AI-Birds, Snap's Next Chapter, Amazon + Globalstar | Kiva Dickinson, Aron D'Souza, Michael Mignano, Wade Foster, Ankur Nagpal, Bailey Pumfleet, Han Wang

29:54Approximately 8 megahertz of the L-band satellite uplink and approximately 16.5 megahertz of the S-band satellite downlink and N53 terrestrial spectrum for private cellular networks, not phone-to-satellite communication that is licensed out. So those assets by and large serve one customer, Apple. Apple owned a 20 % stake in Global Star and had rights to 85 % of Global Star's network capacity for the satellite service it provides to iPhones and Apple Watches. So that's certainly a valuable relationship. But it also meant that Apple had a de facto veto on any Global Star acquisition. And so what he's arguing is that Apple is very happy that Global Star and Amazon are teaming up in this way to provide them with another option.

30:40Apple's like, we love all CapEx that is not our CapEx. They really do. They are the master of that. We'll have to come back to the story of the Apple new Ray-Ban competitors, the smart glasses, because we have our first guest in the waiting room, Eva Dickinson from Selva. Series A lead of Groons. Kiva, how are you doing? What's going on? Welcome to the show. Doing great, guys. What's going on? It is great to see you. I was looking in your background trying to see if we could maybe see our own office. Oh, are you in LA? Yeah, we're looking northwest. Where are you guys? Okay. Okay, we are more east.

31:24Okay. Yeah, so not quite. But great to see you on here. How's your last week been? Was that your first billion-dollar exit? First billion, yeah, for sure. And definitely the biggest exit that we've been a part of. And yeah, it's been fantastic. It's exciting to share the good news and really incredibly thrilled for the company. Spent a lot of time with them this week. And fantastic for the category overall. You guys have been through it. obviously you never lost faith otherwise you wouldn't have invested in Grooms but certainly it's been funny today watching all birds have new life hopefully your portfolio isn't getting any ideas you know hey you know the supplements category is running away from us but we can pivot to AI compute but yeah give us I wanted to have you on just to kind of get an update on like consumer investing broadly Can we start with like some background and how the fund was set up, how you got into investing, all of that?

32:31Yeah, totally. I started my career in investment banking, joined TPG about 12 years ago in the consumer group. And, you know, I think it was an incredibly like formative time of learning what private investing was and, you know, learning about consumer businesses. But what I was thematically excited about was really uninvestable for us. I was seeing these incredible companies line the shelves of Whole Foods and Target. They were getting more and more exciting every year, taking share of the companies that I had grown up with. And, you know, when we talked about them internally in our investment committee, the problem that we faced was they didn't consume enough capital for us to get our kind of two, three hundred million dollar minimum check in on the way up.

33:21and we couldn't outbid Unilever, P &G or Mars on the way out. And so my takeaway was like, I'm in the wrong seat. Like I got to be doing something in the earlier stage, be some part of the early journey of these companies. So wait, double clicking on that trend that you see, I think everyone's familiar with the boom of, I mean, there were a bunch of different factors and I want to know your take on all of them. Like Whole Foods at one point had like local buying. So if you could go in and like talk to the local, like the Whole Foods in your town, they could actually stock you. They didn't have centralized buying.

33:57There was also the D2C boom, online advertising, the influencer boom, celebrities getting the space. Like what were you identifying as the undercurrents of that broader trend of like new products taking shelf state like space on the shelf? Yeah, I mean, this is definitely like height of the D2C boom. And so what was being talked about in the technology industry and in the press were companies, frankly, like Allbirds and Gasper. And I think that was distracting a lot of attention and capital away from what I thought was the most interesting thing, which was when you walk through the aisles of a grocery store, you see completely different things than you did 10 years ago.

34:35You saw companies like Kind Bar taking enormous share and transacting with Mars. You saw companies like Buy selling to Dr. Pepper Snapple for like$1.7 billion. These were not grabbing tech headlines. These were not grabbing traditional venture capital. But they were taking share of an absolutely enormous industry. I mean, consumer packaged goods, depending on where you get the info from, is like 7 % to 10 % of GDP. Yeah, yeah. Buy is such a funny story because, I mean, it has a traditional startup founder in his basement making the product, grinding, unique insight. But during the D2C boom, they didn't have a D2C website.

35:15Their website was just links to Amazon. They never set up a first party e-commerce site because they just didn't need to. Which can make sense. Oh, it makes a ton of sense. For beverage specifically. Yeah, exactly. Beverage is really expensive. Amazon has the logistics for it. But so, yeah, so walk me through the decision and how you actually position yourself to get in at the earlier stages. Yeah, I mean, I made a stop at a company called Circle Up. They had pivoted from helping emerging CPG companies raise capital to actually raising a fund. And TPG invested in the fund. I kind of put my hand up and said, hey, this sounds awesome.

35:52I want to be part of this.

35:56Did we lose you? We lost you. Oh, no. It's called Nutpods that we eventually sold to VMG and then invested in a company called Liquid IV that we pretty quickly turned around and sold to Unilever in about two and a half years. And it showed me what the risk reward can look like. It showed me also that the early stage companies had a real pain point at seed and series A. they needed not just like specific specialized capital firms who understood their journey, but also resources and common challenges for these companies look different from early technology companies. It's like, how do we do online growth marketing?

36:42How do we launch in retail? How do we scale a supply chain? How do we deal with not having all our eggs in the basket of one contract manufacturer? Traditional venture capital firms were not built to solve these problems. And so in 2019, I broke off and basically tried to build a proof of concept of what a modern venture capital firm built to serve these companies could look like. That's what Selva Ventures is today. We're now investing out of our second fund. And that fund has been a part of, like you see, I said, Grooms, but also businesses like OneSkin, Javi, Midday Squares, Array, to name a few.

37:19what's the uh sorry yeah what what what is best in class look like today uh one of your one of your portfolio companies which i won't name because i don't think uh that it's public i i was able to watch them go from basically zero to hundreds of millions of revenue uh in the span of like i think like roughly four years something like that very very short period of time very small team is that Is that kind of what you expect out of a good investment? Or are there still some kind of slow burn companies that kind of find that takeoff moment quite a bit later? Yeah, there are longer journeys. There are shorter journeys.

38:05Gruen's is probably the shortest one I've ever seen. I mean, that business is like 32 months old now since launching the consumers. I think what we do find is that the combination of online subscription, which is for a great habitual personal care or supplement business, can be really, really sticky, like software-like retention, plus the scale of the best retailers out there. I'm talking Walmart, Target, Costco, Sephora. You can build a very, very large business that doesn't consume a lot of capital in a pretty short period of time. And so when we talk short, like we underwrite our investments to five to eight years.

38:51You know, the exits can certainly be shorter than what we've started to see in the technology markets of like longer dated IPOs. But, you know, the off-ramp for these companies is traditionally, you know, like Grooms in the best-in-class way, an exit to a strategic once you're somewhere between$1 and$300 million in revenue. And that's something, if the flywheel online and offline is going, you don't have to wait too long for it. Yeah, it seems like a, I mean, it's a great, it's a great outcome. Whenever you see one of these, like unfold, I mean, growing is particularly fast, but there's a lot of examples of, you know, raising money carefully, deploying it, growing steadily, figuring out the product, the supply chain, and then exiting to a strategic.

39:35We've seen that over and over and over again. Do you ever run into founders who are like, I'm going to take on Unilever. I'm going to become the next Unilever. I'm never selling. I'm going all the way. I think everyone says that until they get an offer from Unilever at over a billion. Maybe, but I'm just wondering. It's funny, actually. When I started this firm, I'd have a bunch of traditional VCs send me pitch decks to be like, what do you think? And the subject line was like, the next P &G or the next Unilever. These are incredibly sophisticated organizations. They are so good at manufacturing, marketing, sales, distribution, regulatory.

40:13What they're not good at is innovating. And so the idea of building a new one from scratch is kind of like doing the scale thing without having scale. Yeah, you need to do the innovative thing and do that really well while simultaneously doing it multiple times and getting the scale and the systems and the distribution dialed. I mean, the one in our space that people naturally talk about is Harry's. They're not called mammoth brands, but, you know, they had a definitive agreement to be acquired by Edgewell, who owns Schick like six years ago, and it was blocked by the FTC. And, you know, now their only path is going public.

40:52And in order to go public, they need to basically build the modern P &G. And so, you know, they bought Coterie. They bought a personal care business called Lumi a few years ago. And like they're probably the closest thing to it. But I think these like pure play, fast growing disruptors are really what we focus on because we feel like, you know, you take share, you cause enough pain for these modern companies and you fit very well into the machine that they have built. Are you expecting chaos in the supplement category going forward after an exit like this? I'm assuming there's at least 100 pretty smart entrepreneurs that think, you know what, I've always wanted to build a supplement company and start going after it.

41:36I'd say like 2022, 2023, when the market crashed, there was a period of time where we talked about how all the tourists left CPG investing and it was sort of locals only. And I would say the space was sort of undercapitalized. Yeah, that's when Groons and I know a number of your other winners like formed, which is fitting. That tends to be a great time to form a company. Also a hard time to raise capital, but the competitive dynamics benefit you. I really hope that we don't have a flood of half-baked ideas backed by non-traditional investors coming in and doing this the wrong way. I feel like there's still a lot of scar tissue from the industry when traditional venture capital started backing B2C brands in the mid-2010s.

42:31And like, obviously, it's ironic with the Allbirds pivot today. But like, I feel like I spend a decent amount of my time explaining to people why we don't invest in businesses like Allbirds and why CPG is fundamentally different than that, that direct to consumer is not the purpose. It's the products themselves and how much better they serve the consumer. Yeah, that makes sense. One thought is that I feel like a lot of VCs are paralyzed right now because they've invested a significant amount of money into the labs, but they want to invest in more companies because they have capital to deploy, but they're afraid to invest in the app layer and traditional software because of just AI disruption broadly.

43:13and I think that that could lead at least some folks to deciding to, hey, maybe we should take a flyer and let's just throw like five million bucks in this CPG company. Groons is a good comp for it. Why not? It's just, so hopefully you don't get credit. If I were those folks, by the way, I'd be looking at the technology stack of Groons. I mean, Chad, who you had on last week, it was a great interview. Like he publishes a lot on LinkedIn, like all of his vendors, all the technology solutions. I mean, there is a huge opportunity to be the picks and shovels of this type of company that will not stop taking share from large CPG and are willing to try out new AI tools or new technology tools.

44:02What outcomes have there really been in that? Because, again, that was like a wildly popular category during the D2C era, too. It was like, I'm going to invest in the brands, and then I'm going to invest in the technology stack. And you had the Klaviyos of the world did well. Obviously, Shopify has been incredible. But it's hard to think of a bunch of other examples that got out at actually a venture scale. Yeah, none obviously come to mind, although there's been some smaller attractive exits. I mean, one of the best tech-enabled Amazon agencies sold yesterday. And so like, you know, you got to think about entry prices and the value creation to the early investors in these companies.

44:45They don't have to sell for billions of dollars to be really interesting investments. But, you know, we see companies all the time that are just like changing the way these companies operate and reducing the number of people that you need to run a CPG business that interfaces with the physical world. So you would think naturally would need more people. Do you guys do any of the infrastructure? We don't. We don't, but we try and be pretty fluent in it. I love the focus. It's basically like there's two companies a year that you need to find and back, and you need to try to be the biggest. We're hunting for them.

45:20Hunting for them. And just focus all your energy on that. So, I mean, on that hunt, you said you wouldn't back just a D2C company with decent growth metrics potentially. The product matters. What does that actually look like? Is it all your own taste? Are you doing surveys and panels and trying to understand where the white space is for a particular product, how durable that white space is? Because every once in a while, there's like a new co-packer that comes online and unlocks a new stick pack or a new gummy format or something. And then there's a boom around that and it feels differentiated.

45:56But really what you're looking at is like, oh, there's just a new piece of machinery that's in a bunch of different co-packers. So what's your process for actually understanding whether a company has a great product and will continue to be able to compound on the back of like the quality of the product? Yeah, I mean, I think you got to be pretty thoughtful about value proposition. And a lot of new consumer value propositions for a CPG product are like downstream of changes in how nutritional knowledge and human health plays out in our world. So, you know, GLP-1s have a bunch of downstream effects.

46:34We don't just invest in GLP-1 companies. We invest in companies that ultimately take share because of a world where GLP-1s are important. And so the first thing we're trying to understand is like, where is the puck going in terms of human understanding and therefore value proposition in the future? We're trying to find signal of great companies with DSC and offline metrics. And then most of our job, frankly, is figuring out what are the false positives. Like what are the reasons why a company might be growing really fast and efficiently that ultimately won't translate to successful omni-channel distribution and velocity and a successful acquisition?

47:12That was the thing with Garoons that I think some people maybe like drove them to pass is that there had been that company on Amazon that had scaled to like 500 million or something like that in sales. I was doing that. Yeah, Goalie. Yeah, Goalie. just basically no value creation at all, even though they had tons of sales. And I think people kind of like wrote off gummies as a category. And clearly that was wrong. Yeah, we were grateful for that misunderstanding. What about the temperature with LPs right now? I imagine that you have a unique value proposition and LPs are watching, you know, so much like K-shaped winner-take-all dynamic in AI, and it's such a complicated industry to allocate towards.

48:03What are you hearing from LPs on appetite for consumer packaged goods investing broadly? Yeah, I think it's a lot of curiosity. Naturally, these are things that they see in their physical world, and when an outcome like Grunin's crosses. But Grunin's not the only one, by the way. I mean, Poppy, Dr. Squatch, we've had actually like a slew of billion-plus-dollar acquisitions in the past year in our industry. And so, you know, they see that in their news cycle, they see it at the grocery store. I think there's a there's a curiosity moment, like what's what's happening here? And how does it work? And I think a little bit of that is trying to square that against what they've heard about the DTC revolution not working over the past 10 years.

48:45And so I would say there's there's a wading into the waters. There's definitely anytime a news cycle of these acquisition hits, like people start to double click and try to understand it and spend time. I think what we found in our meetings specifically is folks appreciate the specialization in recognizing that this is different from technology. So if we were a traditional firm that was trying to spend 15 % of our time on it, I don't know that we could do a very good job on this. Us spending 100 % of our time on it feels like more comfort in their mind. But a little bit distant from what they usually do.

49:26Yeah. Yeah. And, and, uh, I think at this point, like if somebody starts a consumer brand and they are talented, you, they will get introduced to you some way or another. And so you can kind of like meet every, every great entrepreneur, which I feel like would be, is basically impossible and like traditional tech investing. So that's the goal. I mean, there's not a lot of people that, I mean, there's a few firms that we love and respect that do this, But like one of the reasons I started this firm is like there's no benchmark Sequoia A16Z. Like if you're if you're an entrepreneur out there, like there's not that firm that you've always dreamed of partnering with and getting started.

50:07And so, you know, eventually myself and a few of my peers decided like we've got to go create that firm for them. Amazing. Love it. Well, congratulations. Thanks for coming on to break it down and have a great rest of your day. Good to see you. Great to see you guys. We'll talk to you soon. Up next we have Aaron D'Souza He is the founder of Objection.ai And the Enhanced Games He's in the waiting room And we'll bring him in to the TV game Ultradone What's going on? Aaron, how are you doing? I'm great, thanks for having me on the show Of course, thanks for being here Do you want to give us a little bit of your background?

50:43You've done a lot in your time I'm super interested in Enhanced Games And then we can go into Objection.ai At some point But how are you introducing yourself these days as a multi-hyphenate? Yeah, so I'm a lawyer by training. When I was 24 years old, Peter Thiel hired me to lead his litigation against Gawker Media involving the wrestler Hulk Hogan. We won the largest invasion of privacy judgment in history. It's the subject of the best-selling book, Conspiracy by Ryan Holiday, forthcoming movie starring Ben Affleck and Matt Damon. There's a movie. There's a movie. There's a movie, yeah. And since then, I've gone on to found nearly 12 companies, most famous for the Enhanced Games, the quote-unquote steroids Olympics.

51:24But launched today, my new one, Objection AI. Okay. Do you know who's going to play you in the movie yet? I don't know. Hollywood is not like Silicon Valley. It takes them a very long time to make a movie, and it seems to go through a lot of different iterations. Yeah, it should be interesting. Well, I would love an update on the Enhanced Games. the first event is happening in late May. Is that correct? That's correct in Las Vegas. Las Vegas. And talk to me about the scale, the potential value, the goal with that project. It's certainly got a lot of attention. Everyone has a take on it. Oh, yeah.

52:05And the timing feels pretty good considering it feels like more people than ever are enhancing themselves. Oh, yeah, with GLP-1s and stuff, which might be banned. I don't even know. Some of them are banned. Some are not in traditional sporting events. Yeah. But what led you to the enhanced games? And give us the update there. Yeah. I studied philosophy as an undergraduate. And I've always been interested in bioethics. And I read a paper by Professor Julian Savalrescu, who's a professor at the University of Oxford. And he actually argued back in the 90s for an enhanced Olympic Games. And I learned that there were nearly half the athletes in the Olympics admit to using banned performance enhancing drugs, yet less than 1 % get caught.

52:48So there's this like real disconnect. And at the same time, things like peptides, TRT are becoming increasingly normal. You know, even people like Secretary Kennedy, you know, our health minister in the United States is quite an advocate for human enhancements. And so I thought to myself, why should we be shackled by the ideas of the past? And shouldn't we be able to unleash the full level of human potential using the best of science and technology? And that's where the idea of the Enhanced Games came from. Yeah. How many athletes have actually stepped forward and said that they want to participate?

53:24How big is the movement at this point? Obviously, the first Games is happening. Do you want to put it on the similar every two years cadence? Is there demand for more of a UFC-like schedule? How do you think this all works out? The aim is for an annual schedule. We're very pleased that the company is going to go public through a SPAC combination on the New York Stock Exchange in the coming weeks. Obviously, I stepped down from being CEO a few months ago to focus on my new venture, so I can't speak to the exacts of the SPAC. But a$1.2 billion valuation we're very happy about. and ultimately it's not the number of athletes participating it's the quality of athletes so we have Olympic gold medalists we have world record holders and in fact we've already set our first world record in the 50 freestyle no way which was set by Christian Goulombe of Greece in an exhibition event last year he's run faster than any man in history had up to the point in time and he's only been enhanced for a couple of weeks and so to show you how how much of a difference that could make.

54:27Christian was 31 years old at the time, which arguably is about 10 years past his prime for a swimmer. Oh, interesting. Okay. Well, take us through Objection AI. What's the thesis? What led you to start another company at this moment in time? I believe that the fundamental problem that we face in our society is truth. There is no sense of an objective arbiter of truth in our society, and this is something that has caused great societal decay. If we don't have a shared sense of truth, we can't have a functioning civilization. And two decades ago, we would have said the New York Times is the arbiter of truth.

55:06And today, the social platforms don't seem to care about it very much. AI juggernauts don't seem to care about it very much. And so I said to myself, what is the best way to find truth? And truth is not a vibe. Truth is a process. And that process is very well documented in courts. courts are viewed by Americans as being very trustworthy entities, versus the legacy news media in particular has seen a collapse in credibility. 50 years ago, according to the Gallup poll, 70 % of Americans trusted the media. Today, that's down to only 30%. And so the goal of Dejection AI is to create a system where anyone can challenge a claim made in the legacy news media.

55:47Independent investigators will then investigate it, former CIA and FBI agents. And then all that data is presented to an AI jury to analyze, to figure out if the original claims made by the journalists were true or not. Interesting. Yeah, something that's, it feels like community notes have been a good innovation for like truth online but the big flaw is that by the time a post gets like a like solid community note oftentimes like a million people already they didn't know that that it that there was yeah so i mean the logical follow-up is uh is speed a problem here because i imagine that yes yeah i imagine that uh if you have to run a whole jury process and have discovery and argue argument like the original claim could be baked into the society's mind share before you have a chance to chime in.

56:46That's the fundamental problem about news media today is that false information spreads six times faster than true information. And so the incentives for generating clickbait content is very pronounced, and we've known this for a very long period of time. And so by compressing the legal process, which often takes 10 or 20 years and cost$10 million, as we learned in the Hulk Hogan lawsuit, down to something driven through software and artificial intelligence, down to a couple of days, we can adjudicate factual disputes much, much quicker and much cheaper. The whole process on objection can cost as little as$2 ,000 and can be done in as little as 24 hours.

57:25So is the business model to sell directly to people that want to contest claims that are made on the Internet? Yeah, exactly. So if the New York Times writes something inaccurate about you guys and your wonderful podcast, you can file an objection. Then human investigators will investigate the story line by line, source by source. They'll call everyone quoted in the article. And then they'll present that information to an AI jury. And the original author, of course, has the opportunity to respond and say, hey, my reporting was good. It was high quality. But, you know, we live in the era of data.

58:01and I think it would be wonderful if every story published by the New York Times included the long form recordings of each interview. I've done thousands of media interviews. Journalists always record them but they never publish them in full. And so being misquoted or anonymous source, these are the tools that in particular print journalists use that have seen a massive degradation in trust. Have you been following the Satoshi Nakamoto story recently? That feels like a... Yeah, exactly. A textbook example of something where it's been disputed, but it's very hard to disprove if you're being accused of being Satoshi.

58:40How do you process that particular story? Yeah, and so courts are a very good methodology of finding truth. I think there are only two solid methodologies of finding truth. One is courts, and the other is the scientific method. Sure. And so if you take a scientific method approach, anonymous sources should never be allowed, right? So you can't say to a scientific publication, a source told me this. You have to be able to replicate the experiment over and over. Or courts is the alternative method of truth finding is where you have two adversarial parties often arguing antithetical points of view.

59:15And what is truth? It's a really important question. It's almost the core question of philosophy. Well, in the court setting, it is who has made the better argument, who has presented more evidence, and how coherently has that argument been made. And now with the magic of artificial intelligence, we can do all of that very quickly and very cheaply. Yeah. How do you think about tuning different models to actually give you unbiased results? It feels like every different model has slightly different flavors and things that it likes and dislikes and might see things different ways. It feels like a subjective How do you get it to be impartial?

1:00:00Yeah, so that's a great question. That's exactly how we face these issues with human juries and human judges. Human judges are extremely infallible. According to a paper from Professor Posner, who's one of the leading scholars of law and economics at the University of Chicago, AI applies the law 100 % accurately. Human judges only do it 52 % of the time. because human judges can be swayed by whether they've had lunch or not, whether they're having a bad day, whether they have a savvy lawyer in front of them. And then in the same way, we use a jury-based system. Five different models prompted to act as if they were different personas of people based on a statistical sample of how everyday Americans behave themselves and demographic samples.

1:00:52and the models have to find a majority opinion to pass a verdict. Are you thinking about integrations with social platforms? Doherty mentioned the community notes system. How do you think about distributing findings once you actually have reached a conclusion? Yeah, so this is the principal flaw of courts. So courts issue a judgment. but they have no distribution mechanism. And we have something called a fire blanket. So we have an algorithmic posting system on X so that every single claim that is under investigation, we immediately fire off a tweet that says, this claim is under investigation.

1:01:38Please see the full case file. And then when the similar claim is retweeted at some later point in time after adjudication is complete, we then say that claim is false or that claim is true. Please see the full analysis that was done. So it intercepts the spread of disinformation as it is happening. Very cool. Jordy, anything else? Excited to follow along. Yeah. Well, thank you so much for taking the time to come chat with us. And I wanted to confirm, though, you are not being held hostage right now. I am not being held hostage right now. I'm in an undisclosed location. Okay. because as someone who is often subject to negative media reporting, I like to not show where I live.

1:02:23Makes sense. Makes sense. Well, good luck, and we will talk to you soon. Thank you so much for taking the time. Thanks for having me on the show. Have a good rest of your day. In other news, Uber is now going back into robo-taxis. Uber commits$10 billion to robo-taxis in strategy shift. Uber's committed more than$10 billion in buying thousands of autonomous vehicles and taking stakes in their developers, breaking from the asset-light gig economy business model to avoid disruption from robo-taxis. The ride-hailing app has aggressively increased deal-making over the past year, announcing partnerships with more than a dozen providers, including China's Baidu and U.S.-based Rivian.

1:03:05Here's the kicker. Yes. As well as plans to launch robo-taxi services in at least 15 cities in 2026. So they're going to be putting Rivians on the road autonomously. Or maybe they'll do a deal with one of the other robo-taxi providers. Rivian seems like, I mean, great cars, really beloved. Everybody says, like, the autopilot in Rivian is great. Yeah. But so it's also great in Tesla. And we don't see robo-taxi scaling everywhere. Truly scaling yet. I don't know anyone that's ridden in a robo-taxi. But things are moving quickly. And, you know, but it is 2026 right now. So that is a very aggressive timeline to actually launch robo-taxi services.

1:03:49But maybe that means more of like partnership with existing robo-taxi providers. We'll have to dig into it. And we'll talk more. Market likes it. Stock is up 6.8 % today. Yeah. I mean, the stock has been trading down as Waymo and Tesla expanded services. It sold off. Uber sold off. And when Waymo raised$16 billion, Uber sold off again. And when Zoox planned major expansion, the company Uber sold off as well. The stock is down 23 % in the past six months. And investors are starting to get concerned about autonomous vehicles. So Uber is responding. Well, without further ado, we have Michael Magnano from Union Square Ventures now in the waiting room.

1:04:31Let's bring him in to the TDP and Ultradome. There he is. Let's hit the gong for you. So we're going straight to the gong. Oh, man, I love that. I love that gong. Thank you, guys. Thank you so much. Thanks for having me. Reintroduce yourself. Tell us where you were and where are you now. Yeah, of course. I'm Michael McNaughton. I am a GP at Union Square Ventures. Most recently, a partner at Lightspeed. And then before that, co-founded Anchor, which sold to Spotify. And I'm very excited to see you guys again. Thanks for having me. Yeah, what motivated the move? Did you just want to go to New York?

1:05:10You want to work with Fred Wilson? What was the reason to jump over to Union Square? Well, I'm from New York. I've always been here. I've always lived on the East Coast. I've always been somewhere in or around New York City, and I've known USV forever. So I built Anchor here in New York City, pitched them back in the day for our seed and Series A, and they passed both times, unfortunately.

1:05:38Sorry. No, I love it. I love it. Got to know the team, obviously. And after we sold Anchor to Spotify, became an LP in the funds and stayed close and just always love their approach. You know, USV is famously thesis driven, right? They've always been sort of willing to bet early on the stuff that looks weird, the stuff that looks funny. and I've always really appreciated that approach. And so late last year, they approached me and they said, hey, how about coming over to USV and joining us and joining the partnership and had to do it. So here I am. Yeah. How should I think about USV right now?

1:06:16How big is the firm? How big is the partnership? It feels like a unique firm at a time when many firms are just going for insane scale. Yeah, building platforms. But it feels like Benchmark and USV are two funds that could have built platforms and, you know, have made the decision to, you know, stay true to venture. Yeah. Yeah, I think this has always been USV's approach, right? It's been a famously small partnership for a very long time. Small in terms of the number of general partners and small in terms of the fund size relative to lots of other bigger platforms, as you mentioned. And I think they've made it work because what we talked about a little bit earlier, they've always been willing to have a thesis and have a point of view and go really, really early when, you know, a lot of the other big firms and the big platforms are chasing more consensus deals to their credit.

1:07:10Right. I think the platforms have been very, very successful at using, you know, large, large, vast quantities of capital and speed as a weapon to to back the winners. But I think what USV has always done well is said, you know, we think the world is going in this direction. We're going to bet really, really early. Sometimes we'll get it wrong. But every once in a while, we'll be one of the first investors in Coinbase or Twitter or Mongo. And obviously, for a fund of this size, that ends up having a massive impact. And I think, you know, I think that that is the right playbook to be playing moving forward.

1:07:44Obviously, AI is going to create a proliferation of startups like we've never seen before. And I'm not actually sure it will be possible to see them all. And I think the only way you're really going to be able to have a great bet is if you know what you're looking for before you see it. And that's always what USV has done. Yeah, I still am surprised when I look at the App Store charts and see that we really just have LLMs in the charts and not much else. And it feels like that has to change at some point in the next year or two. And yeah, fully expect that. So where do you want to focus your time?

1:08:26What's exciting? Yeah, you know, I've always been somebody that's been excited by great products. You know, I'm a product builder myself. You know, we mentioned Anchor. I have another startup, which we've talked about on here, Oboe. So I'm naturally drawn towards products, great product-oriented founders. When I was at Lightspeed, I led deals like Suno and Granola, which I think for me have always fit that bill, but also I think take a little bit of an interesting approach that's relevant to building startups in AI moving forward. You know, Suno is a company that's taking a form of media and it's leveraging the democratization of that media to not just, you know, optimize an existing workflow or optimize an existing market.

1:09:12They're doing it to build a completely new format and unlock a whole new form of creativity for people. And so I think we're going to see a lot of that moving forward. I think AI is going to unlock use cases and applications that we can't really dream up now. Maybe to your point about there only being like three LLMs at the top of the app store now. In fact, I believe that software is starting to represent more and more a form of media, right? It's so easy to create this stuff. I'm sure you guys, just like me and everyone watching, are building apps and creating agents every day. Think about what happened with video.

1:09:45Think about what happened with podcasts, with text. It wasn't just about building the media. It was about the enablement that goes under it, the platforms, the payment rails, the distribution. I think there's probably going to be a massive amount of enablement that goes into supporting the massive long tail of software about to see. Well said. Wanted to get your take on synthetic podcasts. We saw Pump launched a synthetic, an AI-generated podcast today called Best Stocks. I'm assuming they talk about the best stocks. But wanted to just get your kind of overall take. Nick, we've seen some other AI-generated podcasts start to top the charts.

1:10:28What's your perspective there? Not a venture opportunity, but certainly I would say interesting. Yeah, look, I think it's super interesting. Back in the day when I was building Anchor inside of Spotify, we actually had a partnership with WordPress, where if you had a WordPress blog, you could tap a button and it would immediately turn the text and the blog content into a podcast, right? So it was a way for the creators to get more and more distribution. And, you know, I think it was great. We saw a lot of demand from the creators, but it may have been too early on the timeline in terms of the quality of the content and sort of that uncanny valley.

1:11:06And I think that has a lot to do with it. Yeah, basically the voice models just were not very good. Voice models were not good or they were using. Exactly. The voice models were not that good yet. that obviously they're a lot better right now. 11 Labs creates phenomenal voice models. I don't know what this podcast you mentioned is using. But I think the closer and closer we get, not only to great voice quality, but also a human-like experience for these hosts, whoever they are on this show, I think the closer we can get to getting it to work. I mean, one of the reasons people tune into TBPN is not only to kind of hear the news and hear people speak, but it's to watch you guys, right?

1:11:43They like you guys as the hosts, and they like the personality and the human aspect of you guys. And so I think the agents that are hosting these shows are probably going to have to get closer and closer to that on the spectrum before it really takes off. Did you explore any version of that WordPress to podcast workflow that would sort of on demand hire a voiceover artist? Was there any demand for that or did you ever explore that? We did explore that. And it was one of those things where, again, at the time, it was going to be challenging to spin up the content for that quickly enough such that we were confident that there would be enough demand to make it worth it.

1:12:27I think with media and with publishing, people really want immediacy, right? When we were first building Anchor, actually, the way that we went from being a social audio app to a podcasting platform is we found that all the people that were publishing just inside the Anchor app said, we want a podcast. We want this to be on Spotify. We want this to be on Apple. But there were no APIs at the time. So the way that we did that was we actually had human beings sitting in our office and manually creating RSS feeds and posting those RSS feeds to Apple podcasts in real time. And fortunately, we were able to do that quick enough, you know, within a couple of hours that it satisfied the creators.

1:13:07But I think, you know, at the time, your idea, what you're talking about now, the gap just would have been too long. It would have been days or weeks before the content was ever delivered. Yeah. Have you, have you reflected on the, just like, there's such incredible model progress and such incredible AI functionality progress. But I would have expected to have found, in the same ways that the AI-generated podcast seem to be working and we've seen some audio stuff work, I would have expected to see like a sub stack that was AI-generated text that got really popular. And that seems somehow easier from just an AI capability perspective.

1:13:50There's less steps in the chain, but it hasn't really broken through in that way. And I'm wondering if you've ever like grappled with that question of like why that hasn't, why even like the simpler, more condensed text? I don't even know if I follow, I mean, maybe I do, but I don't know if I follow any Twitter accounts that are like fully AI generated news headlines. That you know. This was actually gonna be my question to you. I mean, it's possible that that already exists and maybe these publications or these writers or these journalists aren't really admitting it yet. I mean, I have to imagine a lot of the content we now consume on the internet is AI generated.

1:14:25We may just not know that it is. Yeah, it feels like a lot of it is like we're in the centaur period where the human plus AI works well. And so you see people doing research and then processing it and adding their own spin. Yeah, Alex Heath was talking about how he's doing a ton of AI in his writing process. He can say that because people are paying to subscribe to sources to get net new information that doesn't exist anywhere else. But I think that that only makes human-powered content that more valuable. Again, not to keep bringing it back to you guys. I think this is a reason that people like this.

1:15:00And I think there's a reason that live sports viewership is at an all-time high and people are going to concerts. I think people crave this farm-to-table human experience. Yeah, I feel like you've done a good job of that with your content. Can you sort of reintroduce it for folks who might not be fully familiar? Because the production quality is extremely high, but it feels like you're sort of playing a different game in terms of release cadence, making sure that everything is special and not formulaic. But how are you thinking about the content that you're putting out these days? To be clear, I think you're referring to the podcast I was running at Lightspeed called Out of Office, which I'm no longer the host of now that I'm not at Lightspeed.

1:15:47However, the philosophy that we and I had for that I think holds true and I believe in it, which is there's so much content out there on the internet. And the tools have gotten so good at clipping and mass producing video content or two people sort of sitting in the same setting, having the same style of conversation that you really have to do something different. You guys did something different. There are a number of podcasts out there that are now doing something different. But one of the ideas that I had that we ran. What's that? There's five total. There's about five, maybe six. Yeah. The thing, the idea that we had for out of office was, you know, why does this conversation have to happen in a studio?

1:16:29Or why does it have to happen around a table? How great is it to watch an episode of comedians in cars getting coffee and watching these guys behind the wheel of an awesome car or in their favorite coffee shop or watching something like Anthony Bourdain, Parts Unknown, and watching these two iconic people travel around a really iconic, familiar place. and we thought, hey, can we bring that to the tech podcast? And so that's out of office at Lightspeed again. No longer something I'm involved with. But stay tuned because I do think I and maybe USV will have something similar in that vein coming soon.

1:17:07I actually just posted a little bit of a teaser photo on my ex a few hours ago. Go check that out if you haven't already. But look, I think the theme is to break through right now and break through the noise. is you've got to do something that looks a little bit different. And you have to do something really, really high quality because the tools have gotten good enough. There's a baseline quality that you just get out of the box. So if you want to stand out, you actually have to go above that. Yeah. Jordy? How many investments do you expect to make? What percentage of investments do you expect to make in New York versus the West Coast?

1:17:40Listing off some of USV's iconic investments, of course, the Coinbase's, the Twitter's, et cetera. Many of them were on the West Coast. Are you going to be over here a lot? USV has always been a sort of one office, New York-centric firm. Obviously, the namesake, Union Square Ventures. But USV has also made a lot of money all over the world. You just rattled off some of the locations. When I was at Lightspeed, I was based in New York. And even though I was based in New York, I was making investments all over. I think most of my portfolio actually was on the West Coast. Obviously, I had Suno, which we just mentioned, in Boston, Granola, and London.

1:18:23But a lot of them were in the West Coast. Going to continue to operate that same way. We'll invest wherever the opportunities are. But the home base is going to stay in New York City. Yeah, what's the secret to getting good deal flow at the early stage? I feel like that's such a differentiator for Fred Wilson throughout his career. you are teasing some sort of conversation. You've clearly been working with him for a while. What do you think makes USV particularly differentiated in finding great companies early on? I know it sounds simple, but I think it's a willingness to be wrong and actually make predictions.

1:19:05I think so much of this industry is focused on momentum and focused on consensus-driven bets, which, again, can be very, very lucrative. if you get in them and you have the capital to fund them. But I think what Fred and the team have always been great at is having a conversation which lasts maybe not just one partner meeting, but lasts weeks or months or, in some cases, years. And along the way, making some bets against that idea and against that thesis, which may look wrong for a very, very long period of time. You mentioned Coinbase. USV famously made that really early Coinbase investment, but they had invested in crypto before that, right?

1:19:44They had a strong thesis around crypto before that. And it took some cycles for them to gain conviction in what they ended up seeing in Coinbase. So I think it's about engaging in honest discussion with your partners. It's about not being afraid to have a crazy idea, to like the weird, and to keep that conversation going for a long time. How are you thinking about wearables these days? It feels like the Meta Ray-Bans are having like an inflection point. There's, you know, Apple's coming to market with stuff. There were a few startups that got in the game. Yeah, and the Hellside Aura, Whoop. Yeah.

1:20:23I've seen. It feels like it's maybe undercovered, but have you processed any of that as like an interesting endpoint for technology broadly? I think there are a couple interesting aspects about it. Jordi, you mentioned earlier at the top of the app store, we're seeing the same three apps now for, you know, the past two years or whatever. However, hardware, which I think has often been thought of as too hard, right? You always hear the term like hardware is hard. And so venture capitalists often ignore it. We might want to rethink that a little bit. Hardware could be considered a wedge now, right?

1:20:55Maybe it's actually a defensible layer where if you can actually have a successful product there, that's going to be a lot harder for a massive foundation model lab to go and replicate than leveraging their distribution for a new form of data collection or whatever the case may be. So I think there could be some interesting edge model or edge data out on a hardware device that is pretty defensible. So that's something interesting I like about it. I think the other thing that's interesting about it is you guys have talked a lot. I've talked a lot about how context is king and finding these edge data sources is really, really unique.

1:21:33And that's obviously something that has been pretty fascinating about Granola, right? They're capturing these conversations that you can't really capture any other way or aren't being captured any other way. And so that provides a lot of value to the model. Hardware offers a similar opportunity, right, where using hardware, whether it's the glasses or maybe something else completely, a pen, an aura ring, you're capturing an edge form of data that the labs can't really see unless they have that hardware. So I actually find hardware to be quite compelling at the moment, and I think it actually is an area of opportunity.

1:22:12How are you seeing startup culture change in terms of who's hiring? I'm thinking about hardware. I was reading this report, maybe it was in the Journal or the Financial Times, about how more and more students are choosing like electrical engineering, mechanical engineering over computer science because of the rise of AI that can code very effectively and the job market changing. And that feels like potentially a boom for hardware startups. Like you have to imagine hardware gets less hard if there's just more hardware engineers and it's easier to staff up for that. But I'm wondering if you're noticing any other changes around location or work-life balance or in-person We did the COVID thing.

1:22:53We did full remote. Then we went back full in person. Then we go in hybrid. How are the startups that you're interacting with looking different than a decade ago? I think the big difference that's happening right now, and this feels like a very recent thing, maybe past few months, maybe past six months, is I think we're going to see a swing back towards the mission-driven company. I think we've seen a lot of capital flow into the ecosystem over the past few years. And I think that that has turned a lot of missionaries into mercenaries. And by the way, it's worked phenomenally well. Again, we keep talking about the same two or three startups that are massively, massively valuable.

1:23:33But I think we're also seeing a lot of churn and a lot of attrition at these companies. And I think it has to do with this mission-based thing. If you think back maybe five years ago or 10 years ago and some of the behemoths that emerged from maybe the previous software and startup cycle, these companies were very, very mission-oriented. They were all aligned around the same mission. They stayed incredibly lean, incredibly focused on what they were trying to accomplish. They left the politics and the bullshit out of the office, and they just focused on hitting that mission. And I think right now we're seeing a lot of people get burnt out after just chasing huge, huge paydays at incumbents or insanely well-capitalized startups.

1:24:12And I think we're going to see the pendulum swing back in the other direction. White pill. White pill. Yeah. How many – last question for me, just because you're the first person from Union Square Ventures to come on the show. Is that true? Yeah, that's how. How many partners are there? Like how big is the investing team? I guess it's a better question. There's about eight of us. And the whole team in general is under 20. So we're a small team. Like I said, we're all in one office. And we talk a lot. You know, when I was at Spotify, Gustav Soderstrom, who I think you've had on the show, had this famous saying inside of Spotify.

1:24:54Everyone at Spotify knows this saying. Gustav says, talk is cheap. So you should talk a lot. And what he means by that is the cost of getting something wrong is actually way more expensive than the time it takes to actually talk and align and earn the right to go invest or spend or bet on something. And what I've always loved about USV is they have a similar philosophy. They talk a lot. It's a small team, but they spend a lot of time together forming their point of view on the world. And to your question earlier, I think that's what's given them the edge. Yeah. Well, they're very, very lucky to have you and excited to see what you guys do together.

1:25:29It's great to see you. Thank you. You too. And by the way, congrats to you guys. Thank you. Thank you. We're very excited. Come see us when you're in LA. And we'll do the same in New York. I'd love to. I'd love to. We'll talk to you later. Have a great rest of your day. You're the man. Thanks, John. Thanks, Jordy. Speaking of Meta, Meta Platform's CEO, Mark Zuckerberg, reportedly moved his desk into the AI lab, working directly with Alexander Wang and Nat Friedman. Zuckerberg is reportedly coding throughout the day. there's a story in the New York Times about about AI sunglasses I feel so sorry for my AI sunglasses and it opens with this anecdote about trying to use the glasses to identify what is it?

1:26:17Some sort of bird it says one afternoon on a sunny stroll I stopped to admire a bright red cardinal singing its heart out in a tree hey Meta I say what kind of bird is chirping in that tree? My sunglasses make their little ding dong noise, analyzing the world, finally they speak. I don't see a bird in the tree or hear any chirping, they say. I point directly at the bird, which is still chirping. I don't see a bird in the tree where you're pointing, my sunglasses say cheerfully, just bare branches and sky. For several weeks, this is how it goes, the disorienting sense of chatter with a toddler who is drifting off into nap time.

1:26:52Look, it would be easy to dunk on my very expensive, staggeringly incompetent glasses. Critiquing AI these days is like shooting fish in a barrel. And I mean poorly animated fish that keep sprouting human fingers inside that barrel. And so, yeah, it's a very interesting read because it's clearly like there's so many times when you're interacting with an old legacy model in a particular in a particular end point where you're not getting the most frontier intelligence and it can really throw you off. Andre Carpathia had this whole take about people that used early LLMs that would hallucinate and haven't tried the latest coding models to really understand how powerful the technology is.

1:27:30There's this big division. And so you sort of have to do both, acknowledge that diffusion takes time. Actually rolling these models out to a device like glasses takes time. Yeah, and it's interesting with Meta because it feels like they're glasses. Yeah. I'm getting served a lot of videos. Oh, yeah. People are using them as GoPros. Yeah, but they're just cameras. Yeah, as cameras. And I know a lot of people that use them as AirPods replacements. They use them just to take phone calls, listen to music, because then they don't have anything in their ears, and it's just more comfortable for them.

1:28:01They even have clear ones that they can wear on. It doesn't really work on a plane because it's too loud, but when they're just in the office. But actually getting the AI features into these devices is true. They have Humor Bench. Oh, yeah? Hey, Meta, I said one day, tell me a joke. Why did the baseball go to the doctor? Why, Jordy? It had a little rundown in its batting average. Rundown? I don't really get it. Swing and a miss. Yeah, it might be better in this case just to pull from the bank of vetted jokes. Yeah, you don't need to generate one from scratch. Siri did that early on. You can just look up some of the best jokes in history.

1:28:41When Siri initially launched, they had comedy writers write out a whole bunch of jokes. And when you ask Siri, tell me a joke, it would just pull one off the shelf that was handwritten by a human about Siri, aware of the context, aware of everything. I'm now, I have triggered Siri, so I'm in trouble on my laptop. But it is just very interesting. Apple, of course, we touched on this. Mark Ehrman has a report on the upcoming AI smart glasses. They'll come in several styles and colors. A few of those frames seem like they're direct competitors to Wayfarers. will be very interesting to see the ecosystem integration with Apple.

1:29:19And Dan Primack has a take here. He says, Apple isn't burning mountains of cash on GPUs or investing billions into frontier labs, but it still may win the AI race. Success from the sidelines. Sort of a contrarian take at a time when Apple doesn't really want to talk about AI right now. They're clearly in this rebuilding moment. They really don't want to talk about AI. But they have partnerships with OpenAI, Anthropic, and Gemini at this point. They're using AI tools all over the company, and it feels like they internally are starting to feel the AGI, but at the same time have a very unique strategy of not overreaching, and it could really work out for them.

1:29:58So it's an exciting time and an exciting story to follow. Well, without further ado, we have our next guest in the waiting room, Wade Foster from Zapier. Zapier. Wade, how are you doing?

1:30:10Wade Foster:Good. How are you guys doing? We're good. Great. Thanks so much for coming back to the show. How are things going? Give us the general update on your company. And then there's a million different things that we can talk about. But I'd be interested to know what you're learning from your vantage point as CEO of Zapier. well yeah you know we launched our zapier sdk last week that's probably the most exciting news this uh you can install it inside of you know cursor cloud code codex any coding agent you might use uh i use it i'm not an engineer and i basically built like an entire ai chief of staff on top of this team thing and uh you know it runs in the background i don't have to close my laptop uh or i can close my laptop lid and not worry about uh things breaking so So it hooks into all 9 ,000 different tools, and I have it running all over the place.

1:31:00Wade Foster:So that's probably the exciting news out of Zapier's land. Do you feel like the vibe coding boom generally is more heavily tilted towards internal tools, dashboards, reports, chief of staff, B2B applications? Because it feels like clearly people are consuming a lot of tokens. They're spending a lot on inference. The models clearly work. The coding agents clearly work. At the same time, I open up the app store and I'm not seeing the, oh, wow, somebody vibe-coded a Call of Duty competitor that's blowing up on the Steam charts, right? We're just not in this, everyone's talking about the new app or the new Uber or the new game that's been created much faster with a leaner team.

1:31:49But everyone has a story of, we pulled forward our roadmap at our company. Yeah.

1:31:55Wade Foster:We've definitely had a renaissance of internal tools internally. We, our marketing team ran a hackathon recently. It's about 50 people on the marketing team. At the end of the week, there was like 80 plus new internal tools. Yeah. Most of these were like dashboards and, you know, data visualizations and things like that, that in the past marketers like especially are pretty analytical. Yeah. But they may not be like the best at like using SQL or running these things. Totally. Getting access to the data has been clumsy and hard. And now it just feels like they've got a jetpack on where they can just go build these like really quick, quick, simple tools that allows them to make like hyper specialized, hyper customized campaigns because they can quickly run an analysis and see, oh, you know, across the last month of like social posts, these are the ones that took off.

1:32:41Wade Foster:Here's the common traits about them. So let's figure out how to like tweak our next campaign for the next month to be more like this and less like that. that type of analysis would have been really difficult for a marketer to do solo, you know, in the past. Now it's, you know, less than an hour of work. Yeah, I think Terrence Tao had a, the mathematician had a point about this where he was saying that like in his papers, he will use AI agents to create new visualizations that would have taken him multiple days to make. Like now he, but it's not like it's saving him time because in previous papers, he just wouldn't have put the visualization in.

1:33:20And so it's this element of like people just doing more with more. But I'm wondering like, as you see, you know, so many companies have done this where there's a hackathon for the marketing department or some organization. And the prompt can often be like replace yourself. And I'm wondering if you're seeing that actually happen or if it's more like the competitive dynamic is everyone's getting more leverage, doing more things, and just out of the 20 ideas that they have, they're able to do 15 instead of five?

1:33:52Wade Foster:Yeah, I definitely think it's more the latter right now. Across Zapper, 100 % of our employees are using AI daily. I would say this is like individual AI usage. People are individually more productive. But if you were to ask me, is the company more productive? Institutionally, are we more productive? a lot harder to say yes to that. And I think that this is where like the company is on the cutting edge. This feels like the next thing that everybody's working on is how do we actually accelerate the institution? And this gets harder because you actually have to rethink how the company works. To give you an example, we've gone from a world where code was expensive to now code is cheap.

1:34:30Wade Foster:In the world where code was expensive, you had all these processes up front to talk to customers, to align on roadmaps, to figure out what the right thing to build was. Because if you chose the wrong thing, that was a really expensive mistake. And so all these humans and all these processes exist up here. But when code is cheap, you can get rid of all that stuff. And that takes redesigning jobs or rethinking the human, where the human actually exists in the first place. And so you can keep running the same. If you were in YC and you were about to talk to your customer, just cancel the call. You don't need to do it anymore.

1:35:04It's all good. That's an old, no, I'm kidding. It's a very good point. It's like, hey, show the customer. Yeah, exactly. That's the thing. It's like, you don't need to show.

1:35:13Wade Foster:Yeah. It's like a non-tech founder in the past would like show up and like give them a survey or like ask them generically about the problem or something like that. And you would get some signal from that. But now a non-technical founder can literally show up and say like, would like, can you just use this? Yeah. And see what happened. And that gives you a lot more information in the past. And so it's just a much different learning cycle. And I think companies that are not in YC, companies that have been around for a while, have all these systems and processes that have been built up that have to be torn down and reassembled to work in the AI area.

1:35:48Wade Foster:And that's where I think the institutional AI really kicks in. Yeah. How have you been grappling with the idea of work slop? There was this great cartoon early on where the joke was, take these three bullet points and turn it into an email and then on the other side, somebody says, take this email and turn it into three bullet points. And people should have just been sending three bullet points to each other the entire time. And I feel like we are hyper in that world where I can send you a full essay, a deep research report, a book if I want, a dashboard, a visualization, a hundred page deck. And at the same time, sometimes the human intuition of knowing, okay, this was a good post.

1:36:26Let's just do more of that might be the right intuition. So how have you grappled with, you know, fine tuning the team on, okay, it's great. Everyone should be using AI, but let's stay away from work slop. Yeah.

1:36:38Wade Foster:My co-founder Brian says you can delegate the work, but you can't delegate the accountability. So, you know, if you're slinging this stuff, you got to stand behind. It's okay to present AI author work. I do that often. Oftentimes I will say, hey, I was going back and forth with my AI friends. Yeah. Here is what I found. I have read this and I actually like stand behind it and agree with this. But if you're just going to like, you know, throw a random prompt in and then pass it off as like your own work. Yeah. That ain't, that ain't cool. And so like we kind of try and coach people on that, the kind of etiquette around that kind of stuff.

1:37:12Wade Foster:Um, yeah, yeah. I had that early on. Uh, I was working with somebody and they sent me something that had clearly been, you know, hydrated by an LLM. And I was like, you, you can just send me the prompt because I can actually imagine it exactly what it wants. And like, yeah, if there's a fact that we need to look up, we can use it for knowledge retrieval. But I actually don't need all of the hydration. You can just, you can just send the prompt. You guys are a remote company. Do you think that gives you an advantage in the AI era? Do you, you know, what are the ways? I think the thing, the biggest advantage we have is every last bit of work exhaust is documented.

1:37:52Wade Foster:So all of this stuff is inside of Slack. All of our meetings are recording. Every last inch of work that happens, there is a written trace of that. And so we can put chatbots on top of that. We can put LLMs on top of that. And that creates a whole bunch of institutional knowledge that accelerates the work. So a new person coming in can literally figure out, is there a standard operating procedure for this? Is there a skill for this? Is there whatever to do this thing? And you don't have to go like chase people down in offices and tap on shoulders and sort of hope that the campfire wisdom like finds you.

1:38:26Wade Foster:So I do think remote companies have a big advantage because they do tend to have so much of the work has a digital exhaust. Yeah. How, how are you thinking about like the, the volume of content? Is it just like stuff, everything in one large context window? Are there like, Are there agents that are going around and creating little roll-ups of things? And there's processes for teams to boil things down? Because even just the explosion of Zoom call recording and note-taking can produce so many transcripts. And I've been sort of shocked by the slow speed of just search in every tool that I use. Because there's so many more words that when I search for receipt in my email or something, I get every single email possible because they all have so much text now.

1:39:17Wade Foster:So what we have done internally is we've built a company brain out that has canonical data. So it exists at a couple different layers. There's a set of company data. This is like our company strategy or company values or ideal customer profile, things like that. That is curated by me and a handful of other people that sort of say, hey, this is the truth in these areas. Then there's team data. And so every team internally has a similar version that kind of cascades down. And then finally, folks have individual content that they've sort of used a supplement that that's private to them. So anytime someone is like talking to their AI, they have this, you know, source of truth that has been curated.

1:39:55Wade Foster:Then, you know, folks know that they can pull in other additional content at any given time. And so for me, you know, if I wanted to pull in a meeting transcript or a Slack thread or something like that, I would point the SDK at both those documents and say, hey, I want you to go review those against the backdrop of this company brain. And so it sort of helps you manage the context window because you can take, you know, bite size chunks here and here, you know, have the like company brain is like the backdrop to all that stuff and get better insights versus just saying, look at all of Slack or look at all the meeting notes, which that would just not be very effective because there is a lot of noise inside of that.

1:40:32Yeah. Have you seen more? Have you seen acceleration either at the lower end below your average customer size or bigger companies onboarding? Like, I imagine that there are a lot of tailwinds right now for you. But what's where are you seeing faster adoption or more interest?

1:40:53Wade Foster:You know, I think small companies tend to just move faster. They just have less bottlenecks on this type of stuff. That said, I have been somewhat surprised over the last couple of years how fast enterprises have gotten onto this stuff. You know, they still have more procurement hurdles and, you know, boxes to check to really take off here. But they're not asleep at the wheel, which is kind of the meme is that, you know, these enterprises move slow and don't make decisions. I think they just have more considerations that they have to do But I do think smaller companies generally just are moving a lot faster on this stuff Are you feeling CPU poor?

1:41:34We've been seeing with the explosion of agents and just more work being done on the internet I could imagine a lot of your customers submitting way more requests There's an explosion in demand And maybe there's two ways One is just scale up your servers But the other one is maybe work on optimizations and rewrites. Now, the company has been optimizing for over a decade now. So I imagine things are pretty optimized. But how have you been processing just increased demand relative to, you know, just infrastructure?

1:42:06Wade Foster:I think the most interesting place we've seen this is new usage patterns. So agents, you know, interact with APIs different than how like a human might set something up. and you're also able to deploy them faster and at a larger scale. And so when we're hitting, I don't know, let's use Slack, for example. The way you might set up automation across your Slack channels looks a lot different than it did pre-AI because you have new use cases and new patterns that you want to use that just want more data. It just wants to look at more stuff. And so I do think that that's probably the most interesting changes.

1:42:47Wade Foster:The usage patterns are different in the AI world. Yeah. For some of the more basic workflows, what is the pitch for using Zapier instead of vibe coding something that takes my LinkedIn post and emails it to me or dumps it in my Slack? Because that feels like something that you could effectively one shot, but then you get caught in who's going to maintenance that? Are you serving that? What happens when the API changes? I feel like there's a lot of excitement about, oh, I can Vibe code this in an hour. And maybe people aren't thinking about what it takes to maintain a service. But how are you reiterating the benefits of being on Zapier?

1:43:31Wade Foster:I think the fact that we maintain all this stuff is a huge advantage. The other thing is it's storing your tokens very safely. A lot of these Vibe coding tools will say, hey, just copy and paste your API key here. And it goes into a plain text file that might get shipped up to GitHub and might get shipped somewhere else that's like, ooh, crap, we're in trouble. And so Zapier just tends to be a safer place to do this. Plus, we have all the vibe coding capabilities as well, too. So, hey, take my leads from this service and add them to that. Zapier is able to do a lot of that as well. So it tends to be a little bit safer, for a more reliable place to run these types of automations.

1:44:09Yeah. Timelines around fully autonomous companies where you can have an agent that you just kind of prompt it to, you know, this has been an idea that's been around for a while. You know, you just say, like, figure out a way to make money selling software, and it just goes off, and it maybe makes some software and figures out how to market it. this already kind of exists in kind of the trading world, hedge fund world, where they're creating algorithms that just go out and figure out ways to make money. But on the software side, I'm curious, given everything you know about automating business processes, when you think it's really possible.

1:44:52Wade Foster:To your point, I do think for some very simple things, it is pretty close. trading, whether you're actually good at it or not, TBD, but pretty dang close there. I also have a former Zapier employee, this guy, Nat Eliason, who made it. We love Nat. Yeah, he's an open claw that started making money on courses. And basically, open claw was just figuring out how to go do it. And so, it's pretty close. so you're starting to see there's also use cases like I think of like a simple app in the Shopify app store where you could just say like go compete with this app and then it's like pretty obvious where you need to advertise and you can figure out the equation between like you could just make a clone of the app and then figure out the equation of like what is the minimum amount that I need to charge in order to be able to spend enough on user acquisition to like eke out some like very thin margin that covers the cost of inference.

1:45:57Wade Foster:Totally. The thing I don't know is if it's truly going to be a zero person company or not. Like it's does still feel like you need a human to like supervise this thing to like, you know, when, when it goes off the ropes to actually have the initial idea. Uh, so, you know, I don't think we're at this dream where you like poke the AI. There's always going to be, yeah, there's always going to be qualifiers. Even the one person, $1 billion company that was in the New York times. It's like, okay, like he hired his brother and And then also he had a bunch of contractors. And so there's never like even if you do create like an autonomous business, like if it's working at all, then suddenly there becomes an economic incentive to maybe, hey, maybe I should spend an hour on it.

1:46:36And then like, oh, is it still really autonomous if you spend an hour a week on it?

1:46:41Wade Foster:Man, that's the thing I think is also going to make it hard for these things to be crazy pervasive is that if someone is being that successful, there's another person that would look at the thing you're doing and say, well, I can do that too. And let me go compete against your margins. Yeah. How do you think about terminology? Jordy always gives this example of a company that had workflows and they just rebranded it agents and they didn't really change anything under the hood. And it feels like there's immense pressure from investors and marketers to just use the latest lingo. And sometimes that's useful because the capabilities change.

1:47:16And so the nomenclature should change. But we're in this weird continuum where there's there's workflows, AI workflows, agentic workflows, full AI agents. Like, are these meaningful definitions? How have you been processing all of this since you're sort of in all categories?

1:47:33Wade Foster:We definitely have a slide that explains the difference between all three of the things you just said. Yeah. Practically speaking, though, they're all agents. You know, I'm sure like somebody who is, you know, like a master wordsmith would like, you know, try and debate the finer nuances of it. But what I see in the market is practically what people are saying is an agent is a thing that does automation for me. Whether it's purely deterministic or whether it's purely agentic, it's kind of an agent at the end of the day. Now, you know, deterministic workflows have some advantages and disadvantages.

1:48:09Wade Foster:agentic workflows have some advantages and disadvantages. As you start to go build these in production use cases, you probably should understand the difference between these things. But practically speaking, we see customers showing up and they haven't really figured that stuff out yet. They're just like, I have a problem and I want that problem solved. Yeah. What about vibe coding? Like how broad do you think this trend goes? I saw someone, I think it was, who was it? downtown Josh Brown joking about how he's getting sent like a new vibe coded thing every single day. He gave the example of like somebody that makes an app that like displays their Spotify playlist nicely.

1:48:53And he's like, you're not my child. Like I'm not going to spend time on this. I don't, I don't owe you any time. But of course. Yeah. I'm just wondering how you're processing. Like It feels like this year, vibe coding is going very mainstream. A lot of people are toying with it. What do you think retention looks like? What do you think the knock-on implications are of this? Any security concerns? Just how are you processing this idea that the number of people that will have created software will probably 10x this year?

1:49:24Wade Foster:I think it's very exciting. You have all these folks who have had ideas or are coming up with ideas for problems to solve, etc. But in the past, those ideas have kind of been locked away because they can only implement like a small piece of it. And now with Vibe Coding, it feels like you can do a lot more stuff. And so I think that's why we've seen such a surge is that everyone feels like they can get these creative ideas out of themselves. I do think as an industry, there's going to have to be a lot more stuff built out to make these things work in all the various use cases. I think one of the places Zapier is like very helpful is, you know, connecting to all your data sources and doing it in a secure way.

1:50:04Wade Foster:Like we'll handle all your authentications clearly. And so that's a way that we can sort of, you know, play a really important role in how people go build out, you know, these infrastructure. And it can move less from vibe into more trusted, you know, infrastructure. Yeah. Yeah. Makes a lot of sense. Jordy, anything else? No, this was great. Well, thank you so much for taking the time to join us. Have a great rest of your week, and we'll talk to you soon, Wade. Good to see you. Have a good one. Before we bring in our next guest, Kevin Warsh, who is the new Fed Chair nominee, had a financial disclosure and has a ton of startup investments.

1:50:42He's in SpaceX and a whole bunch of other companies. Zero Hedge says he's a one-man EC machine. Brave, the browser. Cognition. Wow, Clay is a banger. The tech-enabled patent law firm, DYDX. Omharm Protocol. He's in a lot of stuff. I think some of this is through funds that he's invested in, but still, it's a whole bunch of companies, and it's interesting to dig into. He also owns a horse racing stable, something like that. Did you see this? It also includes his role as a general partner in Vicarage Stable, a horse racing operation. Let's go. He is invested in all sorts of things. So he will be disclosing all of that.

1:51:30I'm sure people will dig in more as he goes towards the committee vote. But without further ado, we have our next guest in the waiting room. We have Ankur from Kerry. He is the founder and is here with an exciting M &A update. How are you doing? I'm doing great. Looks like we both have some money to announce in the last couple of weeks. Yes. Great to see you. First, let's start by you. Start with the gong. Let's start with the gong. Let's hit the gong. There we go. Now we can continue. But let's talk about the company that you built, the background, maybe your background even before starting this company that was just acquired.

1:52:17Yeah, absolutely. Before this, I used to run a company called teachable.com, which helped people make money online for the first time. So it helped people create online courses and coaching and had a successful exit there during the pandemic. Took a couple of years off trying to figure out what I wanted to build. And then I started Cary about three and a half years ago. And the mission with Cary was we helped people first make money online. Now we'll help them grow their net worth, be smart about taxes. I think I came on the show last to talk about taxes, so a lot of that stuff. And now here we are, three and a half years later, excited to share that we've been acquired by two separate companies.

1:52:53Yeah, explain that. It feels very uncommon. How did this happen? Who's going where? How did this deal? It's my second exit. I want to get acquired twice this time. At this rate, the next company, you might get acquired by four different... Yeah, twice the diligence. Nothing like running two simultaneous M &A processes. But no, the overall company is bought by AngelList, which is a platform I've had a long-standing relationship with. My first company raised money on AngelList. I hired a lot of my team there, ran my funds there. They saw the marketing engine we've built. They saw our ability to build and launch financial products, and they wanted to buy us to build and launch new financial products to make the private markets more accessible.

1:53:38As they did this, we obviously have built what I consider to be the best solo 401k self-employed retirement platform. They didn't really have a logical kind of use for that asset. So we partnered with a company called Lettuce Financial that does tax, bookkeeping, payroll, insurance for self-employed people. And that's where we're selling the retirement platform. So it's almost two separate acquisitions. So yeah, what does that look like from a customer perspective? I mean, maybe walk me through the actual go-to-market for the company while you were building it, what the last couple of years looked like, because then we can go into what the experience for the customers is like.

1:54:16Yeah, so our go-to-market, which I think we were quite good at, is we never spent a single dollar on paid marketing. I think a lot of it was content education, teaching people about the insanity of the tax code. There's something fitting, by the way, in announcing our acquisition on April 15th. It was completely unintentional, but it's funny how that happened. The amount of times I would read one of your posts and then send it to my CPA and just be like, please make sure that we're doing this. Was the marketing strategy GoDirect? Content education. What platforms worked the best? We had a personal finance newsletter with over 100 ,000 people, which worked quite well.

1:55:01Twitter has always worked well. I do think there's some percentage of people who follow me on Twitter are going to be relieved to see less tax content. A lot of buddies be like, dude, Leo. But look, I think I'm excited for the next chapter with AngelList where we will apply a lot of the same go-to-market to launch new products for the private markets. Meanwhile, Lettuce, the platform stays exactly the same. Customers will only get a better experience since Lettuce allows people to also do taxes, payroll, health insurance for self-employed people. I'll continue to stay involved there as well with more of my day-to-day time on AngelList, but most of the team ends up going to Lettuce.

1:55:42What was the founding team like? Who did you hire? How big did the company get? Talk to me about the shape of the company. Yeah, absolutely. We got to about 25 people, about 4 million in ARR. The founding team had four other co-founders. The business was at a point where it was growing quite nicely where we got to about$250 million in assets on the platform. However, looking at all the things happening in the market today, you look at it and I'm like, three and a half years in, four million ARR. It's a good business. But you look at some of these modern AI companies and you're like, wow, it's a good business, not sort of the crazy multiples you see these days.

1:56:24How do you think about the, I mean, I guess like level set for me on the boom of the solopreneur, like what have you seen there? What has stuck out to you? And then I want to talk about that in the context of AngelList. Yeah, so we stumbled upon this in my last company where for the first time we saw, you know, tens of millions of people start a business online. I think marketplaces made it very easy for anyone to get attention. And when you got attention, you could start to monetize this. So we saw that first happening at Teachable, but then kept investing in the creator economy and seeing the growth of that.

1:56:58When it came time to launch the new business, which was Carrie at the time, realized that so many things were provided to you by your job, like your retirement plan, your insurance, all of that. As a self-employed person, you may be financially doing quite well, but you have to go figure it out for yourself. and the solo 401k specifically was one of those too good to be true sort of personal finance things where it's a private 401k plan just for yourself. You can invest in literally any asset you want. Your money grows and compounds tax-free. You can use it all to contribute to a Roth IRA. You can borrow money from it.

1:57:34You can get a tax credit for setting one up. I was reading this and I'm like, this is insane. Why isn't anyone doing anything about it? So that's where we built Cary where we kind of help, I think, 4 ,000 separate people set up their retirement plans and custody their dollars and all of that. And then within the AngelList ecosystem, how do you see, it feels like there's a continuation of the solopreneur boom. Vibe coding is obviously a big piece of that. I definitely agree with a lot of people that have come on the show and talked about this idea that maybe we're not at a one person, $1 billion company yet, but just the ability to build something that is software with a smaller, leaner team sort of lends itself to maybe lower capital requirements, not needing to go the traditional venture path because it might look more like a lifestyle business.

1:58:25But I'm wondering how you think the financial markets will react to that dynamic. Yeah, absolutely. I think we're seeing a lot. The rise of the one-person business is only becoming bigger, but the AngelList acquisition was also, we spent a lot of our time educating people on investing and being smarter with money. And I think there's still a big opportunity to make private markets more accessible. I mean, we're seeing this with companies going public later and later. So much of the wealth creation, people are locked out of. I mean, you can buy prediction market contracts and how long a handshake will last.

1:59:00You can gamble on shit coins. So I don't really buy the whole, we're protecting people angle. So the idea with AngelList is we're going to spend some time and figure out sort of how can we take some of this wealth creation that's historically been only for a very few people and make it more accessible, but in a way that's actually fundamentally good and investor-friendly products. Yeah. In a perfect scenario, do you think that the public funds that invest in private market companies should be basically valued at 1x nav because we've seen a couple things go out and they get really highly valued and it feels like some sort of like okay the market's not efficient here.

1:59:45Yeah, I mean, I think there's a couple of very, very big IPOs coming up that will determine, I think, like, I know I'm going to sound dramatic, but the future of the private markets in a lot of ways, because there is such a massive disconnect right now where you have so many fundamentally good companies in the public markets getting completely slaughtered. Meanwhile, all kinds of private businesses are, you know, continuing to rip. So I think there's going to be a reckoning. I think companies are staying private longer. But I would imagine Anthropic and OpenAI have unofficial markets that make them more traded than a lot of public securities.

2:00:24So the line is blurring. There are effectively stock prices for a lot of these late-stage private companies. So I think it's a really interesting time. And it's, to me, inevitable that we allow private markets to access more of this. Yeah, that makes sense. Well, I know you guys have some big news. coming up. So I have a lot more questions, but we'll save it, uh, for when you guys are ready to announce, which, uh, I'm looking forward to, but, uh, congratulations to the whole team. I think it's super, super smart to, you know, make this decision and I can't wait to see, uh, what you do at Angelus.

2:00:57It's going to be incredible. Yep. Appreciate it. Thanks for having me on guys. Thanks so much. Have a good day. Up next, we have Bailey Pumfleet from Cal.com. Cal.com is considering going closed as AI agents overwhelm the open source ecosystem. We're excited to have Bailey join us. How are you doing? What's going on? Welcome to the show. Hey, how's it going? It's been a busy day. Yeah, very busy day. Please introduce yourself and the company. We've reacted to some Cal.com posts in the past, but it's a fascinating company. So take us through the shape of the business, a little bit about yourself, and then we'll go into the decision today.

2:01:37Yeah, Cal.com has been an open source scheduling software. We've been around for like five years. The whole company is around being open source. In my company, we're huge fans of the open source space. And yeah, today we made the announcement that we're actually going to be closing source, which is a very tough decision and quite a controversial one. Before you continue, your audio keeps kind of cutting in and out. Do you have headphones? Or I don't know if you're kind of covering your mic at all. I don't know if you're picking that up. Yeah, it sounds a little muffled. Hearing every other word.

2:02:11Oh dear, is this any better? That's a lot better actually. Whatever you did, what worked? That's a Zoom magic. Yeah, thank you. But yeah, so we started this company as huge believers in open source and unfortunately a lot of shifts in AI. The whole perspective has completely changed. AI is now able to break code at completely unimaginable speeds. It's the one thing that nobody's really talking about. We've seen little drops about Anthropics mythos model, and nobody has really taken the time to just kind of understand the ramifications which things like that can have on not just open source, but broader application security.

2:02:57What was the business model in the prior era? If you're an open source software, how do you keep the business running? Fundamentally, the business model has not changed at all. So we've always been open source and the code has been open. And that's mainly for the things that the average person would need to be able to run their own scheduling service on their own domain or something like that. As a business, we actually sell this software commercially. and open source is actually something that benefits us commercially because we can go to people and we can say, hey, you can look at the source code and you can verify that we're not doing anything sketchy with your calendar data.

2:03:41What, if any, are the benefits of being open source during the AI agent boom? Because I imagine that you are getting automated pull requests and vulnerability reports. Are there any silver linings, things you put in the pro column before you realize that the con column overwhelmed the pros? Yeah, I think a lot of these things remain to be true. With open source, everybody can audit your code. So especially with AI producing a lot of slop nowadays, the one thing that we have going for us is that we have code which is written by humans and reviewed by humans. And so it definitely creates a lot of trust in that sense.

2:04:25It's just that, you know, for us, there's a lot of pros really. And we also just genuinely care about open source. We care about trying to do the right thing with this business. We're going to make money and we're going to do our thing whether we're open source or not. And I think one of the common misunderstandings about this discussion is that it's some kind of like business decision that it leads to greater profit for us. But, you know, really, we just always try and do the right thing by our customers and by our community. Yeah. Talk about reputation farming attacks. I haven't heard that much about this.

2:05:04What is going on in the open source ecosystem with reputation farming? What do you mean by reputation farming specifically? Specifically, like agents that are trying to go and make low quality contributions in order to, you know, earn some sort of reputation in the open source ecosystem so that they're more likely to see other pull requests accepted. Yeah, I asked for clarification there because there's actually a lot of different sort of like branches of reputation farming here. You have, for instance, these people out here who are just trying to use AI to attack open source, to just get cheap bounties.

2:05:47You get people who are using AI to make small contributions to open source. And maybe they might be trying to get open source bounties for developing code. Maybe the end goal in sight is to get a job at one of these companies or something like that. But just like how we see a ton of AI slop on LinkedIn and things like that, we're seeing a lot of AI slop on GitHub. Granted, for us, that doesn't really affect everything that's happened today. We don't make any decisions because there are some AI slop pull requests. But that is something I'm hearing a lot from other projects where it's really hard for smaller teams to deal with just the sheer amount of review workload.

2:06:32How are you? Is there, I mean, it's weird because you're already open source. There's always the possibility that a paying customer would say, you know what, I'm just going to self host this. So in some ways, by being open source, you don't really face the competitive threat of, oh, I'll just vibe code this software and I won't use something off the shelf or I won't pay you. But is there a business threat from AI? Is the business threat greater now? that, I mean, obviously the model capabilities are greater, but being closed source sort of incentivizes people to say, hey, if I want to get it for free, I should Vibecode it.

2:07:10Yeah, I think the question about can Vibecode and replace my SaaS startup is, you know, the hot thing right now. We feel pretty confident in our defense of that, whether we're open or closed source, because scheduling is so fragile and so nuanced. Like if you say, okay, I can build a basic scheduler in a weekend, I probably believe you. You probably can. But to build something the scale of Cal.com that actually works in all these enterprise use cases and things like that, it's a lot harder, to be honest. And you're going to run into so many little hiccups that just AI sort of vibe coding can't currently sort of hit.

2:07:50I have to imagine that the models will be able to to create a good scheduling app, but it's more about, and I wonder if you agree with this, but it feels like it's more about the potential of just a company prioritizing their time because there's, if there's so many other things and you're spending your time rewriting and maintaining all of your custom in-house vibe coded tools, if you just pull something off the shelf, even if it just costs a couple of, you know, 20 bucks a month or whatever, you're just going to have more mind share towards oh the scheduler broke or it's down or there's a security vulnerability we need to patch it instead you can go focus on whatever your actual business is and so i imagine that the prioritization is a big factor there too yeah i mean it's the same way nobody's going to build their own stripe for payment processing or you're probably not going to build intercom for your customer support you absolutely can do that but is it really worth it at the end of the day.

2:08:50For us, sort of our business model has never really relied on sort of gatekeeping anything, but just more that bringing something out of the box and just deploying it and getting selling is the most important thing to pretty much every founder. Yeah, well, thank you so much for taking the time to come and explain it to us and break it down. Good luck with the decision and the reaction from the community. And one of my favorite.coms. It is a great.com, Cal.com. Thank you. still underrated even after all these years building on it well have a great rest of your day we'll talk to you soon have a good one goodbye up next we have Han Wong from Mint LaFi he's the co-founder and Allbirds the markets are closed but Allbirds is down almost 10 % after hours off of the high?

2:09:40It ended at 17 okay that's pretty high that's a very big departure from a$10 million company,$20 million company yesterday, and now it is over$100 million company. That is a massive gain. So we'll see where Allbirds goes. They have a lot to prove, but we're rooting for them. And we're also rooting for Mintlefy. Let's bring in Han from Mintlefy. How are you doing? Welcome to the show. I'm doing really, really well. Hey, thanks so much, guys, for having me on today. Thanks so much for coming on. Love the energy. Are you fired up? You seem fired up. Oh, all birds proved to the rest of us that you can just do things.

2:10:17Yes. And that's the energy that we aspire for. Indeed. Yeah, it was a wild story. But we're not here to talk about all birds anymore. Tell us about you. Give us your backstory and then tell us about the company and then we'll get into the news. Yeah. Yeah, absolutely. So my name is Fawn, one of the co-founders of Mitlefy. We founded the company in 2022 really off of the simple idea that we wanted to help enable other builders. I've been a programmer since I was 11 years old, and it's just basically been defining elements of mine. There you go. I'm so curious. I know, I love it. and basically wanted to build a company enabling other developers because it was who I was and who I fundamentally am.

2:11:11And now, cut forward a few years, Mitlefy now powers the docs for companies like Anthropic, Microsoft, Coinbase, OpenClaw, Perplexity, and now well over 20 ,000 others. That now reaches over 100 million people every single year, which is something that we're incredibly proud of and of the impact that we have. Yeah. So, I mean, fundamentally, developer docs are taxed on an HTML website, essentially. But what's usually unique is the harness and the environment that migrates any changes to the code or API endpoints onto those docs. What was the first approach to actually deliver something unique or more enterprise-ready than, say, some of just the open source reverse engineering the URLs and just statically serving it, which is what I think most people will be familiar with?

2:12:09What was the V1, like, okay, we have product market fit. This is better than the status quo. And then we'll go into the AI era. Absolutely. Initially, when we started, it was off a very simple idea that anything on the market wasn't fundamentally built for developers. Which is so ironic because it's the developers who are often maintaining the content. It was developers who are often reading the content. And everything that was out there was just kind of like static site builders that were often adhering to a different audience, it felt like. And having personally suffered from so many bad docs out there.

2:12:49I was like, hey, look, like if we're going to go build one, let's go build the one that we always wish we had. And let's make that great. And it turns out those opinions that we kind of made into building this platform really resonated with a lot of the industry. It started initially with a few of our YC batchmates, then grew to about a quarter of the entire YC batches. And then it grew eventually now to serving some of the largest companies, including Microsoft and others. So how do you think about integrating AI? I imagine that there's more docs than ever that need to be written. Those docs need to be read by AI.

2:13:32But then also whenever a new endpoint is created, even if it's created very quickly, it needs to be documented equally as quickly. how have you integrated AI into the product? Yeah. Many different ways, but I think it's really important to first take a step back and understand the changing role of all of this. When we started the company in 22, it was a very simple idea. Docs are almost like, you can think of them almost like an instruction manual. When you assemble, let's say, Lego bricks and Lego kit, you have to read the assembly manual in order to figure out how these things work. It's written by people.

2:14:13You don't have to. You can do it the old-fashioned way. There you go. But I get your point. But now it's really important to understand that 50 % of the viewers for the content is actually not humans anymore. It's actually AI. So we basically took a look across all of our data a few months back and we're like, look, we felt the change in AI agents being the end users, what did that actually look like? So we went through, proxied our viewers, identified what are agents and what are humans. And out of curiosity, John and Jody, I'm curious, if you guys had to guess, what percentage of traffic do you think now comes from agents versus humans?

2:14:55Didn't you just say 50 %? Oh, I said 50%. I was like, trick question. It's actually 500%. 50-50 is an interesting place to be because I feel like it was probably 0 % a couple years ago, and it'll probably be 99.999 % very quickly. Yeah, I would have actually probably been higher. I would have guessed higher if you hadn't said. Yeah, totally. Yeah, literally. It went from 15 % about 12 months ago to about 50 % a few months ago. And I would say now it's closer to 60-70%. Sure. And so we have a very strong opinion that by the end of the year, to your point, it's going to start to look like 90 plus percent.

2:15:35Yeah. Right. And if for no other reason than a developer will just be using a chat app to interface and ask the questions that they want and have the chat app go to the particular website, pull everything in, contextualize it at their level based on their memories, based on those particular questions they asked instead of opening up 12 tabs and scrolling through a bunch of different endpoints. Exactly. Exactly it. And also, if you think about even what the developer work cycle is nowadays, right? It's like, I'm getting here to ask Cloud Code to go, you know, Vibe code my product, Vibe code my app.

2:16:12Sure. And if, for instance, I'm like, hey, let's go and use, you know, let's figure out how to implement Stripe or a new feature or an integration, it's going to, the first thing it needs to do is actually crawl through the content to figure out how it's done. Yeah. Right? And so to kind of tie it back into the, you know, Lego kit assembly manual, well, if you're task and they had to go do it, the first thing it's going to need to do is to read that instruction manual. And so we view the change of content out there, not so much like a platform or like a website, like it used to be, but like a crucial part of infrastructure for agents to understand how to implement, how to understand the world and how to even maybe even surface like, uh, you know, whether or not you use your product because of what it knows.

2:16:54Uh, how do you think about that? I mean, you have some huge customers, like the biggest companies in the world. At the same time, like random people that don't even have a business are like vibe coding software. And do they need to be releasing developer documentation? Will they be your customers in the future? Is the long tail going to get longer or fatter? Or do you think you're going to stay focused on enterprise consolidation, like the really, really high traffic, the important endpoints that get pulled off the shelf from agents consistently? Yeah, that's a really good question. What we're seeing is a bit of both.

2:17:33So while it's more important to optimize for the heavy hitters, because those products are getting more usage than ever before, and to make sure that they're really well optimized for agents, it is also equally important for your next two-person YC startup or your billion-dollar one-person company to make sure that their product is also out there. Because in the age of AI where agents are making the decision, how well it understands how your product works, right, is what informs of it of, like, whether or not you use your product at all. And so discoverability, right, making sure that your content is out there available is just more important than ever, especially in the age of where software is, you know, starting to get commoditized as well.

2:18:21Yeah. Tell us about the round. what happened yes uh we raised uh series b how much how much uh 45 million dollars congratulations i'm honored i'm honored i got the gong of course of course uh last question for me uh how how did you answer the saspocalypse question obviously you guys are using a bunch of AI, you're building agents, but I feel like that must have come up during the round and I'm sure you had a good answer because you put it together. Yeah, of course. So what we fundamentally view is that the SaaSpocalypse is fundamentally more of an enabler for us than anything else. As more products gets built because software is actually, you know, the cost of barriers is cheaper these days.

2:19:16Things like discoverability, how well you expose your content out there is just far more important than ever. And this influx of companies being created, companies being out there and making sure that they expose their information out to the world and agents has kind of what we've seen in our data, this massive tailwind of adoption for our product. And now if we're here saying that most of the viewers of your products or how to use your product or agents, having every company enable that is what we've seen as kind of just this big tailwind and explosion of usage across the board. And especially since Mitmify, the company, have taken this developer-first approach into building docs, it actually was really beneficial for agents in actually spinning these things up faster than ever before as well.

2:20:07Yeah. So more companies pulling it off the shelf and not a lot of pressure. Yeah, there's more software companies and then also an agent can just use Mintlify to build docs. Yeah, very cool. And I think the really important thing to note is like, even from the ground level, we've only seen the adoption of our product go up significantly in the age of agents. Interesting. And I think as more companies, you know, have more competing priorities and just more things to do, in the SaaSpocalypse, I actually think the build argument is actually stronger than ever before to preserve focus for what really matters.

2:20:47Yeah, that makes sense. I got to say your customer page is absolutely stacked. You got Coinbase, AT &T, Cognition, Browser Base, Anthropic, Fidelity, CalSheet, Decagon, Lovable. Yeah, it's everybody. PayPal, Perplexity. Yeah, and this is a good definition of like, you know, you're an AI company, if you're accelerating revenue right now in the AI revolution. Yeah. Axe. Wow. Amazing. You're going to run out of company soon. That's the bear case. You're going to run out of customers. Yeah, I like the design language there where the logos, we'll have to pull it up, but the logos have these generated images behind them that have this very cohesive style, but they all have slightly different flavors to each logo.

2:21:31I really like the design. I thought it stuck out. Awesome. Great to meet you, Han. Chat loves you too. Thanks for putting up with our sound effects. All right. Thanks so much guys. We'll talk to you later. We'll see you soon. Have a good day. All right. We got to close with the farm people who are doing schizo edits now. It's in the timeline. 4 ,000 people liked it. I missed this. You missed this one? Look at this. I imagine this says - Oh, this is the pig pageant? I've seen people direct their pigs with the little guides. We have some SD Kid here. Oh, it is SD Kid. There we go. No, I don't think it is.

2:22:10It sounds like in that genre, but the background is crazy. The rotoscoping is extremely sloppy, but that is the essence. All right, enough of that. I'm going to give you a headache. Enough of that. Enough of that. But TBPN is going to wrap for today. Yes. But our friend, Dwarkesh, has Jensen on his show. Yes. And I'm going to listen to that on my way home. Go over there and listen to that. Thank you for tuning in. We will see you tomorrow at 11 a.m. sharp. It's been an honor and a privilege. Give us five stars to Apple Podcasts and Spotify. Sign up for our newsletter, tbpn.com. And we will see you tomorrow.

2:22:43Goodbye. We'll see you soon. Love you.