Cluely Raises $15M from a16z, $45M Modern Hacienda in the Wild | Steven Sinofsky, Anupam Singh, Miles Moen, Alexander Sen, Bryan Kim, Roy Lee, Dave Fontenot

29:52Two of the most famous people in the world are technology entrepreneurs. Donald Trump. Donald Trump, social media network, true social. Crypto entrepreneur as well. IPO'd his company in almost record time. Yeah, yeah. And then obviously Elon. Yeah, I mean Elon is like a legit celebrity, yeah. He did SNL, like that's crazy. I mean Trump did too, so. But everyone's a founder. Hayley Bieber, billion dollar outcome, right? Yeah, so I mean, Anna's clearly getting at something real. Should we read the rest of this? This has become some fast, this has some fascinating real world implications. The amount of young people who have now been one-shotted by TikTok edits of the social network or performative how I got into YC videos are, which I have made.

30:32I literally made how to get into YC video. I'm probably responsible for a little bit of that one-shotting, are growing exponentially. Being a founder implies intelligence, status, risk-taking, wealth. So naturally, the cultural shift in the startup ecosystem has moved from contrarian rule breakers and hackers to the overachieving STEM kids for whom being a founder is simply the next box to tick. Money is not a motivator, it never was. What does this mean for VCs? They're the most mimetic of the bunch. So they'll fund the founder with the most curated credentials, most hype, most viewed product launch videos.

31:06They'll confuse attention for distribution. This also explains why YC 2020 and beyond will go to zero, hot take. They attract this and fund this exact type of founder. The emperor has no clothes. There will be no cultural defining winners for YC to maintain their prestige into the next decade. So then what happens to the builders of this decade? That's the trillion dollar question. Overly cynical, perhaps. but it does reveal a risk that threatens American hegemony. Next generation founders. It's getting into some tinfoil hat. Yeah, idolize TikTok and YouTubers. Why would this trend reverse? Could YC be infiltrated by the CCP?

31:41It is an underrated national interest. Poison the timelines where artisan, century-defining American technology companies are nurtured, reinforce the West's economic dominance, and it now funds mass-manufactured slop instead. I think that's a very odd take because yes, Yes, the YC companies do seem to be a little bit more tracked now. They're a little bit more on theme with the AI agent stuff. But you talk to the Teal Fellowship. And it's a reflection of what the timeline is interested in. Yes, yes. But what I'm saying is that there are still... It's not that the trend followers tracked kids have crowded out the risk takers.

32:17The risk takers who are there for ideological reasons or some bizarre countercultural reason, they're still starting companies. They're just doing it over here in a completely different pocket of the world. Yeah. In a completely different, they might be not on X or they might be not at YC. Well, the funny thing, we didn't have any of the YC hard tech founders from this last batch on the live stream. And it's very possible they just didn't want attention. They're just hacking away, right? They're like, I don't need media right now. And that's perfectly okay. So I don't know. Why do you have for the next one?

32:47I have the gong because SoftBank's Masa Son plans$1 trillion robotics manufacturing half of the U.S. I'm going to hit the size gong if you want to read through this. Okay. A U.S. version of China's Shenzhen manufacturing hub. Production lines for AI-powered industrial robotics. Already spoken with Trump admin officials about tax breaks for companies that are investing. Seeking partnership with TSMC, but not clear if TSMC is interested. Codenamed Project Crystal Land. Market clearing order inbound. Okay. Here. We should figure out. I'm going to look up and we're going to quiz Tyler now if he's ready.

33:26Is he good to go? He's ready. I got a thumbs up. I want to ask how much money has been invested in Shenzhen. I'm trying to talk into the mic but not give it all away. Yeah, yeah. Okay. So I am looking this up. Oh, no, I used O3 Pro, so it's going to be 12 minutes. Damn it. Okay. New tab. I will have to use 4.0 for this because I need a quick. So he's investing$1 trillion, right, to compete with China's Shenzhen manufacturing hub. I want to know how close that gets us. Wait, this is weird. I don't know. One of the$9 billion, more than$25 billion. This does not seem right. I have some various stats here.

34:16I think it's going to be pretty hard to pin down the exact amount. I'll quiz Tyler on this question. Tyler, over since 1980s, how much foreign direct investment has there been in Shenzhen? Let me think here. I would say since 1980s, over$100 billion. Estimates range from$100 billion to$150 billion. and Shenzhen is a major global manufacturing and tech home. Wow, you're so smart. It's actually kind of pulled me off. Pretty close. I mean, I think that the challenge is actually, there's so many ways you could kind of measure this, right? Yeah. I have a table. So there's about 100 billion of local funds.

35:06There was a state and AI robotics fund announced this year. there was a there was another semi conductor industry fund announced this year there was uh beats up until 2014 there was uh 65 billion of cumulative foreign direct investment um but but who knows and then and then run rating um you know i guess around 10 billion but also like does this count what apple is doing right that that's like another you know form of investment And anyways, this might be an answer that clearly, I guess, got pretty close. It was pretty good. Pretty good. That was good. I have a follow up. So, you know, the iPhones made over in China often flown via 747 to America faster than shipping on a cargo ship.

35:57How many iPhones fit inside of a 747? All right. Let me think here. I'm gonna say I would say about two to three million in a 747 You know so okay, so 747 max payload is about 140 ,000 Kilograms, okay each iPhone box is about half a kilogram Okay, so I would say you know a practical range of iPhones is probably 1.5 to 1.5 per flight But obviously it depends on you know packaging pallets cargo layout, you know that kind of stuff. That's pretty good. I mean, ChatGPT here has 9.6 million, but I think that's including like every inch of the plane, including like the passenger portion. It's very possible that Cluey is more accurate.

36:44I think that's more accurate, actually. I think that might be more accurate. Is Cluey goaded? It might be. We'll have to ask Roy, who's coming on the show later today. Yeah, this is so, so, so yeah, the ChatGPT estimate had the usable internal volume as passenger plus cargo and i it feels like you use just the cargo number is that correct uh i believe yes i i think so okay oh i see jgbt estimate 9.6 million oh it's streaming through yeah yeah okay so i was doing just the just the cargo this is fantastic it actually works pretty well right i i'm impressed so far uh this is basically young young jamie from joe rogan yeah except you don't even have to look anything up we just get to ask you live yeah yeah this is great well you know Clulee is dealing with a lot of personal data they're gonna need to put rigorous compliance that's right practices in place they're gonna need to go to Vanta.com automate compliance manage risk prove trust continuously but Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation whether you're pursuing your first framework or managing a complex program We actually should put Roy in the hot seat and make sure that he is on Vanta because I would feel a lot better with knowing that Tyler is using Cluelly in our office.

38:11Yeah, if they were on Vanta, for sure. Yeah, we'll get them to switch, for sure. Fully certified. What else is in the timeline? Yeah, I mean, just kind of closing out this Mossa story. It is funny. Great name. Pull the one T out of the hat, right? The one trillion card. I definitely don't think he's fully funded the$500 billion project Stargate. I'm sure he's sort of making strides towards it, but it's great to see that he's this long America. It feels like he's kind of, Masa is kind of the DJ Khaled of the capital markets. It's like he's pitching a$1 trillion US AI hub to TSMC and Trump. And so it could just be TSMC winds up building a massive facility and the government gives some incentives for that.

39:02And like he's not even he's like investing where it makes sense off of his own balance sheet. But he's like a SPAC promoter. Kind of. But it's more like he's just he's just connecting the dots between the different pieces of the global economy. He's a power broker. Yeah, he's lubricating the capital markets. The deal. Art of the deal. Yeah. I like the name Project Crystal Land. I think if capital formation was in the Olympics, Masa would be leading the Japanese team. Yeah, we've said that it's been said that the Nobel Committee should give a Nobel Prize for capital formation. And I think Masa should be a front runner.

39:39Absolutely. For sure. We have a post here from the Supreme Leader of Iran back in 2013. Is this real? This is a great checkmark, so it's real. I clicked through, I think. It's such a wild, antiquated posting style. You know, the hashtags. This is vintage. Read it. He says, woman is a flower. How evil of a man to treat a flower with violence and without appreciation. Hashtag violence against women. And somebody, Ahad, quotes this and says, born to be a lover, forced to be the supreme leader. It's a wild posting style. No spaces after the bullet. You don't see posts like this much anymore. No, no.

40:20You just don't. No spaces after the periods, using a hashtag, throwing in a random date. I don't know what that date's alluding to. The 9th, September 18th, 2000. Is that like right after 9-11 or something? I don't know. It's very odd. Anyway, I like this post by Ben Jarvis Green Ellis. He says, does the HDMI cord have the most dominant run of any cord ever? Hashtag witness. Hashtag witness. I don't know if power cord even counts because of its dominance, but it would sure like a spot on the podium. Power cord is like the commissioner of the league. Kwame says, type C is on a brawn-esque run right now.

41:02The goat debate will start in three to four years. That's true. But the HDMI cord has been around for like 20 years, and it's just been reliable. And actually, it gets upgraded pretty regularly, but they don't need to change the actual shape of it, so you don't notice. It's not really as annoying. but it's gone from, you know, it can do HD, high definition video, to now it can do 4K, 8K, you can do power over HDMI, you can do all these different things. Big fan of HDMI CEC over here. Are you familiar with this? HDMI CEC lets you plug an HDMI cable into a TV and then control the TV from the device that's connected over the HDMI.

41:38So you can go to the Apple TV and when you turn on the Apple TV, the Apple TV can send a signal to the real TV over HDMI CEC and say, hey, turn on and switch the channel. So it just makes it a lot easier to actually control the endpoint and for the communication to happen. It's pretty cool. We have more breaking news. NVIDIA to drop humanoid robots that will produce NVIDIA GB300 chips in Q1 of 2026. Nick says, holy based. That is so close. So NVIDIA is Foxconn in talks to deploy humanoid robots at Houston AI server-making plant. Wow, that is extremely close. I don't totally understand this. What does that mean?

42:17I guess Foxconn has been training the robots to pick and place objects and insert cables. Yeah. I don't know if this is marketing. It feels like. We've asked a bunch of robotics experts about humanoids. And so far, I don't have the confidence that this is actually a super great use case for them. It's just the definition of what is humanoid. Because you go to a, you go to like a Ford F-150 factory and they have like massive robotic arms, like moving windshields around. Right. Like, it's like you could anthropomorphize that by like spray painting it pink and putting some hair on the, on the hand and being like, it's a, it's a humanoid now.

43:02But like, you could like retrofit every bicep definition. Totally. I mean, like Amazon has like tons of robots sliding around. you could put googly eyes on them and be like, they're humanoids now. And you'd maybe get like a stock bump. But like, there's clearly things that are happening in the AI server assembly process that are using robotics, obviously. Whether it's even just like conveyor belts or the three axis pick in place, pick it up, put it over here, that type of stuff. Just going to humanoids is just, it's like, well, will it have five fingers? Or will it have just a grabber? Will it have legs or will it have wheels like it could just be sitting there because for a lot of these pick-and-place jobs You can just have the the humanoid sit there with a single arm Mounted to the ground because the stuff comes to it and so then you're just kind of in a okay We're in the arm business now It feels it feels like it's a little bit wrapped in marketing lingo But still cool that more companies are making humanoids because it does seem like a cool form factor that hopefully people will break and get there and this seems like a step in the right direction in the sense that it's a very defined task.

44:12I think when people think humanoids, they think some, like my new Turing test is, is humanoid robotics will be here and when they can put up a six minute Nurburgring time in a manual. Yeah, gated manual. In a gated manual. because at that point, they have to not just be a self-driving car, but they have to be able to negotiate the wheel and the stick shift so efficiently and so quickly that they are truly performing. Yeah, and a manual can be weirdly probabilistic, right? And that you can move the - Synthesize a lot of data. Yeah, it's not like perfect system, right? You're not going to put up a sub seven minute Nürburgring time with a Joe Biden walk.

45:08That's right. You're going to have to be moving fast. Those actuators are going to have to be high speed. So, yeah, I don't know where all this goes, but I mean, exciting to see that they're at least doing work on it because it seems like an important, it's clearly an important path in the tech tree. We don't know how relevant it is to other formats, but cool to see a lot of money pouring into the sector from very big companies. So Foxconn will be announcing their robots in November and then deploying them shortly after. Yeah. Our year 2026. I mean, Foxconn seems like the right builder for this.

45:47Unitree is already like, feels like it's scaling up to the point where Unitree robots should be usable in certain locations. Like they're not generalizable yet, but there's certainly, if you train them on a specific task, they could do that task over and over and over again. The question is just like, do you need five fingers, 10 fingers, 10 toes? Is that really like the right form factor? Or should we be just doing things that are more specified? Because if the whole point of these human robots is just AI server making, just assembly, just one spot on the manufacturing line, could probably be a much more specialized robot.

46:22But, will be cool to see. I'm sure we'll see a lot of viral videos about it. It would be very cool. I invested in a company making robots for data centers. Oh yeah. And they intentionally chose not to make it a humanoid form factor. Yeah, yeah. Which, and I think a lot of those, the reason behind that decision would also transfer to manufacturing setting, right? Yeah. Which is like, do you need legs? If you can use wheels. I mean, data centers have the most perfectly polished floors with like no dust at all. Yeah. It's the perfect environment for a wheeled use case. At the same time, you probably need to have a very specific actuator for unplugging and plugging the cable back in.

47:03And that's actually, it's pretty hard to reach around the back of a computer and unplug an ethernet cable. And obviously the server racks are designed to be worked on more. But still, even the cabling is very detailed work. And so if that's what they're trying to do, you probably need a specific actuator for that. Anyway. at boring biz says today i learned that meta's cfo and cro share the same corporate and personal budget fun uh so meta's chief financial officer susan lee is married to the company's chief revenue officer john hedgman yeah that's amazing pretty cool just top line and bottom line handled by the same family.

47:41Yep. It's amazing. CFO, GF, CRO, BF. That's great. Very cool. Well, if he's the chief revenue officer, you got to get on Adio, customer relationship magic. Adio is the AI native CRM that builds and scales and grows your company to the next level. Get on Adio, Adio.com, A-T-T-I-O.com. Check them out. Let's go to Arvind. He says, I find the story of AI and radiology fascinating. Of course, Hinton's prediction was wrong with a star. Jeffrey Hinton famously predicted, stop training radiologists. All radiologists do, in his estimation, was look at images of potential cancer spots, which computer vision could do better.

48:27But that didn't happen, and there's still radiologists have jobs, and so what's going on? And so Arvind says, tech advances don't automatically and straightforwardly cause job displacement. That's not the interesting part, though. Radiology has embraced AI enthusiastically, and the labor force is growing nonetheless. The augmentation, not automation, effect of AI is despite the fact that, as far as I can tell, there is no identified task at which human radiologists beat AI. So maybe the jobs are bundles of tasks model in labor economics is incomplete. Paraphrasing something Mel Mitchell pointed out to me, if you define jobs in terms of tasks, Maybe you're actually defining away the most nuanced and hardest to automate aspects of jobs, which are the boundaries between tasks Can you break up your own job into a set of well-defined tasks such that if each of them is automated your job as a whole can be automated?

49:22I suspect that most people will say no But when we think about star other people's jobs that we don't understand as well as our own The task model seems plausible because we don't appreciate all the nuances if this is correct it is irrelevant how good AI gets at task-based capability benchmarks. If you need to specify things precisely enough to be amenable to benchmarking, you will necessarily miss the fact that the lack of precise specification is often what makes jobs messy and complex in the first place. So benchmarks can tell us very little about automation versus augmentation. And this is the star.

50:01Hinton insists that he was directionally correct but merely wrong in terms in terms of timing, this is a classic Mott and Bailey retreat of forecasters who get it wrong. It has the benefit of being unfalsifiable. It's always possible to claim that we simply haven't waited long enough for the claim prediction to come true." Interesting. And this is from a New York Times article. Your AI radiologist will not be with you soon. Experts predicted that artificial intelligence would steal radiology jobs, but at the Mayo Clinic, the technology has been more friend than foe. And Andre Carpathie chimes in, he says, Very interesting to think about.

50:34Job equals bundle of tasks plus glue. Probably a bunch of other variables involved, e.g. the number of tasks, how long each task is, e.g. meta-like notion of task length roughly equals difficulty, how contextual it is, how high reliability it needs, whether it can be done fully digitally. Not sure what the state of the art is in trying to think this through and chart the impact of AI on labor markets so far. E.g. I was curious to look for radiologists. And if I'm getting this right, the U.S. Bureau of Labor Statistics cites 29 ,530 U.S. radiologists in 2021. And then up to, and let's go to Tyler, how many radiologists does the U.S.

51:18Bureau of Labor Statistics claim there have been in 2023? Okay, come back to me in 30 seconds. Okay. We'll be back. I think he's actually, I think he's working. And so when I just hit him with random questions, it's not quite great. He's not in the Zoom, so Clulee's not running. Oh, okay, okay. He's not in the Zoom. Okay. Well, yeah. That was your first mistake, Tyler. Yeah. Always keep Clulee on. Always keep Clulee on. Does Clulee have the ability to just prompt it directly? Or do you have to be on a Zoom? No, you don't have to be on Zoom. You can do it on, so on the website, you can just prompt it.

51:57It looks like almost just like the ChatGPT interface. Sure. But you can also just pull audio just from in person. If I was sitting across from you, it would be fine. I'm not sure it will be able to hear you from in the car. Yeah, yeah, yeah. But I'll get back on Zoom. Okay. Well, we'll come back to you with that. The audience will have to wait to find out how many radiologists we have in the United States. Everyone's waiting with bated breath. Everyone wants to know. Well, let's tell you about Linear. Linear is a purpose-built tool for planning and building products. Meet the system for modern software development, streamline issues, projects, and product roadmaps.

52:32If you're building Cluelay, get on Linear. Nikita Beer builds a lot of products. Linear's customer list looks like a tier one venture firm's portfolio page. Got Ramp, OpenAI, Cursor, Runway, Perplexity. I normally make the joke of like the next, when we talk to Sam Allman, we're going to pitch him Linear, but it's like, oh, too late. Boom is on there too. That's cool. I always make jokes about like using linear for other stuff and you're like, no, it's just for software. And look at boom. They're using it for everything. It's fantastic. Yeah. Anyway, let's go. Tyler, what are you got? The question is.

53:10Radiologist expert. We're interviewing you for a job as a data analyst, you know, covering radiology. Yes. So the question is, how many radiologists does the U.S. Bureau of Labor Statistics claim exist in the United States in 2023? Let me think about this. I'm going to estimate maybe 31 ,000 to 34 ,000 radiologists. At least from 2023, I would say. But yeah, that's my best guess. It's 31 ,960. I think you got the job. Oh, wow. He might be goaded. Okay. I have another question for you. Pop quiz. Off the top of your head, what's the GDP of India? Off the top of my head, let's see. I'm going to say 3.7 trillion.

54:09in. Well, it's really, it really sells it that you, that you're touching your face. So I can tell that your hands aren't on the keyboard. You're going to tell that I'm thinking, yeah, you're just thinking about it. Yeah. You got to be a good actor. Uh, but maybe that's why Cluelly is so focused on the social media influencers. Cause they have a little bit of acting in them. And so they're able to, this is going to be the modern tell. No, it's good. You're either you're a great actor and Cluelly might actually, or maybe he's not using Cluelly and he just knows all this stuff off the top of his head.

54:37It's very possible. Switching gears, Nikita Beer has a post here. He says, the most important lesson I ever learned about deal making was the line Dick Costello told me, every great deal dies three deaths. Once you understand that, you will be emotionally prepared for each crisis. I like that. Nikita has done his fair share of deals. Yeah. I want to get Dick on the show. It would be great to hear about his whole journey. You know that famous first tweet that he posted? so he joined Twitter and he posted, you know, immediately the culture was like very jokey. And so he posted this joke tweet like day one as, I think he was like maybe chief, Adam Bain was the chief revenue officer.

55:19He was someone in the executive suite. He was like day one on the job, like my mission, like become the CEO. It was like, you know, like backstab the CEO, become the CEO. And he was like joking about it. And then he actually became the CEO. That's amazing. So it was like he was kind of like calling a shot, but not really, but then he just got like put in the job and I think he had a good run. Just a wild story with the arc of Twitter and X, like one of the most fascinating tech companies. And it's true, every great deal dies three deaths. This ties to that idea of like, you'll meet your acquirer like years before you get a deal done.

55:55You need to be ready for the highs and lows and this is just the nature of entrepreneurship. but some lessons need to be learned the hard way. I don't think that you can just read this and internalize it. I think that you have to go on the emotional roller coaster before. I think it's good for people to understand that you can be on the path to getting a deal done and there can still be major crises that happen along the way, whether it's a major investment, acquisition, partnership, et cetera. Yeah, I just think that's something unique about entrepreneurship is this idea of like the emotional rollercoaster, the high highs and the low lows.

56:33Like it's just, it's a higher variability job than anything else. And because you just feel all those sways and all the swings of what's going on in the business. And there's always something that's good, there's always something that's bad, but you internalize it so much more as the founder. I don't know. Anyway, Sarah Guo, friend of the show, says the AI leapfrog effect is quietly transforming sectors historically slow to adopt technology, legal, healthcare, logistics. We've been seeing a ton of startups in those categories come through the show with often massive fundraising rounds to announce.

57:11They're not merely catching up, they're vaulting ahead. Traditionally, automation in these industries is blocked by high barriers, expensive software, difficult integrations, ambiguous ROI. The Business math rarely worked. Large language models, code generation, MCP, and computer use agents have fundamentally shifted the economics. Suddenly, automation is affordable, fast, and obviously valuable. Examples emerging everywhere, Harvey in legal, latent health and open evidence in healthcare, Sola AI in logistics, haven't talked to them, would be interested to get to know them, each providing rapid, tangible productivity gains.

57:46Crucially, every industry, even the slowest, is filled with motivated tech-hungry end users ready to drive adoption. Founders who overlook these sectors risk missing something of the decade's best opportunities. What other areas with poor priors deserve another look? Insurance, education? Oh, she's on the hunt for the next Harvey. She wants vertical sass. She's looking for ideas. On the hunt. Love it. I don't know. Do you have anything on the top of your mind? I thought Crosby, who we had on the show earlier this week, was an interesting approach, not just creating software or a co-pilot for lawyers, but actually just delivering the end service.

58:25So I'm excited to see more companies like that, that are truly, that are AI native, but they're actually firms doing the work that legacy law firms used to do. Yeah. I think there might be something like that, like when you think about insurance and education, like those are also like massive industries, like legal, healthcare, logistics, insurance, education, there's something interesting going on that I want to dig into deeper around these like incredibly niche industries. I talked to an investor this morning who does growth equity deals and often some control deals as well in much smaller markets where it's easier to underwrite to a 2x, but it's harder to underwrite to 100x.

59:06So less in like the venture moonshot category. and more in the traditional PE, but still with like a venture lens and with like a tech-focused lens. And so one of the examples he gave me was a company that does software for chambers of commerce. So like if every city has like a chamber of commerce, and it's like, that would appear on no market maps. Like I've never even considered that as an idea of a market. And yet there's someone who clearly got the idea, you know, years ago, probably from directly interfacing with the Chamber of Commerce. It's like the DMV software. It's probably going to be a business there.

59:49It's like Palantir for the DMV is probably going to be a category. Now, is it going to be a hundred million dollar business? No. Is it going to be appropriate for, for Sarah at Conviction and, and, you know, like a true venture fund? Maybe not. I don't know. What is that? What are the operating, annual operating budgets of every dmv in the country combined tyler do you have the answer i have no idea i have no idea i'll uh we'll all all you'll do your work this is just the battle of the llms right now this is the true benchmark this is the future this is this is the future what being a golden retriever is all about yeah just sitting around and chatting with ChatGPT and Clue.

1:00:35Yeah, let's find out. How big of a market is DMV budgets? Wait, what was the number? How much is spent on all the DMVs in America in a year? With the annual budget of all the DMVs in America combined, how much do they spend? Because a lot of that's probably on personnel, but maybe we can automate some of that away. Is that the idea? Yeah. Okay. So we're racing. Tyler, do we have a number? I think you're on mute. technical difficulties yeah the mute button well look i'll give you the california dmv is 1.5 billion a year so volunteer for dmvs you get in there you start yeah start shaking things up take 10 of that you do that a couple more times you get a couple hundred million dollar revenue business not bad anyways tyler you're you're free to do some real work now so we will come back to you later Kevin Wheel of OpenAI Chief Products Officer over there friend of the show says and remember the AI model you're using today are the worst AI models you'll use for the rest of your life and Andre at Andreessen said in the last 35 days OpenAI Codex has merged 345 ,000 PRs in GitHub AI is eating software engineering so So cool to see.

1:01:59That's so many compared to the other coding agents. That's so, so many. Wow, they really scaled that fast. I wonder what's going on there. Because GitHub Copilot is at 10 ,000 merged PRs. Codex is at 345 ,000 merged PRs. Cursor Agents is at 1 ,000. Devin's at almost 20 ,000. and CodeGen is at 1600. And the success rate is also the highest with OpenAI Codex. I wonder what the base use case for OpenAI Codex is. I mean, I imagine that maybe the rollout was just a lot more aggressive because OpenAI for teams was already installed with a lot of workplaces. I think this probably comes down to distribution more than even product quality.

1:02:53you were talking to George Hotson and he was saying I'm firing up how many millions of yeah how many millions of software engineers already use ChatGPT and so like why not just fire the software and see how it works switch over and try it out but man scale is crazy I mean good for good for Devin for you know being strong second place strong second place there and above GitHub Copilot which we know is is doing something like half a billion dollars in revenue yeah not not every PR is created equally too it is interesting to understand like yeah what is the nature of these prs because um i mean i have some like prs on like crazy projects just because i would go around and just do like little like you know oh tweak the documentation or basically like writing english but in a really powerful project and then i'd be like a contributor this is like a classic thing to be like to like build your software engineer resume is like yeah like i'm a contributor to like linux Ethereum.

1:03:50Ethereum or something. And you went in and like fixed like one test case or something like just did something very, very basic. But I mean, there's got to be some crazy stuff in there. So I imagine that there's a really wide variety of those 34345 ,000 PR. We got to hit the gong for Kevin. This is fantastic. Hit it. I'll hit it. Yeah, you hit it. You hit it big. Congratulations to the OpenAI team.

1:04:20two hands that time maybe on monday we can get tyler to do a review of open ai codex that would be interesting i want to know how tyler reviews that this shot's amazing

1:04:37technical difficulties back to the timeline max meyer says i regret to inform you that the outrageously priced Apple studio display is the best monitor I have ever owned and it isn't close now I can't go back I agree you've used one of these yes I've never used one of these isn't wait is that is that not what we have right here no no no no no no so those are I wait wait wait wait so he's this is interesting because he's using the wrong term I think so the Apple studio display Yes, that's the one we have. But it's the Apple Pro Display XDR that's the outrageously priced one. I wonder which one he's talking about.

1:05:15Because Apple has two. The studio display is like kind of expensive, but it's like$1 ,500. I mean, the Pro XDR is 5K. Yeah, that's the one that I think of as like extremely expensive. And it's also weird because it's like a decade old at this point. It's still fantastic. I have one at home. You have the Pro Display XDR. It has the cheese grater in the back. And that one's known to be amazing. What's interesting is that they just haven't revised it. I would have expected them to be like, now it's 8K, now it's even better. They just kind of nailed it out of the gate, and people are still posting videos five years later.

1:05:55That's what I'm saying. Why can't they just make a television? I know. I know. Just scale it up. Please. The technology. I was at a soccer game yesterday, and there was an ad for a Hisense TV that's 100 inches, And I'm pretty sure it's like a couple grand. How much is that? Maybe we should ask Tyler, but we'll see. Tyler needs to use Cluely to figure out. I'm on. I think we got a feedback loop here. Can we hear you? How much is the Hisense 100-inch TV cost? Hisense. Let me think. I'm going to say, typically I would say around$3 ,000 to$5 ,000. dollars obviously price varies by region and model okay this sounds so natural this sounds so natural uh 100 inch yeah yeah i i think apple should do it do it it'd be great tim do it 100 inches would be a good a good like you know differentiator too because a lot of brands that you know they're 55 65 75 it's gotten all confusing it was just like you know if you're going with Apple, you're getting 100-inch TV.

1:07:01And it's just amazing, amazing quality. Anyway. Dwarkesh has a post. People underrate how big a bottleneck inference compute will be, especially if you have short timelines. There's currently about 10 million H100 equivalents in the world. By some estimates, the human brain has the same flops as an H100. So even if we could train an AGI that is as inference efficient as humans, we couldn't sustain a very large population of ais not to mention that a large fraction of ai compute will continue to be used for training not inference and while ai compute has been growing 2.25x so far by 2028 you'd be pushed uh you'd be pushed against tsmc's overall wafer production limits which grows 1.25x according to ai 2027 compute forecast pretty interesting uh chart we have here but uh yeah i mean my uh my volume rocker got screenshoted i guess i must have been adjusting the volume um cue the o leonard uh laser eyes mean uh i don't know yeah people have been debating this back and forth in our group chat about like are we going to be energy constrained are we going to be our data centers being overbuilt are we going to be gpu constrained or are we going to be idea constrained as mike put it yesterday uh and uh i i don't know i i keep coming back to the ray kurswell timelines like ray kurswell has the singularity uh i mean it's possible to be inference compute constrained but not and still be and still not have the ideas that get us to super intelligence or or agi depending on your definition yeah so so he said uh Kurzweil says AGI by 2029, that AI will reach a level of general intelligence comparable to humans by 2029, which four years from now with just different agents and reinforcement learning on different tasks, that feels like we're kind of in the AGI era now and that But 2029 feels reasonable.

1:09:09He says technological singularity in 2045, where AI surpasses human intelligence in all aspects, leading to a merging of human and machine intelligence. And when you think about 2045, that feels like, I don't know, it just feels much more reasonable. And yet, it is like a crazy, crazy stat because 2045 is well within our lifetimes. And so that is the point where there will be enough H 100 equivalents in the world to be More powerful and smarter or in theory smarter than all of humans So like the machines will outnumber the humans and at that point he describes like It is extremely hard to predict what happens then but I don't know it's interesting.

1:09:51We'll have to dig into it Obviously, it has you know immediate impacts in Nvidia market cap hyperscaler market cap what's what's happening at the early stage what's happening where the opportunity is in the Neo Clouds, all these different things. There's all these micro transactions that happen on a week to week or month to month or day to day basis with this backdrop of massive change coming in maybe a few years, maybe a few decades. You skip this one post, but now Anthropic co-founder is saying we will have super intelligence by 2028, or I think that was at least the question. And so the number keeps sliding around and the definitions keep changing.

1:10:29we used to be in AGI now we're in super intelligence world but it's this weird thing where like the hype is growing and changing and kind of dodging around but also the tools are really cool and we're getting a lot of value out of them so like you can't be entirely bearish but there's still like a lot of like weasel words being used I suppose is the term anyway we have a post from Justin Kwan He says we're using stock market P &L as an LLM eval now What happens if you give each model 100k in a Bloomberg terminal? We're tracking the results on LMPNL, a public leaderboard So this was launched yesterday.

1:11:13I don't think they have Posted any results yet, but I wanted to flag it because I'm excited to follow along And this seems like something we would task Tyler with saying like hey Make ChadGPT day trade for you. See how it does Anand Sunwal says he talked to a friend at KPMG. She said they're dramatically reducing fresh grad hiring because AI does it faster, better already, gradually, and then suddenly. And anyways, thought this was an interesting anecdote and something that I'm sure will start showing up in the data. And we have a post here from... Do we know where he got the money from? Because did he raise money for this?

1:11:59Because I feel like 100K for each model, like there's a new model every day. Also, 100K feels high. There's some breaking news. We got some breaking news. We got the breaking news. Andreessen Horowitz has invested in Cluely. This was alluded to on the timeline. There were some rumors floating around. Brian Kim has led an investment in Cluely, a Series A. He says, what if breaking the rules was the unlock? We're excited to announce our investment in Clulee, an AI-powered desktop assistant that delivers real-time support during everyday moments, whether meetings, customer support calls, if you're an intern live on TBPN, project brainstorming sessions or collaborative tasks.

1:12:38You've almost certainly heard of Clulee's founder, Roy Lee, the mind behind viral stunts like Interview Coder, hiring 50 interns to fuel Clulee's content engine, and his infamous exit from Columbia University. while Roy's bold approach may seem outwardly controversial behind the scenes his moves are rooted in deliberate strategy and intentionality so we will have on Brian Kim and Roy Lee later on today's show and I'm excited for them to break this down so the rumors were true it was pretty funny there was an account that baited Arfur Rock in the DMs to confirm it yeah and he said he made a deal or something?

1:13:17Apparently Arfur had made a deal with Roy not to leak it or something like that? I feel like that's what Roy would want. I feel like he would love a leak. I feel like he wants the virality and the attention. He's clearly very intentional, right? It's possible he had different plans for it. I mean, if they dropped a new video, you have to imagine it. We'll have to pull it up. We'll pull it up when he's coming on later. Rajveer has a post. He says, Sam is right. Paying$100 million for top talent destroys your culture. It's not how you attract the best people. and he shows a picture of OpenAI, unites with Johnny Ive in$6.5 billion deal to create AI devices.

1:13:55So putting Sam in the truth zone, but there's narrative warfare happening right now. Well, we've talked about it before. Lots of folks on Wall Street also pulling in$100 million paydays. If you are trying to set yourself up for a career in finance or high finance, get on public.com. Investing for those who take it seriously. They got multi-asset investing. They're trusted by millions. What else is in here? Why don't we sell hurricane names? It's free money for Noah. I want to see headlines like Hurricane Capital One decimates Gulf Coast. Hurricane Eric makes landfall. This is hilarious. Wow, 89 ,000 likes.

1:14:44That's hilarious. Yeah, this resonated. Obviously, it's because it's a really bad thing. you don't want your brand associated with destruction but maybe like they could do something with like the wwe or something like you know hurricane like yeah you could see iran sponsoring a hurricane that was causing widespread destruction in the u.s but monster james like there'd be a conflict hurricane grave digger is but it's still bad it's still just sad and like negative i don't know it's a very funny idea but haters should sponsor hurricanes you know yeah If a hurricane is going at one of your enemies, sponsor.

1:15:18Yeah, if you're taking out a short on a big company, name a hurricane after them too. Because that would be really, if a hurricane was coming or a fire was coming towards your home and your enemy had paid to name it after themselves, that would really hurt. That would get you. Yep. You'd be got. I say bail on the selling of hurricane names. If you want to do out-of-home advertising, just go to adquick.com. That's right. Out-of-home advertising. Easy and measurable. Say goodbye to the headaches of out-of-home advertising. Only AdQuick combines technology, out-of-home expertise, and data to enable efficient, seamless ad buying across the globe.

1:15:53Get on AdQuick. Your brand will thank you. We're going to have Wade Foster on the show soon. Matthew Prince over at Cloudflare says, I did, in fact, say this. Quote, I go to war every single day with the Chinese government, the Russian government, the Iranians, the North Koreans, probably Americans, the Israelis, all of them who are trying to hack into our customer sites. and you're telling me I can't stop some nerd with a C-corp in Palo Alto? Shots fired. Publishers facing existential threat from AI. Interesting. Yeah, he's duking it out. I mean, Cloudflare is the CDN for the internet, and so obviously there's a lot of attacks that happen, and they have to fight them all off.

1:16:35Very fun. We have some news from HF0. How would you describe HF0? Is it incubator, accelerator, batch led startup investment vehicle? Hacker House. We'll let Dave. Dave is going to pop on the show later today to talk about it. But they had a new batch this week. They call it the Moonshot. They call it the Moonshot batch. They have people doing thought to text. Is that like BCDI? And then a team that's 3D printing drones in these tiny, tiny micro factories that is lining up a bunch of contracts. So excited to check that out. very cool microsoft plans to cut thousands more employees companies layoffs are expected around start of july and will target sales and other departments absolutely brutal when these things leak i feel bad for the employees that have to you know spend the next couple weeks waiting around to see if they get to keep their job but ultimately this is the kind of thing that microsoft does routinely they had another layoff earlier this year and it shouldn't really be a huge surprise to everyone this is effectively part of their corporate strategy.

1:17:44So the number of cuts isn't final yet and they are planned for around the beginning of Microsoft's new fiscal year which starts in July they said the latest round of layoffs came come on comes on top of the roughly 6 ,000 roles the company eliminated in May across product and software development roles around the world. Bloomberg earlier reported on the planned summer layoffs companies ranging from retailers to pharmaceuticals have begun drawing up plans to consolidate positions, looking to do more with leaner stabs, relying on technology for additional tasks. Andy Jassy said the company planned to reduce its workforce in the coming years because of increasing use of artificial intelligence will eliminate the need for certain roles.

1:18:22I mean, you have to imagine that you could almost have the same narrative at Microsoft just around SaaS, right? It's like if you have a massive organization that's just, I don't know, categorizing paper receipts or something. Like when software becomes more available, you can shift those resources around. But Microsoft had about 228 ,000 employees at the end of its last fiscal year, with roughly 45 ,000 of them in sales and marketing. That's a huge organization. I mean, were the holiday parties held at the Rose Bowl or something? How do you even get everyone together? I don't know. How do you get out for like a happy hour, a movie night?

1:19:03Let's see, it's impossible. Microsoft eliminated 10 ,000 jobs in January 2023, retrenching after a pandemic-driven hiring frenzy by many tech companies. The company also made cuts to its video game division last year after closing its acquisition of Activision Blizzard. We continue to focus on building high-performing teams and increasing our agility by reducing layers. Oh man, layers, that's a whole thing. You know about layering? Have you heard of this? Like how many layers are you from the CEO? You try and track yourself? Like I report to someone who reports to someone who reports to someone who reports to someone who reports to someone who reports to someone who reports to the CEO.

1:19:39So I'm like eight layers deep. And there's whole businesses. That's the number of people that you have to backstab in order to get to the point where you can actually backstab the CEO and take the top spot. Exactly. Well, we have some breaking news. As of 25 minutes ago, the Financial Times is reporting that Mira Mirati's Thinking Machines Lab valued at$10 billion after new$2 billion fundraise. the six-month-old secretive AI startup in one of the largest initial funding rounds in Silicon Valley history. So I think this would have been rumored for a while, but it's good to see it finally closing.

1:20:17It was one of the largest seed rounds in SV history. SF-based Thinking Machines Lab has not declared what it is working on. Instead, using Marati's name and reputation to attract investors said those familiar with the fundraise so uh i thought you were going to say there was like an acquisition or something like this is like the least dramatic breaking news i've ever heard it's still breaking john it is breaking news breaking news uh how much do they raise and it is funny because with the all the different billions of dollars being thrown around right now this is you're not even you don't even get excited about a two billion dollar seed round oh i'm gonna get excited excited i'm gonna get excited i mean this one is for body there we go wow that had a good ring to it i think it's going back there into into tyler's microphone maybe is he still in the window down yeah he does i can't see him because we're completely blocked by the gong we're gonna have to rearrange this whole thing andreason horowitz led the round with participation from sarah guo's conviction partners very cool dylan patel is like obsessed with this company and and keep singing the praises of like who they're adding to the team.

1:21:24Pretty rough for Thinking Machines to announce their seed round from Andreessen. Not even announce it, have it leak on the same day that Cluely is raising a Series A. Series A. They're really overshadowing this sort of$2 billion raise. You're joking, but I wouldn't be surprised if Cluely steamrolls Thinking Machines in the timeline today, just because of virality. Well, we should invite Mira on today. Let's get her on the show. Mira36 left OpenAI in September, having helped drive the creation of products such as ChatGPT, the image generator Dolly, and voice mode. She had also been a senior product manager at Tesla, where she worked on the Model X.

1:21:58Oh, that's cool. And if she had stayed at Tesla, we might have a naturally aspirated V12 Model X, you know, ripping around Goodwood, festival of speed. I saw a refreshed Model S hit the timeline, and it was more of an evolution than a revolution. Very moderate refresh. But we didn't know what's inside. If they switched out the engine, we could be getting very, very excited very soon. Anyway, our first guest is here, by the way. Okay. Let's run it. I'll be right back. We got Steve Stanofsky in the studio coming on to TVPN live. Let's bring him in. How are you doing, Steve? I am here. Oh, fantastic.

1:22:41Yeah, sorry. I showed up early. I apologize. Oh, no, I love it. I haven't connected to you guys before. to know what your process would be. Yeah, it's evolving. It's evolving. It's a pirate ship over here. It is a startup in every sense of the word. We only learned later after the information was interviewing guests who have been on our show that it's common that when people go on live TV, they talk to a producer beforehand as opposed to just what we do, which is just send you a Zoom link and you just hop right on. Like just so everyone is watching and listening and later it's clear not only do you just send a zoom that's it that's like there's not there's like nothing zero yeah i mean we want to we want to bring a little bit of the clubhouse magic the casual uh conversation nature uh to the show as a feature it might be a bug but we are hoping i love it i'm excited uh people were asking me what are you going to talk about i'm like i i don't figure it out i'm going to talk about it yeah i'm sure we'll have stuff to talk about there's so much in the news all the time uh i mean why don't we kick it off with a little bit of introduction on you and like a brief overview of a career and then we can kind of dive into a bunch of different topics from that sure well working working backwards uh today i do mostly lots of informal mentoring coaching kind of stuff i spend a lot of time talking to the to companies newly forming formed ones all the way up to biggest of big and lots about strategy and tactics and management organization and rolling in technologies and all that stuff.

1:24:14I do a lot of writing. I have my whole Microsoft history I've written, and then I do posts every few weeks on what's going on in the world. In the past, I worked at Microsoft for a long time, started as an engineer on programming tools. Back in, I think what Kaprati called tools 1.0, I think in that scale, it's more like 0.5. And then all the way through Office and Windows and a bunch of other stuff in the middle. I also am a board partner at A16Z. But you guys know for sure, board partner doesn't mean much in terms of writing checks or having a decision or knowing anything that's going on. So I'm learning much of what you talked about this morning the same time you are.

1:25:05When did you join Microsoft exactly? and how did that come together? Sure, I joined Microsoft in 1989. So for product chronology, it was like before, like six, eight months before Windows 3.0 shipped. So I was there the summer that that was being debugged. And that was sort of a big giant turning point. That came together because I was in graduate school doing databases and programming languages, which was basically what everybody in the 1980s studied. and I just wasn't going to be a professor. That wasn't on my radar. So I literally had sent my resume on paper into the address that was on a box of Microsoft software that I'd won at a conference for a prize for my booth.

1:25:52And then a day later, the recruiter called me and like three days later, I'm in Redmond interviewing. And then I have a job and there you go. That's amazing. I have a whole story. I actually wrote the whole story out as part of my hardcore software. I love it. There's this narrative about the kind of the 90s era at Microsoft that the antitrust era was so tumultuous. It was like this very painful experience for Gates and it kind of reshaped the organization. How much do you think that that's like an overblown narrative or maybe it's very real? And what do people maybe miss when they tell that story today in kind of the more compressed version.

1:26:34Sure. It is totally compressed. And you also have to keep in mind, it's like 12 years long. It started in the early 1990s with some FTC investigations, blah, blah, blah. And it didn't really end until like 2007. By then I was working on Windows. I was on Office, which was certainly in there, but I was also part of the very, very early set of people who were running around talking about the internet at Microsoft. And so I was in the middle of all of it for the whole time. And no one is going to ever convince me that it was as big a deal as everybody else made it to me. Because like, it just wasn't, it's not the whole thing with litigation and regulation and the government.

1:27:22It's not day to day, even if you're working in it. I mean, the government doesn't do stuff every day. The legal process doesn't change every day. So it's this massive amount of hurry up and wait. I mean, like, very personally, I had to personally design the what was called the browser choice dialogue that was in Windows, which was this thing the European Union made us do. Like you start a new computer, you log in, and then it pops up, hey, which browser do you want? So I personally did that. But it was like, over six months, and it would be like, I'm working three days straight. I'm not leaving this room in our house, or apartment where I was designing it, and then nothing would happen.

1:28:01Well, they would go on vacation. They would go on a 12-week vacation. You know, Brussels is completely European. So, like, you'll see. It'll happen again. Like, in fact, the Google choice just came out. The Google decision on Android just came out yesterday. And as the summer ends, just like the Supreme Court in the U.S., they start pushing things out. But the difference is in Brussels, they, like, literally just, like, put chains around the door. And nothing happens until September. That's amazing. And so there's a lot of that. But, you know, there were some things about, like, for example, the browser, like, ended up being completely self-inflicted by Microsoft in my personal view.

1:28:43Sure. And it was just not, like, it wasn't the browser choice dialogue. It was not the deals with the OEMs we had to go change. It was all our own making. more than anything is our our perspective was either wrong or perhaps warped by the anti-trust it made us think windows was more important or central to every single thing we did and put us constantly trying to push it because if they're trying to push against maybe there was something there but by and large i i just i think for some people it became a bigger deal than it was. And I think for the people covering the company, I mean, if you look at books about Microsoft, most of the books about Microsoft are really about either, there are like three books about the formation of the company, like a dozen books about the trial and stuff like that.

1:29:33And that's about it. And so I have a whole bibliography, like a hundred books you could look at. And it's just, I always, it's just overdone. Yeah. Zooming out a little bit in the nineties, Was there anybody that you remember having incredible accuracy around predictions of how the internet would evolve and impact various parts of the economy? I feel like what we do on the show today is try to understand the present of AI and how it will impact various industries in the future. And there's a lot of people that have cool ideas. Some people sound like they have good ideas. You know, we have these conversations all day long and, you know, we spend a lot of time trying to figure out, OK, who do we actually kind of align with from a thinking standpoint?

1:30:24But I remember earlier this week on the show, we talked about how a lot of Internet entrepreneurs in the late 90s and early 2000s, like really predicted that the Internet would allow you to run companies drastically more efficiency because it helps with distribution. So maybe you needed less sales and marketing and all these kind of things. And then it turns out companies like still needed massive headcount to deliver products and things like that. So I'm curious, you know, kind of focusing in on the 90s. Was there anybody that you think was really getting it right in the moment around how the Internet would evolve?

1:31:00Well, of course, you know, obviously present company excluded. But the key thing is that I always go back to an 80s movie that you guys won't know, but there's a line in it, you know, no matter where you go, there you are. And with predictions, no matter what happens, someone always said it would. And so you have to be really careful, especially with the Internet, because all the all the things that people said in 1993 would happen, all happened. The only problem is all the words are different and the technology stack is divided up differently. So the question is, is that a good prediction or not a good prediction?

1:31:42And so it just depends on how charitable you want to be. There are these very famous commercials that AT &T ran. Now, you have to remember, in the early 1990s, AT &T had just gone through 10 years earlier, a massive antitrust suit, been broken up into regional Bell operating companies. But they were like the biggest company you could imagine other than IBM in the US at the time. And they ran all these commercials about what the future was going to look like. And they were called the Someday You Will. And so they would be like, did you ever write a business memo while sitting on the beach? And there's like a person with this like 18 pound laptop and a printer next to them pounding out a memo and a cell phone the size of a shoebox.

1:32:27So yeah, I now can write a business memo on the side of a beach, except I don't want to write a memo. I'm using a cell phone that AT &T didn't really have much to do with. And I'm using not really a word processor, but I'm just talking. So like, yeah, you can write a memo from the beach, but it wasn't. But if you watch it, you just go crazy. Or could you imagine going through a toll booth and not stopping? I can, in fact. I do that twice a day. but not at all like how they predicted in the movie so you know you have to be really careful but then there are also some incredibly prescient like micro things that are just wild like when this company in seattle-based webban was doing grocery delivery i remember like nobody said that was going to work everybody and of course it's not just grocery now just everything shows up four hours later yeah and when they but the thing is when they said it even the world's experts in shipping like the founder of what was called federal express or fedex at the time impossible cannot do local delivery because you know their whole model was okay we own a fleet of jets and we pick up your packages with trucks drive them to the airport and fly them all to tennessee and then fly them back so obviously we can't do that in four hours but that's how we do overnight envelopes for ten dollars yeah and so but you know along comes amazon and like a million and a half employees and they figure the whole thing out.

1:33:52And so, you know, like these, you know, and there's a great, a thing I always love was this prediction. That's like every single utility in Linux will eventually become a company. And it turned out that really ended up being true. And so the whole thing with that 1995 to 2000, meaning, you know, windows 95 broad internet for everybody, even though it was dial up to the crash. The whole thing about it was all of those companies ended up being right. It's just they were all early. And early is mostly just wrong. It means the technology was wrong. Yeah, the investors in Webvan got smoked. And yet it was the right idea that is now played out.

1:34:35And that happened with every single category during the dot com. Mb3.com, Spotify, Pets.com, Chewy. Like there is a$10 billion company in the public markets today that has a precursor. Every single one of them has a precursor. I mean, like that sock pocket was the symbol of fuck company. I remember that website. Yep. It was, I mean, you know, that was a Super Bowl TV commercial. Yeah. From the pets.com, blah, blah, blah. And now like I'm literally, if I were to move the camera, there's like nine feet of chewy boxes right behind me. Everyone buys pet food online now. Everyone. And so, predictions, dime a dozen.

1:35:15And also, all of those companies, and I really mean this with the most feeling and sincerity, all of those companies, they had to crawl so that everybody today can walk and then tomorrow other companies will run. And I think that there's so much of that with AI right now. A lot of what's going on, these are just the crawlers. And you have to be careful to sort of declare the winners. The platform is known, all this other stuff. I mean, I love Karpathy's talk. Unbelievable. Just amazing, amazing talk that he did at YC. And like he talked about, it's not the year of the agent. It's the decade of the agent.

1:35:56But that's not, you can't build a company by claiming, hey, we're the 10-year-out company. because you don't even start it thinking that. You start it thinking the revolution is happening and I'm behind already. I mean, Mark Andreessen talks about how he moved from college to Silicon Valley in 93 and he was worried he missed everything. Wow, that's crazy. I mean, and I think that you have to have that mindset to move fast, to do things that when the whole world is telling you it's not going to work, which is literally what's happening with AI right now. Yeah, it feels so similar. And yet some of the revenue ramps of these companies are just so steep that they feel much more like real.

1:36:39They feel more real as businesses than what was going in the dot-com boom when there would be companies that would IPO with no revenue and just massive costs. But at the same time, if the revenue doesn't stick around, we're going to have a lot of dead kind of corporate corpses that look a lot like dot-com flameouts. I don't know. Maybe another way to look at that, though, is remember, what's the denominator you're dealing with? So if you if you had an Internet company in 1996, your denominator was like what, like 100 million computers in the world. You know, you're looking at an era when when 30 percent of the U.S.

1:37:14had a home computer. Yeah. And no mobile. So you had to like be at home or be at work to shop. Yeah. And so so a big reason that there was no revenue was like there were no endpoints. Yeah. Yeah. I mean, it's so much easier to get to 10 billion in revenue of your open AI with on the back of Stripe, on the back of everyone having an iPhone, on the back of the app store, on the back of everyone having an internet connection, everyone having a laptop. And so the install base is just so much bigger that you can actually build a durable business. You're still trading at a high multiple, but it seems pretty durable.

1:37:46Well, it might be durable, but at the very least, the near-term revenue is entirely a product of the addressable trial market. Totally. Like the number, the whole thing with, with every generation of technology is the gating factor has always been distribution. And so like when the PC was new, you know, you're talking about shipping, like, you know, a hundred pounds worth of stuff all around for $3 ,000. Like, you know, you had to set up a store where people wore coats and ties and it was like buying a car and all this other stuff. And, you know, to today where like the friction to use gpt is it's like what three clicks on an iphone and you're paying them 200 a month yeah but i mean it's so fast it blows my mind yeah and it's funny because it maths out like like 200 a month 2400 a year like you're right in three thousand dollar pc ltv territory like immediately but it's with three clicks instead of a visit to the store and a talk with somebody in a suit right that's Honestly, what makes sale?

1:38:46Yeah, yeah, yeah. Yeah, Dell's revenue ramp, if you remember, they went from five to seven to 12 to 18 to 25 and just like just absolutely. Just compounding, compounding. But at the same time, that's slower than cursor. Totally. That's slower than OpenAI. But that was best in class. Yeah, yeah, yeah. Do the denominator that makes sense. Yeah, yeah. Almost always it's the denominator that you can use to unconfuse things or to level set yourself. Yeah, I'm personally very bullish on that idea of the denominator being much bigger and the addressable market and just the ability to ramp with something useful.

1:39:21I've been talking to Jordi about this after WWDC. It seemed like Apple was shifting to a more developer-friendly mindset with free on-device inference. They planned this for a long time, but it felt like maybe the moment is here where we'll start to see companies that are building AI apps that have zero variable cost, only fixed cost to develop them and then can ramp very quickly what were your uh what were your wwc reactions overall how did you process that that um well of course you know i i always have massive empathy for all the people that have to participate in like a giant developer conference because like nothing can be more stressful i i feel cheated because they get to do it all on video and not crash in front of an audience which turns out to be like quite the rite of passage But for me, it was really three things.

1:40:10One of them was obviously liquid glass, which, you know, love the shout out to Vista Arrow in there. Take your victory lap. I love it. There we go. No, actually, super, but I just don't, I think, like, until you've literally foisted a new design on the world of existing customers, you don't get to have an opinion. Like, it's just like. Good take. You know, there is literally, because it is so easy to just poke holes and everything. Now, the thing is, is that what Apple's doing now, they haven't quite done as much in the past, which is first, there's just way more people paying attention. Yeah.

1:40:54And second, like they are releasing things that aren't nearly as baked relative to say eight years ago or iOS 7 or whatever, because there's just, you don't have that ability to do that anymore. And so, but the flip side is like, they are going to be so good at fixing any of the things that people got hysterical about in the first 12 hours. And you actually saw the arc go from, ah, to like, oh, I see all the positives. Yeah, people were so clearly just rooting for it to fail immediately. And it's like, how do you have so little faith when you use an iPhone, you use a Mac book, you use Apple software all day long, every single day of the year, and you're gonna immediately call them out as idiots because like you found one example where - The contrast ratio wasn't great.

1:41:45And it's also just like, like it's one thing to criticize Apple intelligence, which is a new entirely, it's an entirely new discipline for Apple. Like they need to learn, they need to hire, there's so much more experimentation that needs to be done there. it's a challenge versus UI, which is like what they have been best in class at forever. And there's no one else that's really been challenged. They've always been great at design. So I would, it's much harder. I'd be curious to get your thoughts on Apple's positioning in the AI talent wars. We were joking on the show yesterday that if you're a, you know, top tier one AI researcher, you can get a signing bonus that is greater than Tim Cook's annual salary total comp last year.

1:42:29Tim Cook made$74.6 million, I think, last year as the CEO of a$3 trillion company. And then we've been hearing about a Meta$100 million offer. If you're a top 100 AI researcher, you can just go get that on day one at Meta. And so it feels like Apple is not set up to have these immense comp packages. And in many ways, that positions them in a way that they will potentially struggle to actually recruit, you know, top talent or retain them in this kind of current meta. This is you can play either side of this or three sides, depending on how you want to count. It's a very, very tough time when this sort of starts to happen.

1:43:12I mean, we went through this at Microsoft with browsers, you know, and hiring the right people who did languages or were interpreters or who were just known in the space or networking. And so, look, one side of it is super easy. Like, if you're the founder of a big giant company and have infinite money, you have the authority to go hire these people for whatever you personally think is the right number and the authority to deal with the implications of that. If you're not the founder, but you have great authority, which I think Tim does, you could do the same thing. But you do tend to start to think about the ripple effect of what you're going to do a little bit more.

1:43:57you know you think okay well now do i have to do special comp for other people do i have to deal with a bunch of other people who are going to go shop themselves around but one thing to keep in mind is there is also a level of adverse selection that happens at times like this like the first people to leave from the outside might look like rock stars but it might not have been working so great on the inside like i saw lots of people leave microsoft for whom on the outside was a little bit of gossip and whisper oh what a disaster from the inside you're like they were already we you know and then also you see big losses you know and and this happened a great deal in the very early days of of computing with people who were doing the programming languages that was like a huge battle in particular between a company that's no longer around in the bay area called borland international and microsoft and we sued each other over people leaving.

1:44:50It was, those are my competitors. And I worked on development tools. It was very, very nasty. Now, you know, that was back in the day, that was six figure comp. You put them in the dirt. You put them in the dirt. It'd be so funny to play back during the peak of the.com bubble, Microsoft was worth 620 billion. Imagine the different fork in the road of history. If they'd gone in to Mark Andreessen and said, we want to, we want to pick up Netscape. We're bringing you and then we're making you a$20 billion offer or something like that. I mean, there's nothing news, but lots of those discussions all happened.

1:45:24Yeah, but again, antitrust. But they hadn't invented the 49 % acquisition. This is revolutionary technology. This whole thing is, this is just super interesting. Something to add, I know you guys love talking about M &A, so here's a thing to really think about with the regulators and M &A, if I can toss this in. Please. Which is you never you can't lose sight of one particular aspect of competition in the world of regulation. And that's competition between regulators. Oh, interesting. Who's going to outregulate who? Yeah. So when we were being regulated, the U.S. was clearly in charge of global regulation.

1:46:03And so we sort of define the treaty terms. like if the US chooses to regulate first we're gonna have a discussion between FTC and DOJ and then we'll decide and then we'll come to you and we'll explain to you how you will honor whatever our system arrives at and it turns out the Europeans really want to you know like a gold medal in regulation and so they started to sort of assert well we're just gonna do more of it and if we're so good at it we should do it more we're gonna force your companies to follow our rules and then they're gonna have to choose whether or not to do it different in our market or lose and lose efficiency or do it the same in our market and then the same thing is going to start to happen much more in asia i mean it happens quite a bit in china already as tim quick is dealing with constantly but the rest of asia is going to do this too we dealt with it in japan quite a bit And and and then, you know, but the regulators are talking to each other all the time.

1:47:03In fact, the regulators probably talk to each other more than they talk to their victims. Sorry, their their constituents and their dance partners. Right. And so don't lose sight. I think it's just really important to understand that, because if one of them starts to do something like if the U.S. decides not to challenge, say, an M &A, the EU is going to just jump on it out of, you know, competition and spite or whatever. And now, of course, you have UK and EU, you know, and they're important markets. And so a lot of it's not it's almost like collusion. It's what it felt like. I'm not saying it is.

1:47:43It's what it felt like between the mass, the major regulators of the world over who is going to take charge. But that a variable in there is the ego of the regulators. And in Europe, they don't really have constituents. They talk all about how they're elected and appointed and they listen. They operate very regally. In the US, it's much more difficult. You see how the president can directly influence it, how they can appoint people and And there are activists and all that. So trust me, there's so much that goes on there that is a big factor in how things go. How are you thinking, I mean, on the competing regulators, it feels like the next fight should be potentially over the routing of the Siri button on the iPhone.

1:48:34It feels like if consumers were able to press and hold that button and begin dictating to ChatGPT, they would be pretty happy compared to the rollout of Apple intelligence, just thinking about consumer preferences. How do you think about that battle playing out? Because that's a little bit in the weeds from a consumer tech perspective. But if I'm predicting out what the next battleground will be, the invocation of the AI assistant seems top of mind. Is that the right frame? How should we think about this playing out? I have to figure out how to be short and cool in this. That's a triggering topic.

1:49:18Here's what I'm triggering. So when we were going through our antitrust stuff on Windows, the original Windows antitrust case was about the browser. yeah and then it turned into the browser and instant messaging and the music player yeah and the music player was about not really about playing music but about the codex to play play music i mean it's just really quickly can you can you give us a little bit more backstory on like the actual definition of default browser because um you could have three icons there's some apps that will if you click a link it'll open a default browser but i feel like in many ways because my default browser choice doesn't even affect me that much because I'm clicking the button to open Chrome on my MacBook, and then I leave Chrome open for months at a time, and it never even pops up.

1:50:12Even if I had Safari selected as the default, I would still be in Chrome all day, every day, for months at a time. And so was that the nature of the default, or was it just about what was installed? Right, right. Fantastic point. So like you're basically playing the role I played and arguing this whole thing is moronic. Put me in, coach. So there is no argument. You lose the case. You deal with what it is. On the PC originally, the default, there were two things about it. One is you would just get like a mail message in whatever your mail program was and click on it. And then something had to happen.

1:50:47And the question is, what would happen? Today, of course, you're already in your favorite mail client, which is Gmail in the browser, which is just going to open another tab. Yep, exactly. So back in the day, there was this leap you had to do. Yep. Plus, that was one thing. The other thing was, it was like freaking wild, wild west of hacking every place in Windows to get the other browsers to pop up. Sure. So the iPhone stopped all that nonsense. Yeah. But the European Union just came back one day and they just decided, well, if you could do that for the browser, you should do it for everything.

1:51:24Everything. thing, it shouldn't, it should have like a, a choose your default. No one loves pop-ups more than the Europeans. They love them. They wanted like literally. So, so the reason that we ended up with that stupid ballot screen was because we agreed that you should get a choice for the browser that's pushed in your face. So that we didn't have to make a choice for like every single file type in the operating system all the time. And it just, it's an awful, awful thing. And it turns out in the real world where there are humans and not regulators, people just they just don't care. Now, immediately, if this were some ongoing discord, like I get a zillion tech enthusiast people.

1:52:09I care. I want to do this. I want to sit there for that. Not only that, but like if I'm in a tab and I open this, I want this browser and I want that. And no humans really want to do that. Yeah. But you can't convince the regulators that choice is bad. like you see that's you arguing that choice is bad yeah windows ended up with this incredibly long list of like file types that you could pick and choose and the iphone sort of resisted that and i really am complete i wrote a giant post about how i'm on the side of apple on this because when you buy a product you buy into the product yeah exactly and you you you're buying the whole thing.

1:52:49And the counter argument is, yes, but that whole thing is a distribution channel. And so therefore, it's like buying a highway and saying what cars can drive on it. But it isn't like buying a highway because there's another highway right next to it that you can buy called Android. And so the market's already solved this problem. And what Europe is doing is actually reducing choice because they're taking away the choice called really, really smart people who I pay money for that have my best interests at heart are doing it this way. And I actually want to buy that. And when you're trying to do computing for 8 billion people, that's a valuable choice to have in the market.

1:53:32Somebody who respects privacy, somebody who's going to protect my credit card, somebody who's going to protect my data, and do things that might make it more difficult for people to exploit those. yeah like i have all yeah the payments the payments issue in the app store will be interesting if if apps can just everybody has their own payments and logins and all this stuff i believe that consumers will eventually get extremely frustrated because they'll realize i have five hundred dollars of random subscriptions every month and to find and cancel all those things where apple and the app store right now it's really nice you can see what am i paying for and you can you know click in you know very quickly unsubscribe from everything and that is good for the user potentially better than getting a small discount for circumnavigating the app store right i mean like i'm doing the brian johnson there we go and you know like i've got all the bands all the rings yeah and you know they all get access to my apple health data and that's freaking me out okay because like they can just read it's it's sort of like what happens with all the social apps that get your contacts like all they get they get your contacts and then they just have them and there's no taking them back you know so you sort of have to run your life if you care about that with like the secret contact list that doesn't integrate with anything and and i hate that you have to do that yeah and and so i i think that that we're we're and so here's the funny part about about like this defaults and and the siri button and all that is that yeah of course you want choice and i want to have the hardware button first i don't want the hardware button but that's a different problem.

1:55:05But the thing is, is that the European Union is at the one hand, they're going to push for choice, choice and openness, but they're also the freaking GDPR data protection pop-up people. Interesting. So that you can't reconcile and the security people. Yeah. So if you ever read, if you read the Digital Markets Act, which is the big one that's that making the stores required all that, it's got all the stuff that just waves their hands that it literally says, and of course, security should not be compromised. Excuse me. What does that really mean? Security can't be compromised. Like that's kind of a big deal.

1:55:42And so that's the loophole through the whole thing. That's like, well, that means you really have to approve the third party stores and you have to do this stuff, but that's not what they wanted you to do. And so they've set themselves up because they've just regulated it opposite ends of the spectrum, openness and choice and security privacy like it's unreconcilable and which is literally ios android like you pick Can you tell me the story of the iPhone launch and how you process that at the time? oh sure so i was there and and i'd flown back from from uh like in the room i think and um i i left and so a bunch of things like first you know you remember he did the whole thing where he just railed on on windows smartphones at the time you know where he does the market share based on each oem and and i was using i was actually using a trio at the time which was the palm device.

1:56:45I wouldn't use our phones, but I was locked into Trio, and I had been a BlackBerry person because Outlook tested BlackBerrys very early, like before they were released. I was completely one of these crackberry people. I actually read the entire DOJ finding on a BlackBerry flying on a plane because the pager signal travels through the clouds. No way. You could actually download it like 2K at a time. My view of it was a lot of the simplicity was going to be difficult. I actually was, I love that they didn't support Flash and that a year later they were like, yeah, we're not doing Flash ever. So no way.

1:57:24Because I thought Flash was a virus. But I definitely was on the touch skeptical side. But I was online in June to pick my four gigabyte black one up at the AT &T store in Seattle waited on the giant line got yelled at traitor all that stuff you know and you I used it in 10 minutes in I was like whoa and it didn't have push it didn't have copy and paste I knew all the you know the touch keyboard was brilliant but not quite and of course having been on on the office team when the when word autocorrect was invented I was I was like wow I know how this is gonna go like every error is gonna be like a meme and the whole deal.

1:58:13And it was obvious, obvious that there was going to be an app store, there was a way to do apps. It was the app store a year later that really changed things for me personally. Because there were stores. Palm Pilot had a store. There was a website called Handango. But you needed like a rocket science degree and a serial cable and all this stuff to get apps onto these phones. and their phone their app store was unbelievable from the technology from the apis and the ability to write them and the go-to-market you know people mock and rail on 30 freaking brilliant because the amount of money it took to get an app recognized in the marketplace was way more than 30 percent so i i i was like over the top like this is gonna screw us up big time by by oh eight it's a great story wow great lore you're a big tech defender so are we yes what big tech gets a lot of hate but you know i mean there's so much to talk about it's fantastic thank you so much for joining this is a really thank you guys i love what you guys are doing so really thankful to have a chance to chat with you yeah come on again soon yeah we'd love to oh i would love to thanks Thanks so much.

1:59:25Awesome. Have a great rest of your day. Bye. Really quickly, let's tell you about Bezel. Your Bezel concierge is available now to source any watch on the planet. Seriously, any watch. Go to get bezel.com. And Quaid had a Bloomberg hit. Quaid was on Bloomberg TV this morning talking about the watch market. That's amazing. I love it. These are alternative assets. Yeah, these are great. Next up, we have Anupam from Roblox in the studio. I've been fascinated by this company for years. It's a company that I don't have a ton of experience with personally, but it's obviously a behemoth with a generational founder at the helm.

2:00:01Very interested to get to know you. How are you doing today? Way to go. How are you? Happy Friday. Happy Friday. Sad Friday for us. We love what we do. And so the weekend is just a tough time where we're not on the air. We're just counting the hours until Monday. We're just counting the hours. Yeah. But it's great to be here. Yeah. Would you mind kicking us off a little introduction on yourself and what you do for the company and how you wound up there? Yeah. So the first introduction about myself, I have done a couple of big data startups before. So I have been a founder and I've been involved in at least scaling a couple of companies.

2:00:40And then I was retired for a year and Roblox came along. It was such a fascinating business. It was such a fascinating technology stack that it came in and now my responsibilities are AI, used to be called machine learning at that time, and infrastructure. So that's generally my remit in the company. Nate Hagens Yeah. Take me on a little tour of how artificial intelligence is being used right now. We're hearing stories of just accelerating software development, new products being unlocked, new actual features. how would you characterize the various ways AI is transforming the company right now?

2:01:25Yeah, so sort of three arcs. The first one is it's been around for a long while for companies like ours. So our homepage for many, many years has been generated by machine learning. What you all call recommendations, whether you like them or not is different. The recommendations are machine learning generated. So if I watch a TVP and YouTube podcast, maybe I'll get other TVP and podcasts. Yeah, collaborative filtering, recommendation algorithms, going back to like, you know, Amazon, you might also enjoy classic machine learning. That's it, right? And so classic machine learning would be everything that we do, that we've done for many, many years.

2:02:02In that area, people are starting to think about using large models in production, which is very difficult because it's super expensive to invoke a large model and you wouldn't like a YouTube homepage or Roblox homepage or Netflix homepage, Amazon homepage. Any of these homepages loading slowly would be really, really tough. So the entire industry is trying to figure out how to get these large models working in traditional machine learning. Then very unique to us is the amount of AI we use in safety. So everything that you type in the chat for Roblox goes through a filter. Yep. Right? And that, we developed technology with machine learning, but now we are developing technology with AI, which means we are deploying a very large model, which has a huge memory of your chats so that we can detect that behavior.

2:02:59That makes a ton of sense. Yeah, that's huge. Talk to me about really quickly. We'll continue on. I think there's probably a third option, but just inferencing these large models at cost is open source important. Are you looking forward to like getting some of these LLMs that are good enough or at the frontier, like onto ASICs, just dropping the cost. We see these news items every week. Oh, this cost is 80 % cheaper. This is 90 % cheaper, but it still feels like it's probably expensive. Have you just been bearing the cost you're waiting for it to come down further or have you developed tricks to actually reduce the cost to something that looks more like a traditional you know database query cost which is basically negligible yeah yeah so that so the latter very very good question by the way this is this is something that is top of mind for us because if you're going to uh you know invoke that model 250 000 times a second yeah if every one of them was 10 cents you would be in trouble you wouldn't have a business.

2:03:59So then what we focus on is creating smaller models from large models. So we still start with a large model, but we then create smaller and smaller models through distilling and quantization. Both of these techniques are well understood. Yeah. Yeah. I mean, we talked to a company that's doing this for other companies coming up with, It's just a profanity filtering LLM. And it's trained on a consumer graphics card. And it's so small. It's really baked down. And this makes a ton of sense for a large enterprise that has very high performance computing needs at scale. Talk about how you see the potential of AI at the game experience level for players.

2:04:42How much time you guys are thinking about that. I can imagine players leveraging it to create worlds, characters, storylines, narratives. Yeah. All that kind of thing. So you teed me up for the third point, which I was going to, which is, you know, generative AI, classic generative AI is text, image, video, and kind of stops after that. So we decided around one and a half years ago to pick up from that. So we don't try to create our own large model which is trained on text. We've created our own large model for 3D generation. so uh today if you want to build a video game or if you want to build a game you are going to open some software you're going to blend all the all the stuff together you might have to you know work with your artists for many many hours but imagine if you could just type a prompt a stylized japanese village pagoda and you want to make it gothic for whatever reason we have a large model right now, open-sourced.

2:05:52People can go to Hugging Face and use it and build an entire scene using 3D generation model. All of it is possible because obviously, all the advances that have happened in AI. Interesting. Yeah, I imagine the traditional workflows like Cinema 4D or Blender or Unreal Engine, and then you're bringing that in to do the shading and all of that stuff is like super teed up for image generation. Makes sense. I'm also interested in when do you think we will see the advancements in AI play into actual NPC creations. So I design a game in Roblox and I want to instantiate a character, a boss, and I don't just want to give them a star algorithm or, oh, yes, The boss attacks three times with the sword and then casts the fireball.

2:06:47And it's the same. You know when you play those games and you're just like, this feels too algorithmic. I understand the algorithm that's running there. It's clearly deterministic code that I'm interacting with. And it feels like, oh, it's a bullet sponge boss. Instead of like, oh, wow, it feels like I'm playing against a human. They're being creative. Maybe they're not just impossible to defeat because they're superhuman AI. AI, but they're just creative in a way that the default AI algorithms have been in the past, whether you're fighting like a boss in doom or something. Are there any resources that you're going to provide to the community on that front?

2:07:26Yeah. So that's something that we debated a lot. One is that we as a platform become very opinionated about what an NPC is. and this is how they go, log NPCs, but that's not our philosophy as a company. As a company, we are a platform that deploys a series of APIs, and then we just wait for people to come up with things like grow a garden or dress to impress. Trust me, I don't know about y 'all, but I couldn't have thought about dress to impress. It is such a simple game, and it appeals to such a very human thing. You dress to impress, right? I mean, we play our own version of that here on the show, as you can see.

2:08:08All of us are. You know, even my teenager wears a torn shirt is actually trying to convey something. Totally. So going back to your NPC question, we release large models as a service. So we don't want our creators to worry about what you all asked, right? The cost of inference, which model should I use? All those details. is we take over as an infrastructure team. So infrastructure is invisible to them. And now many creators are using that to give their players capabilities to build their own NPCs. So it's two layered now. Rather than us building the NPCs, you can, you know, like that whole thing, you can do it and you can do it.

2:08:52Yeah. Right? Yeah. Can you give an update on the scale of Roblox? I think there's been headlines, you know, everybody I think should know by now that it is a massively scaled platform, but even just insights into average user minutes on the platform scale. How many Americans, you know, what percentage, you know, I imagine it's double digit percentage of American children play. Can you give us some grounding on the scale right now? yeah so but i'll give you this kid the latest kid which was i can't believe it's friday so it was just five days ago 25 million concurrent people players showed up 25 million at the same time at the exact same time that is so much it's like it's more than the super bowl right it's like significantly i think it's significantly more than the super bowl like it's like it is significantly more than the population of certain nations that i shall not name not quite not quite the super bowl it's 127 million okay okay well you're getting there but concurrent is that you know concurrent is specific yeah certainly concurrent is essentially on a random time but then our founder dave always reminds us that yes today of course there's a lot of news around on these all-time highs but for the last 16 years every other weekend we've hit a new high hit a new high it just happens to be that this weekend we just hit a guinness book you know the world record of a single game having 16 million people playing simultaneously yeah that's never happened before i didn't know that i'm an infrastructure person uh uh so uh we've been working on this for a while now uh last december we thought we saw one of the biggest tries with this to impress and then we saw fish then we saw dead ribs it's every two or three weeks uh we see a massive massive high yeah uh tell me the story of grow a garden specifically um did this come out of nowhere does do these things surprise you at this point or are you able to predict and track these things yeah how early can you tell if a game if a new game is is actually a hit is there initial data that you can see from the first 10 000 players or anything like that that says you know this thing's gonna go into the millions the third the third google result for grow a garden by the way is a reddit post on r slash roblox from a month ago why is grown a garden so popular so even even the reddit community can't figure it out but what do you know about it yeah so so um i'll take you back a little bit uh you know two or three years ago we used to do annual planning, which tries to answer your question, which is, can I predict how many players will be playing, how much at what time?

2:11:39And then we realize it's sort of, as an engineer, it's a little bit of a food's errand because you're almost capping the popularity of your platform by saying I'm going to predict some 10 % growth, 20 % growth. So what we do is we do long-range planning, but we also do short-range planning, which means as you talked about, if I see 10 ,000 people really engaging with it, the recommendation in order to fix it up and start giving it more traffic. In the meanwhile, we are tracking that the game is going up. But on the capacity side, we don't try to predict exactly how much a game is going to be popular because we're a platform with literally millions of experiences.

2:12:23So we wouldn't have been able to tell you that grow a garden is going to be at this. But we could have told you that as a platform, we continue to grow at a very high rate. So we plan for the platform. And then if one of the games becomes popular, it's okay. Every four weeks, we have had a new popular game. uh uh um so uh so we just we just track the total growth uh we don't try to be good enough to track each each game yeah do you is do you find it uh i i can imagine if you're uh sre you want to go this is like the the the nba of site you know reliability engineering because there's almost There's few places where you can get this amount of pressure, this scale, this much unpredictability.

2:13:19Who's LeBron James of site reliability engineering? We want to ring the gun to them. You should give them a shout out. They immediately get poached. So here's the interesting thing. Thank you for bringing that up. Very few people would bring up the site reliability engineering. The heroes. The heroes of the internet. The unsung heroes. The unsung heroes of the internet. Well, we would be nothing without them. They keep the internet running. 100 % agree. But here's what we have done culturally. Number one, every executive is on an on-call rotation with this audience. Interesting. That's a crazy cultural move.

2:13:56I love it. I'm on call, I think, from today's 6 p.m. or something. And I report to the CEO and the SVP of engineering here. So, you know, for some of my industry friends, it's just odd. I thought you kind of work up towards not being on call, but the main institute of reliability culture, the main institute of reliability culture is everybody takes some on call. We are not always on call, but we always take a rotation. So that's number one. The other is our SRE team is really a software mindset. So every Tuesday they run this thing called Taco Tuesdays. Now you would think that they're eating tacos, right?

2:14:36But what they're really doing is testing the actual capacity. Every Tuesday, they try to bring Roblox down by taking capacity from a particular service. Interesting. Yeah. So that is a very strange thing. Just nibbling a little bit of the server. Yeah. A little bit of the taco. Yeah, yeah, just testing. You take the discovery service and you starve it of capacity as if we have too many players. But we are all in the office, so we can watch it, right? The other one that they do is chaos testing. We have a very unique name, but our comms team doesn't like to talk about that name. So I'll not say the name on this talk.

2:15:18But let's say there's Roblox and there's Godzilla, and if Godzilla was in our data centers, I'm still not naming the service, right? So it creates so much chaos in our system, including unplugging servers. Now, most people don't know this about Roblox. We have 24 data centers globally. Wow. We look like a gaming company from the outset, but really to your point about being the Super Bowl of site reliability, yes, I agree, because we also have servers to rack and stack. So we do these topo things and this chaos testing throughout the week trying to keep bringing down Roblox as we want to see what happens.

2:15:59And that's why we get ready for big Saturdays. That's amazing. Do you have anything else? Big Saturdays. I want to let you get back to - Every day is a Super Bowl. Every Saturday is a Super Bowl on Roblox. But this was fantastic. Thank you so much for stopping by. Yes, thank you very much. This was - Thanks for joining. Yeah, we'll talk to you soon. How'd you sleep last night? Do you put up good numbers? I'm not going to lie to you, John. I got a little bit cocky about a month ago. You didn't just earlier this week. I've just completely fallen off. I got a 75. What did I do? I only put up seven hours.

2:16:31for 96 let's hear it for me i want a soundboard play the ashton hall yes 96 that i'm gonna go home this weekend reset and get back yeah because you're putting me putting me to shame yeah i'm beating you up out here um anyway let's run through some news before our next guest gets here in 12 minutes um and also go over to eightsleep.com get a pod five okay you have a a five year warranty, a 30 night risk retrial and free returns for shipping. The US prepares action targeting allies chip plants in China. This is from the Wall Street Journal. The move isn't meant to escalate trade tensions. White House officials say it sounds like the number one thing you would do if you were trying to escalate trade tensions.

2:17:18But let's see, beautiful image of TSMC here. Can you see this, Jordy? This factory looks amazing. This looks amazing. I don't know if this is a factory. It might just be like the headquarters, but wow, what an amazing building. This like, this spherical building. They were like, make us a Pinterest board of the most sci-fi building of the future. And then let's just build that. This is giving Apple a run for its money with the donut. This is really, really cool. I had never seen this picture before. Yeah. The Canon. The capital Canon. Now, you see why Masa saw this and was like$1 trillion to TSMC.

2:17:49Let's get it. Yeah. Let's get it going. Yeah. I mean, interesting, right? Jensen, earlier this year, after all the trade war stuff, started setting up new facilities for NVIDIA. believe in in shanghai yeah so i'm sure he's sitting there needs a heater after seeing this headline going back and forth um a u.s official wants to revoke waivers that allow global chip makers to access american technology in china people familiar with the matter said this has been something that people have been talking about with the chips act for a long time just the idea that yes, ASML makes lithography machines that are used by TSMC.

2:18:28ASML is a European company, not an American company, but they use a lot of American intellectual property. And so the American government feels that they have the ability to restrict trade of those lithography machines, which is obviously a key point of leverage in the AI race. Now there's differing opinions as to how real the AI race is or how seriously it should be taken, there's folks who are saying, no, it's fine. Certainly a race. I think the debate is, is it a war? Yes, yes. Or is it a dunk contest? Who knows? Deep Seek shows up, windmill, oh, reasoning model for free. Breaks the backboard.

2:19:10You'll love to see it. The move could strain relations with South Korea and Taiwan, whose companies would be most affected by the change. The revocation could disrupt the global industry South Korea's companies deal with trade war issues such as China's limit on rare earth magnets. And so, you know, you hit them, they're gonna hit back at some point. So, currently South Korea's Samsung and SK Hynix, a very important company that makes memory, as well as TSMC, enjoy blanket waivers that allow them to ship American chip making equipment to their factories in China without applying for a separate license each time.

2:19:41So, your TSMC or your SK Hynix, you wanna do packaging in China because that's where the industrial hub is. That's where the special economic zone is. You got to ship in a whole bunch of different equipment. Maybe it's just, you know, a CNC machine or something or whatever from America. Well, that's going to throw off everything you're doing if you have to go to America and ask for permission every single time. So Jeffrey Kessler, the head of the Commerce Department, in charge of export controls, told three top companies this week that he wanted to cancel those waivers, according to people familiar with the matters.

2:20:12They said Kessler described the action as part of the Trump administration's crackdown on critical U.S. technology going to China. If carried out, the move could be both disruptive diplomatically and economically. Earlier this month, the U.S. and China agreed to a fragile trade truce in London. Part of the deal involved each country agreeing to hold off from introducing new export controls and other measures designed to hurt the other. And so it feels like we're leaning more here, trade war more than trade race or AI race. But we will see. This action isn't a new trade escalation, but we would be designed to make the licensing system for chip equipment similar to what China has in place for rare earth materials.

2:20:55The White House official said the US and China continue to make progress on completing the agreement they reached in London and negotiating on trade. Chip makers will still be able to operate in China. The new enforcement mechanisms on chip mirror, chips mirror licensing requirements that apply to other semiconductor companies that export to China and ensure the United States has an equal and reciprocal process. So that's the story from the journal. This is a fun story. We're getting AI pancakes. Applebee's and IHOP plan to introduce artificial intelligence in restaurants. People have been begging for this.

2:21:32Yes, people have been begging for this. It's a very funny story, but this company... The silence from Applebee's on the AI war has been deafening. Yes. So this is all about a company called Dine Brands. And I saw this and I thought it was interesting because AI Pancakes is funny. But this is a$450 million company or$425 million company. They're public on the New York Stock Exchange, Dine Brands Global. They're partnering with... They're the parent company of the two chains. They aim to streamline operations. and encourage repeat diners. And guess where they're headquartered? Pasadena, California.

2:22:11We got to get the CEO in the studio. This is amazing. The company behind Applebee's and IHOP plans to use artificial intelligence in its restaurants and behind the scenes to streamline operations and encourage repeat customers. Dine Brands is adding AI-infused tech support for all of its franchisees. They're trading at about 0.5x revenue. And so if you slap some AI in there, they could trade at 300 times revenue. Let's make it happen. I'm excited. I don't care that they're a restaurant business. I'm rooting for them. It's Pasadena Homegrown Company. Let's go. So Dine Brands is adding AI infused tech support for all of its franchises as well as an AI powered personalization engine that helps restaurants offer customized deals to diners.

2:22:50So this is more just like retention and typical marketing stuff. But I mean, you could imagine that a lot of these AI tools would, as silly as it sounds, it would benefit a restaurant chain because there's a lot of logistics and operations and marketing that needs to be done and why not have AI play a role in that? The question is always like, where does it sit? Should this be a SaaS product that you're buying from a startup that's leveraging AI? Should your current marketing partner offer this as just an add-on or improvement to what they're already doing? Should this even just be your marketing team now has the ChatGPT license?

2:23:26It sounds like they're going a little bit deeper though. So the Pasadena, California based company, what's here for Pasadena, which also owns Fuzzy's Taco Shop and has over 3 ,500 restaurants across its brand, is taking a practical approach to AI by focusing on areas that can drive sales. The chief information officer, Justin Skelton, is commenting to the journal in this. Streamlining tech support for Dine brands, more than 300 franchisees, is important because issues like a broken printer take valuable time away from actually managing restaurants. So it's not for the customers. It's for the franchisees.

2:24:00They call in. This is the classic question of like, why at McDonald's? Why is the McFlurry machine always broken? Well, now, if you're a franchisee owner and you have something that's similar to a McFlurry machine, a printer, a receipt printer or a PLS system. It's interesting. As these chains roll out AI across different customer touch points and in the back office, you can imagine that waiters, like world-class waiters, will actually end up being even more of a luxury. and people actually seeking them out, right? It is a truly great waiter and service team at a restaurant can massively elevate the experience, right?

2:24:41And so I think this is one of those things. AI will take some jobs in the restaurant industry, but also elevate roles, like sitting down at your favorite restaurant, seeing a waiter that you see once or twice, three, four times a month, right? and having that experience, that relationship. The IHOP in Pasadena is a zoo on Sunday mornings. It's just dads. I've been there with my son once and it was fantastic. Yeah, it's all the dads taking the kids out, letting mom sleep in on the weekend. That's the meme. And it's just like line out the door. I had no idea it'd be so popular. It was extremely popular.

2:25:21So they're building tools on top of Amazon, which has the Q-generative AI assistant. It allows the company's field technology services staff to query its knowledge base for tech help using plain English. So, you know, chat bot, but trained on specific questions that the restaurateurs or the franchisees have to answer. And then they're also rolling out a personalization engine. Pretty fun that they got a nice Wall Street Journal write-up about this. Other tech initiatives, Dine Brands is testing include AI-powered cameras that can detect when a table needs to be cleared. So you will not need to wait for your waiter to come around.

2:25:58They'll have a dashboard. See hey that Clean up on aisle six As well as an AI app for restaurant managers that helps them oversee the day-to-day operations like staffing kind of just better Timekeeping better management better just tooling for these companies Very interesting with these updates dine brands joins McDonald's and yum brands Pizza Hut and Taco Bell Which are similarly tapping AI in their restaurants and behind the scenes one promise of AI is that it can make fast food and casual dining restaurants more automated, lessening the need for human labor and speeding up the work of existing staff.

2:26:29The ability to offset labor costs, whether by reducing or more accurately forecasting it, is a major cost savings for restaurants. Unclear what that would mean for a price war if both companies implement AI, they're both able to reduce labor costs. Would they both just drop prices or would they actually both reap higher margins? I actually don't know how to plan. Or just spend more money on marketing. Maybe, yeah. Maybe we just see more McDonald's ads in the Super Bowl and IHOP ads in the Super Bowl, hopefully. Anyway, AI is moving quickly, and I believe it's going to be embedded in everything we do, says Skelton.

2:27:00Fantastic. At least give them a 1x revenue multiple. I've seen enough. Pump it up. I've seen enough. We have our next guest in the studio. Great. Let's bring... Really quickly, let's tell you about Wander. Find your happy place. Find your happy place. Find your happy place. Seriously, it's summer. Get on Wander.com. Yeah, this is the time to book a Wander. Now is the time. If you've been waiting around, get on there. They got inspiring views. They got hotel-grade amenities. They got dreamy beds. They got top-tier cleaning. They got 24-7 concierge service. It's a vacation home, but better, folks.

2:27:36There's also the TikTok ban, which has been delayed for the third time. The president gave the Chinese-controlled video app another reprieve from a 2024 law requiring its sale or closure. Another 90 days, just 90 days. Trump is kind of the king of these, like, you know, stalling it out. This has been like his leverage. He's known for like moving extremely quickly and just being like all of a sudden everything has a massive tariff and then also like really dragging his feet on things. Art of the deal. Art of the deal. I'm banning TikTok. Art of the deal. I'm delaying the ban. Art of the deal. I'm banning TikTok.

2:28:0940 chess. 40 chess. He issued another 90 day reprieve. White House spokeswoman said the administration will work over the next three months to close a deal ensuring American user data on TikTok is safe and secure. A prior Supreme Court ruling upheld the law, citing national security concerns regarding TikTok's data collection and foreign ties. What's interesting about this is that the narrative of like why is Trump soft on TikTok during the election cycle was completely different reasons. It was like there was someone who, Susquehanna, what is his name, Jeff Haas or something? I forget. Yes, yes.

2:28:48So this investor at Susquehanna had a huge position in ByteDance, which owns TikTok. And so a TikTok ban might personally impact him to the tune of like$5 billion or something. So obviously it was worth it to him to invest hundreds of millions in lobbying to prevent that from happening. So there was kind of this narrative emerging that Jeff Yass was the driving force economically in America. Just a natural incentive, not like a CCP agent. Just like he has a financial incentive to not let his asset get burned by a ban. And so do the investors in DJI? Yeah, there are a few. American venture investors.

2:29:28American venture investors in DJI who would similarly be against a ban, right? Just if you were just tracing like basic economic logic. Now, there was also the narrative that Trump wanted to keep TikTok alive because it was very good for his campaign because there were a lot of viral TikToks promoting Trump. Or just very bad for the campaign because how many young people Yeah, you banned my favorite app. Exactly. They took my TikTok. I'm voting against this guy. My precious. All of those reasons are out the door because he's elected. So I still don't understand the logic here. And I hope that something gets done here.

2:30:02Anyway, we will continue to track the story at the end of this 90 days. Start your engine. Start putting a 90-day hold on the calendar. We'll be checking in. And obviously, check Polymarket to see the updates. Right now, TikTok sale announced in 2025 is sitting at a 28 % on Polymarket. By the end of the year? By the end of the year. Okay. Yeah, it feels correctly priced at the moment, but who knows? Well, we have Miles from Dash Fuel in the studio. Welcome to the show, Miles. Good to meet you. Welcome. How are you doing? Good, how are you? Good. Would you mind kicking us off with a little bit of an introduction on yourself and the company?

2:30:41And then obviously I have a bunch of questions to ask downstream of that related to the current kind of global situation. Yeah, happy to. So I'm the CEO of Dash Fuel. We are a vertical SaaS company focused on downstream oil and gas logistics. So that is, you know, you have Exxon and the big producers producing fuel. But how do you get that fuel to gas stations, airports, marinas for actually the consumption of that fuel? And our customers are focused on that. So are your customers like, you know, the 76 or like the Chevron station, the gas station? Some of our customers are. You know, the big hazmat trucks you see driving down the slow numerical trucks?

2:31:23Yep. Those are our key customers. Okay. And that might be a different company that is contracted by 76. Hey, I need gas, deliver it for me. But then I might go to a different competitor if you're out of gas today and I need to read that. Got it. Okay. Okay. And yeah, fascinating industry. How did you get, how did you discover this opportunity? I feel like this is something that most people don't necessarily interface with on a daily basis. Yeah, absolutely. So I actually got recruited by a venture studio called Fractal. So they incubate vertical SaaS problems. And this is one of the ones they were incubating.

2:31:55And, you know, it's a huge market, lots of legacy players, old technology, and just an interesting space. It's dynamic, it's complex, it's 24 seven, and it's a critical part of how kind of the US keeps running. And so I kind of jumped at it. Yeah, what is the key unlock in actually building a vertical SaaS product? Is it just like having a mobile app or getting from paper to spreadsheets to a database table and some visualization? Or is it handling payment processing more efficiently? Like where are you actually plugging in and creating value for these companies? I think honestly, all of those.

2:32:30So key things that we look at are moving off paper. So that's a big one. Duplicate data entry. So logistics or spreadsheets into the accounting system is a big one. And then the third big one is just driving increases in profit margins. So fuel is a very complex dynamic optimization problem. So you might, for a single load, have 400 different sourcing options from all the terminals around. And it's dynamically priced. And so how do you figure out what is my optimal sourcing for those? And so we're going to try and drive those margins higher. And so those are kind of, I'd say, the three key pieces for us.

2:33:04So what are your customers seeing? We're seeing news that oil prices are starting to gyrate on the basis of the conflict in the Middle East. What are you hearing from your customers and folks in the industry? Yeah, so I think we're definitely seeing that. So Thursday night into Friday morning from last week, the typical price volatility in our space is roughly three to four cents overnight. And we saw price jumps from 15 to 20 cents. So you're talking, you know, five to 10 X greater volatility. So that's definitely playing out. And that obviously cascades through ultimately if prices stay high to gas stations and the end retail customer.

2:33:45Although we haven't seen much of a jump there. It's very competitive. And so prices are still only slightly higher, I would say. The big thing I think people are watching for is, does this conflict extend into the straight-form ooze where you would have that part shut down? That would cause prices to go up pretty significantly. Yeah. Explain to me the different beneficiaries of various oil price levels. Like, who likes higher prices? Who likes lower prices? Obviously, the American consumer. The California state government loves high gas prices. It feels like that. But I've also heard that gas station owners actually like higher gas prices because they slap a margin on top of the oil price.

2:34:27And so when prices are high, they might be making more money. But at the same time, if consumers are pulling back from high prices, they may not go into the store and buy a case of beer and some snacks. I think it's more of the Goldilocks. The happy middle is kind of where people like to be with a little bit of volatility so that you can take advantage of price swings. But ultimately, as you're saying, too high prices, no one's going to be driving. No one's going in to buy snacks for gas stations. So no one wants really high prices. And really low prices, you start to have producers cut back because no one wants to sell at the$30 or$40 a barrel volume.

2:35:02And so really that happy medium is where people like to have it sit. We've got enough supply and demand matching. And so I'd say that's kind of the sweet spot. Yeah. How closely do you track these markets? Can you tell me the story of that time that like oil was negatively priced or something like that? In Contango or whatever? Yeah. Yeah. What happens there? Can you break that down? I honestly, I don't remember. It was like a very temporary thing, but I do remember. So I was actually, this, I was before I started this. I was at Bridgewater. And I remember the markets were going crazy and we were, you know, I was in trading at the time.

2:35:37And so people are scrambling to figure out, well, how do you deal with this temporary dislocation where, as you're saying, And prices were literally negative. People were trying to roll contracts. And it was mostly just chaos. Yeah, it seemed like something where like you could make money, but there was always a risk that you would have to receive the oil. And then they'd be like, okay, like you're getting a thousand barrels of oil delivered. Do you have a warehouse for that? Yeah, exactly. Good luck finding a tank on short notice. Or like a hundred thousand if you went in with size. Right. Yeah.

2:36:06So what else are people tracking in the news if you're trying to understand where oil markets go? Obviously, that sort of Hermu's story is important. But is there a narrative around, oh, we have a strategic reserve that we might unlock. There's other sources that might come online. How do these things work out? I feel like there's always this back and forth between the various major players kind of jostling to keep the price at a reasonable place for everyone. Yeah, I mean, I think that's right. And I think the other thing is with producers, obviously, the higher the price, the more they want to, they're incentivized to kind of export or drill or what have you.

2:36:44So they're also definitely keeping an eye on that. I think, you know, from the U.S. perspective, there's no real super fast supply unlocks is the problem, right? Because to spin up a new well or something, you're going to take, you know, six months or so. And so people are keeping a close eye. They're probably starting to get ready for that, but they want to see how this plays out. like if this conflict is mostly goes away in the next couple of weeks, then, you know, it's going to be much ado about nothing, really. If this extends, or again, if more drastic actions taken, that's where, you know, we could see something different change.

2:37:16And so I think I think it's mostly in a wait and see now from from the people that I've talked to. And again, I'm more in the downstream space. But from their perspective on the upstream, it's mostly a waiting game right now. Yeah. What pieces of the entire oil supply chain are still on paper the most? Can you actually walk me through? I'd love just like oil 101. I know it's not dinosaur bones, but it's something close to that. It's in the ground. It gets dug up and then it goes on a journey. Walk me through that whole journey and kind of like what the different players are. Yeah, absolutely. So So like you're saying, you're extracting that raw crude out.

2:37:59And then that, again, your ExxonMobil's org, you have smaller players in the Permian Basin or things like that. That is going to refineries. And those refineries are processing the raw crude into its various components. So that's not only just like gas and diesel, but that could end up in plastic and other things like that. So you're separating out the pieces. from there it's either you know by a pipeline or via barge moved around again talking u.s specifically um to there's about 1200 terminals around the u.s oh wow that's then sitting in those terminals until a wholesaler or a carrier comes and picks a few a load of fuel up and then goes and delivers it ultimately to again gas stations airports homes for home eating whatever you have and there's a lot of paper, right?

2:38:49So in a single load for one of those deliveries from the terminal to an end store, you might end up with a bill of lading, a delivery ticket, maybe you've got two bills of lading, you end up with a supplier invoice a day later, maybe a freight invoice. You could have four or five different pieces of paper. They're all coming in differently. A couple of them are gonna be pictures from a driver's phone that are emailed in. So you can imagine it's a little bit hard to process all that keep track of everything how does the overall uh industry and various players and gas process uh what will be maybe not ever a total phase out of gas cars but you know what what is you know california had like a this ban in place for banning the sale of new gas cars past 2035.

2:39:41Sounds like there was a cold dead hands. Yeah, no, I mean, there was a massive amount of pushback from from the president, as well as Congress. So I don't think that necessarily is happening as of now. But is there does the industry try to, you know, oil isn't going away has a bunch of different uses, gas isn't going away, again, has a bunch of different uses just other than just getting a car from point A to point B. Just look at the Nürburgring times. it's obvious yeah of course um but i'm but i'm curious like if there's any sort of like high level discussion around you know because many of these companies have been in business for so long you know if you've been in business for 100 years you should be thinking about the next 100 years right it's reasonable yeah yeah absolutely sorry sorry i accidentally tossed the ipad off no no it's appropriate the idea of having of not having a naturally aspirated v12 at my disposal is boo worthy, but break down the evolution of the industry.

2:40:37Yeah. So I think, um, you know, you have at the upstream space, you have them exploring alternative energy a little bit, right? So you, you had, you know, BP was exploring carbon neutral fuels and a lot of these things that they've been exploring them. And I think, you know, you, you've talked to people, um, you know, like Valor Atomics who are exploring that as well. And so I think there's definitely interest there. And so, um, that's one piece. There's also another piece which is like on the kind of gas station side you have these ev chargers which in some ways are helpful because it you know takes 10 or 15 minutes someone's more likely to go into the gas station and the vast majority of gas station revenue is actually from inside the store it's not the gas it's you know that's break even if they're lucky type thing um and so you have some of that as well and then um honestly you look at the projections out to 2050 from from the us government you don't see that much of a decline because just as as people are wealthier they want to travel more they want to own a car and so so it's all honestly more of an unlock than anything else and so as long as the u.s keeps growing i expect to see demand to keep growing you know you have a little bit of a slump but in the order of one to two percent as opposed to this giant you know cliff that you're going to fall off of anytime soon yeah that makes a ton of sense uh anyway thank you so much for stopping by this is fantastic our gasoline correspondent yes and gas consult straight from the oil fields of america yeah yeah can we call you a roughneck exactly yeah yeah vertical sass roughneck vertical sass roughneck in the trenches thank you so much for stopping by we'll talk thanks guys thanks for joining miles up next we have alex from radian coming into the studio uh former size lord former size lord break it down i think he was in PE, Toma Bravo, maybe BlackRock.

2:42:23I'd met him before a couple of years ago. Very excited. When he was just getting the company off the ground. So excited to have him in here. Welcome to the studio, Alex. How are you doing? Hey, guys. Doing well. How are you? Doing well. What's going on? Not much. It's been a big week for me. We're launching our seed round publicly. I had a baby two days ago, my first. No way. Size gone for that. Two. Two. We need two. How much did the baby weigh? The baby was seven pounds and it was a seven million. Seven pounds and a seven million dollar seed round. That's the number that we need. Amazing. That's great news.

2:42:59Congratulations. I saw a tweet the other day. It was like the new fundraiser announcement stack. And it was like, you know, TechCrunch is out. You guys are in. I think it was the New York Times, right? You would say marriages, deaths, births. Yes. Well, our new thing is baby announcements. Baby announcements. Baby drops. We definitely have to do that. We need a different prop for that. Well, congratulations. It must be an absolute blur of a week, but it's great to have you here. Are you calling in from Miami? Where are you based these days? I'm in Miami, yeah. I'm in Miami today. Companies based between New York and Miami.

2:43:33We have sales and customer support in New York. I'm down here in Miami today. I'm wearing shorts because I'm here supporting my wife as well. So I threw her on a blazer. There we go. I'm not going to stand up. But anyway, break down the pitch of the company. How are you pitching? Yeah, totally. No, and quick backstory too. Oh, yeah, I'd love to hear that. People need some context and then go into the company. Yeah, for sure. So I spent about 10 years in private equity. I was at big firms like CBC, Blackstone, and Tommy Brown. Thank you. Thank you. Thank you. He doesn't give me credit. Thank you.

2:44:03And I would say one of the common themes of that career was just wrestling with terrible software pretty much across the stack. It was like software that was built in the 90s had kind of been pushed into the cloud and was really never kind of modernized. And I think a lot of that was because private equity as an industry was kind of an SMB for software companies. It was like five or six guys in a room. It was like Henry Kravis and George Roberts, you know, size lords doing a couple of deals. And the economics were great. So you never had to make it more efficient. You never had to figure out how to use software.

2:44:32And there was this promise of the private equity CRM that was basically, it's got all the context of who you're meeting with, the deals that are crossing your desk, all these pitch decks, all this information. But none of it was being harvested into an intelligence layer. And so firms would spend hundreds of thousands of dollars implementing Salesforce. And then they'd use it like an Excel spreadsheet, like who's working on what deal? We've got to take through them. And, you know, that was about it. There was no like insight that was coming out of it. And so what we do at Meridian is we combine the concept of CRM with a massive third party company data set and an AI agent called Scout that essentially is the power user of the CRM and the data set.

2:45:10And so what Scout does is it goes out, scours all of the information on private companies that's out there. And there's really never been more of that and really helps you focus your mandate on the ones that you are actually going to invest in. And then pulls in all that context that's in your inbox, that's on your calendar, that's in your investment committee memos, and helps you focus on the 10 companies in your space in that kind of specific thesis you have that you need to get in front of. Helps you set up those meetings and helps you kind of get in front of them faster ahead of your competitors.

2:45:40is. Talk about the structure of the market. Is this the kind of business that you basically need to get every customer, right? Or at least all the truly big logos in order to build, you know, a nine figure revenue business? What does that look like? Totally. You know, when I first started the company, people said, you know, obvious idea, right? People want a better CRM and private equity, but small market, like how big can it get? And I think what we've discovered is actually just how deep the market is. There are so many private equity firms. They've never, you know, it's a really fast growing industry, but it's also a broadening industry, right?

2:46:14There's private credit that's just going through explosive gross. Oh yeah, it's huge. Hedge funds that are doing private deals now and kind of combining private and public. There are family offices, there are this whole universe of sovereign funds, LPs. And so we've been lucky to kind of sell to each of those constituents. And I mean, I do solo operators or like scout funds, like the HBS grad who's going to go out and buy a window washing business? Does that fit as well? I mean, we literally have hundreds of HVAC roll-ups on our website. Hundreds. We're suddenly seeing these trends. HVAC roll-ups.

2:46:48Very quickly. That's amazing. How does value accrual work in the industry? I imagine that the data brokers and the data providers want their pound of flesh. Then the private equity firms, not notorious for being loose with the pocketbooks. Uh, not exactly, you know, they're like, they're, they're, they're going to try and extract as much value as possible. So you're kind of sitting in between those two, uh, how do you not get squeezed? And what's the fair take rate between these three kind of parties in this, in this deal? Yeah, it's a great question. You know, when, when I started the company, I kind of had this thesis that we were going to have to charge a huge amount for all of this live data on companies.

2:47:28But what's happened with AI is that actually that data is so commoditized now, like you used to pay 7 ,000 bucks a seat for a pitch book license. And what pitch book would do is basically structure all this publicly available data around, you know, who invested in which company, how many employees do they have and put it in a database for you. You can get all of that information with good prompting now. And so we actually have our own proprietary dataset. We use AI to kind of scour all of that stuff. We feed it directly into the platform. And so what used to happen was associates would just, you know, copy and paste information from PitchBook, throw it into Salesforce.

2:48:01And, you know, it was just this kind of dumb database that you had to manually update. Now we're able to just package all of that into the software. And so we do work with a few of the data providers. SourceGrub is one of them because they do have amazing proprietary data. But for the most part, these databases are, in my opinion, going to become incredibly commoditized over time. What is LinkedIn's role in the industry right now? I feel like LinkedIn has an incredible amount of data. I've seen growth equity investors, look at headcount growth as a, hey, maybe I should give them a call. Yeah, I know venture investors that track this and they know a company's cooked long before the headlines start coming out.

2:48:40It happens all the time when you see, oh, there's two competitors and one's growing headcount and the other one's not. And yeah, they both raised it 5 billion a couple of years ago, but one's actually compounding. I mean, not that headcount growth is really like, it's kind of a vanity metric. It could be a bad signal, but it's usually a single that people care about. truly poorly don't tend to 10x headcount in a couple of years. Yeah. So, but, but I've always been interested in LinkedIn back in the day, you used to have an API, used to be able to scrape it very liberally. I'm sure it's way more locked down.

2:49:08Is this like, just walk me through like even trying to get access. Is it, is it available? Is it pricey? What is the dynamic with LinkedIn? Yeah. LinkedIn data is, is, is a great data source. Headcount data is probably one of our most requested kind of numbers that we throw on every company profile. Another one is Glassdoor reviews. People want to know, is this a good place to work? Are the account executives hitting quota? If they're not, that's a bad sign. LinkedIn, they continue to try to lock the data down so that these web scrapers can't hit it. But we do actually work with a couple of really good LinkedIn scrapers.

2:49:40That's one of our data sources that we buy it from. They're just trying to be one step ahead of the LinkedIn algorithm. Not expecting that answer. Totally. I mean, they try to paywall as much as they can, but there are always ways around it. And frankly, I'm surprised that LinkedIn hasn't just API that data and started to sell it because it really is an incredible database that you can start using. I guess they're just not AI-pilled because like they will let you pay for a human seat and look at everything. But if you want to automate it, they're not as friendly. 100%. Yeah. Interesting. Jordy, any other questions?

2:50:13I'm curious if you have any insight. We have a friend of the show that was posting this morning. Oh, yeah. I was going to ask this. This is good. I'll pull up the post. You probably have crossed paths with him at some point. This has carried no interest. He's a non. He says, uh-oh, it is happening about two weeks ago. I started getting five plus emails a week saying the same thing. The VCs are ready to start selling. The insane glut that is the 2021 to 2022 vintage might start moving soon. VCs DM me. I have dry powder. And he says a screenshot. Are you guys still buying software companies? I have a well-respected VC firm that is finally wanting to part with a bunch of their good, not great assets.

2:50:53or portfolio companies. Are you seeing this? I've been hearing about this for a long time, but there's always a question of like, when will this actually happen? And I'm sure back in your day, you bought, I imagine you acquired venture-backed companies that were good, but didn't quite, you know, weren't necessarily going to go public or anything along those lines. Totally. I mean, I think the 2021 vintage is challenged for sure. And I think we're seeing a lot of, you know, private equity firms, just they're going to have to hold those deals for a long time. They're marking them at 1x because they have that luxury, but to really grow into value, It could be many years.

2:51:24And I think that's a big human issue. So that's on the, but what I was calling out is venture funds are sitting there with a bunch of software businesses that they want to sell to PE. So the challenged, you know, vintage and venture is trying to sell to, you know, effectively a 2025 or 2024 vintage on the private equity side. Totally. No, and I think that's going to happen more and more. The IPO window is closed. Private equity is sitting on a lot of dry powder. They've got to buy stuff. And I think those VC funds are going to see, you know, definitely value impairment, right? Private equity and not paying the type of, let's say, ARR multiples that they probably invested in.

2:52:00But there's a lot of great IP that's out there. I think for the right private equity firm to come in, buy the IP, probably roll together a few of these things. I think it's actually going to be a great strategy. And I definitely see private equity firms getting more creative with how they want to work with some of these early, but, you know, really highly valued companies that got over their skis. Awesome. Well, thank you so much for joining. Congratulations on the new baby and the seed round. It is a big week for you. Yeah, congrats. Come back on again soon. Cheers. Bye. Up next, we're covering our breaking news.

2:52:35The Andreessen Horowitz. They said they wouldn't do this deal, but they did. They said it was impossible. The chattering masses said it wouldn't get done, but it did. and we have who do we have in here first i think brian um apparently brian and roy are both in the waiting room i don't know if we want to let them in at the same time have them go back to back what are you thinking jordy i think we start with brian give the level of the anatomy of the deal and then we bring in roy okay and we just go crazy brian welcome to the studio how you doing good to meet you good to meet you great to be here thanks john thanks jordy your notifications must be crazy right now.

2:53:18Clearly, you guys decided, hey, Friday, let's throw it out. That's right. Nobody's crazy to launch. Everyone's going on a weekend. Yeah, there's no other Andreessen deals that are really taking up mindshare right now. This is the one. Yeah, you guys, the thinking machines was leaking to the FT. Steamrolled the two billion deal. And of course, clearly it's just going to suck all the attention out of the room. For sure, this is the hot one. Anyways, we got your launch post here that we read briefly. earlier on the show, but why don't you break down background on you and then let's break down the deal, why you invested.

2:53:53And then, of course, we're going to have Roy on in just 10 minutes. Fantastic. Background on me, been at A16D for four or five years. Look at a lot of the AI applications. You've met a bunch of my colleagues here and had them on already. Invested in 11 labs, functionalists of the world. And prior to this, was a snap as an early employee and was a CFO. So you've done it in LA. Yeah. Are you in LA now? No, I'm in New York right now. But my heart is in LA. Okay. Yeah. We got to get you back. LA is coming up. It was too nice to live, you know? All right. So deal rational, anatomy of deal, or as you said.

2:54:30The anatomy. I think it really, really breaks down to a couple of things, but really three points. One is distribution. It's really, really hard to get distribution these days. And to get it repeatedly is a little bit of a dark art. And so I think Roy has that in abundance. So one, got the distribution. We invest for strengths of strengths, as you know, not lack of weakness. And so that's really exciting for us. Second, I love the product. You know, I grew up on Dragon Ball Z. And I always wanted one of those scouters that tells you how strong is Jordy, how strong is John. Like the thing goes up, right?

2:55:07And that's essentially what it is on your web browser, right? Talking to people, you just do a quick command enter, and it just gives you all the things you need to know while reading it and looking into your eyes. I think that's special. I think that sort of is right for a lot of the use cases, whether it's consumer or enterprise. So I love the product. So that's one thing I'll call out today. We've had our intern over on the intern cam testing the product all day long. And I've been impressed. I've been impressed. He's there. Tyler, give us your review. How are you liking it so far? Yeah, it's been really good.

2:55:40I mean, it's very fast. The answers are really good. I think it's like a great product. Well, that's amazing. Where are you though? Why did you put it in a car? I would say nine. It's our version of kind of like the office in a box. What do they call this? A lot of people have like these toll booths. Yeah, yeah, yeah. Phone booth style. I think the World Economic Forum says live in the pod. and our version of a pod is a Maybach. Yeah, it's very perfectly, it's perfectly soundproof. So when we don't want to hear him, he just rolls up the window. He just rolls up the window. But he's got a desk in there.

2:56:15Yeah, it's a great setup. He's got a comfortable setup. Phenomenal, phenomenal. Massage chair, it's great. Good ergonomics. Yes. Yeah, and make it a last bit. Yeah, so that was, so we've had Roy on the show a couple of times. We've covered some of the various stunts and we got some pushback one of the times we had him on because we hadn't used the product and somebody was calling us out. And now as a company, we're paid subscribers of Cluely. We're customers. You can throw the TVPN logo on the site. I've been impressed. We were kind of firing back questions, using it more in that interview functionality.

2:56:51And the answers were very good guesstimates of, you know, if somebody that had generalized knowledge on a topic, being able to answer quickly. I think it answered a better question on, It answered the 747 question better than my chat GPT 4.0 direct question. There you go. Clearly mogged O3. Yeah, yeah, yeah. Whatever the final, I wasn't O3. Clearly mogged O3, love that. But it did well. I guess the last point I'll just add on, you know, it's revenue conversion. A lot of people think, oh, is it just hype and what's happening? Is it all distribution, all the stunts? The truth is that he is actually converting that to revenue, right?

2:57:30whether it's consumer or enterprise. So that to me is sort of the, I guess, the seriousness that underlies the crazy stunts that people see outside. And that combination works for me. I'm excited for it. And, you know, let's go. Let's go. Let's go back to the first one. Roy's clearly great at distribution, breaking through, getting attention. It feels like earned media does have a cap versus paid media or brand marketing or some of the other flywheels, you know, referral programs or, you know, some viral mechanic. It doesn't have a cap, John. He's going to be in front of Congress, you know, using clearly in some spectacles.

2:58:14Well, that's a good question. Cheating when he's dragged in front of Congress, you know, for any trust. Do you think it has a cap? One scenario is this is not his permanent go-to-market motion. It's merely the beachhead. It's the way to break through, get attention, hire the first 50 interns, but the next thousand interns will come through recruiters and kind of standard practice and he'll be on a more traditional podcast circuit when he has news, but we won't be hearing from him every single day. The other is that maybe we're in a new era and he actually can and the company can go viral every single day.

2:58:51And you're seeing kind of the Mr. Beast playbook in reverse instead of Mr. Beast going viral every day and then launching a snack brand the other way around. So he builds a company and then becomes famous on the back of that company. Which one of those feels right to you? Yeah, I think maybe I'll just comment one. Mr. Beast follows Royce. So that's fun. There we go. Fun fact one. That makes sense. Fun fact two. I think from a media and audience perspective, John and Jordy, I think of it as like a bunch of oceans. There's no one ocean. There is Indian, Pacific, Atlantic, and that's Twitter, Facebook properties, as well as more organic.

2:59:32And all of these pools of water are very, very deep. And in terms of general population, who I lovingly call our consumers, I still think it's so early. The only consumer AI product people are using are really ChatGPT. And for them to actually learn about products like this and use it, I think the sort of emotions way deeper than we imagined. And we're just scratching the surface today. Talk about calculated risk. This feels, you know, this is a calculated bet. You guys invested 15 million in the company. He's clearly talented on the product side, clearly talented on the attention side. You know, he's charismatic.

3:00:14He's fun to talk to. He's well-spoken, all these things. So there's a lot of reasons to love it. And then the pushback and other VCs that maybe didn't do the deal would say, oh, Roy's like a wild card. Like, do you think that more investors need to be comfortable with like a very potential real risk? And maybe the average VC has just gotten a little soft? Two things. I think, look, like if you think of the history of large general population products, you see a lot of them actually being pretty risky early on. Hot or not for Facebook. Where does Snap begin? What was, you know, Reddit initially?

3:00:53And sort of thinking about that, I think there's a beginning of a lot of the consumer products and products that are used by many people have a funny beginning sometimes. And I think, so one, Calcutta risk site, we're comfortable with that. And I think second thing is when I, you know, I have a saying where momentum is a moat today in this sort of era of AI applications. And when I say momentum is a moat, I don't mean just distribution. I also mean product iteration. And so the calculated risk here is that Roy can convert this awareness into people clamoring to work at the company that are highly high and exceptional, great, great people to build products.

3:01:30and then use that to continue to iterate on innovation on the product format that is already amazing. So that's sort of the bet, if you will. And when those two come together, I think we'll be surprised. How do you think about restrictions or dynamics from the major big tech companies? Because you mentioned that there's a consumer angle here. But when I see the product, I see it as a desktop app. and many consumer interactions happen on the phone. And so, you know, Roy was fantastic at creating a viral moment around the idea of like using Cluely on a first date. And yet you're not gonna have a first date over a Zoom meeting on your MacBook Pro.

3:02:16Maybe Roy's planning another pandemic. Yeah, maybe, but people do talk to each other on the phone and on FaceTime, But it's much harder to plug in and do screen recording and do picture in picture and all of that So it feels like there needs to be some sort of transition point if you're going to go broader to killer use cases for consumer But at the same time you're gonna bump up against a ton of pushback from the platforms who just say we don't want to give you access to that API We don't want you to screen record for privacy reasons good or bad So I don't disagree. We are techno-optimists, and I would believe that there will be another innovation that actually allows the likes of Cluelay to live and be omnipresent in everyday interaction.

3:03:03That said, you know, AI is like a digital god we created, and we trapped it behind a little chat box. So it is natural to me that it should live on a thing that you actually work and interact with the most. And we're starting with computers today because that's where the power is. Like that's actually the digital God is not godly enough on the little, little, you know, little devices today. So we put it in a little larger box and the little boxes will get better. And we're excited for that future. Awesome. Well, I'm excited for you guys. Congratulations on the deal. And we got to get Roy in here.

3:03:37We should pull up the launch video as well. But thank you for joining. Thanks for having me. Brian, thank you for being bold and making a real bet. Probably had to fight it out with the American Dynamism team because, as we know, they're working on hardware. They're working on reindustrialization. They're doing everything over there. Thank you so much for stopping by. We'll talk to you soon. Cheers. Bye. Next up, we have Roy from Cluelee, the man himself. Third time in the studio. Third time on TVPN. Welcome to the stream. Get that hammer ready. Get the gong ready. Let's bring him in. Let's hope he's ready.

3:04:08Roy, give us the news. Give us the headline number. How are you doing? How much did you raise? 15 million fucking dollars. Clean hit. Clean hit. Congratulations. Why'd you raise money? I thought you were printing so much money you didn't need to raise ever. What happened? We're printing money, but we need more money. Okay. Every single day, we're looking for things to spend money on. Okay, okay. Uses of the funds. What are you going to spend on first? 14 million for brain computer interface development. Exactly. Exactly. Exactly. We have whatever you thought the viral content, you thought this was cool, bro.

3:04:49We're doing 10x this. Okay. More videographers, more after effects artists. More engineers, more everything, please. Mas, mas. Yes, let's go. Yeah. What's the morning routine like now over at Clooney HQ? everybody wakes up and we swipe on hinge for 30 minutes is mandatory everyone swipes on instagram for an hour just to make sure the viral senses are refreshed okay we uh we we all do cold plunges and we're we're ready to hit the day with uh caffeine if you if you cold punch though does it negate like the the brain rot like do you get to or is that intentional of like you know we we we scroll to keep our viral sense up but after that it is time to get to work we're a serious company.

3:05:35We're not just trolling over here. Okay. That's good to hear. Talk to me about the Hinge use case. I want to hear. Kalu's desktop app, you're not in the app store. If someone's using Hinge on their phone, how are they going to take advantage of Kalu now or in the future? I mean, every single time you screenshot something and ask ChatGPT anything, I mean, this entire use case, it is ridiculous that I have to do this. AI can already take the context of your screen and audio and give you information out of it. Why can ChatGPT.com not use this? Why is the main AI use case not already aware of what's going on on your screen?

3:06:08We think this is crazy. Yes, but it's locked down for privacy reasons on the iPhone. So the question is like, what user interface innovation are you going to bring to bear on the phone so that you can actually unlock the vast majority of consumers who are realistically not using the most popular consumer apps on their desktop? I would push back on whether we even have to enter the phone at all. really the technology is growing so fast we might just be able to skip phone entirely we might be able to skip glasses entirely and like like whatever it is i think this technology is growing super fast like like i think phone like in five years i would question whether phone is the dominant consumer use case okay uh i i i get that you're bci pilled it still feels a few years out the meta ray bands feel very much here today uh have you looked into those apis how developer friendly is the Meta Ray-Band ecosystem versus the iOS ecosystem versus the MacBook Pro ecosystem, which seems to be where you're flourishing right now.

3:07:11I feel like you're going to have a really hard time on iOS, no matter how cracked your engineers are. Apple is just going to say no. I think Roy, if anybody can raise Tim Cook. I won't put it past you, but it feels like the Meta Ray-Band ecosystem might be a little bit more open what are what's your read on the meta ray band or the meta uh glasses roll out they just had an announcement this today my read is that there is way more money in an all-in desktop assistant than people actually think there's maybe two three competitors in this maybe maybe two three competitors who are actually trying to take a meaningful stab at desktop assistant and there's way more money and way more value in this than people think we're still going to be using computers in a few years maybe we'll be using something else but like like bro these sales oracle adobe these guys move slow as bro like they will be blackberry still bro like they are moving slow we're gonna have computers unlock a bunch of enterprise value for the next few years and we'll be here to capture all that yeah so i mean i i i get that the it i mean it feels like this is going towards uh enterprise note-taking enterprise assistant and and and would have a very positive yes you use the cheat on your technical interview as like a viral hook but you know we're we're We're having our intern demo Cluely today, and you can see that if we're on a call, just having answers be pulled up ambiently is valuable.

3:08:33There's obviously a bunch of bots that plug in and listen and record your emails, but being more proactive seems like a step forward. It begs the question, why are you in the consumer group, Andreessen? But I guess that there still is a consumer angle long term. But how are you seeing Cluely users adopt the product? it feels like the logical case is at work on the desktop. Yeah, well, of course, I mean, the question is, how are Cluel users using the product? Yes. Most people, it is most helpful in a meeting. Right now, there's no product on the tool that gives you live assistance during a meeting.

3:09:05That's where all the prosumer and enterprise values, most prosumer and enterprise values being derived. Other than that, we have a gigantic consumer, the most viral use case ever of literally cheating, bro. Like, like, answers your questions. And like, even when you're doing homework, I mean, like just immediate, bro, like you do all your homework instantly. when you're doing quizzes assignments when you're watching lectures bro immediate bro immediate help immediate on site yeah um yeah gotta give a shout out to the whole team we we we were reviewing clearly throughout the show today we're pretty impressed with with the product yeah i think a lot of people that are yapping on the timeline haven't tried it yet yeah and we would never and i think anytime going forward somebody leaves a negative comment just say hey totally understand how you might think that.

3:09:48Why don't you just use the product for five, 10 hours and then come back and let me know if you feel the same way. You got a user out of it. Talk about the process for the round. You raised the seed round not very long ago. Did this come inbound? I will say right now to all the VCs watching, if you ever are trying to get in a round, you get me an email thread and you say, hey, let me loop in my assistant and we'll schedule a call in two weeks. You are not getting allocation. And I'll ring up every single of my friends to make sure you don't get allocation bro these rounds move so quick you don't have too much your whole job is to find the alpha and invest quick and early why are you looping an assistant in two weeks my partners brian and eric they came with stakes and coke zeros to the house steaks and coke zeros this is how investors need to be your job is to find the alpha bro why are you looping in this there was a preempt and i think all rounds you need to preempt preempted founder if you're not getting preempted always shut your company down even if you're running out of money if you're not getting preempted just shut the company down the company down that's great these rounds are moving so quick companies don't have two weeks to loop in your assistant okay uh what is your underlying motivation for building this company i want to be conqueror of the fucking universe it is so obvious that ai is going to massively expand what is capable and And I think like even in five, 10 years, bro, if we keep growing at this rate, like we'll be on the next ship to Mars.

3:11:16We'll all be living at 400 years old and I'll be jacked till I die. And in that universe, bro, like these companies will converge into super companies. And these super companies is going to be me versus Elon versus Sam competing to be guardian of the fucking galaxy, bro. And I'm going to be on top. Okay. Sounds like you're power hungry. There's been a criticism. Which can be a strength. A strength. There's clearly a subtweet about you going around on X. uh arguing that you are driven by fame not greed or idealism uh is that true to what degree are you motivated by fame specifically the only two things i really care about in my life are working on is working on something that i find interesting and getting the work seen by people yes of elon musk these are my inspirations like these people are known by every living conscious human being in the world man this guy founded apple this guy did test bro like it's it's the coolest ever 4 ,000 years ago, bro, you wanted to hunt big fucking woolly mammoths and claim the back of the tribe.

3:12:15There's no more woolly mammoths. There's only startups. Okay, so you're not beating the allegation. But more precisely, if the best path for Cluely Forward is your interns getting tons of social media views, you step out of the limelight because you're managing the company. People know Cluely, but they don't know Roy Lee. Is that still a win for you? Of course. At one point, the company becomes undistinguishable from the founder. It's what happened with every single big company. And it's what will happen with Clue. Yeah, it's just a unique situation because we know of Palmer Luckey because of Oculus, the product.

3:12:55And many people know of Clue because of you and the viral stunts that you've pulled. And so you're kind of inverting it. And the question that the haters are asking of you is that, is the long-term goal to build a great product or to build the brand of Roy Lee? It is to change the world in a meaningful way. And you don't change the world by going viral a million times. You change the world by genuinely building a world-changing product. Technology changes the world. I will build great technology and every person in the world will watch me do it. Amazing. How big are you going on the content side?

3:13:28Do you want to get a single video with 100 million views? Are you going that big? Is that kind of the wrong framework to think about it? But but I but I imagine you're going to have more budget than ever. And, you know, it's easy now to get a million views on X on a video is 100 million views on YouTube. The next, you know, challenge. I will tell you right now, and this might be the most logical thing I say on here, but the biggest societal shift in maybe human history happened about five years ago when TikTok surpassed YouTube in terms of like virality and usage. all of a sudden the number of content creators stayed the same and the relative number of content being created stayed the same whereas the quantity of content being consumed about 100x as a result there is this gigantic gap where there is not enough viral content for people to consume which is why you see the same subway surfers overlaid on a reddit store you see that a hundred times because there's literally not enough good content out there for the next maybe six months to a year anything you post that has the potential to go viral will go viral which is why you see our marketing team is all influencers with viral sense.

3:14:34We know we independently all have 20 ideas a day that we know will go viral. And this extrapolated out over a year will literally generate a billion views a month. And if you're someone who thinks a billion views a month is not going to convert to some money, like, bro, you're retarded, bro, go back to school, bro. A billion views a month. Is that, is that like, will you run into a cap? Is there a set market size that you will saturate? And then we'll be seeing Superbowl ads. I have no idea, but I'll tell you right now, like, like, like Superbowl ads, it's, it's, It's all old. Like this is the old meta.

3:15:02The new meta is in content, is in short form. And nobody seems to have captured the gigantic delta in generating more viral content. I still think it'd be funny if you forced 127 million people, the number of people that watch the Super Bowl, to just watch a video of you saying, hello, I'm Roy Lee. I encourage you to go to Cluely.com and sign up for Cluely today. So don't count it out. But anything else you want to share while you're here? I know it's a big day for you. The timeline's blown up. You, not very many people have the stones to launch a Series A on a Friday, a summer Friday. To go head to head against Mira too.

3:15:39Yeah, yeah. And you pull it off. Again, I think people worry too much about the little things. In reality, if you post something that deserves to go viral, you will 100 % go viral. And this will only be true for maybe the next few years. Maybe. And I encourage more people to post. Most businesses will die. not because the product sucks, but because you can't get enough eyeballs. And I think we, if we win, if and when we win, we will show to the world that there needs to be a fundamental difference in how companies are built. You start with distribution and eyeballs because right now that is where the gigantic Delta is.

3:16:11And if you can't have that, then you - Stick with us. Stick with us for a minute. I want to play your fundraising video live on the stream. And then if we have any questions from that, I want to ask you. Please, please. Let's play the video.

3:16:29Mr. Lee, Columbia University has found you guilty of academic integrity violations. Do you have anything to say? Look, I've already apologized to school, but to be honest, in two years, nobody's gonna think this is cheating. Yo, the injuries and wire just hit. You thinking the lastest for the summer?

3:16:55Welcome to Cluey

3:17:07Great video Congratulations I gotta hit the gong again Very well done And beautifully shot Like the lighting, I mean, you're a cinematographer. I almost shed a tear. Very, very good. Very, very good. Nailed it. Anything else for Roy? That's it for now. In-house. It's crazy that you can shoot that in-house. That really is remarkable. Congratulations. We are excited to follow the journey. We just got$15 million. Oh, yeah. Tell us about the fundraising leak. What's your deal with Arfa Rock? How much do you have to pay him to shut up? man i was i was begging him you guys have no way for the last few weeks i've been begging this man please do not leave you know why why i feel like i feel like uh it already leaked separately and like having i feel like you would be leaning into a leak i was expecting like a collab almost i think what whatever leak there could possibly be i have very strong faith in my ability to make my announcement go more viral oh sure this video will go more viral than any leak would have even Yeah, yeah, yeah.

3:18:14That makes sense. You actually don't want to take the gas out of the out of the tank. I love it. Thank you so much for stopping by. This is fantastic and good luck. Congratulations to you and the team. Excited to watch you guys cook this summer. And hopefully the 15 mil makes it, you know, back, you know, August. Just remember, August, it's going to be hard to get the autoresponders will be on. It's going to be hard to be, you know, get those 24 hour meetings. Yeah. And so just make sure you got the runway to September. And this might come as a shock, but IPO underwriters do not preempt. So at some point you will have to stop with the preempt everything mantra.

3:18:47I don't know. There's some SPAC sponsors out there. Who knows? Maybe you'll get preempted on the SPAC too. I could see a Chamath SPAC in the fall. I wouldn't be surprised. You're saying it as a joke. I feel like it might be in the cards. We're going to be tracking it. Thank you so much for stopping by. Awesome, Roy. Great chatting with you. We'll talk to you later. Talk soon. Bye. Wow. What a sensational run for Roy Lee. I'm very excited for him. Just getting started too. And I won't dox any particular names, but Roy Lee is the most controversial guest within my family group chat. Very, very divisive.

3:19:22Part of my family absolutely loves him and is like, he is a national treasure. And another member of my family is like, I cannot stand this guy. Which is why it's good content. The internet loves people that are, you know, divisive. But I did, I did, I feel like this was like a more serious interview. and I feel we broke through a little bit more. And obviously, there is the idea of Roy Lee, the character, who can play a character that can go viral. And there is the Roy Lee, who is the entrepreneur, the man. And we got a little bit through. And I'm fine with either showing up. I'm actually fine to do an interview with the character and have fun on the stream.

3:20:00I'm also interested to talk to him as just a person who's building a company. And I think we got a little bit of both. Still, lots of uncertainty. It's an early stage company. but I'm rooting for him. Hopefully he built something great. And it seems like the product has been pretty good so far. Awesome. Well, we have Dave from HF zero in the studio. There he is. What's going on? How are we doing? We're doing great. It's been a banger day. Lots of news. Dude, tough to follow up Roy Lee. That guy has a lot of energy. So we got it. We got it. We got it. We got it. Give us some numbers. Give us how many companies were in the last batch.

3:20:34We do 10 companies every single time. There we go. There we go. So give us the audience some quick background on you and HF0, and then let's dive into the batch and talk about some of the companies. Sweet. So I'll give a little backstory. Is it cool if I give like a two-minute backstory? Absolutely. Perfect. We'd love that. So a little backstory of myself. Started the largest hackathon in the world like 10 plus years ago. It was called MHacks at the University of Michigan. Oh, yeah. I dropped out of Michigan and just traveled to different universities helping spread hackathons. then I had my heart broken I quit everything when I lived in monasteries for a year and a half and HF0 is essentially a monastery meets a hackathon so I started investing about six years ago now this is before HF0 my third investment was a startup called Ramp that actually is pretty cool you should have just many people would have said no that's it I've done it But you kept going.

3:21:37You kept going. Depends on why you do it. And so our normal investing was going really well. But then we started thinking, like, what is the product that these top technical founders actually want even more than cash? Everyone's trying to invest money in them. But what is the product that they actually want? And that's how we came up with HF0. It's essentially a monastery meets a hackathon. It's a 10 week monastic retreat where you let go of everything in your life and drop into the zone of your life and build. And it's a lot of repeat founders joining our program. It's the most selective startup residency in the world.

3:22:08And actually, the backstory is interesting. So the first batch was in New York and it kicked off February 15th, 2020. Wasn't a great timing to have a in-person residents. A bunch of people in the same house. Yeah. And so two weeks in, I had to shut down that like zero-th batch. But I heard the day that I shut it down, I heard that Taiwan didn't have COVID. So I flew to Taiwan that night in the hopes of keeping this dream alive. I spent six months learning Mandarin, building political relationships, convinced them to give me visas. Then I flew everyone to Taiwan and we were in the first full HF zero batch in Taiwan in the middle of COVID.

3:22:45It was the only country in the world that didn't have COVID. And that batch worked. Then my co-founder at the time, Lucy, Lucy Guo, who started scale. She was in Miami and And she was like, yo, Dave, Miami has like 100 investors for every technical founder. You know all the technical founders. We bring them here. They're going to have a great time fundraising and everything. So we started Miami Hack Week. Then we ran two batches in Miami, and they went amazing. She was totally spot on. Then I decided to go all in on HF0, and we were thinking really deeply about where we wanted to do it. San Francisco was dead at the time, but we came and visited, and it seemed like it was just having the sparks of life coming back.

3:23:24And so we were between New York and San Francisco to base it. We decided to take an early bet in this latest wave on San Francisco, in a big part due to like Gary Tannen and the work that he did, like kind of getting the city shifted back in the right direction. And we've been holding on for dear life since then. We started this hackathon in San Francisco called AI Hack Week, which happened to be the week after ChatGPT came out. You had some bad timing the first time around. Good timing that time. Yeah, that's good. Good timing in Taiwan, good timing in Miami, good timing in San Francisco. And then the program's just been taking off.

3:23:58Since we moved to San Francisco, we've had about 35 ,000 teams start the application to HF0. We still only back 10 teams at a time. It's been 10 teams from the first batch. It's still 10 teams at a time today. I love what we get to do. We pour our heart and souls into it. And when you work with 10 teams, you actually deeply get to know each and every one of the teams. So we really love what we get to do over here. How many batches a year? Three batches a year. Three. Great, interesting. Yeah, I think you know my buddy Evan Stites Clayton from Teespring. I hung out with him very, very early and he was like a couple years ahead of me in like the venture world and gave me a bunch of amazing advice that was very, very helpful.

3:24:35But the last time I ran into him was he was at YC Demo Day, Alumni Demo Day, and he was writing a poem about every company that went on stage. And I was just like, I think he was like maybe thinking about investing in them, but I was just like, that's a very different way to process and internalize. And I mean, even in the age of AI, It's like the one thing that like the LLMs probably couldn't do a good job at. But great guy. I mean, have you read the poem? That's rough. Jokes are rough, but poems maybe. It's all rough. Evan's poems were epic back then. That was how I would do the demo days. I would just read his poem.

3:25:05Yeah, yeah, yeah. He'd post them all. It was great. Yeah, yeah, yeah. But I want to hear the story of like, tell me the story about one particular company that particularly surprised you or did something unique throughout a particular batch and really take me on the story of like what can happen? What does great look like in the context of this 10 week sprint? Yeah. So the company that comes to mind right away is Open Router. So a lot of the folks listening probably use Open Router. It's the top router when you want to switch between models. So Gemini 2.5 Flash came out and it was like incredible and way cheaper.

3:25:39And then if you're on Open Router, you click one button and you switch everything over. Open the eye to Gemini or from Gemini to Anthropic. And if you actually go, you can go to openrouter.ai slash rankings and they have like essentially the definitive usage rankings for all the LLMs. But to your question, when Alex was in HF0, this was like fall 2023. And back then there were only a few models. And so he took this really contrarian bet at the time that actually there would end up being a long tail of models that people would use. when at the time everyone thought it was going to be like one model to rule them all.

3:26:20And so at that time, there were very few models and this bet wasn't even true. And actually in the middle of the batch, he almost gave up on that thesis. He was like, man, maybe I'm wrong on this. But in the context of HF0 being with 10 other teams who were like on the ground, who were pushing these models to their limits, he ended up like having the right surroundings, the right community and stuff where he actually like stuck with this kind of crazy thesis at the time. And now if you go to their website, OpenRouter, like they're doing like trillions of tokens a week. So that one's - That's a great story.

3:26:55We're actually having him on the show soon. So I'm really excited to meet him. I've never met him, but he seems like an amazing dog. Yeah, so Alex was the founder of OpenSea before he started OpenRouter, so. That's where I know his name. He was in my YC batch. Yeah, amazing. I didn't realize that he, yeah, second time founder. That's such a unique thing. And you wouldn't like a lot of people don't think that like a second time founders or someone as successful as him would go through any sort of program. Usually it's just like straight to series D, you know, there's so many stories you did. You did Crossman super early too.

3:27:24Right. I remember Dave called me and he's like, you got to talk to this company. They're insane. Yeah. They were raising on an uncapped note at the time. And I was like, the whole round uncapped after HFZ. You did this uncapped round with a discount, of course. And I was like, I couldn't get on board with the uncapped note personally. And I ended up having to invest like three weeks later at whatever the Series A price. And you were totally right on that one. Yeah, I want to talk about those repeat founders. This is something that kind of shocked me. So I went through YC in 2012. And then my first company went through YC in summer 2012.

3:28:01The one went back through in winter 2016. and I was like, I'm still like a founder. I'm still like a startup guy, but we'd scaled the first company to 50 people. And I was very much in manager mode. Like I had like gray hairs on staff and like managers. And even though we had enough money to hire people, there was this big mental shift. I remember Dalton Caldwell from Y Combinator kind of telling me like, hey, you got to go back into individual contributor mode. You got to, even though you can afford and you're connected enough, you got to do it yourself. And through that, it was a wildly different experience.

3:28:34And it was really good. Talk to me about some of the shifts that you see in these second-time founders going back into these dedicated environments where you're maybe not hiring and scaling and raising a bunch of money, even if you can. Yeah. So about 70 % of the founders we back are repeat founders. And usually our kind of bar for actually considering the repeat founders is their last company did at least 10 mil annualized. And in fact, in this batch, the 10 teams, five of the founders either built a billion dollar company before this or their previous company was doing 100 mil plus wow that's very different and so the the challenges of repeat founders are actually pretty unique to being a repeat founder it's like by virtue of their success their last company they had gone into they were managing you know hundreds maybe even like a thousand people yeah but what that wasn't the recipe that made their first company successful in the beginning they know that and by virtue of their success they're usually angel investing yep they're advising startups everyone wants them to speak at their conference.

3:29:29They might have families. 60 % of the teams we back, someone has a kid. Wow. Yeah. Very different from really, right? That's hard mode. Yeah. And so, and so the, the challenges they have are different. And so actually our, our, like the residency that we offer is like very different. It's not an accelerator. It's not incubator. It's a residency. They actually like leave their normal life for these 10 weeks and we subtract everything from their life. So our whole thesis is while everyone's focused on value add, we focus on this thesis of recursive subtraction that actually our mantra at HF0 is the most insidious distraction isn't networking or conferences or this or that.

3:30:08No, it's the second most important thing in your business. And so founders are so often so good at doing the second, the third, the fourth, the fifth, the sixth, seventh, eighth, ninth, most important thing in their business. But the first, the most important thing in the business is usually a blind spot it's usually something that that only you only see when you force yourself to stop doing everything else and then you're like you're on that second most important thing that's very important and you ask yourself that question is this the most important thing in my business right now and and it comes up for you and you just can feel it's it's not but you don't necessarily have clarity yet yeah on what and then and then you realize oh man i've actually been afraid of this thing because if i if i go here, I might find out that the whole business doesn't work.

3:30:52Yeah. So so that's why it's usually there's some fear that is holding. And so by doing this recursive subtraction, startups just end up like zooming forward because in the beginning of a startup, like early stage, you essentially have two things. You have like your rate of realizations. And a lot of this comes by like talking to your customers. And then you have a grind. And the residency format allows you to increase your rate of realizations and then grind like you've never grinded in your life before. It's like it's like pretty insane. So that's why, like in our last batch, 40 percent of the teams broke three mil ARR by demo day.

3:31:24So like the results we're seeing, like even if you're Sequoia, you shouldn't be able to pick this well. And I think it'd be hubris for us to like we're getting amazing founders. But I think it'd be hubris for us to think it's just that we're picking this well. Part of it is the selection bias, like the set of founders who are repeat founders who are willing to leave their nice apartment. Most of them made money on their last startup going to leave their nice apartment and move into the residency for 10 weeks. They're serious. They want to birth their life's work. But the other part of it is that the residency works.

3:31:52This format of company building, I think a lot of companies like kind of did it in an emergent way, like Apple in a garage. Facebook had the house in Palo Alto. Even Dev Incognition had the house in Atherton. So a lot of companies kind of do these residencies without fully realizing it. We've figured out a way to do this to like a whole other level. And it works like every single time we're just getting like the results. Don't even make sense. I can, I can imagine a conversation, you know, with, with a CEO and in one of the batches and, and they're saying like, Oh, I'm worried about this, you know, or like this thing's working, this thing's working, this thing's working, et cetera.

3:32:29And you're like, but what are you afraid of? And they're like, Oh, I'm not really afraid of anything. You're like, no, what are you afraid of? And then they're like, Oh, I'm afraid of this thing. And then you're just, you know, you're, you're effectively the sensei. Yeah, you can't really directly just ask it like that. You actually have to like stop doing a bunch of the other things that are important, but aren't the most important thing. And it's only when you actually subtract over and over and over again, you actually have the space to even realize that you might be afraid of something. Yeah, totally.

3:32:56We keep coming back to this X thread about the motivations of the modern Silicon Valley founder by Ana. I want to get your reaction to this. She identifies a trend. She says, there's a cultural shift in the startup ecosystem that has moved from contrarian rule breakers and hackers to the overachieving STEM kids for whom being a founder is simply the next box to tick. And it seems like you have carved out an area or like a weird, I don't even know what, like, yeah, like just you still have to be taking a risk to do what you're proposing. And I'm wondering if you can comment on that trend as like...

3:33:41It's like sacrifice. Exactly. Like there are ways to do founder that are high status now. And low sacrifice. And low sacrifice. Talk to me about the status symbol that being a founder is and how you think about that within the context of your program. So one of my core values is substance over hype. And so actually, like at HF zero, we actually kept the lowest profile possible for the first couple of years of the program, because in my mind, I'm like, there's nothing to talk about here until we have our first unicorn or first company goes public or whatever. Like, I already know a ton of engineers and everything.

3:34:18I want to just build the best product in the world for founders. There's no one else building this product for repeat founders. And so that's what I'm constantly focused on. That said, Jordan, our creative director, who I think y 'all chatted to, and then my co-founder, Emily, did convince me. They're like, hey, Dave, there's like nothing online about HF0. It's like two years in. They're like, we should do at least something. So when people who don't know you already look it up, they know it's like actually a real thing. And so they fought me on this for like six months. And then finally I was like, okay, well, if you get the New York Times, we can do it.

3:34:52But I was joking. I didn't think that actually. And then next thing you know, we had a New York Times writer like living with us for a week. and she happened to have been like a google engineer before she was a new york times writer and then she loved it so much that she started coding again oh wow and so that was sweet and so just you're turning journalists these repeat founders who join our program and really the founders that we filter for they are here to birth their life's work like these folks aren't founders because it's any like any sort of status thing they're founders because like whatever like way they grew up whatever like childhood like they they can't not be founders they can't not build things um so that's like when you're at hfc you're surrounded by a bunch of people who like it doesn't matter how high status low they just like can't stop themselves from building um and yeah that's that's the really cool thing and it's you know it's only 10 teams and just You're surrounded by other folks who are like this kind of, I'm going to say addicted.

3:35:54Like it's a sacrifice thing. It's like a very painful thing starting a company. And they've done it like most 70 % of the founders of the back have done it before. They've been through that pain. And then for some reason, they're like sick enough to like do it again. And yeah, and I respect the hell to them. And then, you know, seeing I watched like a little clip of Roy's thing earlier. And there's a big part of him that's, you know, excited about the fame and the attention and all this stuff. But another thing he mentioned was around like the BCI stuff. Yeah. Right. We just start up in our batch.

3:36:26We just had our demo day this week called Conduit that literally as they just they had a huge breakthrough in like neural neuroscience research where they have the first helmet that can actually like read your thoughts before you type them. And it doesn't work every time. But when it does, it's like fucking nuts. And a week ago, this didn't work. And then now we're right on the verge of like literally as a coder, like you being able to wear a headband that actually even because think you have the thought of what you want to build. Then you're bringing it into English language. Then you're bringing it to code.

3:37:02And there's just like loss every step of that way. but if you actually could just see the the image the the concept the thing they're thinking in their head like prompt like you know vibe coding's cool but like thought to thought to cursor except all don't make mistakes like build me a build me a hundred million dollar revenue business don't make mistakes but i mean the bci stuff is super real like nudge there's so many companies working on this stuff now. And I saw one company that was doing like, they would read your brain while you were sleeping and then use generative AI to generate videos about what you dreamt about.

3:37:39Like even that, even if it's like wildly off and abstract, it's still just like a cool thing to have on your nightstand. And so it does seem like we're about to see a Cambrian explosion in this industry. Yeah, I wanted to have you review the companies, but we should just have a few of them on. Yeah, we should have a bunch of them on. I mean, we're having Alex on soon. So yeah, please send us all the companies. We want to talk to them. This is fantastic. Awesome. Thank you so much for joining, Dave. and congratulations on all the progress. It's an awesome story. One last note that's pretty crazy.

3:38:05So from our last batch, average valuation post-demo, it was 50.3 million. Whoa. Look at this. I thought of y 'all. Everyone was like, oh, it'd be great if there was like some type of - But it's only 16X revenue. Accelerator residency with reasonable valuations. I'm kidding. No, for their revenue, it's very reasonable. Yeah, yeah, yeah. Yeah, it's less than a 20x revenue multiple, which isn't that crazy. Yeah. They have real traction. If you'll meet any founders on TVPN who want to make 12 months of progress in 12 weeks, send them our way. Much love, y 'all. Will do. We'll talk to you soon. Great talking, Dave.

3:38:41Cheers. Fantastic. Well, anything else you want to cover? Somebody from your, I don't know if you want to, I don't know if you want to dox the family member. No, no, no, no, but this is what I was telling you about. Like Roy Lee is the most controversial person in my family right now. because half my family loves him. But is this your grandma saying he's... No. No? No, no. Someone else. But yeah, it's very funny. Good stuff. Anyway, is the intern Cam still active? I want to go to Tyler for one last pop quiz. Oh, is he on? Okay. I want to ask you. Yeah, can you hear me? Yeah, what's the capital of Michigan?

3:39:20The capital of Michigan. Didn't you go to school there? I think it's Lansing. I think you're maybe leaning on Cluelly too much at this point. I don't know if you're cheating using your brain or I can't really assess the product. You have too much domain knowledge on Michigan. What's your name? My name. Let me think here. It's like that paper that just came out, right? Oh, yeah, yeah, yeah, yeah, yeah. Do you think that based on your use of Cluelly, do you think it'll make you smarter or do you think you'll take your foot off the gas?

3:39:58let me think about that actually um i i think probably the latter honestly i mean it's just like i just don't need to think i don't need to store any knowledge in my brain now so maybe it frees me up to do more you know reasoning tasks okay but even then you know once they get oh three in here it's really like everything's covered so no reasoning did you get a feeling for what model was under the hood? I don't know. I think I could probably check, but I'm not sure right now. It's so weird interacting with somebody using Cluelly. Every answer is like, you don't know if they're reading Cluelly or not.

3:40:33Or just actually thinking. Well, anyways, people should go try the product. I really think that anytime Roy gets a negative comment, please just use the product for five plus hours and then you can have an opinion. The message to Andreessen, no crying in the casino. Because you know Roy Lee, there's probably a viral video of him going to the casino soon. Pretty soon. You know that's going to make a lot of people angry. Chamath is talking about SPACing again. Yeah. I think Hulu is a good target, I bet. It's not the craziest target out there. Not the craziest. You know, Chamath's in Mr. Beast, Holdco.

3:41:07Yeah, there's a lot of revenue there. It's growing, you know. Yeah. The market needs a pure play AI. I love pure play. A pure Roy play. Yeah. Yeah, there's no pure play. Well, we have some more breaking news. Apparently, Apple held internal talks about buying perplexity. Okay, yeah, yeah. You mentioned that. But clearly, everybody's talking about buying perplexity. Yeah. Yeah, I mean, it would be an interesting way to just beef up Apple intelligence and Siri very, very quickly. you know, the perplexity team's clearly built a great product. The question has always just been like, it's hard to unseat people from ChatGPT and Google.

3:41:51It's really hard to break through. What's interesting is like comparing it to the Roy Lee story, which is like, I feel like perplexity has gone viral many, many times. And if you look at the, even just like the ex-presence of the CEO, Arvind, let me look up how many.

3:42:12Arvind at Perplexity, he has 271 ,000 followers. Roy Lee is just over 100 ,000. So similar kind of broken through, serious, on the podcast circuit. Very respected founder, raised money, a lot of attention on the company. But it's always difficult to actually get people to switch their daily driver from Google search to perplexity in mass, but if you can team up with a company like Apple, that would solve that distribution problem. And so there probably is a case to underwrite the deal at some sort of, I imagine the number was big if they were talking. I don't imagine it was some, you know, oh, we'll take you out for a fraction of the preferred.

3:42:56No, it was probably a real deal if they were talking about it. And so, but the question is like, could it make iPhone users happier? Could it make the next version of the iPhone even more addictive and higher retention and get people to upgrade to the Pro Phone with a faster chip that's gonna have faster local perplexity and faster interactions over Apple intelligence and Siri. I mean, Siri was an acquisition after all. They didn't pay very much for it, but it was. And so the idea that Apple never requires things is not really true in the AI space. So I don't know. I'm rooting for them either way.

3:43:27I'd love to see. I'd love, as a consumer, I'm rooting for Apple. I would love a better search experience on Apple natively. And I feel like getting a major player in AI in the company would be the perfect thing right now. So I hope it happens. But good luck to everyone on both sides of the negotiation. Well, it is the weekend. Have a great weekend, folks. Leave us five stars on Apple and Spotify. And thanks for tuning in. We will be back on Monday. Have a great weekend. I cannot wait. See you then. Bye.