of them started as mail order companies, where the first credit was provided by you shipping a product and asking for cash on delivery. And then the third time you shipped to that same address, you would say, well, this customer seems to have money every time I ship to them. So I'm going to provide them a small credit line. credit line.

1:01:07That was actually the beginning of a lot of credit bureaus, even in the US back in the early 20th century. So, the same thing happened across Europe, across the US. So, the idea of doing cash and delivering and turning that into a credit line based on the performance of that customer always having cash at hand when they were buying those products or paying for those products that they ordered, at that point of time, mail order, catalog orders, right? That actually worked. So, in the end, to us, when online came along, most people were offering debit card payments, And the problem with Evercard payments is the customer ships money and then maybe receives what was on that picture that they ordered.

1:01:42And so basically offering credit, not necessarily for the purpose of lending, but more for that kind of security of getting and touching the product before you paid for it made a lot of sense. Especially in markets where credit cards were not as prevalent as they were in the US, right? Yeah. What was the, like, as you were looking at the global map of expansion, what informed, like, where you go next? Like, walk me through the different markets that you had at the top of the list and how you chose to kind of rank those as you expanded the business. How soon did you know the American consumer was undefeated?

1:02:19And you had to be here. That took a long year. We like stuff. We like stuff. Yeah, we like to buy stuff. Exactly. Funny enough, it took a lot of years before we realized the US market. I mean, first and foremost, like when we started, you have to remember at that point of time, there was one, one globally successful European tech startup, and that was Skype. That was it. Like Skype was the thing. Everything else had been launched out of Silicon Valley. There was no real European tech success of that size that had like global consumer, you know, recognition, right? and hence to us it was like okay where are we going to go from sweden let's go to the massive market called finland that is half the size of sweden because it's very close and we kind of know what it means right so we kind of started with expansion in the nautics then went to like germany which was a huge thing for us figured that out and then went to uk and we ever we always thought that we never like figured out how are we going to come into the u.s market because credit cards are so prevalent.

1:03:17But the funny thing that we hadn't realized was due to the financial crisis in 07, there was a lot of regulation introduced that made it much harder to get a credit card. One thing that kind of sounds sane is that credit card companies couldn't advertise in campuses anymore. They're like most common places to sign up credit card holders. And so there was a lot of that regulation. So it meant that between 2010 and 2020, debit cards in the US grew like, Like a credit card volume grew about 2x, debit card volume grew 10x. And so actually, rather than us figuring out how to be successful in the U.S., it was the other way around.

1:03:55The U.S. market became more like the European markets with much higher debit card penetration, where buy now, pay later and the kind of products that we offered became much more relevant suddenly for this new consumers that didn't have credit cards to the same extent or didn't have as much credit limit as they used to have historically. right yeah what was the the rollout like in finland germany the uk the us um from a regulatory perspective do you have to get like banking licenses their approval is it mostly like regulatory or are you more focused on the engineering challenges setting up servers localization all of them all of it yeah all of the above but i think from a regulatory perspective um we quickly identified that it would be a benefit to be fully licensed over time so about 10 years ago, we did get our bank license.

1:04:42And we were fortunate that EU had implemented legislation that meant that you could take your Swedish bank license and passport it across all the European countries, right? So we run on the Swedish bank license, but it's applicable for all. And it included UK up until Brexit, right? So that worked for a long period of time. But obviously, technology was an important challenge for us as well. What are the knock-on effects of having a bank license. I talk to fintech founders all the time where it feels like they're in the money moving business, but then you find out they have all this float and they're making a ton of money on treasury bills or something like that.

1:05:19Is there an aspect of your business where you can actually drive more profit and bottom line on the basis of having that bank license, or is that less relevant? And then I want to go from there into stable coins and what's going on and how that might be changing the market? So I think that in our case, what you're referencing is kind of, funny enough, something we have underutilized Teclona, which is the ability to generate the deposits and kind of earn interest of those deposits, right? Funny enough. And that's something we're addressing right now where we're introducing more products and services in that area.

1:05:55But where the big advantage for us was, if you look at the history over those 20 years of a lot of fintechs who were credit providers. A lot of them offloaded their balance sheet and did so effectively up until at some point in time, something went wrong in the underwriting mold and it turned out that the losses were higher than they expected. At which point of time, the people that had taken on their balance sheets realized that they wanted to shift the liability back to the fintech, but the fintech was not capital-wise equipped to take that on, meaning that like their valuations either collapsed or the whole thing went bankrupt so and this happened over and over again if you look at a lot of the like fintechs over the last 20 years in our case the benefit of having deposits is we are a bank we're fairly capitalized as a bank we need to hold the same capital that you know any large bank needs to do in comparison to our balance sheet and it means that we're much better equipped like even if our losses go up and down or whatever it doesn't change the sentiment of our balance sheet doesn't mean that we can dry up funding.

1:07:00We're not dependent on the funding in the market. The whole thing that went down in 07, right, which was the problem, we're basically, we only have deposits and that are insured by the government and so forth, right? So it gives us a much more stable way of funding the bank, basically. Then on stable coins, is there any benefit to you there? We've heard stories about lower fees and faster payments. And obviously there's a massive, you know, the frequency of payments across the Klarno network is massive. Are you looking at it? Have you already looked at it? What's your takeaway on stable coins or the genius act that seems on the cusp of passing?

1:07:36We've had fintech CEOs on the show that have said, actually, traditional banking rails can be pretty awesome if you utilize them properly. So I'm curious. I think that I have been one of the crypto skeptics for a long period of time. And the reason for that is when I started playing around with this stuff 10 years ago, I felt that, you know, I always ask myself, like how is this technology going to make my mom's life better? Like how is it going to make our products and our customers experience better? And I didn't really see any benefits of the technology apart from the fact like people have been speculating that it's an interesting asset, an alternative asset to hold, right?

1:08:12Like Bitcoin or whatever. But then I gave it a second chance, a second look now one about a year ago, started looking at the next generation of stablecoin implications and implementations. And I very much agree with the fact that like now the technology is showing strong promise, both in its speed, in its cost, et cetera, et cetera, versus the traditional bank rates, especially, especially obviously in international transfers in corridors that are more like, you know, you're moving money from like UK to US, you're probably not going to make that big of a difference. But if you're moving money from, you know, Brazil to, let's say, Greece.

1:08:47Oh, sorry, that was a euro country. Stupid idea, stupid suggestion. But another one, which with a smaller currency, you're going to have a very real benefit of it, right? So I think that stablecoins definitely are going to be a major thing. I think also like crypto is here to stay and the bitcoins of the world are here to stay as assets, right? That people find are interesting assets in their portfolio. So I don't think that's going to change either anytime soon. Yeah, I'm always interested in where the value will accrue with these technological shifts. Specifically, like with stable coins, you can imagine if you're going from two day settlement to instant settlement.

1:09:22Well, if the customer would accept one day settlement, all of a sudden you have one day afloat that's going to sit there that you could kind of keep as your own margin and maybe put in T-bills and maybe it gets to be a big number. And so I've been kind of trying to understand when their time is money and when time decreases and systems get faster, you can pass all of that along. You can pass half of it along. And there's different things that different companies can do based on the consumer preferences. I think you're totally right. But that also means that in some degree, which is something I tell my investors, right?

1:10:01It's a race to the bottom. Like this payments industry has had a, you know, combined revenue globally, addressable market of like$2 trillion or whatever. It's not going to stay that way. There's no real good reason why it's more costly to send money than send an email, right? There's a little bit more risk involved. There's a little bit more like, you know, security involved. But in essence, you're just like shifting data between two accounts. And so like in some to some degree that like that those kind of solutions will drive. And that also means that the revenue generated from these services will be significantly lower than it is today.

1:10:36What's your broad, not to go too broad, but I'm curious to get your broad market outlook right now. You have this interesting view into consumers, companies all over the world. It's a very interesting time right now. We have geopolitical instability. We have, you know, and effectively war. We have tariffs, trade wars all happening at once. Yet at the same time, we've seen circles went public and is now trading higher than Coinbase. So it seems like the general excitement around... Capital marches on. Yeah, capital is marching on. But yeah, how are you thinking about markets broadly from your view?

1:11:21To me, from my perspective, because of kind of the deep interest that I've taken into AI and try to understand the implications and how to apply the technology, I am very much convinced that it will have implications on white collar jobs. effectively you can have like economical swings that you know recession non-recession it may you know higher inflation higher cost pressure and so forth may have some implications for consumers sentiment of spending and you know i don't like i'm gonna spend a bit more i'm gonna spend a little bit less and so forth but but the true life-changing moment that affects people's true ability to pay back right is to a large degree uh loss of jobs right if you if you lose your job, then it severely impacts your ability to do so.

1:12:07I think that the key question now is to what degree we're going to see an acceleration of loss of jobs among white collar people, which in itself then potentially leads to a recession and less spending in that group of people, which is obviously a group of people that have had a lot of money. So I think that's the area that I'm like mostly because the tariffs so far obviously have been announced and kind of changed and then announced again and changed and so forth. And now it almost feels like if I look at our merchants, when the first tariffs were announced on Labor Day, they really did adjust their marketing spending.

1:12:41We saw changes in the transaction volumes. We saw, for example, some of the larger Chinese companies that we work with would shift their marketing spending from U.S. to Europe and would try to offset more of their inventory in the European market as opposed to the U.S. market, etc. But as now also like there's been more and more like turbulence around these things, things, people start also discounting and saying like, yeah, it was announced this week, but who knows what's going to happen next week. And so we're not going to draw too early conclusions from this. We're not going to change the business of what we're doing until we know it's for real, right?

1:13:12And until it's actually going to happen. So I don't actually see currently in the trends, the direct implications of tariffs themselves in our business. But I'm more just like trying to second guess if I'm going to see a major change in the job. And what are you tracking on the job side? The other thing is you're kind of hedged in some way and that I think people would be more likely to use Klarna if they have, if they, you know, in some ways have reduced spending capacity or, you know, more likely to leverage a product. I'm curious, just getting more specific around this, what is your updated hiring strategy throughout the rest of this year and next year?

1:13:58What areas are you hiring in or kind of shifting away from? So we, about almost now two years ago, after experimenting with ChatGPT and so forth for a few months, our conclusion was that this will lead this will change our ability to get more done with less people and that we believed that in an AI world revenue per employee will be a very key metric because the higher revenue you have per employee the more profit you make per employee means that you can spend more on marketing means you can acquire more customers which you can scale more so if you have a low revenue per customer you'll be less competitive basically and so for us the question then was how do we catch those efficiency gains and so we decided to stop hiring about two years ago at that point in time we were about 5 500 people and since then we shrank to about 3 000 employees that we are currently which is basically through natural attrition because you know in a tech company about 20 will leave on an annual basis voluntarily because they're going for other jobs other opportunities and so by just being very strict in not rehiring, we have been able to shrink down the company and continue to do so.

1:15:08Now, we've hired a little bit. It doesn't mean that we haven't hired at all. We've hired in some areas and so forth. But if you look at the net, you can see that we've shrank from about 5 ,500 to 3 ,000. And then what we've also done is internally, we've been very honest and open about this. And we've communicated to our employees that what's going to happen is the gains, because obviously this means our labor cost is shrinking while our revenue actually has been growing, right? So revenue, as a consequence of that, our revenue per MP went from about 400K to about over a million now, which is kind of equivalent to where Goldman Sachs is.

1:15:43Now, next step, yeah, thanks. Next step is kind of, you know, Apple, Netflix being at$2 million. I sometimes joke because like OnlyFans is at like$30 million. Don't go there. Yeah, or don't look at Tether. don't look at tether will have you looking at uh net income per employee yeah exactly so i'm not sure that's a great uh the comparison you want to have but the thing is that like um the anyway so we are continuing down that path and and the point is that um uh what we communicate internally however is that as the total labor cost shrinks um we are reinvesting that in an acceleration of compensation for our existing employees so as they participate in the efficiency gains we have been able to accelerate both equity and cash compensation at Klarna to our existing employees more than we had in previous years.

1:16:32And that's kind of a way for them to share the benefits with the shareholders of the efficiency gains that this creates. So if you're a new grad and you want to get a job at Klarna and make a career out of it, maybe be there for 20 years and earn a lot of stock and cash over a great run. Come for your job, maybe. Come for your job. Yeah, exactly. What should a new grad do? Should they learn prompt engineering? Should they learn to vibe code or program? Or should they be focused all on networking and management skills and people skills and the things that the robots seem to be mediocre at? I think very clearly is there's one area where we have not reduced number of people at all, and that's sales.

1:17:15In our case, I mean like B2B enterprise sales. So we have people locally on the ground across the world in over 50 offices. you know whether they sit in in china next to to xian and tiktok and so forth or whether they sit in portland next to nike or whether they are in you know somewhere on the east coast or in parts of europe on and those obviously there's no like immediate expectations that those are affected by this right rather the opposite like that human connection in my in customer relationship is super important i've never i've never seen a chat gpt have a power lunch and a steak dinner exactly yeah it's impossible until we get better humanoids um but on the rest of jobs actually it's been quite interesting because for a period of time now we said okay let's stop hiring all other competences than engineering so if you look at that 5 500 of them 1 500 were engineers today of the 3 000 1 500 are engineers so actually the number of engineers have stepped kept the same while it's all the other like marketing you know analytics, whatever it might be that shrank.

1:18:23Now, with that said, however, in the last months, because of the acceleration of Cloud Code and Cursor and all these tools, what you're seeing now is a group internally at Cloud now who are like a few hundred business people that in the last two years have learned how to code. Not necessarily like the most complex things, but inspect code, understand code, interact with code, but doing so with having all the business knowledge, understanding what the customer wants, understanding a lot of things that engineers sometimes have struggled with, like they can't articulate why we're building the product for what purpose, for how it's going to solve customer needs and so forth.

1:19:00And they've been maybe sometimes more focused on just like latency or technical implications or, you know, et cetera. So I kind of see a convergence now of the truly kind of renaissance people that are kind of, you know, both on both sides that can understand the code but can also understand the business those are the ones that are now truly becoming the most productive in my opinion what is your framework for buying ai tools versus building ai tools internally right so we we have been fairly vocal about the fact that you know as part of acceleration of ai clana one of the first thing that we can remember chat gpt came along there was this like, hey, I can now connect all my PDFs to ChatGPT, you know, and like, and people were like, oh, that's awesome.

1:19:44The only problem with that is that like the truth in data modeling and data science has always been like shit in, shit out, right? So if you feed it tons of like internal stuff that is contradicting, inconsequential, not truthful, whatever, it's just going to get lost, right? Just like some of our employees get lost when they start navigating the internal work of Klana, right? So, the first thing that we realized is we can't silo the data anymore. We have to have a common data model. And that meant that we started removing the software as a service providers. So, we closed down Salesforce, Workday.

1:20:21I think the next big one that would go is Atlassian, Gyra, and so forth. But then there's also like a long amount of smaller software as a service vendors that we have utilized. So actually, if you kind of count it in total, there's over 1 ,500 that we have closed down in the last two years. Now, obviously, some of them are like super small, like some plug into Gyro or whatever. Like it's not like every one of them is like a tremendously big software as a service application. And then we've kind of harmonized and standardized our data layers so that we have one data that we can then let the LLM interpret and run on top of, allowing us to do more with less.

1:21:01Right. So currently, there are a few things that we will use 11 labs for voice. We will use like specific people that are really, really proficient. We use a company called FullStory that's fantastic to kind of catch user interactions. So there's some places where we just found like this software is so strong that there's really no meaning in trying to replace it. But a lot of traditional SaaS may have just been a database with some nice UI and some data and some process logic on top of. And that is easier to replicate today. And it's not easy, but it's easier. And the problem is if it's spread out, it's inconsistent.

1:21:42And we need it to be consistent for both humans and LLM to make sense out of it. Well, this is fantastic. This is super interesting. I feel like we should talk regularly about the situation with white collar workers and just the advance of AI tooling. There's so many interesting things where you have a unique insight because of your position. Last question from me before we jump. Do you think cookies hurt conversion rates across Europe? You mean cookies policies? Yeah, just the fact that just generally... Well, generally, yeah, we could spend half an hour, but generally, like the more clicks you need on a website to get to the eventual purchase, you know, the worse the conversion rate.

1:22:27And maybe that's why the American consumer is outperforming. Yeah, it's just purely a technicality. I think the challenge of GDPR, which is the legislation that led to that cookie, sorry to say, shit show, is that the problem was it came from a very good principle and idea. And that is that we as users should own and control our data, not the companies, right? And if you think about it, when a few years ago, WhatsApp, there were some rumors that they have updated the terms of use. and now there was so much more privacy infringing and everyone should shift, right? Now, how many people really shifted to Telegram or whatever?

1:23:07Not enough, right? Why? Because at the center, the way it should have worked is you should have been able to go to WhatsApp with a click of a button, extract all the data, all your contacts, all your messages, and be able to then feed them into Telegram or Signal or whatever your choice of preference would have been. That would have been true customer mobility. And that was the purpose of GDPR is to allow that customer mobility. But because the legislation wasn't implemented thoughtfully enough, that hasn't materialized. Because in that world where you had 100 % customer mobility, the point is if you lost trust in WhatsApp as a provider of your primary chat messages, you would have instantly been able to move to something else.

1:23:53And the fear of losing all your users in such a very mobile world where everyone could move would have forced WhatsApp to be concerned about privacy and not just speak that they are concerned about privacy. Right. So like I think that and that was the original intent. But then it was implemented with these cookies, which actually creates the opposite, which is that like we're just tired of accepting cookies. I don't want to go to a new website because then I have to click again on 55. private. The people that wanted these kind of laws had the right intentions, but the implementations are horrendous and actually caused the opposite, which is less customer mobility, people less willing to shift services and try different things.

1:24:37And funny enough, are hence playing into the hands of the same companies that they were annoyed with for infringing on our privacy. Because as that creates just like the bigger success of these big tech companies. Full customer mobility, the being able to take your data with you and go wherever you want, is the key to creating a better competitive internet, where companies actually have to compete on creating more value for you as a customer, rather than locking you up and locking your data up. And then, you know, creating models that just forces you to use the same thing. Makes sense. Makes sense.

1:25:13Thank you for joining. It was fantastic. Thank you for having us. Thank you for having us. Thank you for an amazing pod. Yeah, we'll talk to you soon. We'll talk soon. Bye. Cheers, Sebastian. Let me tell you about public.com, investing for those who take it seriously, multi-asset investing, industry-leading yields. They're trusted by millions, folks. Millions. And they're the sponsor of Aston Martin F1. That's right. Public.com, head over. We have our next guest in the studio already. And we have John Chu from Coastal Ventures. What's going on? Welcome to the show. Good to be here. It's great to have you.

1:25:42We didn't prepare at all, which is part of the beauty of our show. You just come in, you're live, and we'll do it live. Good to know each other. It's great to have you. I've got a bunch of questions. You're always getting spicy on the timeline. We'll talk through a bunch. But why don't you introduce yourself, quick background, what you're doing at KOSLA, all that good stuff. Yeah. So I'm John. I'm one of the partners here at KV. I do a lot of technical things. So basically anything I've really written code on in the past. So like data infrastructure, dev tools, a lot of AI stuff from foundation models and the app layer, if I really understand what it does.

1:26:22Right. Yeah. Awesome. Let's get right into it. What did you read? What are you reading into the news around DG potentially going to meta and all that? Was that a surprise to you? What can you talk about there? Yeah, so it was a big surprise. I actually say like the funny thought that came to mind when I first heard the rumor was like, should I be buying meta options? Because does that mean like Zuck got Ilya, right? And then that gets announced and meta just jumps. But like, look, it's, it's pretty obvious what's going on here is that there's a giant market in AI, right? Open AI kind of proves that.

1:27:03The pie is huge. And Zuck has a shot on goal. And Zuck is not someone that likes to lose shots on goals. So if he has to pay up to get the best talent that he thinks that can actually compete, he doesn't even have to win. You don't have to beat open AI, he just can be number two. If he can get there, billions of dollars is worth it. So all of this makes sense when you think of it through the lens of like, Facebook is a giant company, a billion dollars,$5 billion. That's a drop in the bucket if you take on this new market that ends up dominating. And the last thing he wants to be is like Microsoft and miss phones.

1:27:37Yeah, yeah, yeah. what do you think about the phone story? At one point, Facebook was working on a phone, and maybe the lesson that Zuck learned is that he should have put more resources into that instead of not try to do it at all. I'm wondering, because there's all these different stories. VR, he's investing way ahead of adoption. Reels, he was a little bit behind, but wound up rolling out a product that I believe produces tons of profit. And so it was totally worth investing in and taking the shot at building out all those data centers to do the recommendation algorithms. So yeah, I mean, what other examples do you think we should be pulling from the meta history?

1:28:18Or do you think that might be relevant here? Well, I mean, there's basically two things to know about meta, right? One is that they are fantastic fast followers. They either buy or they copy. I think like Snapchat is a fantastic example of that where like the stories literally like that was a team that just copied stories and just executed better so like if anything facebook is an execution machine and they are able to pull that off and then the second takeaway is zuck is willing to date bets so like i think oculus uh the acquisition and the whole metaverse is a great example of that yeah and you don't win all of them.

1:28:57Right. But like, I think on the upside, he's won more of them in terms of like, call it shareholder value, then he's lost. So this is just another one of those. And this one, I think he has extremely strong conviction behind more so than you, like the fact that AI is already doing real things in the real world that actually are impacting enterprises and impacting real people gives you a significantly higher confidence bar than the metaverse bet. So you have to have significantly more confidence here. I was talking to a guy at Windsurf who runs their field engineering team. And he was like, oh, you know, I talked to his enterprise and they're like, we're going to buy 2 ,500 Windsurf licenses because this was fantastic.

1:29:42And he's like, why don't you buy 5 ,000? Because you have 5 ,000 engineers. And he's like, yeah, we're not going to have 5 ,000 engineers. Because at Windsurf, we're going to have 2 ,500 engineers. Like it's actually doing real things. Yeah. Yeah. How can Meta benefit broadly from AI, not in the just found in the foundation model game or image generation or things like that? You know, we've talked about just just getting slightly better at AI across the entire organization can have tremendous benefits on just improving delivery of ads and these sort of like basic things. You were working on ML at Facebook back in the day.

1:30:19So I'm sure you have some, some strong insights. Well, so I, I think like there's a few layers here. Um, if you start from the product level, the obvious is that there's new products that are going to be built, right? Um, there's the whole like journey I've and, um, I think, right. And that's some kind of hardware thing. We're all guessing, uh, that's probably a new product. and if you just take the lens of like if this technology in five years gets significantly better i have a bunch of hardware products inside of meta i can add more i have a great research team that knows how to build great hardware i know how to handle the supply chain what can i build there's like random upside there in products right if you go back to the current facebook suite um there's uh meta ai is like already doing a lot of this but you have chat apps you have messenger you have WhatsApp, right?

1:31:12Like you can already do chatting in there. You have a great Lama model. If you have ChatGPT in there, you can just pull them into conversations and already add value. That increases engagement, that increases ads views, right? That's automatic money there. And then if you go back to where I worked, like in ads models, there's a, you're seeing less of it with generative eye today, mainly because everything is doing attention on sequences and ads like ranking models are a little bit different, but people are already applying research to different layers of the ads models. And like, if you have a 1 % uptick in ad score, that's like a billion dollars.

1:31:49Actually, it's like a percentage increase in revenue. So like a billion dollars today in the future, 1 % is going to be like$10 billion, right? If Facebook goes tax. And there's some really interesting research there. Like I have a port code that I invested in actually was a distinguished ad at Facebook or at Meta. and like he's kind of got some technology that's a foundation model for ranking where you really just don't need any feature engineering and i saw another company this morning which is the same that's like that rips out giant pieces of infrastructure has like automatic increase in ad score ranking and you can like fire all your ml engineers you probably don't want to do that but like you can fire most of the ml engineers at facebook like that's gigantic right that kind of stuff is going to happen i mean the time scale is a question but it's going to happen over time Do you think that the UI paradigm around AI's interface with social networking is defined enough?

1:32:40I'm thinking of the example of Snapchat and Instagram stories. It was very clear that Evan Spiegel at Snapchat did enough of the experimentation to define how vertical disappearing video and images would be distributed on social networks. And then Facebook was able to adapt that and adopt that and bake it into Instagram When I think of who what big tech companies which of the mag seven is most closely? Cheating off of chat GPT's homework It feels like it's Gemini and the Gemini app feels much more directly competitive to the chat GPT app I would be surprised if if Facebook or meta or or Instagram tries to stuff an entire chat GPT experience into the fifth panel on the hamburger menu.

1:33:31So I'm wondering, like, is there is the story of Meta's AI rollout really just about ad optimization? Or are we actually going to see it instantiated in different UI and like interaction paradigms? You're going to see it everywhere. Ads is just an obvious place you get a win, right? Yeah. But like, you're going to see it. Actually, someone might stick it in a tab somewhere. If I know anything about Facebook having been there, like if the number goes up and to the right, they'll do it. So like that, that might happen. But like, in, like I said, inside of messenger and WhatsApp, I think there's going to be lots and lots of usage there.

1:34:07And if you think about it, like people ask this question, I was actually a little confused at first. They're like, why did you pay 15 billion for what it sounds like just Alex Wong? Right. Like that's my guy. And then you think about it for a bit and you're like, okay, well what makes open AI open AI? You're like, yes, there's great tech, but Sam, like Sam actually runs that thing and Sam makes it successful. And then like, if you ask anyone and why even Paul Graham is like, okay, if you had to pick someone again, that's like a top founder inside the YC network, that's not Sam, who would you pick?

1:34:38Even Paul Graham says Alex. So like if you're looking at products inside of Facebook or even outside of Facebook, like pick the guy, right? Yep. Yeah. I mean, it feels like a really, really strong, like total reset of Meta's AI strategy, bring in three absolute hitters and turn them loose on whatever the next phase is. Jan LeCun was there for kind of the research journey and the research part of the AI story. Now we're in the application story, and so go get the managers, right? How much should people read into, again, this is all alleged. I don't think it's been confirmed that DG is joining Meta yet.

1:35:17but rumored, how much should people be reading into the CEO of SSI leaving to work at Meta and using that to inform their general timelines, right? If we were on the cusp of developing superintelligence in two years, maybe you wouldn't leave billions of dollars on the table at a company that had the potential of cracking that. even getting a billion dollars-ish liquid at Meta, it's great. But still, I think it's hard not to read into the move if it happens. I have no knowledge here. I don't know Dan well enough to know what his kind of thoughts behind us are and what drove the decision right. But I can say, honestly, for myself, working with Ilya would be an amazing dream and it's kind of amazing to do.

1:36:15At the same time, SSI is a pure research lab. So it's not like OpenAI where Sam is over here trying to build products. It's a pure research lab. And I don't know if that's what Dan actually finds fun and also what he actually excels at versus Meta also has a giant research lab, tons of opportunity and unlimited GPUs and money. And distribution to scale products extremely rapidly. and get data to feed them back into the models. Yeah, it's huge. This is the only thing about Facebook. Like distribution, everyone underweights that. All these gems that come out of Facebook and pitch me, the thing they always miss is like, no, distribution is not trivial.

1:36:55It was trivial for you at Facebook, but it's like not in the real world. Yeah, yeah, oh, you got a billion people to click on your new button. It's like, that'll be a lot harder when you're starting from zero. Do you think, do you think, I want to talk about Apple? Yeah, I was gonna say, do you think Apple will make a big bet here? There were conversations reporting every week. You're talking about Tim Cook going to Meta to get a pay bump? Yeah, yeah, yeah. Because he's barely scraping by it at Apple. He's making under$75 million. Under$75 million a year. He could probably double that if he went over to Meta.

1:37:20Yeah, something there. No, but do you expect Apple to make a big bet here? There was rumors over the weekend of reporting that they had spoken to perplexity and thinking machines. Do you think that they're willing to make a big investment? You know, Siri over 10 years ago was$200 million. That's like an AI seed round for ants these days. So I'm curious. Man, seed rounds are so expensive. You're like, don't remind me. Dude, I've been priced out of almost every round in AI research I've tried to do this year. There have been a few exceptions, but it's not like I'm priced out by a small amount. It's like 2x my price or whatnot.

1:38:02How are you even underwriting an AI research bet these days? are you pulling out a spreadsheet and building the market size or is it just like based on the vibe of this founder I'd pay 50 but the price came back at a hundred like I get in the delivery go to the future and then I figure out if it's actually gonna work yeah it's basically you got to underwrite the team and then the opportunity size sure right that's really all you can do yeah I mean that's basically what we did when we did Sakana yeah I was gonna ask about Takana, the Japanese language model. Very cool company. But that, like today, it's even worse is like, that's really all you can do.

1:38:39Yeah. And then part of it's like, if the price is just not there, given the confidence interval, then like, gotta step back, right? Yeah. At the same time, I mean, the SSI deal, the scale AI deal, like it feels like it's potentially easier than ever to get taken care of as an investor. even if you overpaid and did a seed at 300, like the Mag7 has money to spend on multi-billion dollar acquisitions. Like a lot of these crazy looking deals might get rolled up, but I don't know how many of these there will be. Like you have to be backing a founder like an Alex Wong or like a Nat and Dan to even have that on the table.

1:39:19Yeah, if you back a generational entrepreneur, you'll have a little bit of a safety net. But they have to be... I mean, yeah, it's funny. We don't typically underwrite acquisitions, right? We only underwrite this company IPO. And I think most investors do that. If you think about it, too, there's like seven of the nine, seven. So like, yeah, thoughts on goal is like not really an investment. You're going to put like 200 mil post on. Right. Yeah. That being said, like I made this, it was a half joke. We were looking at one of these hot rounds in the seed rounds. I was like, you know, we should do like we should just invest now and then sell at the next round.

1:39:55It'd be like the best hire on the planet. And Vinod like looks at me dead. And he's a part of me. He looks at me dead. And he's just like, you'd be a great hedge fund manager. We don't do that here. I think it's a good strategy for angels, to be honest. Angels are the only, the check sizes are lower. You can get a quick 10X and sell in three months. Yeah. Would you rather have a$4 billion SPAC or a$14 billion aqua hire on your ROI? Yeah. I got pitched to SPAC recently. Wait, why are they pitching you a SPAC? They're like, hey, got any portcodes? Then he wanted us in the pipe, but I was like, wow, they're SPACs now.

1:40:31They're coming back. Wow. What's your updated thinking on memory as a lock-in and a moat for different model providers and app layer companies? It sounds like you've been pitched by a bunch of plaid for memory type companies you've been posting about this. What's your kind of framework there? I still don't have a clear view to be honest I think it is a mode so um you look at all of the the model providers they need different ways to keep their customers coming back to them versus going to other model providers and memory is one way to do it um if you have enough history on the person you can tailor recommendations tailor responses etc etc right you've seen this play it isn't like the first time someone's done this.

1:41:22Like, I think the best example of a data mode is effectively Google and search. It's literally like, oh, I get so much search data that I can do better ads and make a better product. And then like no one on the planet can get more search data than me because I am the best. You all come to me and you give me more data. Right. So it's kind of the same play at the same time. Like I see why users want this memory thing of like, oh, I can like shift memory from open AI over to like Anthropic. But then like I have the same question of like, well, why would open AI let you do that? Like, why is Google going to let you take their data and give it to like thing, right?

1:41:58Just like not going to happen. So yeah, I still don't see what the outside function is where like an Anthropic or an open AI would be game for that. Yeah. Yes. I mean, it seems like the entire dynamic is that it's a good idea, but customers won't demand it. It's the same, it's the same dynamic as, uh, you know, should I be able to, we were having this discussion back in like 2010 about like Google's walled garden. Like, can I export all my Google data or all my Facebook data and just give it to Google to get better ads like that potentially, but no one's demanding that from the customer side. And so it'll never happen from the, from the B2B side.

1:42:34Yeah. Are you worried in general about, you know, there's been a bunch of different app layer companies that are kind of B2B players that are building products that ultimately will just potentially be built into things like the Google suite right and and happen where you can get sort of these sort of fast revenue ramp something like a Calendly but then ultimately just get added as a functionality to core Google workspace or teams functionality yeah so this is kind of funny this is like um one of those questions that keep come keeps coming up is like started with like GPT wrappers where it's like oh my sorry it was gonna die now it's like alright what's inside the space of these model companies.

1:43:11And the answer is like, you need some story for defensibility. And I actually don't understand why this is so hard for people to grok because like, this has been the history of technology forever is like, yes, there's tech, tech eventually gets eroded. Like what stays along that gives you defensibility? Is it like community? Is it some kind of open source thing? Is it like workflow lock-in? There's like these moats that have existed forever that everyone knows about. And for some reason, because the word AI is here, like entrepreneurs and VCs just like forget about the existing notes. They're like, yeah, this is like something different and new.

1:43:42And it's like, no, it's not just like technology like before. Right. So that's really the story is like, yes, things will get eaten. And like, yes, you're already, I've seen plenty of companies get killed that are shallow GPT wrappers. And I'm seeing some actually deep, like technical notes getting overrun by OpenAI and Anthropic. And you just need the story on like, okay, how do I get to something defensible? And can I get there in time? What do you think about these new generation of startups? Some of them are in AI where it seems like the efforts of the team are like 90 % marketing. And they're really, really good at getting a ton of attention.

1:44:26And they're being very noisy. And Gary Tan was kind of framing it as like maybe they're trying to win in distribution and then kind of hold on to the attention until the AI models pick up. It feels like it's almost like a necessary evil of the current market and how fast things move that if you want to accelerate revenue and kind of live up to the standards of the day, you have to be very, very noisy online. But do you see it as like a counter signal or do you think that there's something that where maybe this is around to stay it's not necessarily a counter signal like um it's the same as like when when we pick companies if you're too early you're wrong right same thing for founders here is like some of these founders are fantastic especially like these like younger founders are fantastic at getting attention yeah um and And I think historically with any company, any movie, any like pop star, any type of hits, eventually the attention wanes, right?

1:45:34So really it's a timing thing of like, I need the attention to translate into customers. And if my product's not there, I'm going to see churn. So how long can I capture attention? Does that overlap before my product actually doesn't have churn? If I have enough timeline in the middle and I timed it right, then I win. yeah then i lose right yep it's like pretty much that simple it's like it's simple to say it's not an easy game to play but if you play it correctly you have a giant company yeah yeah the the organic attention getting it seems like a a really great go-to-market or like early beachhead but it doesn't feel like a sustainable flywheel versus some of the other permanent go-to-market motions or like viral loops that drive you know meta to beat earnings every year for decades or something like that.

1:46:21It can get you into one, right? Exactly. Yeah. It's an opportunity. Yeah. It gets you the shot on goal where as a lot of entrepreneurs that are maybe those PMs that say, I'm going to get distribution and they just put out a press release and they're like, where did all the customers go? I don't have any clicks on my button. So yeah, it is kind of a necessary thing. There was a market map a while back showing all the different venture funds, investing in all the different labs you you guys at cosla are open ai purists so passed on i'm sure had the opportunity to invest in all the different you know anthropic xai cohere all these different players um talk about kind of that was that always an intentional decision is it let's stick to ventures roots and have our bet or was it sort of um uh sort of a lucky ability to just be in you know, the breakout, you know, winner, that's the lab and the consumer app.

1:47:16And so, you know, talk through that decision making process, and maybe kind of insights from from different partner meetings. I don't know if we've ever like discussed this intentionally until we kind of came to the like, open eyes it thing. But I think it happens kind of organically where you you work with a founder, you work with a team, you have giant breakthroughs and giant wins, you own enough of a company where like their success is basically your success and uh it kind of looks like they're gonna win it anyways like i mean i i we had this back and forth on twitter i know but like keith was kind of like yeah i don't think another transformers based or another like foundation model company is going to be a great vc return outside of acquisitions and it's kind of like yeah how are you going to beat open ai like it's quite hard they have distribution they have marketing brand they can have a shittier product and they probably still win this right um so like and they have money money is actually very important in this game because of gpus so it's it's both the like we bet on this horse we love this team we work very great with them very closely with them there's like symbiotic value here and also like they're probably going to win this.

1:48:31Yeah. Yeah. Um, what are you, uh, last, last question, what are you expecting out of, uh, are you expecting to see ads in various LLMs consumer apps this year? I'm sure you've been pitched people saying I'm going to build an ad network for LLMs. I'm sure you have opinions on that as well, but what are you, what are you expecting to see out of the next year? Big question I have is will the average American pay for their favorite LLM three years from now? and I've been trying to figure that out or will it just sort of, you know, go closer and closer to free over time? I don't actually know. Like, I think what do you mean?

1:49:10You're a VC. You should have a crystal ball on your desk. You know, I like technology and I think there's many ways to monetize this and people will try all of them. So like, are there going to be like ads? Yes, people will try it. I mean, like Aravind over at complexity already said he was going to try it, Right. So we're going to see that at some point. Yeah. It's going to be the model that works. Who knows? But it is like a nice, tried and true model. Right. So there's a good probability that it does work for some set of products. I saw somebody somebody had a pretty viral post last week saying, I'm so sick of being asked to sign up for different services across the whole Internet.

1:49:49And then what's his name at Antonio at Coinbase was like, I'll tell you a model for this, but you're not going to like it. Anyways, thank you for joining. Appreciate all the all the insights and welcome back anytime. Yeah, this is fun. We'll talk to you soon. Thanks, guys. This is fun. Cheers, John. Bye. Until you can buy ads in ChatGPT, you got to stick to ads out of home. Out of the home. Out of the home. Using adquick.com, advertising made easy and measurable. Say goodbye to the headaches of out-of-home advertising. Only adquick combines technology, out-of-home expertise, and data to enable efficient, seamless ad buying across the globe.

1:50:31We have our next guest in the studio, Tom Mueller from Impulse Space. I'll be right back. I will handle the intro. Cool. Tom, how are you doing? We are bringing him in in just a second to give you more background on impulse. I met Tom, oh, it must have been two and a half years ago, legend in the space community, worked at SpaceX. Also a car aficionado who engaged with my post about the Coogan test, which of course is the challenge to have a humanoid robot step into a Dodge Viper ACR and lap the Nürburgring in under seven minutes. It's been done by a human in seven minutes and one second and a few milliseconds.

1:51:19It's an extremely difficult challenge requiring physical endurance of the G-forces, but also a bunch of dexterity and obviously image processing. And I was talking to a bunch of AI folks about how long it will be until humanoids are actually at that level where they could do something like that. And it took, I think the consensus was around 2045, something like that. Anyway, Tom, great to meet you. I don't think it'll take that long, will it? How long do you think it'll take? This is a great question. When you take fear out. Yeah, yeah, maybe that's a big one. I don't know. You've done a lot more laps than I have.

1:51:53So please, what do you think? Haven't they already shown that fighter aircraft can beat a human in dogfight? Yes. So the difference there is that I'm adding the challenge of the Dodge Viper ACR must be stock. And so the humanoid must operate three pedals and the stick shift and the steering wheel. And that's a lot more challenging than just plugging into the steering column and controlling it like you do in a Tesla or Waymo. And so I think that adds a wrinkle. but I don't know, 2045, 2035, it's clearly going to happen. But that's a good way to say that you're bullish about humanoid robotics. Yeah.

1:52:40Do you anticipate that humanoid robotics will have an impact on your business in the next decade in terms of maybe manufacturing or something even more narrow scope? Certainly manufacturing. Sure. I think for manufacturing, robotics are developing fast. But I think humans should explore space. but it's really expensive and dangerous to put humans in space. So if you really want to use the economies of space, I think it's going to take a lot of robotics. So I'm bullish about robotics. We got to put a Nürburgring on Mars. And then ship the bike car on the moon, which I'm all about. Yeah, I would love a track on the moon.

1:53:16Let's use taxpayer dollars to put a racetrack on Mars. I would love that. Anyway, sorry, we jumped straight into it. Would you mind kicking us off with an introduction for both of you and a little bit about the company? Yeah, sure. Yeah. I'm Tom Mueller. I'm the CEO and founder of Impulse Space. I was one of the founding members of SpaceX, where I led the development of propulsion systems for Falcon 9 and Dragon and a little bit into Starship. Yeah. Left there in 2020 and started Impulse in 2021. And we have our four year anniversary here coming up next month. Congratulations. That's great to hear.

1:53:56It's amazing. And it's been doing great. And Eric, would you mind introducing yourself? Yeah, sure. Hey, I'm Eric Romo. I'm the president of the company. I've been with the company for the last two years. I've known Tom for a better part of, God, Tom, 22 years, I want to say now. 22, 23. So I was an early propulsion engineer at SpaceX in the very early days. And then went and did a bunch of non-space, non-propulsion stuff for a long time. And found my way back to Tom a couple of years ago. Amazing. Yeah. Yeah. I mean, as I describe it in kind of layman's terms or in terms even a venture capitalist could understand, you can get to LEO pretty easily, but you need impulse to get to higher orbits.

1:54:33Is that still kind of the bread and butter of the business? Am I describing it even remotely accurately? Yeah. At SpaceX, I think we really made LEO very accessible, lowered the cost, made it, you know, you get flights going there, you know, every other day. I want to do the same for everything beyond Leo. So make it much easier to get to high energy orbits, to get to the moon, to get to Mars, get to outer planets, anywhere in space. Like, let's make it much easier, much more affordable and, you know, really start the true space age. Yeah. When you were at SpaceX, did you have an idea of how valuable Leo would be?

1:55:11I think a lot of folks, Starlink took them by surprise, basically, in terms of not just how awesome that service is if you've been on a plane and used it, but also the business that was to be built there. Did that take you by surprise and then is there something that you can use as a concrete example of what is exciting about the higher orbits? Well, yeah, I'm old enough to remember the first time we tried to do Leo Constellations, you know, back in the late 90s, and actually worked with some companies that tried to do it. and you know it wasn't the right time uh technology has improved a lot especially low cost launch you know better bandwidth uh easier to manufacture spacecraft all the things so i kind of realized it was the right time when it was you know when when spacex first started talking about it it really clicked i'm like yeah we need to do this yeah uh and so in terms of the higher orbits.

1:56:09What are I mean, obviously, going to the moon is very concrete. Are there specific obvious business cases or or anything that would be, you know, inspiring or or immediately relevant that will just somebody with geostationary orbit is still probably the most important orbit because it's it's the 24 hour orbit over the equator. So you you just point your dish at that and and it's always there in the sky or your antenna um or you can cover the whole you can cover the whole earth with with three satellites wow so making access much more affordable to that i think is a huge help but then i'm i'm you know i've always said i'm really all about a lunar economy starting and even mining asteroids and things like that so those are the those are the things i'm most excited about question uh i'm sure you guys have gotten asked in every single pitch meeting, so I'll ask it.

1:57:06Why, do you think this is on SpaceX's roadmap? Are you guys partnered with them already? Do you expect them to want to prioritize this in the future? Or, you know, what does that kind of partnership or competitive dynamic look like? Why don't you cover that one here? Yeah, I mean, they're pretty busy, right? And so this is the equivalent of the Silicon Valley startup question of like, why doesn't Google just do that or Facebook just do that? better, right? They could, they can do anything. All right. They're technically an amazing company that's shown they can execute on pretty much anything they sink their teeth into.

1:57:37I think the question is, will they, and where would this be on their priority stack? So, you know, one of our products, Helios is really good at getting large payloads to high-range orbits like medium earth orbit and geostationary orbit. SpaceX can certainly do that with a Falcon Heavy, but they've been super vocal about the fact that they would prefer not to. And there's operational reasons for that, cost reasons for that. But so this is a space where we think that we can actually help them and other launch providers get to a thing that they're to some degree ignoring right now. And then on the small spacecraft side of things, we make a small, highly maneuverable vehicle called Mira, which is largely usable for Space Force.

1:58:13So we announced some contracts with them late last year and a partnership with Andrel to go to market with Mira. And that's the thing that's pretty far afield from anything that SpaceX is working on today. That Starlink is awesome, but it kind of just sits there, right? as a constellation, whereas Muir is designed to move very quickly as its core competency. Can you talk about the lunar economy and how you imagine that developing? Does it start with tourism or scientific exploration? Do we need an ISS-like base that's kind of taxpayer-funded on the moon first? Or do we go straight to mining or helium-3?

1:58:49I remember the movie Moon. but what what are what like what are the rest what is the tech tree look like as we build out the moon as a capability possibly all of the above like mining helium three certainly could be a good starting point because it's uh even right now on earth if you don't flood the market it's on the order of 20 million dollars a kilogram so you can certainly if you can get it on them i don't know how the part of getting it like you gotta mine apparently millions of tons of regolith to get it. But if you have it there, transporting it back is like one 500 to that cost. So it's definitely, it's economically doable, assuming that you have, you have a, you know, a can of it.

1:59:27But I think if, if, if you're doing megastructures or, or, or massive manufacturing in Leo pretty fast, you'll figure out that it's 23 times roughly more efficient to bring mass from the surface of the moon than from the surface of the earth, just because the gravity well is so much lower. and I think that's when the real use of the resources of the moon will happen just to get material to lower orbit. Yeah, I've always thought that the moon was also just a great staging ground for Mars. You imagine if you're up there, the moon's always facing Earth, so something goes wrong, you know, the sci-fi version, oh, there's a crack in your suit.

2:00:08You can just dive in an escape pod and you're aiming in the right direction, whereas Mars, you have to deal with the transfer window. It's a whole different problem, and if we can get really good at living on the moon We're probably gonna be able to get good at living on on Mars I am interested in kind of these questions about when the what one of the great stories about SpaceX is that it feels like it's a relatively smooth handoff of capability in terms of Launch capacity from NASA to SpaceX from the taxpayer to the private sector At the same time, I love that NASA put a helicopter on Mars because I can't imagine that would ever happen no matter how crazy the VCs are.

2:00:50I don't see that happening and being underwritten against any sort of 10-year, 20-year life cycle of return on investment. And so my question for you is like, what type of space projects should we be pushing into the, you know, the NASA world, the government funded world, the even the nonprofit world versus what is commercializable today? And do you see both as important going forward that we that we kind of have this, you know, taxpayer funded R &D, the really fringe scientific experiments, and then we commercialize things when they're ready? Well, I think two things really that I'd like to see from NASA mainly is answering the big science questions like, is there other life out there?

2:01:37And also let's explore what's out there. Like I just posted a picture this morning from the new Vera Rubin Observatory. Amazing, you know, just amazing stuff that you should continue to do, things like that. and but i think as far as develop they should develop technology that's that that takes a lot of money to develop that maybe commercial companies aren't going to invest in now like i think the next thing you know jared isaacman was talked about is is nep and i'm always touting that too as the next step in propulsion so rather than build a giant rocket that competes against people that already build giant rockets why don't they go develop the next thing you know next the next technology.

2:02:14I think that's part of the best use of taxpayer money through NASA is to advance science. How do you think about that? Oh, sorry. I was just going to say, fundamentally, I think NASA needs to figure out if it's a jobs program or if it has goals, right? I mean, I think that's a thing. Like, SLS made no sense in the commercial rocket economy. Maybe it did when it started planning 20 years ago or whatever, but for the last decade or more, it's made zero sense. But yet it's still kind of hanging on because Congress wants to fund jobs in Alabama, which is, you know, maybe a noble purpose. But, you know, is that the purpose of nasa i think is the thing that they need to figure out yeah how do you think about extraterrestrial life uh do you ascribe to the dark forest theory or uh like uh or do you think it's promising do you think uh you know we will have an answer to this in a few years do you think we're getting closer what what is your take overall i yeah we need we need to get we need more information definitely i i hope it's just life is extremely rare it's not you know some some great filter that's that's still out of us yeah yeah that'd be good what are the what are the the kind of upcoming events in the next one to two years in commercial space that you think can be major catalysts for the industry overall that maybe you guys are working on or that you want to see from the industry broadly well something that we're not working on but something that that spacex is developing is starship i i i really started this company realizing that Starship and other, you know, reusable rockets are going to be bringing massive amounts of cargo to LEO and it needs to go other places.

2:03:47So that's what's the premise of this company. So I'm excited about Starship finishing development and becoming a commercial, you know, commercially viable spacecraft. And then what we have really, you know, we have the mirror that Eric talked about earlier and energy is really for government and defense. And then And we have Helios coming on online here next year. We're developing it right now, which is basically adding a third stage to Falcon 9 and other mid-sized rockets. And that's what it can allow us to get, you know, more easily to high-energy orbits. So those are the two big things for us right now.

2:04:24Yeah. Super excited about lunar landers, Mars landers, all kinds of cool stuff. It's going to be awesome. I was just going to tack on, I think you're seeing one big trend is that Space Force, the rhetoric of counter space or space defense defense activities has really ramped up in the last six months. In the public sphere, there's not been a lot of investment in space on space defense. But in privately, SpaceSource has been talking about that a lot more. And then they started about six months ago really publicly, not coincidentally with the change in administration. And so the question I think for companies like ours is, do dollars follow that rhetoric?

2:04:57Is there increased investment in that sort of thing? And then on the commercial side, I think you have an awful lot of operators who are trying to figure out how do they carve out a niche and survive in a post Starlink world? How do they figure out what their space is as a commercial comp sat operator that, you know, once you have the dominant Starlink player in the market? I'm interested to hear about the challenges of developing rocketry or engines at different scales. I've heard this story that one of the reasons the Raptor engine was so successful was that it wasn't a mega project in the sense that you didn't need to bring in a crane necessarily to work on it.

2:05:37There were more smaller modular pieces. Is that true? And then I'd love to know about the challenges of scaling up and making bigger rockets. I was tracking, obviously, I want Starship to work as soon as possible. It feels like it's a really, really tough challenge, and it feels like it's a harder challenge than the Falcon 1 or the Falcon 9, which, you know, It did crash several times, exploded several times. But SpaceX is a much more mature company. And so you would expect that it's not just the first time that they're doing it. It's actually something fundamental about the larger scale engineering that's going on there.

2:06:17So can you share anything about that? We've become a lot better at developing rocket engines over the last 20 years, certainly. when we started SpaceX, there wasn't really, it wasn't like startups would do rocket engines. That's why I had such a hard time hiring people in the early days because nobody believed in it. And now there's many teams all over the world that are developing liquid rocket engines. And it's never easy. It's not, but it's doable. And modern tools are affordable and available. And certainly modern manufacturing techniques like 3D printing is almost a cheap code for rocket engines.

2:06:53but when you're doing something like raptor which is extremely high pressure full flow stage combustion very advanced engine the most advanced engine ever it's just it takes the difficulty to 11 so not surprising there's been setbacks also one of the hardest things to do uh and in launch vehicle recovery is getting back and upper stage you know you saw we had pretty good success with falconine getting back to first stage and certainly the flying the boosters back and land that awesome landing but second upper stages are coming in with a lot of energy so really hard um they're gonna solve it um it's it's just a matter of time yeah this time well it is rocket science after all you know i'm saying people don't people don't use that phrase it's not rocket you know this isn't rocket science anymore because we've gotten really good at there's a lot of stuff that is rocket science how many times i heard that come on tom it's not rocket science It is rocket science.

2:07:50Well, I'm sure you have a lot of rocket science to do. We'll let you get back to it. We really appreciate you taking the time to chat with us. This is a lot of fun. Yeah, thanks for the insights. Have a great rest of your day. Next up, we have Derek Thompson. We have some breaking news from Truth Social. Where news breaks. News breaks, apparently. Trump says, Iran has officially responded to our obliteration of their nuclear facilities with a very weak response, which we expected and have very effectively countered. there have been 14 missiles fired, 13 were knocked down, and one was set free because it was headed in a non-threatening direction.

2:08:24I am pleased to report that no Americans were harmed and hardly any damage was done. Most importantly, they've gotten it all out of their system, and there will hopefully be no further hate. I want to thank Iran for giving us early notice, which made it possible for no lives to be lost and nobody to be injured. Perhaps Iran can now proceed to peace and harmony in the region, and I will enthusiastically encourage Israel to do the same. Thank you for your attention to this matter. That'd be good. It's a good way to, it's a strong way to end your post. Now, if you want to come across presidential, just bolt on, thank you for your attention to this matter.

2:09:00That's great. Well, we have our next guest, Derek Thompson, the author of Abundance in the studio. Welcome to the stream. Great to be here. Thanks, guys. Congratulations on the move. Being a businessman now. You're a businessman now. I'm a businessman. I'm an entrepreneur. I'm a CEO. You're a CEO. I mean, I'm everything. You name whatever piece of the corporate suite that you want to name. I guess I am that. I'm CTO. I'm COO. There we go. It feels great. He's about to become a hyper capitalist. Vibe shift. Yeah. Yeah. Talk to us about the move. I mean, we're going to go and do a bunch of stuff, but maybe just start with a little bit of background and then the decision to move.

2:09:40Yeah, sure. So I've been running for the Atlantic for 17 years, which is honestly like freaking disgusting to even contemplate. How old were you when you joined? I want to say four. I mean, I'm much better about myself. The answer is 22. So I joined in September of 2008. Wow. After graduating from college, I was like a comms intern for a while. I'm currently the comms intern for my Substack. So it's right back to that old position. and I wrote for 16, 16.75 years at the Atlantic and it was great and the truth is like I've always had like a little bit of wanderlust I've always felt like I had a bit of an entrepreneurial bone on my body there was never a great time to leave because the Atlantic is really a wonderful place to work they let you write about my position in my case just about whatever I wanted tech science economics social mysteries and conundrums and finally I just thought you know abundance the book has become a lot of things um it's certainly become something almost like something that could be a full-time job like if i just wanted to reply to and engage with like the political feedback that would be its own job and part of this was like this is an opportunity to do something on my own to do something that's really independent not put the atlantic in a position where they're like a non-political organization with a writer who's doing some things that are explicitly political, but also like, I just kind of wanted to know, and this is what I said, I think my original essay, um, maybe you guys felt like this when you started this, um, this show, like, what does my mind feel like when I wake up in the morning with a new job, a new task for it, right?

2:11:16Like you work at like a, maybe like a VC, right? Or you work at a, maybe a private equity firm and like you wake up and like, how do you process the world? You process the world the way it's most remunerative to you at that private equity firm or at that VC. And when you change your job, you change your mind. New tasks, new questions, new mysteries are unlocked. And I wanted to do that for myself. I wanted to wake up with a new set of questions. And so that's what I'm doing right now. Was there a price that could have kept you at The Atlantic? Many people say everyone has a price, but it feels like legacy media is not set up to retain superstars.

2:11:50And I imagine... No, look, The Atlantic paid me well. And the book paid me well. I'm really lucky from that standpoint. There's not really a number, right? Like, you know, $1 billion. Okay. Now suddenly you're talking about something completely silly, but like I wanted the freedom, the freedom to say, all right, like I'm 39 years old, right? If I stay for another 11 years, I'm 50 years old. Like now I'm basically on the other side of the hill. At some point you want to like be in your sort of, in your, your electric prime and be like, what is my mind like? What is my life like when I give myself an entirely new task of starting my own company, of thinking about the world from the perspective of I'm writing for myself, not writing for this August institution.

2:12:37So no, I don't think there was a number. And I also think it's important to say like the Atlantic is one of these places that as much as independent media is growing, and obviously I think that's great. There are a couple empires that are going to the times is growing like crazy. The Atlantic is hiring like crazy. They pay the writers over overall, I think quite well. So this was not really about money at all. This was about being in a fortunate position to take a risk that had nothing to do with money. Yeah. If you just cared about money, you'd become an AI researcher and you join Meta, get a hundred million.

2:13:06Well, we got to pitch him on private equity because he sounds like he doesn't want to be a private equity guy, but I think scale to sub stack, start acquiring, rolling up other independent journalism. Could happen. Could happen. I do want to talk about advice from Ezra about this. Obviously you wrote the book with him. He's been at the New York Times. He's been started his own thing what was his advice to you as you kind of stepped out in this next uh next yeah that's an interesting question no one's asked me that question before so i told ezra three days before the book came out uh we had just done here i'll set the scene for you yeah we had just done um freed zikaria's show and it was one of the first interviews that we did together and we were in the hut you guys are based in san francisco we're in la you're in la okay cool so in new york hudson yards is that like enormous like skyscraper mall right and we go like get some food and we get some lunch.

2:13:54And we sit down and I say, Hey, by the way, just so you know, I'm starting to think a little bit about doing something independent. This was before abundance became what it was. This was before a single book had been officially sold. So I kind of think I want to do something on my own. And he said, you know why? Like the Atlantic's a wonderful perch. And I honestly told him, I told him what I told you. I was like, like the job of like being a content creator, right? It's a weird and wonderful job. Your job is to wake up in the morning and think, what do other people want from me that lives in here that I can put out here and give to them?

2:14:29It's a really weird and wonderful job. And I feel like that work requires a certain amount of churn and experimentation. If you're stuck somewhere, and I didn't feel stuck at the Atlantic, but I was there for 17 years. If you were in one place for 17 years, there's a certain amount of experimentation that you're essentially choosing to block yourself out of. Right. And so I told him, like, I really wanted to, to experience like, you know, what, what different kinds of questions, what different kinds of writing styles and podcasting styles, maybe eventually like, you know, um, video podcasting styles would, would eventually be unlocked by going independent.

2:15:03I think that's a lot of the, the appeal of Substack for people in a certain place or, or Beehive or some of these other places. Patreon is really getting into the game here as well. Um, it's this opportunity to be like, all right, what is it, what does it feel like to be a content creator for myself. I'm not answering to anybody else. And I'm taking on all of the responsibility and the stress that comes with, you know, being my own boss. I feel like that's an experiment that more people should run on themselves. What about the timing of the launch? You always hear about the big bump that comes from like a big podcast appearance or something like that.

2:15:38Is there a world where you think you should have been doing email capture in the lead up to the book so when you did the interviews for the book, people could just go subscribe immediately. Is this the ideal time? It seems like it's really well timed because the book has really, really done extremely well. And so you're in this capturing moment. It would have been probably much riskier to launch this a year from now. But did that factor into it? Did you think about doing it before or later? How'd you land on this time? Look, here's a framework that I thought about. It's not like the only framework in my head, except for one.

2:16:13obviously having a number one New York times bestselling book is a good opportunity to do something new because a certain number of people are already paying attention to you. Like there's a certain amount of obviousness there that I don't want to be like cynical or, or conceited or vain, but I think it's just flatly true. And there was a certain calculation that was like, if I do want to do this in the next 10 years, does it really make sense to think, am I going to have another, another number one bestselling book in five years? Like almost certainly not. Yeah, let's get a polymarked on it. Like it is incredibly fortunate and frankly, just strange thing.

2:16:49I mean, I don't think there's no reason for you to know this. Our book did not hit number one on week one. It didn't hit number one on week two or three or four. It hit number one on week five. Like a discourse built around the book allowed it to sort of like hover in the two to three range and then spike to number one in a way that my book agent said she's never seen this before. Like no book. She said, she said, I've had a lot of books hit number one. None for the first time in week five. So it was a really original, not original, really maybe unique experience in terms of the way that it hit the discourse.

2:17:22Was that a directly - Because of the popularity of the Democratic Party. Yeah, was that a product of like the podcast appearance strategy I saw Ezra going, and both of you guys going on like a really wide swath of podcasts and bringing the message to places. And I think a lot of authors, when they launch, they think like, I'm going to do like five big things on week one and then like, you know, I'm done. But it felt like you were kind of taking the message to all these different pockets of the internet and really capitalizing over a long time. And it didn't feel like it was necessarily like, okay, you got 50 % of the audience in week one and then 10 % and then 10 % and then 10%.

2:17:57It felt more like you're having different frames on the same conversation again and again and actually building that momentum. Yes. So if we were doing like book publishing economics 101, one. Yeah. Okay. The book publisher controls very little distribution themselves, right? There's very few channels of exposure that the book publisher controls themselves, right? Even if you love reading nonfiction, my guess is that you aren't going to randomhouse.com to check out their nonfiction titles every Tuesday when the new nonfiction titles come out, right? What do you check? The New York Times book review, podcasts that you listen to, other sites that review books, right um tv interviews those are the channels of exposure when you're a book author you're basically in a position of begging you're in a you you go to fridzakari and you say please please can you talk to me about this book and you go to barry weiss and you say please can you talk to me and you go to the far left podcast you say please go to the right wing podcast you say um you know ben shapiro please you're asking please please please of everybody and their assistants and the assistants producers and the producers assistants you're begging the entire world made this and And almost every author is doing that.

2:19:07Every author is in this position of begging because often other people control channels of distribution if they don't. Now, in Ezra's case, he arguably has, with the Ezra Klein Show, the medium of exposure that is most successful for launching books. So that's a unique experience for me and my co-author. That said, two things. One, I would guess that if you really looked into the numbers, what sold this book early on, despite the impression that you gave that this book was on a lot of different podcasts, what sells books is still television. And the reason why is not because television is bigger than podcasts in a weird way, it's because television is shorter and it's a worse product, but a worse product gives you a better amuse-bouche for the book.

2:19:56If I talk to Lex Friedman for four hours about abundance, exactly how many more hours do you want to spend inside of abundance. That's a great point. If you talk to Freed Zakaria for 10 minutes about the book, it's this little amuse-bouche that makes people think, ooh, that sounds like exactly what I want. The future of the Democratic Party, the future of liberalism. So my feeling is that while podcasts are important, because in a weird way, the product of a book is even more the conversation about the book than the product, than the book itself. The best way to sell a book, weirdly, even in this like podcast heavy moment is still television.

2:20:34That's my, that's my unempirical vibey bet right now. That's fascinating. Yeah. I love that. Yeah. I mean, I noticed that with, I mean, Ezra on his, I believe it's his own branded podcast channel, put out a 20 minute video essay outlining the thesis for the book. It also, that video did very well. I believe it went viral and got maybe millions of views, that feels like number one bestseller on that week. But I think that happened on week one, not on week five, which is interesting. Which is weird. Yeah. I mean, so what happened was John Green, the author of The Fault in Our Stars, who's become a nonfiction writer, wrote a book called Everything is Tuberculosis.

2:21:16And that book, I don't think I can share the inside numbers that I have. But he smoked it. We did great week one. that book kicked our ass we weren't even we weren't even close like i don't know if you guys what kind of sports fans you guys are like that book was like the miami heat that book was the cleveland cavaliers we were the miami heat like every single blowout it destroyed it so like we weren't even close number one number one um and that does go to show that that with book selling it's all about channels of exposure yeah and there's already a celebrity owns their channel of exposure. And so the book publisher, by giving them a lot of money, is essentially buying their channel of exposure.

2:21:56And that's partly what they're paying for. But yeah, the strategy, you know, beyond strategy, there's this question of like, how do you sell? How do you sell a book? And there's also this question of like, I don't mean to get like weird existential, but like, what's the job of a book? Like, what's its job? And I think the job of a nonfiction book is to get people talking about it. Of course, you want people to buy it. But like, Or start your presidential run. Or start your presidential run. Honestly, you should do it the other way around. Because the best way to get attention for yourself, like if you dudes, if you guys run for president as a ticket, you'll get a ton of media exposure.

2:22:32A ticket? We haven't even thought about that. That's hilarious. Then you'll build. Then people will buy our book. It's totally free. It's totally free media. Because all these people, the New York Times is like, what are these guys doing? And the Washington Post is like, what are these guys doing, right? You get all this free media exposure. Then everyone knows your name. Then you go to a book publisher. then they give you a lot of money and then you have this this um platform that's been co-constructed by media organizations that were conspiring to build it even though they didn't even mean to so you should not write the book then go for president you should do it the other way around but the the product is the conversation um and so the reason to go on podcast the reason to go you know on on shows like this yeah because like you want these ideas to get out in the ether like No one should write a book if they don't want to talk about it for 50 ,000 hours, because that is the experience of writing a book.

2:23:22You finish it. And then if you're lucky, you talk about it for the rest of your life. Okay. Well, talking about things actually in the book, this has been fascinating so far. I've really enjoyed this. But one of the narratives that I keep coming back to is the story of California high-speed rail. I grew up in California. I have friends and family members who have worked on that project, actually. And I think for a lot of technologists and Californians, the comparison between SpaceX and the California high-speed rail story really draw a very clear conclusion in people's mind, which is that the private sector is more efficient than the public sector.

2:24:00And I'm wondering how you, do you think that that narrative of, you know, one was literally rocket science, but at the same time, we had put things up into space. It wasn't entirely a new capability. The other was a train, which, you know, we've had trains for generations, and we tried to build a train, and we tried to build capability to get to low Earth orbit. One feels very successful, and the other feels like we didn't get anything out of it. And I think for a lot of people, the takeaway is let's not have the government even try and do things. Let's let the private sector do as much as possible.

2:24:35And that trade-off between if there is an economic feedback or there is a profit motive or a payback, that will act as a better incentive to actually get the result that people want, which is transportation or the ability to get to space. How have you grappled with that anecdote specifically? I'm a liberal who believes in a large role for private sector innovation. There's no question that SpaceX is an extraordinary company. that has done things for rocket technology and especially for rocket efficiency that no other company in the world has done, no other individual in the world has done. Yeah.

2:25:15How many people has moved to space? A lot. How many? Like - Dozens? A thousand? Hundreds? I don't know. It's just the ISS. Do you know how many people are in California? This is our Iran population question. Yeah, we're getting out. 50 million, right? Yeah, yeah, yeah, in California. To build infrastructure - Yeah. that is a public good for 50 million people is a different challenge than building technology whose proof point requires several dozen. These are entirely different challenges. And that's why we need - Yeah, I would say like SpaceX has provided internet infrastructure for billions of people.

2:25:53Yeah, the satellites kind of count as the people in this case, because that's the benefit that people really want out of. The public benefit is that people in Iran right now can access the open internet. And I think the value of Starlink is sensational. But when you're talking about, I mean, when you think about broadband technology, broadband technology is already being built by a lot of private companies. So it's not as if the creation of Starlink is taking away what has historically been a public service and is creating a private company out of it. Like broadband companies exist, like Charter exists, Comcast exists.

2:26:34So I think it's great that there are private companies that are working on that. When we're talking about questions like public goods, when we're talking about roads and bridges, things that are frankly quite difficult to monetize. And it's not entirely clear that we want to directly monetize them. Do we want to charge poor people a certain fee every time they use a road or a bridge? Or does it make sense to live in a rich society with a progressive tax structure such that we can tax the rich a little bit more than the low income and use that money to create a set of public goods that makes life better for everybody, that allows, frankly, low income people to use those roads to drive into areas where they can provide services for high-income people such that everyone turns out better off.

2:27:20I see a large, large role for the federal government and for state and local governments to look out for public goods. And to me, when I say public goods, and you look at the actual budgets of the federal, local, state governments, overwhelmingly, what we're talking about is health care, subsidized health care. We're talking about education, and we're talking about infrastructure and transportation, things like roads, things like bridges. I think there's a huge role for government to play. I like having both. I like having a mixed economy, right? Muscular government. And also even more, I would like to see in the book is probably about this, even more purposeful pro tech, pro science strategy, because the truth is if you look at what SpaceX did for rocketry, why haven't we done that for Alzheimer's?

2:28:05Why isn't there like a SpaceX for like chronic inflammation or SpaceX for carbon capture technology. So I see a huge role for both of these things. I don't think that the success of government in some places proves that the private sector can't do certain things very well, or the success of SpaceX means that the government should get out of the business of building bridges. Yeah. Yeah. I agree with you on the roads and bridges thing. I guess my worry is like the free market solution to how do I, the real problem that you're solving is like, how do you get people from LA to San Francisco? How do you move people around California?

2:28:37And And the free market solution is a one hour Southwest flight that costs a hundred bucks or a Camry that drives you there and maybe costs$20 in gas or something like that. And so the free market's kind of solved like if you have six hours, you can do it in 20 bucks or if you have one hour, you can do it for a hundred bucks and you kind of choose those. And we kind of came in with like, we want something that's cheaper for two hours, but there wasn't this like massive market pull. And so that it was more like the decision to, to, uh, to create that interim solution, uh, just, just never really happened.

2:29:13But I mean, we, you know, just, just, just stop there. I love Southwest. I love flying on airlines. I, at no point in my life. And I'm like, what we need to do is like nationalize, you know, like I do that. But like who built LAX? Totally. Right. Like you need public infrastructure or infrastructure that is partly publicly financed. in order to create a kind of platform technology, a platform for the private sector to succeed. In the same way, if you look at the history of Elon Musk's companies, what is the total number of government contracts and loan guarantees that SpaceX and Tesla have benefited from in order to become the companies that they are?

2:29:52I'm pretty sure it numbers in the billions, especially when you count all the NASA contracts. I see a huge opportunity for a public private partnership to give it some of the outcomes that we're most proud of. Yeah. With the California high-speed rail example, the question, I guess, is, was it a mistake for the government to pick high-speed rail as an important project? Like, was it a misranking? Should have spent more money on roads and, or was it just the execution? I mean, billions of dollars were spent on a rail system. I mean,$33 billion were essentially tacitly approved by a bond that was voted on by the public.

2:30:30And several million dollars were actually spent constructing rail that isn't connecting anything to anything right now. So clearly, if you were looking back, the total number of man hours spent and dollars allocated on high-speed rail has not been worth it for people who live in California. I don't want that failure to indict the good that the public sector can do any more than I think you would say that the failure of one SpaceX launch proves that SpaceX is in fact totally overrated. Switching gears a little bit, how much are you digging into general research funding? We had Andrew Huberman on the show last week talking about his concerns around NIH.

2:31:10Tell me his concerns. Can you actually, I didn't see that. He had a long interview with Jay Bhattacharya. and essentially the proposal is to cut government funding to university labs and scientific research labs by something like 40%. And that's a pretty significant chunk. He was very worried about this. There are some unnecessary costs potentially that he highlighted, but in general, this is the research that cannot be underwritten by the private sector because it does not, If you're just doing research on how do we fall asleep or how does the brain work at a very abstract or what is DNA? You can't patent DNA as soon as you discover it.

2:31:57And yet, as soon as you've discovered it, boom, there's a huge, huge, massive boom in biotechnology. But it won't be captured by a single patent holder on the underlying. Yeah, so that just feels like an area that you could really dig and dive into in an area where it's okay to be kind of skeptical of this hyper capitalist viewpoint. And, you know, venture backed companies can solve everything when some of the smartest people in the world are. It's broad. It's a ton of funding of a whole bunch of different research areas. They all can't be underwritten by the private sector in any way versus being very prescriptive.

2:32:33We need to train from here to here and we are going to hire the people to build it. that feels like much, much harder for the government to get right. I agree with the way you guys are thinking about this. And maybe it means I agree with the way that Andrew's thinking about it as well. I think what's happening at the NIH right now under HHS is horrendous. I think if what you wanted to do was to reduce America's annual deficit by$18 billion, which is what we're cutting from NIH, you could either cut NIH, the fountain of scientific discovery for not just the US, but the entire world. You cut NIH by 40 % or you could cut the corporate income tax cut by 0.4%.

2:33:12Why not just kind of reduce the size of the corporate income tax cut by a microscopic percent and still fund scientists to discover the basis of Alzheimer's and chronic inflammation and multiple sclerosis and pancreatic cancer? You said it really well. Knowledge can't be patented, period. Like when the first scientist who looked inside the Gila monster's mouth discovered that there was a little hormone there that seemed to suppress its appetite and regulate its insulin, you can't patent a lizard's tongue. But our knowledge, our subsidized knowledge of the liquid on top of the lizard's tongue gave us GLP-1 drugs, which now appear to be something like a miracle drug for everything.

2:33:55I mean, they don't just help people lose weight. They don't just help people reduce inflammation. They literally seem to, there's a study going on right now that they might reduce rates of Alzheimer's because Alzheimer's might be caused in part by inflammation in the brain. And if GOP-1 drugs work by not only reducing obesity, but also reducing chronic inflammation, they might ironically or incidentally be a neurodegenerative drug that works through the gut brain interface. that's fucking crazy. And we do this by looking at a lizard and no smart VC. Like Mark Andreessen's brilliant, but he would be an idiot if he told scientists, if he gave scientists tens of millions of dollars to just look in the mouths of a bunch of reptiles.

2:34:37That's dumb. That sounds really dumb, but it's really wonderful to have a publicly funded organization that asks scientists who simply discover the truths of the universe with the hope that if it's true, someone along the line can find a way to make a product with it that makes money and saves people's lives. So I think we need a really aggressively funded NIH and NSF. I think what's happening right now in the Trump administration is really, really sad. And Jay Bhattacharya, if you're out there, I would love to talk to you about what's happening at NIH because I know you a little bit. We've talked a little bit.

2:35:11I know you don't want this deep down. And I would love to talk about how we could amend NIH policy to truly make Americans healthy again and make America great again, because cutting NIH by 40 % ain't it. Yeah, I think the frustration from Andrew was seeing the increased spending and trying to reckon with the two decisions there. Let's switch to something less politically charged. Can you give us the highlights of your conversation with Mom Donnie? You released that this morning right before we got on the air, so we haven't had a chance to listen, but I'm sure it'll be entertaining. look um let me uh give you my reaction as a like a person and then it's like a political thinker um he's unbelievably charming and engaging and um and thoughtful like you know when you know when you're talking to someone who's listening or who at least is very good at performing the act of making you think you're listening to them like i only know this guy for 37 minutes but um i loved our conversation.

2:36:17I really, really loved it. I think he's wrong about a lot. I think rent freezes are a terrible idea to increase housing supply. I think fundamentally, you cannot try to cap the price of something whose supply you're trying to increase. The way I put it in the podcast is I was like, you want grocery stores in New York to be better. If you pass a law that capped receipts at grocery stores at$50, what do you think grocery stores would start doing? They would stop stocking the shelves because you can't make money selling nice stuff to people if it's illegal to make$51 on the receipt. So in the same way, if you cap the price of a good and you're trying to increase its supply, you're tying the hand behind your back that you're trying to use.

2:37:03He was respectful, but we disagreed about that point. We disagreed about the fact that I think I'm persuaded that while I'm not against public sector unions in the big picture, If you look at why New York City subways have among the highest transit construction costs in the world on the per mile basis, one part of it is that their union staffing levels are four times higher than the international average. You cannot build anything efficiently if the first rule of your company is your staffing levels have to be four times higher than all your competitors. It just doesn't work. So I asked him whether he was interested in talking to public sector unions, which are a Democratic ally, of course, if he was interested in building public transit efficiently and increasing public excellence.

2:37:47Again, he disagreed. That said, I was really impressed by his insistence that he wanted to be a politician who was outcomes first and process second. And there was a certain amount of ambivalence that he seemed to communicate to me. Again, maybe he's lying. I don't know him that well. He's a nice guy on the phone. There's a certain amount of sort of procedural ambivalence that I really like in a politician. Like if Jay Bhattacharya came on your show, right, or if RFK Jr. came on your show or whoever else in tech policy, someone who's doing AI policy, and they told you that the outcomes that they wanted were the outcomes that you wanted, they were ambivalent about what process it took to get those outcomes.

2:38:30but they were going to try their damnedest to make sure that when they left office, those were the outcomes that everybody could agree about, right? There's something that's very refreshing and non-ideological about that perspective. And I heard a little bit of it. That said, I'm - Is that tying into changing his mind around how police should be funded? We talked about him changing his mind on two things. One, he told the New York Times he's changed his mind in the role of private developers in housing. I think he used to be more on the side of like the government should build up the houses. I'm more on the side of the government should fund some of the housing construction, but like we've got private companies that are really good at building housing.

2:39:05Let's not do what the private sector is already doing too much. But the other thing he changed his mind about, as you said, is defund the police. He talked four years ago, five years ago about defund the police being a slogan that he believed in. And now he, really his policy is fund the police. He thinks that police don't solve crimes fast enough and that in many cases they're forced, and this is clearly true, to act as social workers rather than police officers. And so if you've got a police officer or say 100 police officers who have to basically be street therapists and social workers all day long, how are they going to solve crimes?

2:39:38How are they going to stop crime? And so he said he's changed his mind there a little bit as he talked to more police officers. And the end of our conversation was, I thought, a really refreshing point that he made that he thinks that his job, he sees his job as surrounding himself with people who aren't too fast to say yes to all of his ideas. He likes being around people who challenge his ideas. And I think that's a useful thing for really just, I mean, anybody, any politician or executive. And so there were many things in the conversation that I was impressed by. But also fundamentally, I think it's just important to say there are very big differences between socialism and liberalism.

2:40:14And he's a socialist and I'm a liberal. So we're not going to agree about everything, but you have to be able to talk to people you don't agree with. What are you hoping to see out of tech, the intersection of tech and Washington in a post-Elon White House world? You know, we got to get beyond this insanity where the administration attacks on progressive thought become attacks on the kind of talent America needs to train and attract and retain to keep our dominance and to keep the really valuable, information clusters that we have here. If we attack Harvard because we're pissed off at them and force all of the international students to leave, or if we encourage some of the smartest people around the world who deep down want to be at Caltech, want to be at Harvard, want to be at Duke, and we push them to be in London and Beijing and Shanghai, we are losing so many great ideas, so much agglomeration, the interaction effects that happen when you put a lot of smart people together in America.

2:41:28Like, it's important that Silicon Valley is in California and not in Brazil or France or Poland. Like, there is a special sauce to America, I think. You want to attract and retain these people. I want us to rationalize our talent and immigration policy and not have it be a kind of, unfortunately, like assassinated bystander in this war between the Trump administration and a progressive movement that they think has gone too far. I think we really need to get back down to a reasonable basis on immigration policy in particular. And then I would like us to add back the funding to NIH and NSF. I'm sure Andrew talked about this a little bit.

2:42:10Is it possible that there are so-called indirect costs, money that I work at Harvard, let's say, you're NIH, you give me some money for my Alzheimer's study, you also give me money for Harvard. Is it possible that Harvard money, that those indirect costs are too high? Yeah, sure. Let's have a public debate about that. Let's have public testimony. Let's not cut NIH by$18 billion and say, well, hopefully we're only cutting the fat and none of the muscle. So better immigration policy and better science policy, number one. I wish I had a good answer on AI. I don't know what the hell the next five years look like.

2:42:46I feel like the smart people that I follow change their mind every six months. Sometimes they accelerate, sometimes they decelerate. I wish I had like this sort of crystal ball that could tell me what like optimal AI policy coming out of Washington looked like. Right now I don't, but hopefully in the Substack, I'll be able to get some answers there. Well, we're excited to follow along. That was great. Thank you for joining. This was great. Thank you so much. Yeah, we'd love to have you back. We could talk all day. All right, great. I'll do it. Have a good one. Awesome. Later. Bye. Sorry, I went way over.

2:43:14I thought we were five minutes over. We have a light system in the studio, but the golden retriever does not concern himself. The lion does not concern himself with who's in the waiting room. We have our next guest from Juniper Square. Thank you for waiting. Great to talk to you. And we got some news. We got some news. We got some news. What do we got? Happy to be here, guys. What do we got? We got two things. We got a$130 million Series D. Congratulations. I've been waiting to do that all day. Glad we got it out. Yeah, and then we launched our AI product that we call Juni AI. Congratulations.

2:43:59That's our second bit of news. Now that we got the important stuff out of the way, who are you? What do you do? Yeah, well, I've become a fan of what you guys are doing. I appreciate it. I'm glad to be on the show. It's really interesting and unique, so keep it up. But yeah. Yeah. So my name's Alex. I'm the co-founder, CEO of Juniper Square. Thanks. And Juniper Square is an 11-year-old technology and services company. All of our customers are private markets GPs. Sure. So venture capital firms, private equity firms, real estate investors, crypto, private credit, you name it. If they raise an outside fund, then we help them with fund administration.

2:44:42We help them with technology to manage that fund reporting. Basically everything is not the investing. And we've got close to 2 ,500 customers now, trillion dollars that were responsible for stewarding and about 750 employees. 2 ,500 customers. Is that not all of them? How many investors are there? I feel like I saw that number on your website. You need to update it, by the way. It says 2 ,000 on your website. You're at 2 ,500 now. Yeah, we got to get those numbers out. But I mean, I feel like there's, I mean, we try and talk to, I thought we talked to all the VCs. We've had, we've done hundreds of interviews.

2:45:15I thought we talked to them all. Apparently there are more. There's some left. How big is this market? It feels like, like, you know, you talk about private markets investing. That feels like a small market. Yet Series D is at a billion point one. Like you're building a big business. How big is this market? Right. So the, you have to kind of take it by asset class, right? Because VC is relatively small as an asset class, both in terms of the capital that it manages and the number of GPs or investment firms that are out there. Private equity and real estate are respectively probably five times as big as venture.

2:45:55So there's thousands and thousands of, you know, just think of all the middle market, private equity, kind of LBO, growth focused firms out there. I think of all the different real estate investors focused on apartment buildings or retail or whatever. There's probably about 20 ,000 GPs globally that matter. Was the Zurb crash hard for you? Was that a slower growth period for your business? Because it seemed like it was really rough for all the VCs and a lot of investors pulled back. But it's not as like they really shut down their funds. Yeah, it's more like net new fund creation slow. So yeah, was that a real driver or are you just going after the existing players and so the new ones don't necessarily matter as much?

2:46:39No, we definitely felt it. And you're right that what we're sensitized to is in a given quarter, a given year, how many new funds were created. Our market share is so low that we still can sell to existing funds and people switch their providers all the time. But it's always easiest to get the new funds that's raised. So we had the double whammy of, you know, big focus for us is commercial real estate. So when COVID hit in spring of 2020, you know, it just took out retail, it took out hospitality and hotel and commercial real estate was, it was really tough slaying from like 2020, you know, really until 2022.

2:47:16And then the rate hike started and commercial real estate's, you know, heavily CapEx intensive, it's super rate sensitive. And then we had the big bust in, you know, public markets correction. We had the big overhang of everyone investing at 100x revenue multiples in 2021, 2022. We're going to bring those back. Yeah, we're working on it here. The sober days of 100x revenue multiple, which we're kind of back to now. Yeah. There he is. How do you get through all those times? It's recovering. It's coming back. How did you get through the tough times? Did you have to do layoffs or change the direction of the business?

2:47:53Were there any pivots in the story? Specific meditation, track.

2:48:02yeah all of the above you guys we we had to do some layoffs for sure like everybody i mean we were going into you know the fed started hiking rates in like late 21 but by spring of 22 is when we had that huge correction yeah in the public markets and by that summer it was clear that companies like us and pretty much every high growth late stage tech company was spending too much, right? We were all investing for revenue growth and we all had to pivot to invest for growth efficiency. So we did a round of layoffs then like a lot of people did. It was the first time I ever do that as an entrepreneur.

2:48:38It's not fun by any means, but it was necessary. But then the big thing for us is, you know, we started out as a technology company, but then we added in 2019, we moved into the world of fund administration and we became a full service fund administrator. And fund administration is like a really critical market for private capital. There's probably 30,$40 trillion of private capital out there in the world. And almost all of it needs a third-party administrator. So we kind of moved into this market that was like 50 times the TAM of our previous market, concurrent with this big correction happening.

2:49:16And so those forces ended up working out that the company did really well. and we continued you know we grow we've grown really quickly uh through a really challenging market period yeah um so it wasn't without its bumps and bruises but it's overall it's been really well yeah what talk about uh the new ai products uh i imagine it's a difficult one to build you can't really like lps and investors can't really tolerate hallucinations you know and uh i can imagine it you know a product yeah i don't mind if you leak my fund returns to my direct competitor no problem less of that but if you're trying to understand fun performance and things like that or reporting you can't like oh i'm sorry i misreported tvpi yeah yeah yeah so that is the crux of the challenge is um i mean as you guys know you talk about it a lot on the show you know generative ai is probabilistic right so next token prediction is is probabilistic in its nature and the kind of work that we do of reporting on financial assets is deterministic, right?

2:50:23Like one plus one always has to equal two. It can never equal, you know, one or three or anything like that. And so the real trick is, and if you guys ever know, if you've ever like taken a pile of invoices or something and uploaded them to ChatGPT or any LLM and been like, hey, tell me the, you know, invoice amount for invoice number eight, it's terrible at financial data extraction. And so the big hole that we have in the market that we're filling for our customers is if you're an asset manager, if you're a financial services firm, like a venture capitalist or a real estate investor or whatever, one of the key things you have to do is you have to report to your regulators.

2:50:59You have to report to your LPs. That reporting is like, you know, you're putting your brand and reputation and your ability to raise more money in the future and everything behind it being accurate and correct. And so the trick is how do you leverage all of the power of generative AI? It's like all these summation use cases, document extraction, the chat interface that we've all become so accustomed to. How do you fuse that with a set of deterministic tools that do all the precision stuff, the precision math so that like you're calling functions in Python to run some code to answer a question?

2:51:34You're not trying to have a next token prediction be the answer of that. And that's basically what this platform is. It's multi-model. We have a bunch of different models running inside of it. and the idea is think of it as like a model orchestration layer where if you're a private markets gp there's a lot of work you got to get done you want agents for different things different models fit different use cases and you need a way of orchestrating all of that work ensuring it's compliant ensuring it's accurate and that's what this layer does anything else this is great yeah we yeah we gotta have you back on we're running behind today as you saw But we really appreciate you stopping by.

2:52:15Yeah, congratulations on making it through the fire and come back on when you have more news. Yeah, we'll talk soon. Happy to be here, guys. Good to see you and the team. Cheers, Alex. Bye. Really quickly, how'd you sleep last night? I'm on a run. I'm hopeful. I'm on a run. 94. Let's hear it from me. If you beat me last night, I'm going to be extremely. 86. Oh, let's go for John. Go to H3.com. Ever since I got sick, I've been trailing you by like 10 points. Off your game. Well, we got Jesse from Decagon in the studio. He's certainly been on his game. How you doing, Jesse? What's going on? Welcome.

2:52:53Welcome to the stream. Good to see you. Thanks for having me. Yeah, thanks for hopping on. Give us the news. What does the company do and what's the breaking news with you? Yeah. Hey, everyone watching. I'm Jesse. I'm one of the co-founders of Decagon. If you don't know about us, we're essentially an AI customer service agent. And so what this means is, you know, imagine next time you're booking a hotel room or something, right? There's an AI agent that's there that you can call or chat with live and it knows everything about you. It can book rooms for you. It can upgrade. It can answer questions about your loyalty points and so on.

2:53:25And so essentially what we do is we work with these large, usually consumer businesses that have a lot of customers and a lot of contact center volume. And we're there to both drive a ton of efficiency for them because now their human agents don't have to deal with these more mundane issues. But also, we are able to give a better customer experience for the actual customers. So that's what we do. And then today, we're very excited to announce our Series C, co-led by A16Z and Excel at a 1.5. Let's go. Give it a bigger hit, John. Oh, there we go. Talk to me about the early beachhead for the company.

2:54:08I feel like LLMs are so generalizable, and yet go-to-market is so specific. Every vertical industry has slightly different customer service needs, whether it's a Shopify store that needs access to an inventory management system or refunds or Stripe versus a SaaS product. What has been most interesting to you? Where has adoption been the strongest early on? Yeah, it's a good question. So I would say that with most of the use cases out there, the ones that have had the most impact from AI has been ones where there is a little bit of complexity, right? Where you need the AI to go and take actions for you or look up data for you.

2:54:49And it's kind of adding this extra layer beyond what traditional chatbots have been able to do. Because I'm sure both of you have used chatbots in the past. And usually it's a pretty frustrating experience. because it's horrible. Yesterday, I was trying to cancel. I was trying to cancel spectrum internet at our old office. Good luck, John. Couldn't do it on the website. This is 100 % deliberate by them, by the way. Couldn't do it on the website. Of course, like, you know, this is a crud app. I should be able to click a button and update the - You have to wonder if it's intentional. It's 100 % intentional.

2:55:21They broke one password, so I couldn't log in. Finally, they say, hey, text us. And I'm texting with their chat bot. It texts me natural language, but then asks me to click a dropdown for two different options every single time, finally get through all that flow. And they're like, you need to call us. And by the way, we're closed today. It was awful. It was the worst thing. Yeah. Is that some maybe friction in the sales process where people are like, I get it that you want to make this more frictionless for our customers, but you're going to increase our churn 10%. I don't know if we should sign up for this.

2:55:52Are there actually trade-offs there? I'm super interested. Yeah, no, that is one use case, right? So to your question, what use cases are good? It is where you need a little bit of complexity because in the old days, the reason why you can't do that today is because it's really easy to game that system. Oh, sure. They made a decision tree. Then people, someone would make a Reddit post about, hey, here's how you actually cancel easily and everyone will just do it immediately and it impacts their revenue, right? So the nice thing about LLMs now and the thing that it really unlocks is this level of nuance.

2:56:19You're able to create much more complex flows and sort of situations that are tailored to specific users. And because of that, you're able to cover way more ground, right? So we have customers, for example, that use us to process refunds. We have customers that use us to, you know, either book flights or reorder a credit card or things like that. And that's really powerful because it just wasn't possible before, right? Before, you're kind of in this more dumb decision tree that you have to get forced down a certain path. And maybe that path is kind of what you want, but you want something slightly different and you're just forced to go down the path.

2:56:59You get stuck and you get escalated, right? So that's one of the main benefits of these LLMs. Last question from my side. I know we're going crazy on the timing. But in my previous jobs when I've hired a really all-star customer service agent, the thing that's always impressed me is when they're able to expand out of their role as just customer service, just doing specific flows, and actually turn into sort of a salesperson. and actually turning the interaction into customer education or upselling or cross-selling or, you know, hey, yes, you said you wanted to cancel, but I think you want to cancel for a reason that we could actually alleviate by switching you around or giving you a discount or doing something like that.

2:57:41Some of that can be deterministic. We've seen this with like, are you sure you want to cancel? We'll give you 20 % off. But the real best customer service agents tend to do it with more touch is that something that you're seeing already being able to be handled by llms do you think that's coming with more agentic workflows um how do you see the role of the customer service agent or human evolving over the next couple years we work on with our customers a lot on that exact concept which is what is the definition of customer experience over time because when you start you you're generally working on the low hanging fruit like yeah you know processing transactions or looking up information and FAQ, right?

2:58:19Right, exactly. But if you end up having a really, really good AI experience, and if you just think about it with the products in your life, right, you're just going to interact with it more. Like, it's not that you're going to, it's like, oh, I'm done with this. If it's a good experience, you're going to interact with it a lot more. And that opens up a much broader surface area for what customer experience means. You can start being proactive instead of just reactive. You can start looking at revenue generating use cases like you just mentioned, right? And it doesn't just have to be a cost center, which is what a lot of enterprises look at customer service, right?

2:58:46It's like, oh, it's this thing that it's not really strategic to us. We're just going to put it off to the side and try to save as much money as possible. Yeah. AI is changing a lot of that. And, you know, a lot of the customers that use Decagon are super excited by this concept that, hey, if you have a good system there, it almost becomes essentially a concierge for your product, right? It becomes like a new UI that customers can use to interact with you. Totally. Because if it's really good and it can do everything, like, yeah, maybe people are just is chatting with it instead of using an app or kind of going through a bunch of, you know, dropdowns and stuff like that.

2:59:16So, yeah. What's happening in the call, sorry to interrupt. What's happening in the call center BPO market right now? Or is there already pretty massive impact from companies like Decagon, Intercom, Sierra, et cetera? The, yeah, the call center market is definitely thinking about this space a lot, as you can see, right? When AI first came out, a lot of their stocks went down by quite a bit because it's one of the most obvious use cases for AI. I would say the best BPOs out there are really figuring out how to become AI native and either partnering with solutions like us or spending a lot of effort in building their own.

2:59:56I think partnership would probably be the realistic path because when customers, when end customers buy these services, they usually kind of see BPOs as a service provider. So there is room for partnership. But that's typically the way it will happen. Yeah, I mean, just the advent of software as a service or just websites in some ways reduce the amount of humans in a sales interaction because you could go through the Shopify checkout by yourself and yet there was still a role for the humans. So I'm sure that will be evolving. But it's exciting stuff. Awesome. Thank you so much for stopping by.

3:00:31Congrats on the round. Congrats on the round. And congrats on picking up Max Skarnecchia a couple months ago, star sales leader. I went to college with Max. He's an absolute Chad. He's telling me about you earlier. So shout out to Max. Thanks for coming on. We'll talk to you soon. Cheers. Have a good rest of your day. Next up, we have Yapsine. Welcome to the studio. The chat has been waiting. Yapsine. How are you doing? What's going on? uh you are on mute no audio yeah i do love we got the video we got the weight set up in a weight racker looks like hello test yeah we testing how you doing testing good I haven't set up my DAC yet so no you're good you're good what's happening let me just ask let me just check like let me just look at something on sure yeah well we'll give the basic background on you for the audience while you figure that out no He needs no introduction.

3:01:34He needs no introduction. Founder of Dingboard, a web-based... Of the hosts of... John and Jordy. John and Jordy. Hey, how's it going? Nice to meet you guys. My name is Yasin. Apologies. Good to meet you. Hello. Thanks for the introduction. I appreciate your time as well. And I'm actually a huge fan of you guys. Someone I used to work with introduced me to you guys. So I found a lot of enjoyment. That's great. That's great to have you on. How have the last few days been? A little bit stressful, to be honest. I've slept three hours every single night because posting on Twitter has been so fun. And unfortunately for...

3:02:18Fun or stressful? Yeah. So both fun and stressful because I can't stop posting. And that's not good for me. But I will basically just not stop posting. But if you keep posting the way you have been, could you potentially replace your salary? you were making in the seven figure mark you got to really ramp it up a little bit i used to go on 4chan and he's like people will be like yeah i'm like making like a million like a two mil at these ai companies i'd be like no way there's no way you guys are making two million of these ai companies and i started making like seven digits at an ai company i was like wait like these guys weren't actually lying like you actually make that much money at these ai companies um and it's like pretty fun work too as well so yeah for what it's worth like i didn't work on ai i worked on like bugs that really annoyed me on the app and that's why i joined actually so i joined x because uh There were a lot of bugs that really annoyed me.

3:03:01I fixed quite a lot of them. So I'm pretty happy with the work I've done. X is the product I use that I find the most bugs, yet I'm not just, I don't abandon the product. Like, that's how addicted I am. That's a super interesting thing. And it's because there's like, when it's a very large app, right? So the scale is huge. And you guys are like right tail users, right? Like you guys are famous. You know, the podcast of the Tech Bro Podcast. Sorry, the host of the Tech Bro Podcast. You guys are famous. So the amount of notifications you guys get, the amount of replies you guys get, The amount of DMs that you guys get is an insane amount.

3:03:31You guys are so far into the right tail. But how many people are there like you? That's why it feels buggy for you. But for most people, like an engineer who's creating a test account and clicking around, it's not going to be buggy for them. And for what it's worth, this is true across all the apps. Power users will always experience a lot of bugs unless they have someone to really talk to them and understand what their pain points are. And for what it's worth, X does a really great job. I mean, it's actually remarkable how well X has run as a company. and like just working there i got to learn a lot about like how to get engineers motivated and like really like get shit done like uh it was it was awesome work there it was like super how how uh do you wish you had the opportunity you were working remotely the entire time and that was maybe counter to the broader x culture do you think it's the type of organization where if you're going to work there you should just 100 be in the office or do you think yeah so the person who hired me like was like dude like let me tell you you're gonna have to get to the office my wife uh so at the time i got the offer in between i was interviewing i found out that my wife was pregnant so i let them know and congratulations because uh yeah so so kid is born now he's seven months old um so that's by the time i joined i was like okay well we're gonna i'm gonna work remotely for a bit and then see if i can move but as the kid got older my priorities started changing a bit and i was kind of like considering and talking to my wife about like different ways i could come to politics i could totally work remotely like and be productive i think after an XAI merger, the information attractor got so strong and the talent at XAI, they're shipped so much.

3:04:57I used to be able to read every commit. I used to actually just sit there on my email and click archive over and over again and read every single commit that went to the code base. But after the XAI people joined, there was literally no way. I mean, I told one of the XAI guys this. It was like, you know, when you have like dinosaur shit, which was the sort of code base and you get extreme pressure which is xai talent you get diamonds that's how you get diamonds i was so bullish after that like it's like they're really like they're really pushing on it it's like uh really awesome i wanted and honestly i really wanted to be there like i wanted to go i think it was like probably i was talking to my manager at the time i was like um can i like maybe come like twice every quarter or something and he was like super i mean for what's worth by the way like my manager was reading between the lines totally surprised totally surprised he looked super depressed in the meeting.

3:05:47I felt so bad for him. But if you're watching this, it's okay, dude, don't worry about it. Seriously, some people, probably the best thing I think working at X is I got to work with these engineers who've worked at Twitter for so long. And when I joined, I was like, oh, these guys are onboarded onto everything. But that wasn't the case. My manager literally could just figure anything out. It felt like he was already onboarded onto everything, but he could actually just read the code, look at the logs and figure out what the problem was within 30 minutes and it didn't matter what it was. I was like, this guy must have been here for years.

3:06:18No, he's like actually just that good. So, I mean, I really loved working with him. He was like a really awesome guy. I think he was surprised. I guess like, I guess I kind of kind of guessed what happened to be honest, but it doesn't really matter. But anyways, like it's been really great working there. So I was going to go to visit, actually we were planning to visit Palo Alto. My wife is going on a trip with her mom and the baby. And I was like, okay, like that's one week. I know I'm not just a baby. I didn't want to go on the trip because I had work to do. Not anymore. so probably gonna go and it's gonna be pretty fun um but like uh i was gonna i was planning for that week to go to pal pal alto but uh but also my manager was like can you come like next week and i told my wife and she's like literally like you know baby in her arm like you know trying to like scarf food in her mouth for like the few seconds that she had just like literally dying so first-time parents right so we're kind of getting used to it but yeah no it's a crazy it's a crazy change what um why don't you give some background on what you were doing prior to x ding board and then i want to talk about the future is that's okay sure so uh you guys can i'm probably going to like work a bit more on dingboard um there's a bunch of bugs that i want to fix i just didn't have the time with my full-time job uh so dingboard.com is the best app ever dingboard.com if you want to make a meme in seconds 15 seconds meme i was something happens i was a paid user i loved it it was amazing thank you very much for paying actually very very good you know like your money actually helped pay for this like this is john coogan no way two percent of this is no coogan You should give them naming rights.

3:07:43It is actually crazy that mobile meme making is so in the dark ages. And I feel like you just pulled it forward. I used to have an app called Photoshop Mix or something, and it was pretty good. And then they just completely deprecated it. And then they put ads in it, which I was dealing with. And then they just shut it down. And it was pretty good at dropping out the backgrounds and doing it. It wasn't great, but it was okay. And then they forced you to go over to Photoshop Express and Lightroom. And so I'm using two different apps and neither of them are good anywhere near what you need. And so I was always a fan of Dingboard.

3:08:17I was actually in the process when I joined them. I was in the process of rewriting it so that I could deploy on iOS and Android. It's actually like, if you guys are a nerd, there's something called Sokol. Sokol is made by this Swedish guy, I think. He's retired, kind of. He works three days a week. and on the four days that he has he's working on this cross-platform gl um graphics library like uh i guess like transpiler so you write in one gl language it'll uh it'll produce the the metal version for ios it'll produce the version for android and it'll produce the version for web so i can use the same code and the code can be basically like it can basically be the same code base deployed to all apps and then john coogan and jordy can make memes in seconds and get more followers on x.com i'm so ready this is the future i was promised yeah it feels like an incredible founder market fit for you to make a tool that helps posters make memes i mean like the reason i made it was because i was like making memes on my software engineering diagramming tool and then they like started adding a watermark and i was like this yeah you fucking ruined my meme making tool i am going to war yeah two weeks and then i remember i remember you had you had sponsors this is i love this there were sponsors when you would open a new ding board file it would just have a semi-analysis ad from dylan i mean it's still there it's still there oh it's still there so i I mean, somebody analysis Dylan Patel, you guys should go check out his Substack.

3:09:32He sponsored me for three months and he stopped sponsoring me and I was too lazy to remove it. He's not a Substack, but he's actually on Passport with Ben Thompson. But anyway, he's the man and we love him. But the big question is like, Dingboard, you're super popular online. Why didn't you go and raise like$15 million from a grocery store? Dude, man, that sounds so lame, man. like raising money with managing people. I'm so not down. I'm so not down. I can serve. Like literally dude, like this, I have two plugged in right now and I don't have enough power going to my house to like plug this one in, which is why it's unplugged.

3:10:08Yeah. But like I could serve literally a fucking million people. Okay. A fucking million. So we're going to blow your mind. We're going to blow your mind with this, but sometimes you can raise 15 million and you can do what's called a secondary transaction where some of the money goes directly into your pocket. Oh, interesting. Yeah. So you could raise like 50 on 500. And it's, you know, take 40 million straight to the bank. Yeah, just to kind of set yourself up a cushion. They call it a cushion. Safety net. I feel like you want the opposite of that. I would never invest a founder who asked for that.

3:10:36Give me an example. Tell them we told you. He's joking. He's joking. Yeah, I'm joking. Roy Lee. Never happens. You don't want a founder with money. You want a founder with no options. You want a founder like Roy Lee from Cluey. New dad. New dad or new dad. He's got to figure it out. Got to figure it out. I've got to figure it out. I need a two-acre lawn and a zero-turn tractor. That's what I need. I need this. I will get it no matter what, maybe the next 10 years. But like Roy Lee, for example, CEO, I think of Cluley, he got kicked out of his university, a disgrace, tarred and feather. He has no choice.

3:11:08I wish I got an allocation. I didn't even know he was raising. I would have written. I need the money, right? Like I got a kid. I would have written a check without even thinking. I don't even care what the fuck he's doing. Yeah, you're a Roy Lee defender. Okay, okay. Back to Dingboard. Here's the bout. I want a number for Dingboard. How many users did you get? How much money were you making? Give us some sort of number so we can ring the gong for you. I think the peak MRR was like, I mean, I don't remember. I'm not raising. I don't want funding. Probably like 10K. 10K! Indie developer, let's hit the gong for you.

3:11:39Let's fucking go! Let's go! Honestly, bringing that for your fans too, because you genuinely, you're an internet celebrity and you have real fans. The question I have for you is, because you're a Roy Lee defender, what's your line between balancing engagement, sort of like rage baiting versus, you know, you sound like much more even keeled on the show right now. All my rage baits, here's what makes people so mad. Sure. That I actually believe what I say. I'm not saying it to piss people off. I really believe it. Like, I actually really mean it when I say it, which makes them even more angry. It makes them just like, but like when you're honest, like, I mean, like, you know, I'm a big believer in like honesty and stuff like that.

3:12:23And like, like when you're honest, like you kind of like also the same amount of haters that you gain, you kind of gained twice as much as like people who like are fans of you. And I think that's what really matters. And sometimes the haters can kind of drown it out. So like, I mean, the reason I'm a big fan of Roy Lee is because he posted that video right around the time I got fired. I got like an email, which was a bunch of like really scary legal words, like non disparagement. I had like to like Google what this meant. Like I even know what disparagement meant. i'll go okay wait i'm not gonna sign anything i'm gonna wait for a lawyer and um and then roy lee did this thing and i realized like you know what like roy lee's a legend he's got skin in the game like roy lee has to win so you know what i'm just gonna be retarded i'm gonna be a retard i'm gonna be a retard online and i'm going to have to win and then i'm going to i'm obviously capable of winning but if i have like you know like the one thing i don't understand about people like elon who like work super hard like super smart and like you know elon's done he's got like compounds he's got like you know tons of children you could spend all day all day with like he's got all the toys he could also want to play with all the engineers he could talk to like if i was him i would just like goof around with fun toys but like he's actually trying to get to mars like you know like i feel like i feel like sometimes it's like so i think about being elon i could never be him because i would give up at 20 mil i'll be like all right i'm done dude like this is me like i did my part like i'm gonna chill with my zero turn tractor and my two acre lawn um so i know what you want yeah yeah exactly right so but like i think that's what happens to a lot of founders is they get enough money like palmer lucky palmer lucky it was like a chip on her shoulder he was like he's like fucking jason cow i'm gonna fucking get that guy like i'm so pissed i'm gonna do it again and he did it again 100 did it again um by the way i love jason i love jason i love palmer i love both of them i'm just and you know what jason is like the it's like it's like you can't have batman without the joker right you can't have palmer lucky without jason he's the joker okay i see it he's a joker for palmer lucky for me the joker would be the middle manager who i pissed off by complaining about android bugs uh okay we're getting into the story we're learning what actually happened we're unraveling it don't worry about it i never said anything about my no don't worry about it yeah yeah what's next yeah i mean yeah so what's what's next all in on ding board building out the team okay so ding board is doing great i'm gonna you know keep on growing it and like try to get more money because i need some land and here's why i need some land i've been building things this is the ding bot the ding bot the ding box see this what does it do i 3d printed this what does it do this this is an arm esp32 attached to a motor driver attached to a stepper driver uh sorry sorry stepper motor okay some 3d printed things that i made with ding cad by the way ding cad.com it's currently down because i had no time to work on it but i have like a local host version ding cad is coming back breaking news obstacle sticks no events oh here's here's the genius of it okay it is i need cash flow you know hardware is hard because you buy an iPhone and like you have an iPhone it's like good for like 30 years right yep but with popsicle stick robots they'll break in a month you have to buy a new one oh there we go there we go get them in the boardroom you know what it's like you have a Dobby the elf robot that's like cleaning the house and like you know it costs you 30 bucks and it makes a mistake you just go in and just fucking eat it you like kick it across the room because it's like 30 bucks you can just buy another one it's like it's like that's the feature you're not worried about getting you're not you're not worried about you know uh getting paper clipped by abusing your dobby robot uh because it'll come back to get you at some point oh no you just program it to be happy that it gets kicked you know it's like dobby's like you know please you know please sir you just program it right like oh this is like a soft problem by the way like the no we're not gonna get run away i think the real problem with like ai and stuff is like youtube shorts have you seen these kids all like like i saw a kid like uh so uh uh he'd dinner so with my family this kid was trying to get to the bathroom.

3:16:04He was 12 years old and he was on his YouTube shorts and he was trying to find the door now, not clicking around, but reaching around. It was actually kind of brutal. But I think that's the real AI risk for what it's worth, YouTube shorts specifically. So you're in retirement. Is there anything that could get you to come out of retirement? Could you - Who said anything about retirement? I was coding yesterday. A$100 million offer? I mean to get you back into a mega corporation. yeah it's gonna take 80 million 100 million 120 million what gets you to take the job i think like i probably like find it really hard to say no that's like 2.5 probably like but i don't think like the companies i would work for have that kind of liquidity hanging out i'm thinking they do well what are those kind of companies what are those kind of companies we could run a process right now big tech so i know you don't want to work there but there has to be a number what's the number all right i'm gonna just start doing an auction here i go okay I got a message for 500, 500.

3:16:59Do I have a 600, 600. All right. Our internal goal was to get you hired on this stream or at least get you to raise 15 million with 5 million in secondary. I could raise money like super easily. And like, I could also get a job super easy. I mean, I have friends. The thing is like, I'm a very honest hard worker. Like the places I've worked at, I could, I could go back. Like I have like a really strong network. So I don't really need a job. And I mean, I'm not worried about getting a job, but if I wanted to join in terms like for For me, mission alignment is super important. I really need to use the product to actually enjoy it.

3:17:27Sure. So I sold some Dingboard t-shirts on Shopify and I'm aligned with Toby's awesome. And then Cohere I think is a really interesting place to work for because they're on the come up. But mostly, I mean - He's in Canada. Hold on. Yeah. Breakdown Cohere. That is not the common narrative. I mean, I love Aiden and I think it's an incredible founder, but it does seem like there's got to be some sort of like geopolitical strategy there for that company to really play out i think i just think they're based dude like i just yeah i just think they're fucking based yeah so it's just pure pure culture they'll figure it out yeah just like i mean like you know like it's it's a lot closer than palo alto or like california yeah canada okay and i think they're fucking cool like they got it they need a poster in residence you could basically you could a thousand x their their impressions do you listen to death grips uh yeah yeah it goes it goes it Yeah, yeah, yeah, yeah.

3:18:18Because Aiden's a big Death Grips guy. That just bumps Coher on top of Shopify. Yep, yep, yep. Wow. Yeah, he did an interview in a Death Grips t-shirt. I saw it and I was like, this guy's a killer. But what's wrong with just becoming a full-time poster, getting a sub stack set up, getting a nice little revenue stream? I think I'm going to try to become like an actual billionaire. The Dingbot thing is not a joke. I'm actually going to actually do it. That's cool. And I'm not going to race. I think I can do it. I probably can find people to do this for. There's a lot of things I'm really interested on the way, which is doing reinforcement learning for control of robots, certain types of robots.

3:18:56I think I figured something out after listening to a popular lucky podcast about don't solve for things in hardware when you can solve for them in software. And I'm pretty good at software, and I'm sure hardware isn't that hard. Clearly, it's not hard enough to start. And I feel like it's easy, actually. You proved with that that it's easy. Yeah. Explain that with the reinforcement learning for robotics. Are you thinking like you're going to do stuff in simulation and then transfer the knowledge back with reinforcement learning? Are you doing the thing that Dylan Patel was posting about with like the robots on the harnesses, trying to walk and then reusing that as reinforcement learning data?

3:19:32Like what are you actually thinking in terms of reinforcement? So I don't think like humanoid robots is what I would do. I think I would do robots for things that annoy me. For example, dandelions on my lawn. And I would do whatever is possible. I have a pretty good solid network like talk to some of the guys from eth zurich sure i don't know how to pronounce it yeah uh and just be like hey like just give me feed me the papers and like some shitty python code i'll clean up and make good um sorry researchers uh you guys this is extremely natural he's out of the game someone's got to build the robot that picks up the leaves exactly exactly what now because it's like i'm so fucking sick of like going outside and just you know just like poking out all these dandelions i'm i'm actually so sick and tired of it and it's so easy to build this like literally 3d printer like you could even 3d print the wheels like you can get like five bucks off of aliexpress and esp32 cost like three bucks i just impulse buy them i like get my ex ad revenue payout and i just like spend the whole thing on aliexpress my my house like constantly has like these random aliexpress packages and it's like maybe build like a minecraft sorting system in real life like yeah i feel like i feel like if you could build it you could sell it at home depot for like 300 bucks you know kind of new i could also sell for 25 bucks on subscription and you get a new one every month oh there we go yes people are kicking their dobbies yeah it's like it's like if it breaks it's like i literally will tell the users hey i'm just you know a dude and uh there's a cnc shop here uh that does wood uh so like they do wood cnc so i'm gonna make the robot chassis out of maple wood because there's a lot of maple here in ottawa and i'm gonna laser you know laser on like made in ottawa on it um except you know modulo the the motors inside that will be made in china until i figure out how to make motors which shouldn't be that hard um but i guess like the the the core insight i have after doing dingboard and like honestly working for a lot of these big tech companies is like there's a lot of little businesses on the way to something like i'm going to try to do this i'm going to realize like holy shit like all the cad software fucking sucks and i'm going to build it because i can and like you know with gemini 2.5 pro just like ripping on my credit card what about grok uh gemini 2.5 pro is the best model to use for coding if you write your own coding frameworks because it's cheaper, it's faster, and it's good enough.

3:21:37I'm not giving all of my work away. I don't need a PhD level intelligence because I'm not a PhD level guy. I'm trying to write React. It's not that hard to write React. I need a model that can listen to me and understand what I mean and doesn't avoid using too many if statements. Stop. Just stop using if statements and yeah. Okay. But here's the thing. You want to build a bunch of stuff. You need some cash flow. Why not set up? Why not set up some type of subscription? You have a lot of fans. A lot of people want to give you$10 a month so you can hack forever. I mean, honestly, that was the Dingboard subscription that I signed up for.

3:22:09It was like, support this guy who's doing something cool. Yeah, the issue is Dingboard is so hard to sign up for. How do I even sign up for Dingboard right now? Go to Dingboard.com. Okay, first of all, Jordy, never doubt my conversion levels ever again. I am a conversion. I can convert anything to anything. I am a conversion god. Okay, go to Dingboard.com. Wait, you're so good. You're so good. You don't have a buy button because you want to inspire yourself to grind harder. Go to dingboard.com. Okay, I'm here. I'm here. Sign in first. Do I have to sign in? I don't even have to tell you what to do.

3:22:43Legend. Legend. All right, I'm signing in. I'm signing in. Okay, I'm here. What do I... Ding scribe. There we go. You just want to know what ding scribe even means? You can't not ding scribe. $12.99 per month. You got to get your numbers up, dude. This should be$12.99 per month. I actually don't have the code. It's like on my server. You got to do what Google does. It's 500 a month, but they give you six months at half price. And then it just auto converts up. Yeah, that's the real ticket. Or I could just level the price. Yeah, but the price should be cheap on day one. And they just get more and more.

3:23:15It should go up by 10 % every month. That's some cool shit, dude. I'm not doing that. That's some YouTube shorts. Fry some kids' brain shit. All right. I'm a ding pro subscriber. I'm dinged up. No, dude, you're a ding subscriber. You have to get that dude, Jordy never get that wrong again. You're a ding scriber. Ding scriber. You're part of the crew. I'm ding maxing. Okay. Talk about the rest of your stack. Gemini 2.5 pro. Are you cursor guy? Github copilot cursor? Never. No. Never say cursor. Okay. I have a custom framework I've written for NeoVim. So basically what it is, is I have a server that creates a facade for all the LLMs that I have.

3:23:52So Claude and before Gemini was good, Claude and OpenAI's models, they'd go down all the time like this infrastructure is hard to serve at scale like i don't blame them it's like that's a hard problem to solve uh so i have like always i always have a model available like even if both of them are down like i can hit deep seek over open router right like it's a facade can't you do that in cursor by just switching the model router switching was there a hockey so for me it's like i have a weird i mean let me let me let me let me flex on you guys a bit so this is my Let's see, flex away. Whoa, there we go.

3:24:26Nice now, by the way. I got a homie from China to make it for me. And so basically, space, I don't even remember, space I is LLM provider, so that's Gemini 2.5 Pro. Space K is Gemini Flash. And it all renders in a single markdown file. And I've created commands for these models. So basically like depending on the fence, so three, three apostrophes, and then there's going to be the command name. So like edit file, it can, I can give, basically give it a syntax to edit a file for me and then I actually manually. Sorry. Is this open source? Like, you know, files up, like, like, like can I install your, your setup?

3:25:08So no, it's not open source, mostly because people are going to start posting issues and I'm just not interested in fixing them. So I'm just going to do it for myself. I'm like, I'm pulling up the ladder behind me. Like I'm, you know, I know how much this is. If it's not open source, you got to raise for it. It sounds like we got ourselves a real cursor competitor here. It is a cursor competitor and it's probably going to kill cursor if I open source it, but I'm not going to sell it for money. I think what's going to naturally happen is like when people see me do use it, they're going to have ideas of their own and then they're going to build it for themselves.

3:25:34Okay. Like the whole point of it's a Neo Vim plugin kind of, it's like also a server and it plugs into a bunch of, I'm a good software. So part of the part of the problems with, so Neo Vim is great because the software, software is configurable as a feature. So like the best software you can get is like software you can read and understand and actually change. The best software for you is software you've written yourself. Is there something annoys you? You can actually go change it. And that's why I really like to like write my own frameworks cursor. I'm sure is a great product, but I like, I'm just so particular about certain things.

3:26:05Like for example, like I want to like implement, um, uh, tree sitter across multiple files and automate adding files to my context. So I have a, I have a file which has a bunch of file paths and those get added to my context based on the workspace that I'm in. And I manage it manually by literally going to the file and deleting the lines. And what I want is to automatically create an AST, so an abstract syntax tree, to be able to jump around with an LLM or train a model to actually go through the AST. And this is all super obvious to me. It's like, fuck, man. What about you're going to make any products in the parenting space?

3:26:37Oh, yeah. So actually, I had an idea. So I used an e-ink display. By the way, before you try to hack this, I never update the software and I never write anything private on it. so go it's i'm gonna i'm gonna put it public um so i have an e-inc reader uh slash like tablet and what i wanted to make is to teach my kid to write uh instead of texting him uh i want to make like a kind of like a radio e-ink display kind of like an etch and sketch which you can write on but it synced to your parents etch and sketch so my wife i will have one i will have one and my kid will have one yeah and he'll learn how to write because we'll write to each other over over distances right right i think that's like a really cool idea so i'm probably gonna build that But if you wanted to make like a gajillion dollars, it's so easy.

3:27:17Just look around and like come up with good ideas and actually just believe that you can do it. Like literally just go do it. Like it's not that hard. It's actually not. Get a 3D printer. Watch some YouTube videos. YouTube is a great product. I would love it if they stopped putting shorts because there's so many great videos we can learn. Stop with the shorts. Stop. Get ready. This is going to be a YouTube short. We're going to edit this. Yeah, exactly. This will be on TikTok, baby. Yeah, exactly. We got to get the news. Yeah, Minecraft speed run. Yeah, we'll put subway surfers over you. anything.

3:27:45AI, slop. How many kids do you want to have? So as much as God allows us to have. So just leaving it up to God and yeah, that's how many kids. Good answer. That's great. What else, what else do you have to say to the people while you're here? Thanks for doing this podcast, guys. I know how hard, I know how hard it is to run a podcast and all the tech behind it. Let's just say that. And like the doing live and stuff, like it's, it's a pain in the ass. So keep it up guys. I know like, thank you. and and and the the the live the live uh the video team at x are like generational talents like super super good people like so like just like find someone to like like find a way to reach out to them so when you have any issues and like actually talk yeah i didn't fully realize that you were the bug guy until you posted over the last few days otherwise we would have been hammering you with messages every single day but this has been great come come back on whenever you want to talk about your projects yeah we'd love to just hang out and chat yeah thanks i really appreciate you guys and uh yeah so keep it up thanks we're excited and and i'm excited to be ding maxing just absolutely offer offer a premium tier like i don't know like 10 grand a month or something like that and maybe we become you know oh i mean if you guys want to ask this i'll sell you right now like spit handshake if you want to put an ad on ding board your podcast i'll uh let's say 12 grand 12 grand okay we'll talk to our we'll talk to our cfo all right all right and uh our Our CFO will talk to your CFO.

3:29:08But can we get the semi-analysis deal where it's perpetual? It's 12 grand once and we get it forever? Because what if you sell this to a billion AUM? Is it a one-time? It's a great option. There's a lot of upside here. It depends on how good your podcast is. Yeah. Okay. Okay. Well, we're going to keep working on it. Hopefully, Dingboard keeps growing for the next decade and we can be massive together. Yeah, absolutely. All right. We're excited to follow up. Awesome. We'll talk to you soon. Godspeed. Have a good one. Bye. Hi. Next up, we have to talk to you about Wander. Find your happy place. Find your happy place.

3:29:41Book a Wander with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, and a 24-7 concierge service. Yep, Scene should get a Wander. He should. He's going to be traveling around, interviewing at all the top companies. Wander would be a great choice for him. We have had a great show. We were going to play you an interview that we did earlier today with Shervin Peshavar, but there are some security issues, unfortunately. And so we're working to potentially blur the background and we will get you that interview as soon as we possibly can. But security comes first. It comes first.

3:30:11And so we will obviously realize some of the key takeaways. And honestly, it's such a developing story. I feel like already the story has moved and we should just have him back on again soon from a more secure location and make sure that everything is above board. But thank you so much for watching. Thank you for tuning in. and we will talk to you tomorrow. Have a great day. Bye.