Jacob Peters, Logan Kilpatrick, Roy Lee, Erik Torenberg, Jack Whitaker, Trump Laying Groundwork to Blame Powell for Downturn, Boeing to Sell Some of Its Navigation Business in $10.55B Deal, History of Thoma Bravo, Oil Companies Have Bold Plans to Fix Their Water Problem

25:56But it was just a creative mess. They would just have fun, and they were under FTC investigation and SEC investigation. They weren't even trying to cook the books, they were just so sloppy and messy, they didn't care about the books. And so the accounting was all off and there were all these different lawsuits going on and Strauss came in Wind up buying the company for like kind of zero dollars. It was a crazy deal We should dig into all that but basically he gets the he gets the company and immediately starts running it like a professional Business because he's a beast of an operator and so profit 21 this first tech Acquisition did really really well, so they exited with a 4.7x return on investment And this was an early win in software with minimal debt and just hands-on operational fixes.

26:41And it really cemented the firm's new direction. Middling outcome if you're a venture capital firm. But if you're a PE firm, that's going to do a bunch of those. Yeah. And it feels repeatable at this point. Avoid losing money. It feels repeatable. There are other companies that are in Profit21's segment. And that playbook can be run again and again. So Orlando Bravo, he's just 30 years old. He gets promoted to partner in 2001 on the strength of those efforts. Being promoted to partner at 30 in 2001 just sounds amazing. Goated. Sounds amazing. No, the timing is perfect to go on this generational run in software.

27:18Oh, yeah, yeah, yeah, yeah. I mean, timing is a lot, but it might not be everything, but he really nailed it there. And so software deals started piling up. Orlando Bravo's influence within the firm grew. By 2005, Carl Toma and Orlando Bravo had recruited a trio of new talent. Scott Crabill, Holden Spot, and Seth Borough joined the investment team to focus on software sectors. All three of them are with the firm today as managing partners. One of them came from Summit Partners. The other had experience at Morgan Stanley. And the last one was also at Summit Partners and History of Investment Banking.

27:56So a pretty common path into private equity from banking and other deal-making organizations. So as they're growing the team, they can also scale up their deal flow. In 2007, they rename the firm because they love Orlando Bravo. The firm is renamed Toma Cressy Bravo. Overnight success. A year later, 2008, co-founder Brian Cressy decided to depart. Cressy spun off with a team to form Cressy and Company, refocusing on healthcare investments because Toma Bravo, of course, is doing tech stuff and he's doing healthcare. And they're just slightly different strategies. So they split the firms and then the firm renames to Toma Bravo LLC, officially born 2008, an overnight success.

28:41And today they're buying Boeing assets for$10 billion, running kind of the same playbook. It's a boring software company that's just profitable, growing. A series of software companies that they can then go bolt on other stuff to. And just continue to grow that. They're not dipping their toe in the super hot, high growth startup market. They're going for just, hey, there's some great software assets. They're installed a bunch of places. It's critical. They built a bunch of stuff. This business is good. We can make it better. Yeah. And I'm sure they'll look at venture assets that are eventually, maybe they don't reach escape velocity get to, you know, 100 million of ARR, but not quite breaking out.

29:25But they will be very sort of aggressive in terms of valuing these businesses. And certainly, you know, if Toma Bravo is buying your company to bolt on to an existing sort of roll up, very unlikely they'll give you a revenue, you know, 100x revenue multiple, could very well be a 1x, could be less than that. It could be more if you're growing quickly and have a real path to generating cash flows. As they scale up, they start doing more and more ambitious projects. They eventually scale from mid-sized PE firm into just a complete heavyweight in tech buyouts. In 2012, they do a $1.3 billion take private of Bluecoat Systems, a cybersecurity company.

30:11Bluecoat had solid technology but needed a strategic overhaul. And Toma Bravo takes it private, improves operations, and later sells it for$4.5 billion in 2015. So three years, they go from$1.3 billion to$4.5 billion. And of course, at this point, they're also able to use leverage because they have better access to the debt markets. And so they're making even more money. I'm sure you're trying to pull up a size gong of some sort. I was trying to pull up a lever up button. But they pull up this... Risk on. Risk on. But they basically got a pattern. They buy a public tech company. They optimize it privately and then sell or IPO at a much higher valuation.

30:46And this is the whole take private strategy is like, it's very hard to completely turn the cruise ship or the battleship in the public markets because your stock's just going to get completely hammered while you go down for a couple quarters because you're saying, hey, there's a business over here that's really, really damaging us. It's low margin, but it's going to hurt our revenues. And if somebody's looking at you on a revenue multiple, they're going to be upset about that. Or there could be a number of different things that could look like red flags in the public market. But if you have a couple of years to rebuild in private, you can just take more risk.

31:20You can go risk on. Risk on. The other thing that they're doing, just because a large part of our listener base is venture. By using leverage, they can get a much higher return on their equity. So if they're buying a company for a billion dollars, they might, in some circumstances, only put up you know,$200 million, right? So when they turn that$200 million into a$4 billion outcome, it can be very meaningful. So there's one more funny deal we got to go into. By 2018, Tomo Bravo is regularly doing multi-billion dollar deals. They're in the same conversation as Silver Lake and the other big tech buyout firms.

32:07In partnership with Silver Lake in 2016, they buy SolarWinds, this IT management software company. Are you familiar with SolarWinds at all? Okay. So SolarWinds is IT management software. They will go in the network, watch for what deployment's happening. It's like kind of middleware for deploying software at a large enterprise. But there was a hack where it was a very, very bad SolarWinds hack. Did you hear about this? And so basically someone injected something into SolarWinds. So then that was what's called a supply chain hack. So every company that had SolarWinds was now vulnerable. And it was like a really, really brutal attack.

32:46And it was funny because I was doing it. I was trying to understand like the history of SolarWinds. And I found the founder who had built the company, sold it, like exited like years and years ago. And he had gone on Shark Tank to pitch a cooler like a like a like an actual cooler that you could put like beers in because he was like yeah it's truly golden retriever mode mindset so he he started this cooler called the coolest and it was it was a cooler that you put ice in but then it also had a fan so it like provide you air conditioning when you're hanging out with your buddies watching the baseball game or something.

33:26He was clearly like post exit founder, just like hanging out with the kids. Like, yeah, wouldn't it be cool if like we had a better cooler? And so he goes on there, he pitches it. He explains that it's like a thousand dollars or something. Cause of course he's like, Oh yeah. Cool. Or like, what does that cost? Like a thousand dollars. But he tells the sharks that he's like, yeah, I founded solar winds. And you'd think that would be like total bull signal, right? You're just like, okay, this guy built a billion dollar business. Like he's probably can figure out like the cooler market. Um, but all the sharks pass.

33:55They all say no, they all say they're out anyway. Uh, fascinating little side side tangent anyway. Yeah. The, uh, unfortunately coolest cooler shut down after a five year saga and 20 ,000 people that backed the Kickstarter didn't actually get the, Oh, that's brutal. Um, yeah, very rough. But, uh, anyway, I mean, solar winds wound up being a great company and, and Tomo Bravo, I'm sure made a bunch of money and then they had the hack and I'm sure they rebuilt from there. But, uh, All of that is to say that Orlando Bravo became known as an incredible dealmaker. He was on the cover of Forbes in 2019.

34:30They called him Wall Street's best dealmaker. Pretty great. He was the first Puerto Rican-born billionaire in finance. He was on the Forbes 400 list. The Financial Times nicknamed him the King of Sass. We love that. A lot of people vying for that at a lot of venture capital firms. But you're going to have to go up against Orlando Bravo, the king of sass, the best deal maker in the world, to really, really own that name. Bravo's war chest grew. They got a fresh$12.6 billion fund. Now they're eyeing a$10 billion plus deals. And we just saw that with Boeing. They did a$10 billion deal. How'd they do that?

35:07They raised a massive, massive fund. And there's a bunch of other interesting deals that they've done. But they're doing mega deals. Yeah, just to give you a sense of some of their bigger deals, they did Proofpoint at$12.3 billion. The Boeing Digital Aviation is actually their second largest, which hasn't obviously closed yet. Oh, yeah, Anaplan, too. Anaplan,$10.4 billion. RealPage,$10.2 billion. SailPoint,$6.9 billion. Medallia was at$6.4 billion. Acquisition binge. Darktrace,$5.3 billion. Sophos. Click, Imperva. Absolute size lords. You don't need to do that many of these deals to put up some very large numbers.

35:54Although there is something unique about the Boeing deal that was announced today is that this is kind of the start of a new strategy for Tomo Bravo. We've seen a lot of traditional private equity deals where they buy a private company, do an LBO and grow and buy and build, that buy and build strategy. Then they start doing the take private turnaround strategy where they take a single digit billion dollar public company, take it private, change the strategy, try and sell it later or take it public again, exit the position. But they also are starting to do carve outs in 2023, 2024. And that's what the Boeing deal is.

36:36They're carving out a piece of Boeing's business. and then packaging that up as an entirely new business. And I mean, it is a different operational challenge because it's not operating completely independently. And so you're going from being a Boeing employee to being an employee of a new company, a digital aviation software portfolio that's owned by Tama Bravo that will eventually have a brand and be cobbled together from all these different assets. Anyway, let's move on to our next story. The Oil Patches Manhattan Project, How to Fix Its Gargantuan Water Problem. I thought this was interesting.

37:14I had no idea the ratio of oil to water when they are fracking. So for every barrel of crude oil that's pumped, they are now producing four barrels of water. And so this is happening in West Texas and New Mexico around fracking. The Permian Basin is the location. And so there's a big question about what do you do with the gargantuan amounts of noxious water that they produce because it's gross. And you need to filter it or treat it if you want to put it back in the water supply. So typically what they do is they just pump it back into the ground. And so if you go a few slides forward, there's a graphic of what they do.

38:02They're taking oil-rich shale, which is how you frack. So you blow, I think, like air and water into the ground. You suck up the oil that's there. The oil is very, it's not like a clean pool of oil. There's a whole bunch of other stuff in there. You pump all that up. This is about two miles down. Then they kind of filter out the oil, take the oil out, and then they pump the water back down, three miles down. but that's causing earthquakes and so people are upset about that they're playing god we don't like that we don't like earthquakes because they're pumping so much water like a word yeah um yeah so we can kind of go through this so so they're trying to evaporate the water which is interesting yeah so they're basically producing all this wastewater it's just sort of like sitting there in a pond and then they have a system to effectively uh try to get it uh just disappear it disappear into the atmosphere um so there's a picture here i don't know if we can pull it up but uh you can see it's a fascinating image so there's these evaporators hum on a huge saltwater pond in the middle of the biggest oil field in the u.s they are part of an experiment by exxon mobile to address one of the challenges facing frackers in the permian basin and all these crazy economic projects to to process all the side like you know waste products from this.

39:22Um, but it's all part of, uh, you know, the goal to help unleash American energy. And, you know, I'm, I'm generally in favor of figuring out new ways. I mean, fracking was in many ways, like a miracle in terms of America's energy independence. Uh, very controversial. You know, the meme that's like America's like, you know, struggling. And then we just discover like, you know, an immense, like, oh no, we're out of rare earth elements. And then it's like, oh, we found the largest deposit ever in Utah. Companies in recent years have made strides in treating brine and recycling it for their operations, but more water flows back to the surface than they can use.

39:58Meanwhile, the volumes of liquid they can pump in the ground have shrunk as companies have started to run out of underground space and regulators have imposed limits on disposal to prevent earthquakes. Enter the evaporators. Exxon, which is the largest Permian producer, started testing the machines about a year ago as a pilot project. The device is manufactured by Colorado-based RWI Enhanced Evaporation. A little hard tech company we've got to get the founder on. They blow air down on the pond. The company's probably been around for 200 years. For sure. They've been evaporating for about, you know, coming up on - Who knows?

40:35Who knows? Maybe RWI is just like, you know, seed stage startup. They just went through YC. You never know. Don't judge. One machine costs$46 ,000 and consumes about as much electricity is wet, dry vacuum cleaner. RWI estimates that between 20 % to 50 % more liquid gets vaporized that way compared with natural evaporation. And it's interesting because like if this water truly is like underground, is it something that when it's brought up in the atmosphere, it's going to wind up, like if it gets purified and evaporated, does it wind up going into clouds? Does it wind up getting seeded? And does it wind up making the land in that area more verdant because there's more water?

41:12or was it already considered in that equation because it's kind of groundwater? Or was it buried so deep it doesn't matter? I mean, these things, like the primary thing here is environmental concerns, right? Totally. If you're a farmer trying to grow organic food or you have cattle a mile away. Oil water raining down on you. It sounds like acid rain. It sounds bad. But I'm sure that RWI would argue that there's some chemical process that means that the water itself is potentially just as pure as rainwater. It's hard to tell. I mean, evaporation is pretty effective. It's separating things out. That's how it's done in a little bit of a flash.

41:50But this is like a very long-running trend where looking at the production side for groundwater in these areas, specifically for oil and gas production, for a very long time, if you owned a piece of land, you were legally allowed to pull as much water as you wanted out of the ground. Yep. And what happened is regulators, governments, states, et cetera, realize that no water is a shared resource. There's effectively rivers that run underground that are, you know, tied to aquifers. Those aquifers can stretch across multiple states. Right. They're massive. and so um you know uh one private company uh that it seems like broadly you know we need a change to regulation and energy you know across like you know we've had a number of nuclear founders on but uh the nuclear thing is huge for desalination yeah it seems like almost impossible and a perfect match because you have such concentrated energy you can just throw the nuclear power plant right next to the desalination plant you have all the energy it can run 24 7 like it's perfectly matched as opposed to a lot of other systems yeah and one of the yeah the issue with this fracked water is that it's salt water yeah um yeah and and that means that you can't just go spray it on plants or you can't put it on crops it will just kill everything yeah it's bad uh so this is interesting last year permian drillers discarded roughly 5.5 billion dollars billion barrels of water by pumping it back down into the ground, right?

43:24So 5.5 billion barrels a year. That's what I want you to keep in your mind. 5.5 billion barrels a year. So they're doing a, so an injection well typically - There's 42 gallons per barrel, by the way. Okay. So an injection well typically guzzles between 10 ,000 and 25 ,000 barrels of water a day. An evaporator working in optimal conditions optimal conditions, vaporizes about 32 barrels an hour. And so now Exxon is trying to build a commercial-sized desalination facility operating in the Permian by the end of the year. Last month, they commenced a pilot project, but the testing that's happening, the desalination technology, is doing 20 barrels a day.

44:10That's the pilot. So 20 barrels a day is, what, 700 barrels a year compared to - Not making a dent. 700, compared to 5.5 billion. Yeah. Like we're off by so many orders of magnitude. But at the same time, like you have to start small and you have to, and you have to, you know, run the test to see, can you actually purify out all the toxic chemicals in the water? and does the actual scientific process work? And then it becomes an economic equation of what happens. So they are scaling up. They're going to do a$25 million desalination facility with an initial capacity of 10 ,000 barrels produced a day.

44:52And there's plans to build a plant that's 10 times bigger next year. Yeah, this is the challenge. So generally with oil production, you think it's just about how do we get this out of the ground? It's how do we get enough water to support this process? And then what do we do with that water after the process to dispose of it? And you're being regulated on both sides, right? I have a portfolio company called Resource Monitor that helps oil and gas companies basically stay compliant on the production side. So they're producing all this water, and they get budgets from various groups on how much water they can actually produce.

45:30But that's only one part of the problem here. Yeah. So the goal is to bring the cost of purifying a barrel down to 75 cents, which is slightly more than what it costs to flush it down the well. This$25 million desalination facility is going to do 10 ,000 barrels a day, which is$3 million a year. But remember, we're$5 billion we're producing last year. And then they want to build one that's 10x bigger. So that'll get you to$30 million. but you're still two orders of magnitude off. Two orders of magnitude, a hundred times. They're going to, like, if they build this 10X plant, they will only be doing 1 % that they're pumping down into the crowd.

46:14It's actually crazy, the scale of this thing. Anyway, you know, they're trying to scale up, and there's a bunch of companies that are working on this. Anyway, if you want to go hang out in Texas, book a wander. Find your happy place. Find your happy place. Book a wander with inspiring views, hotel-grade amenities, these dreamy beds, top tier cleaning and 24 seven concierge service. It's a vacation home, but better. And in most of those homes, you're going to find really, really nice lighting. And that's the topic of our next story. Offices are ditching harsh fluorescent lights. New tech is on the way.

46:46I think this is going to be extremely controversial because the new tech, it doesn't sound Lindy. And so we're going to have a debate over whether or not the path to better lighting is more advanced technology or do we need to go back? Do we need to go back to incandescent? I was hoping they were going to say every office will be required to have a warm fireplace. A warm hearth. A hearth. We're going back to hearths. I was joking the other day, I think on the topic of lights, I think Apple to really differentiate should come out with an iPhone UV that sort of gently tans your face. That sounds like a real product, the iPhone UV.

47:24The iPhone UV, it gently tans your face while you use it. Yes. And I think they've got a hit on their hands, at least in Los Angeles. Yeah, yeah. They do the blue light filter. They can take the blue light out. Why not add some UV light in? Yeah. I love it. From faux skylights to circadian tuned systems, lighting upgrades are a priority for companies trying to lure employees back to the workplace. Faux skylights. I know. The faux skylights are crazy, but it's going to be really big when I build my Batcave that's three stories deep into the ground. Yeah, it's going to be really key. You're going to want 24-7.

47:57It's going to be like Las Vegas where you can't tell what time it is and you just get lost. It's going to have a kind of severance vibe. That's what I'm going for in my basement when I build. Glaring fluorescent lights in the office are on the way out. The technologies coming in promise to do much more than make everyone look better. Improved and potentially more healthy lighting is high on the list for companies and building owners trying to lure employees back to offices after an era of remote work. They are investing in new technologies such as faux skylights that mimic natural light, complete with a virtual sun and moon, and adjustable illumination systems designed to sync within play circadian rhythms.

48:35I wonder how long until this becomes a culture war issue. Because you know the air conditioning thing became a huge culture war issue? Well, I mean, incandescent light bulbs were banned. Yes, they're extremely hard to get in California. I've been struggling with this for a long time. Trump sort of like re-legalized them, I guess. I didn't hear about this, but that sounds cool. I like an incandescent light bulb. I'm a fan. I like a tungsten bulb. We've known for a long time that natural light is better and makes people feel better, so it's not a completely crazy idea, says a professor at Wharton.

49:07So aside from the psychological benefits, research studies have shown that light can have an impact on non-visual brain function during cognitive tasks, particularly those that involve sustained attention. Office lighting revamps are expensive. installing some of these technologies can add 20 to 30 percent to the cost of a project those in the industry say and it could take time for them to become mainstream i'm assuming that's 10 to 30 percent more expensive than the existing light solution you know stanley kubrick shot a film entirely using candlelight and he had to collaborate with nasa to make a new lens that could absorb even more light so he could shoot in darkness, essentially.

49:51And now, modern filmmaking technology has gotten so good that the cameras that we shoot on can basically shoot in the dark. And so, I think for the next set, we might want to consider just a warm hearth, a nice fire. That's right. Some candlelight. There's something there. There's something there. Some nice, a candelabra that we just light and that lights our face. It'd be a very different look from your typical new show. But I think that's what would make the show special. Well, once we go to 24 hours a day, we're going to obviously have eight sleeps on the floor. We'll be able to sleep in cycles.

50:21For the sub-a-thon. Yeah, yeah, yeah. So there's something there. So programming the day, playing into post-COVID wellness trends, office designers are exploring so-called circadian rhythm lighting to sync with the body's circadian rhythms, the biological clocks inside our cells that time when we sleep and wake. Expect to see illumination that can be tuned by intensity brighter or dimmer, and color temperature cooler or warmer throughout the day to mimic the light outdoors. So I'm in favor of this. I want Huberman's take on this. I want Brian Johnson's take on this. I want Soulbrah's take on this.

50:52Marc Andreessen notoriously has gone to war with Huberman and said, I'm going to use my phone. That is one of my favorite interactions on the internet. Their bromance is fantastic. It's amazing. It's hilarious. Anyway, yes, you got to be using, at the very least, Warm white 2700 Kelvin bulbs. You can't be walking around in the bright fluorescent. Lighting's important. I mean, the issue with LEDs is there's the flicker, right? Which is not visible to the human eye. But if you use it on a camera, it'll drive you insane over time. I firmly believe that. Efforts to develop such lighting took off after researchers in the early 2000s discovered photosensitive cells in the retina that detect light generally below the level of our awareness.

51:43These photoreceptors, independent of vision, can affect biology and behavior. Research has found. So, I mean, really what it depends is not just the one size fits all. It's like, do you want an insane work culture? Do you want to drive your employees insane? Well, then maybe you should go with the craziest, most strobing LEDs constantly. So when your team is in the office or they're ripping their hair out hadrian on a lights out factories yes lights out offices only light is from screens from your email machine just yeah blasting anything to get you to work harder anyway um you know i imagine that the lighting is pretty good if you're at um if you're at a uh an auction for a championship bowl we should do that story before our guest tops on in six minutes.

52:32The vicious sport of landing a prize bull at auction with herds thinning, finding the right stud is a high status game. That's why you come to this show, folks. That's right. Tech, business, this is the business of the technology of prize fighting bulls. With herds thinning, finding the right stud is a high stakes game, but how do you tell a$45 ,000 bull and an $8 ,000 bull apart? I know a lot of you have been asking that question. Well, we got the answer from you, courtesy of the Wall Street Journal. When Randall Grimmius shows up at an auction, everyone knows there's a prize bull in the barn, John.

53:06The 56-year-old rancher has an eye that is the envy of America's livestock industry, never more so than now. The U.S. cattle supply is thinner than it has been in 75 years, sending prices soaring for bulls with the tools to repopulate a herd. Ranchers travel to auctions across the country in search of the right bull, and they can't afford to mistake a dud for a stud. I had no idea that there was a population crisis among the bulls. I only see people yapping about humans. I had no idea. That led to a crisis. So when Grimmius' private jet touched down here last month and he strolled into a barn at TD Angus Ranch, all eyes were on him.

53:43Which one did he like? Grimmius would keep them waiting. I love this article. After 283 bulls were scooped up, his prize came charging out onto the show ring. Number 284 had what he likes, a longer neck, a chest that looks like it can add more weight, and a backside with a nice arch. We're talking about bulls here, to be clear. Grimmius figured the bull could fetch as much as$60 ,000, and he was prepared to pay up. Sales like this one can make or break a rancher's year, and as a result, they are filled with the kind of tension and gamesmanship often associated with art auctions or the NFL draft.

54:22in North Platte that day. The average selling price was about$10 ,000 up from a year ago and double what it was seven years ago according to the rancher running the sale. He and his wife, Dana, spent hundreds of thousands of hours preparing his roster of bulls and putting together the catalog. They drew ultrasounds on the bulls to compile stats such as marbling score, which determines the potential tenderness of a future kind of state. They're doing more diligence on bulls than the average VC does on a$100 million check. For sure. No, this is a mature industry. This is very, very serious. Their paycheck comes once a year during the sale.

54:58Ranchers pour through the data and analyze images of a sales catalog. Rare is the bull that hits the boxes based on his genes, frame, and testicular fortitude. Scouts also look for intangible traits, such as if he looks powerfully made or has the requisite sass and swagger. Founder mode. They are in founder mode to spread his seed. Sean Lowry is a sixth generation Nebraska rancher and is part of the Mildale Ranch, one of the oldest running ranching operations in the state. He looks at the lineage of the bull, who the mother and father were, and certain physical traits, not too big with near perfect feet for wandering around his 3 ,000 acres and large testicles that won't freeze during the winter.

55:40It has to have that look. You just know it. He can't be too tall, can't be too small. We like them long, we like them deep. this is in the Wall Street Journal folks uh this is an important industry it's fascinating up close auctions get messy the bulls occasionally defecate and if a tail swings at the right time it flings it into the crowd like a killer whale splashing the onlookers at SeaWorld we gotta we gotta go yeah we gotta go we gotta go and uh see one of these bull auctions uh you know I I follow an account that does breakdowns on horses uh for horse racing and the technology that they use to do due diligence on horses, on thoroughbreds, is far more advanced than anything you'll find in a dock's end in Silicon Valley.

56:23They match the gate. They have these computer vision algorithms that show the different proportions of the horse. It's incredibly detailed, and the horses can get well up into the high six figures. So the bidding started at 5 ,000 before quickly climbing to$15 ,000,$20 ,000,$40 ,000, and soon$55 ,000. Your man has to beat$55 ,000. Come on, Randy, he said, trying to get Grimmius to bid. Grimmius didn't bite. He knew his prize was later in the day. It's interesting. You know, the same dynamic plays out in venture, right? It's sort of like, oh, the second that Sequoia issues a term sheet, then every other investor is coming around and being like, oh, yeah, we always wanted to do this.

57:07and it's an interesting dynamic where in theory once he's starting to bid everyone else should just sort of like pile in and try to get the one that he wants so even some of the ranchers are sending other people so low reason people know him so he sent his foreman into the bleachers to do the bidding so he wouldn't get recognized because people know oh he's got a good eye he's the sequoia of bull bidding yes yes so he says sometimes i will get emotional i will get pissed and it will cost me thousands of dollars i didn't intend to spend i already said in this business these cowboys are awful prideful yep well um i would love to get one of these uh ranchers on ramp because time is money save both easy to use corporate cards bill payments accounting and a whole lot more all in one place and if they're flying out to different ranching uh bull auctions they're going to need to be expense expensing flights booking hotels if they don't have the private jet already.

58:04And ramp travel would be a great fit for these guys. So if you are a rancher, get on ramp today. And coming into the studio, we have Roy, who has been tearing up the timeline. We're very happy to have him join. How are you doing, Roy? Welcome to the show. What's up, brothers? What's up? Brother. Great to have you. How have the last few days been for you? Can you take us through the brief prehistory here of what happened with Columbia, and then we'll get into what happened with your launch. Yeah, I mean, earlier this semester or earlier this year, I built this tool called Interview Coder to let you cheat on your technical interviews for software engineering jobs, like Leacode-style interviews.

58:45I filmed myself using this to get a job at Amazon. I recorded it, posted it online, and Columbia saw Amazon reported me to Columbia. Amazon is mad, Columbia is mad, and this ends up being a spat that I sort of publicized on my Twitter. It becomes this big spectacle. And I eventually get kicked out of Columbia or suspended for a year and blacklisted from most big tech companies. What was the original plan? Was the original plan a publicity stunt or did you actually want the job at Amazon? I was never intending on taking the job at Amazon. The impetus for everything was I did a bunch of legal questions thinking that I would work at a big tech company one day.

59:24I really hated it. And at a certain point, I kind of just thought I really want to build companies is what I want to do with my life. So this is just seems like the optimal route for everything. I can make this big protest against LeetCode. And at the same time, I can do something that I suspect will be super viral and just give me like a distribution channel to build off. Yeah, that makes no sense. It's crazy. I mean, there's there's many people in venture that have spent, you know, 10 plus years trying to build a following here. And, you know, the growth of your account alone in the last few months has just been insane to watch.

59:56And, yeah, absolutely gives you an edge as you now, you know, come in. So take us through the product, what you launched and how that video came together. Yeah. So, I mean, Clue is supposed to be the, I guess, the ultimate experience layer for AI in a world where models are truly multimodal. And you can sort of have an AI that remembers the last 10 years about your life. If nobody's going to be on a chat bot, chat to the T.com, where are they going to be on? They're going to be on Cluey. And this is what we hope to build. And this is what we are building. We sort of filmed this launch video to be like a vision for the ultimate end state of what does true AI in everything look like?

1:00:33Like a true AI maximalist life 10 years out in the future. What does that look like? And it ended up coming together really well. And I think it resonated well with the world. Went very viral and very controversial. Yeah, very controversial. How did you, you went against, uh, you know, the norm you launched, uh, 2 PM Pacific on a sun on, on Easter. Uh, where, where did you just launch, uh, when, when you were ready or was that, was that planned at all? Yeah. I mean, it was four 20 and it was sort of in brand with like me being like the, the little like punk college kid, like I launched on four 20 might as well.

1:01:06Okay. Yeah. Yeah. Yeah. Um, yeah. What about, uh, what, uh, has any of the pushback been, correct. Can you steel man any of the pushback? Is it bad to cheat? I think there's two worlds of, or like there's two big arguments for the pushback. The first is, this is dystopian and this sort of destroys what it means to be human. And I disagree with that pretty fundamentally. I think every time technology has advanced our capabilities, people have said the same thing. This destroys what it means to be human. But ultimately, like that never ends up being the case. We just become more efficient as a species.

1:01:42And The traits that make us human are not the traits that sort of like AI can fill in the blank for. The second large area of pushback is this is cheating and cheating is unethical. I think it's sort of the phrase cheat on everything is left intentionally vague. I mean, what exactly does cheat on everything even mean? You can cheat on a test, but you can't really cheat on a sales call. You can't cheat on a meeting. You can't cheat on a conversation. in. What Cluey or what AI allows is sort of like an unfair advantage that it's so unfair that it feels like cheating. Imagine you're in a sales call and you have this tool, this genius tool that knows everything about you, your company, your clients.

1:02:19Anytime a technical question is asked or an objection is asked, immediately it just knows and just can give you the right answer. Your human intuition just looks at that and thinks, wow, this is not fair. This is cheating. But I mean, ultimately, this is where the world is headed. And really, this just makes us 100 times more efficient. Yeah. I love it. I mean, in many ways, business is like, you know, feels like a sport in some ways, yet there's sort of laws that govern, you know, corporations and, you know, different markets. But there's nothing wrong with giving yourself every possible advantage to win.

1:02:49And so you should absolutely just use every tool available to win. Yeah. Talk through the product. Version one seems like maybe a web app or an iPhone app, but then it seems like there's a, it seems like there's like a, like a augmented reality vision in the future. Um, but how are you thinking about that? Because, you know, there's a monopoly in Apple products. There's a monopoly on foundation models now. Like, uh, where are you seeing yourself breakthrough? Are you sitting on top of different systems or trying to build something from scratch? Right, right now the, the vision, um, the, the, the version of the app that we have is just a desktop app, has complete access to your screen and your system audio.

1:03:30So it sees everything you're seeing and it hears everything you're hearing and can help with that context. Ultimately though, what we want, the end state of the product is a chip inside your brain that lets you use AGI to think. That is the ultimate end state. And the way we get there will be sort of like variable, like if the models get significantly better, significantly fast, the chips get better, significantly better, super fast, then it's sort of like, like the way we get there is subject to change. Right now, we are building a really, really good desktop app that can be the ultimate player to your computer.

1:04:02Yeah. What would you, how do you respond to like the criticism that like, maybe you're over optimizing for distribution ahead of product, you're going viral, like you're frustrating people, you're being controversial. And like, maybe you should just be heads down and like go build something great. And then, you know, the, the, the, the business will build up slowly over time. And this is like, oh, you're like clout chasing is like the criticism. I think there's a lot of people that say distribution is the final mode. In fact, everybody says this distribution is the last mode. Anything can be built.

1:04:37And if you truly believe this, then you would behave exactly how I'm behaving. That's a good point. That's a really good point. But specifically, there are different modes of distribution, right? You could have found a distribution angle that frustrated people less, right? Or was less edgy, right? And so is the need to be edgy, is that a product of the way our social networks work or the way our distribution moats already exist? Like there's already a, like the humane team was able to get distribution and attention by kind of aping an Apple ad and they were able to get a distribution. You went a very different direction.

1:05:19Is that, are you a, like a product of modernity, I suppose? Yeah, yeah. I would think so. I think I have a very unique and I think a very strong viral sense. I've been into it about what makes things go viral and what things will go viral. And for me, that's like pretty edgy. And I would probably think and most people probably agree the video wasn't nearly as controversial. It wouldn't do 10 million views in two days. I agree. Yeah. Talk about how the round came together in the process there. I remember you first started going viral and I imagine it It was just your DMs flooded by everybody pushing back, being like, this is wrong.

1:05:57But I think you were smart to pick a common enemy. People generally are sort of anti-university right now. They're sort of anti-leet code. But talk about how the round came together. And yeah, I'm curious. because that kind of like you basically did a, you know, indirect roadshow where you had the attention of everyone in tech for, you know, a period of time. It was a really messy round. And I don't think I can advise most people on fundraising. The round in total lasted less than 24 hours. Yeah, I remember that. You just posted. I think I saw one post and then you're like, OK, the round's done.

1:06:38We are our pitch and our entire deck changed twice in the middle of the round. And the only thing we came in really knowing was, I'm super hot right now. We can build a bigger company. And right now is the best time to get money. In reality, in the future, there were only two worlds. One, I fell off and I can't raise again. Or two, I succeed massively. And all of a sudden, the extra 5%, 10 % dilution here, if we fuck up, it means nothing. How do you think about flexibility at the product level? It seems like you have a very clear vision of how to unlock AI for individuals broadly. It sounds like starting in the workplace.

1:07:17But are you flexible in terms of what that sort of exact implementation looks like? What kind of different niches? I can see the application in sales, et cetera. But how are you thinking about sort of adapting now that you have this flood of customer demand and attention? Now you need to like, you know, turn it into a durable revenue and things like that. Yeah. I mean, this is something that there's like a really, really hopefully viral experiment that I'm going to try. That will be the first of its kind. I'll just announce this right now. We're probably going to hire about a hundred interns and we're going to turn them into like sort of the ultimate content farm over the summer.

1:07:52So be like the most motivated high school and college kids. I really truly believe if you're over the age of 23, you probably don't have the viral sense that you need to go like consistently massively viral. We're hoping to have this gigantic content farm, and every single one of them will be assigned a different use case for the product. You guys are going to advertise this product for sales calls, and you guys are going to just ship out content for sales. You guys are going to advertise this for meetings, and you guys are going to advertise the deep research feature. And the consistent virality and attention on the different use cases, that will make it what's lasting.

1:08:25I mean, do you have a suspicion as to which use case is most profitable in the midterm? Because, I mean, the Riz app already exists for messaging people on dating apps. Deep Research allows, you know, basically everyone to already cheat on their research papers if they have to write research papers. There are co-pilots all over the place. And you're kind of pitching like an omni-co-pilot, at least in the short term. Um, are you still, is your formula like go viral and test everything and then find a new niche and then double down on that? And then maybe we're talking to you next year and you're like, yeah, we, we, we found it.

1:09:07We cracked it. And we are, we are cheat on sales calls and we're, and we're, you know, great for SDRs and we've kind of dropped all of the other things or, uh, or do you really want this to be everything for everyone on day one? Everything for everyone on day one is a bit unrealistic. Right now, the two things that we're really zoning in on are our meetings, like virtual meetings like this and also sales calls. But it's entirely possible that every single use case flops doesn't go viral isn't helpful, except for one random use case, in which case people like quintuple down on that. But the end state is, you know, like AI for everyone, for everything.

1:09:42I want to go deeper on the actual, like, hacking your way into products to make using the product easier. Talk to me about getting an iPhone app. Is there a way that you can hack your way into the camera button with a shortcut or the action button on the iPhone? Meta Ray-Bans, is there a deal that you can do to take over for that? It feels like most of the hardware providers have very sharp elbows around their ecosystem. Some of that's breaking with the latest FTC lawsuits, and maybe they're more open to it in the future. Also, the failure or the setbacks to Apple intelligence have kind of raised questions about, hey, should that Siri button be remappable to other things?

1:10:29How are you seeing the landscape around hardware evolve? Because that feels really, really important to you. if you don't want to go build it yourself, burn$100 million trying to build the next Apple product? Yeah, I don't actually think that we'll, I doubt that we'll ever be integrated into the iPhone. I don't actually know that that is the best modality, even if we could. I think most people, I mean, friend.com, Omi, like everyone is sort of betting on the idea that there will be this new hardware element that is like a companion. I think that's, if we were to delve into hardware, that would be most realistic, like the next step.

1:11:04But what I really would like to do is sort of like beat out Neuralink and just sort of just directly skip all that and get the chip inside your brain. That is the true end state is what we want anyways. Yeah, don't ask the Neuralink guys about timelines then. Can you talk about what the general, you know, while you're at Columbia, what the vibe was on campus. I'm assuming everybody's using ChatGPT, but do they know, like, do they have a sense? Like, is this, is there an impending - Tyler Cowen, like, has completely embraced AI in his, I believe, graduate level economics courses where he says, yeah, go have ChatGPT write the paper, then dig into it and teach me something from it.

1:11:45It seems like he has a very positive view on AI in education. What's been your experience? Yeah. And getting into that more, is there a general fear about AI on campus? Are people worried, oh, job displacement, things like that? Or are they just excited to leverage it to grow and be better at their jobs, etc.? There is definitely concern, but I think the most interesting thing here that most adults probably don't realize is exactly how many people are using AI. I mean, the percentages get even more skewed at the higher level schools. But I will say out of every single person I've met at Columbia campus, there is not a single undergraduate who hasn't cheated on at least one assignment using AI.

1:12:27And the vast majority of CS majors at like - You should tell them to kick everybody out. Kick everybody out. Kick everybody out. Just clear the slate. Yeah, whistleblow. Columbia whistleblower. Columbia cheating is okay. It's trolling that gets you in trouble. Yeah. Don't have fun. Don't expose technological changes. You will be blocked. Who are your, who are your inspirations? Are you like a Nathan Fielder type guy? I feel like there's someone that you look to in your, in your virality. Who inspires you? Bill Burr and Nathan Fielder. Those guys are, are, are so funny. Dave Chappelle. I mean, I was big in that.

1:13:07There's a lot of Asian content creators, college-wise, that are funny too. That's awesome. Last question, Jerry? No, I love it. Who are you? Have you put together a team of former classmates? Have you gotten anyone else to bail with you? What does your team look like today? Yeah. So over half the team is friends that I met from community college, actually. And the other people are my co-founder who dropped out. and also the guy who's probably going to be valedictorian of columbia there we go that's are you are you worried about any like serious backlash from columbia because i saw the thing where like they told you don't share this document you shared it it feels like like they find out you have five million bucks they sue you like i just want to go out on the lens i will i will i will give money to the roy lee defense so we got you we got you i i i'm anti-cheating in the literal sense but I like building and I also like trolling and comedy.

1:14:09So I'm like 50-50 on you, I guess. Why are you worried? I'm not as reckless as people think. I read the documents very carefully before disclosing the confidential and it was not legal. It's the worst case to get expelled. And for me, it's probably the best case. Yeah, yeah, exactly. I think I would be okay with you getting kicked out of Colombia, but I would be upset if they really tried to punish you seriously. I'm going to go out on a limb and I'm going to say, I believe there's a real chance that you eventually go back and give the graduation speech at Columbia. You know, you're going to have a, you're going to have a crazy arc and I'm, I'm excited to witness it.

1:14:46I hope so. I hope so. Yeah. Good luck. Love it. Hopefully there's not too many crazy biz dev deals between you and success, because I feel like if, if it comes down to it and it's like, in order to win, you have to partner with Apple, you might have to clean it up a little bit, but good luck to you. I'm sure you'll figure it out. And thank you so much for joining the stream. This is a very exciting story. And I love when anyone blows up the internet like you did. Yeah. Congratulations on all the momentum and the round and excited to have you back on soon. Yeah. Fantastic. Thank you guys. We'll talk to you later.

1:15:16Cheers. Bye. Later. Next up, we got Jacob from Superpower announcing a$30 million Series A, B, C, D. Series A. Series A. That's good. And also a Sousa Ventures. Back to back Sousa. Back to back. It's Sousa Day. It's Chad Day. Shout out. It's Chad. Yeah. Chad Day. He's easy. April 22nd. He's just - Will forever be known as Chad Day. I'm a Chad. I invest in Chad's. That's his investment thesis. The end. The end. We got Jacob here. Let's bring him in. Jacob, how you feeling? How you doing? Should we ring the size gone for you? Preemptively. Congratulations on the round. How's it going? What's the latest?

1:15:54Jordy, gentlemen. Great to see you guys. Great to see you. What's happening? Give us the breakdown. What are you announcing? And give us the high-level pitch for Superpower. $30 million Series A. It's our biggest capitalization to date. Capital is no longer a constraint, and we're building a super team to reinvent healthcare. Okay. Crazy. Give us a backstory. How does it start for the consumer? No, no, no. I was going to say let's go back a little bit. Yeah, let's start with the backstory. 2023, you had a rough year. You want to share that? Oh, yeah, that'd be interesting. How that kind of catalyzed Superpower?

1:16:29Rewinding the clock. It was 2022, actually. And that year, I almost lost my life to reverse health care incentives and a system that doesn't necessarily help people be proactive and take control. So that was when I was building my last company, I was diagnosed with an autoimmune disorder. It's one called Crohn's, which those might be might be familiar with. Fifty million Americans have autoimmune disease, two thirds of which are undiagnosed. And in my case, it led me to being hospitalized for close to four months, had multiple surgeries, lost part of my stomach, got stuck with a multimillion dollar bill.

1:17:01And you realize really quickly that, you know, the thing that health systems are designed to do from a business model perspective is make the most money. So that's billed pharmaceuticals or surgery versus getting at the root cause of like what's actually driving complex diseases in my case. So I was really, you know, more seen as a way for them to make a lucrative billing, you know, cash versus versus actually getting to the root of what was ailing me. Talk about, yeah, talk about how that led into Superpower. I remember we had, I think we had lunch just after that period. I remember you were like, it felt like you were almost still in a daze because you just like almost, you know, spent a year almost dying.

1:17:40but you knew from you know basically right at that moment kind of what you wanted to do it actually comes back to twitter so i posted about my hospital story and it went really it went mega viral and as a result i got connected to a handful of these like high-end concierge doctors who basically work with the tech billionaire class yeah and they charge like 50 to 100k for their clinics and what you do when you're a patient of these practices is they test everything in your body leave no stones unturned they'll like sit down with you on a whiteboard for four hours connect the dots across every little thing in your health, and then pair you with a full-time team to basically put your healthcare on autopilot.

1:18:15So if you're fortunate enough to be in a financial position or in the know to be a patient in one of these practices, you're basically never going to die of chronic disease. You're going to look like the memes of Bezos and Zuck that are super yoked. I know the doctors that probably do their peptide regimens. You think it's just peptides? You think it's just peptides or is there a secret juice? It's a lot of jiu-jitsu. too. I can't reveal too much. But you realize really quickly that there's a gap between healthcare for the best and healthcare for the rest. And obviously being a technologist and a brand builder, it just became very obvious that there was a big opportunity here to democratize once I went on my own healing journey.

1:18:51Do you think that founders should, like, I feel like there's this balance between founders need to be performant, you need to have high energy. You need to be able to oftentimes go 20 meetings in a single day, whatever that looks like. You basically have the life, you know, the intensity in many ways of a professional athlete. What do you think the sweet spot is for founders in terms of, or investors in terms of caring about their health, but not letting it sort of like take over their entire life? I went through a period in college when I had a bit more time where it felt like 60 % of my brain power, which is like going to lifting and eating and things like that.

1:19:32And obviously, it's not sustainable if you're trying to run a company and things like that. So I'm curious what you think the sweet spot is in terms of wanting to be high performance, wanting to be healthy, wanting to not be accelerating your aging, things like that. It's quite a quandary, right? Because we obviously pride ourselves on being a team. We even joke that our office is the world's healthiest office, yet sometimes we still have to put in the hours and we have the engineering team closing up on the floor for the evening um but uh yeah it really i think is an opportunity to usher in a new paradigm shift where we can put health at the forefront of the conversation you know i think we're actually about to have our first board meeting here uh this upcoming quarter and we plan to have some statistics on the health of our founders and our team in our in our board so we want to kind of set the lead the vanguard and what it what it looks like to to have health be like a tracked foundation just as you would you know your amplitude or your stripe data or all the metrics that matter for business yeah i was joking john and i uh gave a uh sort of humorous talk in miami last year and we were talking about the case for like vc platform teams should just be like basically like doping yeah Like helping their portfolios.

1:20:50What's your creatine style? What's your creatine? How much are you working out? What's your eight sleep score? Looks like we need to add a little bit of tea in the mix. I think someone's going to launch a fund with this thesis. Maybe it'll be me. I don't know. Unironically? No, I think it's a very real. I joked about this, but I think it's very real. A lot of venture funds, especially if you're niche, you need some type of edge. Some people are like, oh, we're going to help you with sales. We'll help you with recruiting. will help you with your next round. But just saying like, we're going to help you with your one rep max.

1:21:21Help you with your one rep max. Exactly. Thousand pound ventures. Talk about the sort of opportunity in AI specifically for superpower. You guys have started with biomarkers. How do you plan to leverage that to kind of unlock value on the data side and all that? So in the limit, something that we deeply believe is everyone will have a healthcare super app on their phone. And today, a consumer health experience is deeply fragmented. You know, consumers have to run around town to a bunch of different places, and it's all disconnected. And the front door to health is, you know, something like ChatGPT or Google or WebMD.

1:22:04But the problem with those platforms is they don't really know much about you, and they definitely don't know everything about medicine. So what that sort of culminates in with what we're building is what we think will basically be an AI doctor in everyone's pocket. So we're kind of aggregating all of your medical records and health data, making it super easy to test your whole body and combining that with all the world's medical knowledge, which does not necessarily exist in foundation models today. Each foundation model is sort of trained on a select aspect of the healthcare and medical literature universe.

1:22:38So there's a lot of creative ways to get a full picture on what's actually happening at the edge of science and medicine. When you put all this together, you have a recipe for a really unique company that we think one day will be something that a large majority of Americans own. And healthcare is ripe and it's demanding of better consumer experience. We've seen the consumer experience being reinvented for every other aspect of our lives, but healthcare sort of the last domino to fall. And we want to be basically the predominant company to bring this to the market. How do you think about the interaction with LabCorp?

1:23:12It's an$18 billion company, kind of, you know, stocks up and stocks down, like kind of, you know, up 60 % over the last five years, obviously. When you're doing lab testing, like you kind of got to build on top of like the railroad that's already there. But how does that relationship evolve over time? yeah labs are a super commoditized business what's really happening what's exciting to pay attention to is innovation in testing so today it's actually kind of a cumbersome process you have to have like a nurse come to your house you have to go to a lab corp they test your blood they send it off to another facility it takes like you know eight or nine tubes to get a full enough panel and picture and then a week to come back uh and there's probably like a dozen or so friction points there so something like theranos over the next few probably is not that far off and thankfully we're building at the application layer so we have to be best in the world at building a trustworthy brand uh low-cost customer acquisition and a really amazing ai doctor product and what what is your takeaway from the theranos story do you think that there's any like sort of misread on that historical uh like anecdote like we've we've done like a deep dive and we were like there's more nuance the issue is don't apply build fast and break things to consumer health yep healthcare doesn't move at the speed of code.

1:24:30There's a certain reverence you have to have for the human body as, as a healthcare founder, which makes the, you know, the Zuck adage a bit tougher to apply. Yeah. What about the, what about interactions with the FDA? I assume that there's oversight around, you know, you get the lab markers done with LabCorp or Quest or something, and then you're interpreting that and you're at the application layer doing AI data analysis, all sorts of good stuff. But at a certain point, actually making a recommendation about someone's health is probably regulated. What does that look like and how will it evolve over time?

1:25:05Yes. We're not quite in a world yet where AI and the algorithm can make a medical diagnosis or a recommendation. So the system that we've architected and the paradigm that we're in just as an industry is basically AI with human. So anytime it gets to a point where a human has to intervene, we do plug in a doctor on that, which helps us avoid any sort of messy FDA regulation. That makes sense. Do you see a world in the future where the average company, maybe outside of our tech bubble, is like giving budget to employees specifically for preventative health, right? Like sort of health insurance broadly is a pretty standard benefit, but feels like preventative health is like, you know, potentially the next place that employers want to invest?

1:25:54Undoubtedly. So one of the challenges in health care today is because the average American is transitioning jobs every two to three years, and that health care is tethered to your employer, that means it's just a game of bag passing, where the insurers don't really have an incentive to underwrite anything that's a bit more long term or preventative or optimization focused, like getting your testosterone checked, checked out. So in that world, the insurers don't necessarily They want you to get access to things like superpower and have it be covered. So that's where employers will ultimately step in and potentially start to cover these types of things as supplemental benefits because they're going to drive clear employee retention acquisition in a market where it's only more competitive to get the best people.

1:26:35What's an underrated supplement right now? It feels like magnesium. Magnesium's hot. Atiyah takes six different types of magnesium. I take four. I also take six different types of magnesium. It takes six different types. I'm only on four. You're only on four. Those are rookie numbers. What do you think the next kind of magnesium? I've been surprised to see creatine in the timeline so much just because it seems like something that - It's very Lindy. It's Lindy. People have been taking it forever. Methylene blue is in the timeline. Methylene blue, I'm 50-50. Yeah, yeah, yeah. I don't know about that.

1:27:07But what's your take, Jacob? I got one for you. Oral BPC-157. I've been on oral BPC-157. This is the Wolverine peptide. Oh, yeah? 40 tip. It's from the Wolf-Loreen political. Yeah. No, this is what it's sort of, is it semi, like there's an interesting thing right now where there's like substances that are banned by various like sports organizations and leagues that as a CEO, you can just take. So like if a pro athlete is like wanting to take something but can't because it's very effective but it's so effective, it's maybe made illegal. If you're a CEO or a capital allocator, why not get on some PPC 157?

1:27:51Interesting. In Silicon Valley, performance-enhancing drugs are encouraged. Yep, that's true. Yeah, it's an interesting time right now. I mean, it feels like these sort of psychedelics have long been a part of Silicon Valley culture, but I think that performance drugs, not just exogenous testosterone, but I think they will just become more and more popular where it's going to move beyond, okay, founders are on, maybe they're on caffeine or nicotine. and creatine, but on onto the like really optimizing, you know, peptides in the way that, you know, you mentioned it, people like Bezos are probably doing to some degree.

1:28:27Better sleep. Yeah. Makes no sense. But anyways, great to have you on, Jacob. Very, very exciting. Congratulations. Anything else? Anything you want to plug before we end? gentlemen i uh i appreciate the the time but maybe before i leave i'll give uh people the quick sneak preview of our ai doctor product that uh we're launching very shortly so jordy if you have any questions for the algorithm maybe how to boost that testosterone there we go what's the yeah how much bpc 157 can i take before i explode what is the ld50 of bpc 157 because i'm going to the max i want to look exactly like wolverine john has the john has the highest natural t levels of i think uh yeah but imagine how much higher they could be if i was on the wolverine stack if john if john 2x his t it would be potentially a world record so uh i mean obviously foundation model plugged into that but then also built on top of your own data so you can query your own lab results and get customized uh recommendations that the idea exactly and then as soon as it as soon as it like makes a firm recommendation kicks you over to a doctor before it starts violating fda rules basically exactly the way to think about it is the world's best doctors will spend hours and hours with you on web boards connecting all the dots in your health but that computation doesn't scale at a at a higher price point so we're democratizing it that's awesome love to see it love it Awesome, Jacob.

1:30:00Congrats to you and the whole team. Cheers. Well, speaking of AI, we're having Logan Kilpatrick from Google on. There we go. I'm very excited for this one. A lot of different stuff. Logan has been on a tear. He has. Great poster. And as soon as he is here, we will bring him into the studio. But maybe in the meantime, we'll tell you about Bezel. Go to getbezel.com. Your Bezel concierge is available to source any watch on the planet for you. seriously any watch and i mean well we're also on it we did not get a chance to talk about eight sleep go to eightsleep.com slash tbpn uh get a pod for ultra it has a five-year warranty through 30 night risk-free trial free returns and free shipping and in and we will go back to the show now let's bring logan into the studio how you doing logan let's see we're bringing it in how you doing?

1:30:53I'm doing great. It's been a great busy last six months of AI stuff. So I'm trying to stay alive. Just the last six months. I feel like every six days is huge in the AI world. It really is like the best place to do content around or just read about or listen to podcasts about like the AI world is just so fertile, regardless of what you think about like PDoom and acceleration and all that, just in terms of like the applications, the deals that are getting done. It's fascinating. So yeah, what are you watching today? What are you watching this week? What's most interesting? Yeah, that's a good question.

1:31:28I think continued momentum of 2.5 Pro on the Gemini side. I think obviously a bunch of new open AI models, which has been awesome to see. It's also been back to your point about how much fertility there is and like different AI stuff. I think if you look like two years ago, it was really just the models and like the model cadence of launching was like actually quite slow relative to like products. And now we have all these products, which people are really excited about. And like the product innovation actually happens a lot faster. So it's like acceleration across the model category, but also across all the product category.

1:31:59And like, there's just, there's too much to keep up with at this point. It's impossible. Are you generally feeling the acceleration? Because I feel like we are seeing acceleration on the product side and certainly in the fragmentation of models and the specialization of these models. But in terms of just like massive order of magnitude breakthroughs, I feel like when And we went from three, GPT-3 to GPT-4. That was a huge leap. We passed the touring test. ChatGPT was huge. And then since then, it's been more incremental. Extremely valuable. Extremely great. I love it. But I haven't seen as many of those viral moments.

1:32:36Studio Ghibli accepted probably. Yeah, I actually think I'll push you on this point, which is I think we've had more of those large moments than I think people appreciate. And I think it's just like, it's actually hard to appreciate some of those moments if there isn't a product experience that brings it to life. And I actually think that's been a lot of the gap. It's like if you look at like multimodal, like the fact that the models can like with better, better than just from a multimodal input perspective, better than humans are at like most vision tasks. Like the number of products and like things that that unlocks is like truly mind blowing.

1:33:09If you look at two years ago, what you would have had to do to make something work in that ecosystem. and like that is from an order of magnitude of impact like massive maybe it's not as big as like text but it's still this like huge massive and like we're continuing to see that with like you know now the models are really good at tools and now the models can actually generate images and audio and video and all this stuff so I think we're getting those I also think our expectations have just gone up so much that we're like if the thing isn't you know brushing my teeth for me then my call it's not it's no longer yeah yesterday we read a post from uh near cyan and uh it was a it was a screenshot of the definition of AGI read it out so it's a reminder of how far AGI goalposts have moved and and it's a screenshot it says an AGI could beat you at chess tell you a story bake you a cake describe a sheep and name three things larger than a lobster and it's funny because like okay everything we're doing everything except baking you a cake yeah and And even then, it'll give you a great recipe, and it'll give you a recipe for any type of cake you could possibly imagine.

1:34:12A sheep-sized cake. It can do anything. Anyway, I want to talk about this idea of the Pareto frontier in AI models. Sean, previous guest on the show, mentioned that Google has been very good at delivering not just high-quality models, but affordable models. Has that been a deliberate strategy to try and create the best model at every different price point? Or is that just a natural outgrowth of the engineering culture and the scale of being a hyperscaler? Yeah, that's a great question. I think a lot of this has been an intentional decision. I think if you look at specifically 1.5, the Flash series is really where we landed this point the most.

1:34:51I think recently with 2.5 Pro, it's been the first time that we've actually had truly one of the most intelligent models available. And relative from a cost perspective, it's still super affordable. but I think it also just like why we've been able to do that and why we focus on it is like goes back to like you know Google controls from a product perspective like all the way to how the models are delivered to how the models are trained down to the silicon so like you can make decisions assuming a bunch of those things are going to be true which is like a lot of folks don't have the flexibility to make those decisions and and the beautiful thing is like I think this this point goes underscored, which is like, what does this mean?

1:35:29It means that builders have the freedom to do stuff. It's like, not that like Google has this really great advantage. And the, you know, thing we do with the advantages, find out how to milk money out of people. It's like, we have this advantage. And what does it mean? It means the world gets cheaper AI models and they get to go and build the products. And like the margin for people building AI products actually goes up every, like the farther you push the Pareto frontier, the more money builders get to make, which is like such a interesting and like unique thing about this AI moment that I actually don't think has been true in a lot of these other platform shifts that have happened in the past.

1:35:59Yeah, I want to dig in there. Google's in a unique position in that it's a consumer tech giant, but also a scaled infrastructure provider with GCP. And you could see... Not to mention a B2B player. I mean, I don't know a single startup that doesn't run on Google 100%. And so you, like Ben Thompson was just talking yesterday about, he's really pushing OpenAI to just go full consumer and let Microsoft handle all of the B2B stuff. Don't let the API load take away from serving your consumers. With Google, there's probably some sort of tension there, but what are you excited about on the B2B side and this idea of like, oh, if I build a wrapper, is Google going to just steamroll me?

1:36:50This is kind of like an old meme, but now it seems like the best time ever to build on top of the 2.5 or any of these different models that have great cost and scores and benchmarks and whatnot. So what message are you sending to kind of the developer community? Yeah, so two things. One, our reaction to your first comment, which is around like, is there value in doing both consumer enterprise and some of these other things? I think my push for companies, like if you're in the position to be able to do that, and like there's a bunch of nuance to this, because obviously OpenAI as an example is like extremely well capitalized, et cetera, et cetera.

1:37:27Like they have the means to do this and do it well. But for companies that can, I think part of my core worldview is a lot of the reason that ChatGPT has been as successful as it has is because there was an API business built around it. And if you think about it through this lens of what the API business did was like, allow the world, allow this like massive proliferation of AI products to educate the masses that people are interested in this stuff. And then like at the end of the day, who has the world's best AI product? I think like, you know, some people maybe argue it's ChatGPT, maybe it's another product, but like you sort of wet the appetite of the world that like, oh, I can actually do these things and it's, you know, in my product surface or et cetera, et cetera.

1:38:06And all that's enabled by the API. And then when you look for like, okay, what's the best way for me to use this product? Maybe it's one of those consumer products made by one of the larger labs. And I think that that playbook actually works a lot. And that's why I'm extremely bullish for people who are sort of doing both of these things. I think to your question about like where's the value for builders today, I 100 % agree with you. I think there's so much value to be created at the application layer. If you look at, I think about these like three curves at the same time. There's like on one hand the cost of AI going down into the right.

1:38:37You know, cost of AI down 99 % over the last two years. The intelligence of AI is up into the right. Like models continue to get smarter and better with test time compute and scaling and stuff. at the same time that the models are getting smarter, the costs are going down, consumer understanding of AI is going up. And then in parallel to that, as consumer understanding is going up, as the models are getting better and cheaper, consumer willingness to pay for AI products is also going up. And I think there's this, like, it's this beautiful, like, you actually could not ask for a better set of four lines on a graph than those four things if you're building a product.

1:39:12It's cheaper for you to build a product. Your customers are willing to pay more. there's more customers and the tool that's actually enabling the value creation is just getting better for free. Like you don't have to do anything. It just gets better for you. And like, I don't think there's been a time in human history where for builders, all four of those things have happened at the same time. So I'm super excited. Like I literally wake up every day excited because people are building companies and all this is happening for them. How do you think about balancing benchmarks and capabilities, right?

1:39:46Google's consistently been a leader, yet at the same time, every consumer at this point has experienced a new model coming out, performing well on benchmarks, just sort of like broadly, not talking about any one lab, but then being sort of disappointed with the actual experience with the model. And I'm sure you've spent a lot of time thinking about this yeah this is one i think for folks who haven't thought about this um evals are just really hard it's like such a hard problem and actually like if you abstract evals out of like ai and into everyday life like you become surprised maybe i'm like too eval pilled at this point but i think about just like random problems in life are truly eval problems and like we look at them as this like very human thing um and actually they're evil and like the the sort of really weak version of this is if anyone's had a job and like have gone through like performance review, like performance reviews are an eval.

1:40:41Like, are they a super scientific eval? Like, no, they're actually a pretty crabby eval in a lot of ways if folks have gone through performance reviews. And yet like that is the core foundation of like how human, you know, career growth goes in a lot of ways. Like there's a lot, like evals are generally a really hard problem. For AI, they're also extremely hard. And I think to answer the question specifically, Jordi, about what does that mean for people who have this disconnect between capabilities and what the evals say. I think this goes to why there's so many vibe evals. And if you look at Ella Marina as a good example of this, Ella Marina is basically capturing vibes.

1:41:19It's like, how do people feel about this thing? It's not scientific. They're not actually evaling that the model is saying things that are true. It's like, how do humans feel about this response? And I think that's incredibly important. Is that the only thing that matters? Certainly not. But I think more and more you see long-term happening where like the vibes are really important in addition to the actual quality of the models being important. And I do think you could do both of these things. In some cases there's, there's tension, but I think you have to do both if you want to be successful.

1:41:48Totally. How do you think about the different buckets of foundational research that are happening? Uh, do you believe in the, the data wall, the pre-training scaling rule, like kind of hitting diminishing marginal returns. We've talked to a lot of folks who are extremely excited about reinforcement learning, reasoning, going a lot further there, program synthesis, these kind of topics. Even just more tool use. Let's bring more tools in. Google has a million tools and a million interesting APIs, both internally and some externally. It feels like there's a lot of low-hanging fruit there. but what excites you on the research side these days is just let's build an even bigger data center you guys probably already have the biggest ones in the world but you could obviously go bigger or or is it more algorithm based how are you thinking about the future of the foundation model landscape developing yeah i think two points one i think there's never been more opportunities to push the frontier like i think all those examples that you just described like because the models are evolving more than just being models and like they're actually there's like systems that have tools and all this other stuff together um i think it it just increases the number of opportunities for scale and for the models to get better so i think that's like one of the positive things we'll continue to see a lot of growth specifically because there's so many dimensions that you can make the models better at um but i think if you look at like a an example of this in practice like 2.5 pro is actually an example where it wasn't just like rl scaling that made that model better um yes rl was part of the story but like there was also a bunch of pre-training innovation and principally there was a ton of um of post-training and post-training and pre-training innovation so i think like it's not that um companies are going to continue to see value created across all of those and the really magical thing is like and this is why like i don't i don't subscribe to the like pre-training is you know dead and all that stuff because the the more work that you can do at the pre-training level those capabilities as you do post-training and as to give the models RL capability.

1:43:53It's like the capability is like amplified almost exponentially. So like if you can make the model 3 % better at the post-training level, when you actually finally do all the reasoning work and the model like has to reason through really complicated problems, it's like many multiples of bang for your buck because of that. So like, I think we need to continue to do the innovation across all levels. And like, that's exactly what we're doing right now. I know you probably can't talk about specifics, but one of my favorite Google stories is that crazy anecdote about the V8 JavaScript engine. I think it might be apocryphal, but it's like a bunch of Google engineers go out to like Iceland or something, spend like a month building a new JavaScript runtime.

1:44:37It builds Node and it winds up being like the foundation of like the Chrome browser. I'm interested to hear, what is your take on just lower level optimization? Obviously, Google's already doing the TPU, but in terms of just maybe even just squeezing extra performance out of inference, we saw this with DeepSeq. It seemed like they did not just one or two breakthroughs in terms of kind of cost performance optimizations, but they did a ton. is that type of work happening at Google? Do you want to see more of it? Is it an exciting area? Or is it just like, oh, yeah, that's just something that we're going to have to do.

1:45:17It's going to have to happen at some point, but it's not as critical as some of the other stuff that's happening in the industry right now. Yeah, you should ask our inference engineers who are working like 24 hours a day. I think 2.5 Pro has actually been a fundamental example of this. All of this stuff matters. I think if you're not doing it, especially with large larger models and especially with models that have lots of demand, like there's no world where you can get away with not putting a large order of magnitude of investment into, into inference. And like, I think credit to our, our team who's like actually working around the clock right now to make it so that people can keep scaling with 2.5 pro because there's so much demand and we're having to like, you know, as a, uh, an artifact of the constraints that we're under, like we're having to innovate and like solve new problems and come up with things to like find ways to make 2.5 pro more scalable from an inference perspective.

1:46:06So it's, uh, it's awesome to watch that happen. Yeah. Where, where do you want to see more, you know, knowing, uh, 2.5 pros, you know, capabilities probably better than anyone else. Where do you want to see more developer activity? Yeah. I think the thread right now that has the most excitement is around coding just because like developers love coding. There's like so much the whole vibe coding thing is is a real phenomenon um i i think i continue to be interested in like all the multimodal stuff like there's just so many products and like i think back to early in my career when i trained computer vision models and like deployed them and like the amount of time that it took and the amount of resources to do that relative to it to today where you can literally just write a prompt and send images or videos to the model and have it do those tasks like with basically, you know, near or better accuracy than you would get from domain specific models is absolutely fascinating.

1:47:05Like, I think we're, we haven't actually seen that wave of like multimodal startups that are like building on top of this stuff. And I think that includes like audio things. Like I think the audio ecosystem is like still pretty nascent. Like I think the foundation is being laid for that. Like, you know, people saw this with Gemini Live, people saw this with the ChatGPT version of the product that came out, but I don't think we've seen like across other product services, people actually invest in like real-time audio and real-time video and image stuff. And I think that's like the next iteration of the UX of how people are going to interact with AI models.

1:47:38And the foundation is all there. It's just like the lag of how long it takes people to build interesting products. Like it just takes like 12 months or something like that for that to happen. Yeah. I mean, speaking of like building new products, I love the Paul Buchheit story of building Gmail. Is this like almost like April fools project. And then you see a, you see a seed of that in the notebook LM project. What's the culture like around the idea of like 20 % time does that even exist anymore? And then like, if I joined Google, could I just go and say like, Hey, I'm going to go like build an AI native email client.

1:48:15It might destroy Gmail, but you know, we'll figure it out. Or is that something that like, you know, we're going to need to think about cause like Gmail's mature, but there's still like a, you know, right now Gemini is like kind of being vended into Gmail, but you know, maybe there's an entirely new paradigm at some point. How do these like side projects and 20 % time work at Google these days? Yeah, that's a great question. I think my, my 200 % project is, is on all the Gemini developer stuff. So I do think people are doing 20 % projects, which I think is great. Like, I think if you're, if you have freedom to do that, like you should, and that's how Google's going to come up with innovation.

1:48:53I think for Notebook LM specifically, it came out of Google Labs. And like Google Labs, the whole point is come up with new product services. So not just Notebook LM actually, but there's like a whole like WISC, if folks have seen WISC, it's like a video image generation platform that came out of that. Actually AI Studio, the product that I work on, came out of Labs originally and Josh Woodward's team. So I think there is like a whole lot of innovation that's being seeded out of that group. And again, it's like the big, bold bets that you wouldn't see coming out of potentially other product areas because it just takes time to build these products from scratch and oftentimes time that they don't have.

1:49:32But I'm super happy because Josh Woodward, who leads the Google Labs team, now also leads the Gemini app team. So I think we're going to see that like fusion of all of these like new ideas and product spaces because him and his team sit so close to like the blank slate. Let's build any product to solve problems that people have. Now that they also run the Gemini app, I think we're going to see like a ton of explosion. You should have Josh on. He's one of my favorite people. Yeah, we'd love to. How do you think about getting, helping, how do you think about almost like prompting new consumer behaviors?

1:50:05Because I feel like every day on X, you'll see somebody that's like, you know, I saw somebody post, oh, I'm doing, you know, basically like run it, you know, doing a prompt every day about like a specific industry. Like, you know, pull together, you know, the most important headlines and news from this industry. And I'm like, oh, that's really valuable to what we're doing. We should just sort of like get into the office in the morning, adopt that and do that. And that's something I wasn't really thinking about. People are so used to this sort of idea of just getting...

1:50:36They'll adopt a model for one specific thing, and then they just do that. Maybe they veer out of it a little bit. But how do you think about getting users to just be more creative? Obviously, it's helpful when a developer builds an application to prompt a new consumer behavior. But how do you think about that kind of at the application layer at Google? Yeah, I actually, my personal take on this is I think this is a bug of this current AI moment where like the, if you look at like, what is the ideal case? The ideal case is the models and the products like pull out what they need to from the user in order to create value for them.

1:51:14And I think if you look at like how all AI products basically work today, you as the user, the burden is on you to create value with this tool. And I think that just like one inherently like limits, you know, the number of people who are going to get value. But two is also it's like a shitty experience, honestly. Like I hate like this is my biggest problem with AI tools, which is like you have to go for most AI tools. you have to go to and like make this like sizable investment in order to get anything out of this. I think like a couple notable exceptions to this is like deep research. I can just fire off the like random question that I have about how something works.

1:51:51And then I'm given a 50 page research report inside the Gemini app. And then I can one click turn that into an audio, like a notebook LM audio overview. And like, that's great because like as a user, I don't need to make this like large order of magnitude of investment. The model does all the work for me. And I think as we see more experiences like that, like I'm super excited to like, I'm almost, and this is maybe too like absolutist of like a product perspective, but like, I will not build a product that like, we have to try to convince a consumer to change their behavior. because I think actually the promise of AI is that these tools are going to be able to like pull this context out of you and you should just go and build that product.

1:52:27And like, I think the models are actually good enough in a lot of ways to help you do that today without having to like rely on a user, hopefully having the right behavior for to make that product successful. It's an interesting thing where I feel like users are so trained on like software doing a specific thing that this dynamic of software and the sort of application being able to act like a smart friend or a smart coworker that knows infinite more information about a bunch of different subjects and then can help you accomplish tasks is very different than this sort of, and I just feel like even as, you know, somebody who's, you know, I've now, I'm not 30 yet, but I've basically spent 20 years using software that behaves in a very specific way.

1:53:19And I need to just like totally reimagine how to use software. I think the push needs to be though, that like, you don't have to reimagine. Like, I think about this all the time. Like the thing that I would love is the AI tools that I have today interact with me. Like I already interact with software today, like shoot me a text, shoot me an email, call me on the phone. Like I'm already doing that all day. Like you could, there are ways I think to bridge that experience gap where like you don't need to convince, you don't need to go into a new flow. And I actually think if we've, if we fast forward like 10 years, I do think there's going to be a lot of those experiences, which like look eerily similar to the way that they do today, because it's just like so ingrained in like human culture, like how like texting is a great example of this.

1:53:59Yeah. Yeah. Like I want a push notification from Gemini that says, Hey, you're talking with Logan later. Be sure to ask him about this funny story. And then it's like, boom. And I didn't even have to, I didn't have to. Everything in AI, we're kind of reinventing from first principles. Like even just like the idea of like the cron job, we're like adding that back in. And now like, there'll be like a whole news cycle from like, Oh, like the AI apps got cron jobs this week. Like that's incredible. And it's like, yeah, but this has been around for a long time, but it really does transform the experience.

1:54:28I mean, speaking of, I want to know more about some of your specific workflows that you're enjoying. You mentioned deep research into Notebook LM. Is that something that you're able to do within AIstudio.google.com and run it all there? Or is there like a copy paste step? Because I've had to do that before where I've been like, okay, I got a deep research report. It's not reading it to me here, at least in ChatGPT. I need to take it over to Speechify and get it to read it to me there. But obviously, AI Studio is a little bit more prosumer, I feel like. I mean, the temperature is there. There's still some buttons that are almost like developer terminology.

1:55:08So walk me through some of your favorite AI use cases in AI Studio and what you're getting the most value out of so people can just kind of copy your prompts. Yeah, John, I think this is actually a great reminder for folks. AI Studio is a developer products like so we're building it for developers and like the use case that we're trying to build for is like showcase all the models raw capabilities so that you understand what the models are capable of so that you can go and build great products yourself like we're actually not so like deep research as like a great example of this like deep research is built on top of a bunch of capabilities that the model has which is like native search functionality tool calling etc etc um and that's available in the gemini app so if you're you know you want the like polished consumer experience or even prosumer experience honestly like the gemini app has that functionality it has audio reviews it's like a fully baked product ai studio is like give you the rawest possible experience and like some consumer ai enthusiasts like that experience which is why they come to ai studio but like generally we're trying to showcase like frontier capabilities show you the art of the possible so that you can go and build the products that you really like.

1:56:14So I do spend a bunch of my time from like a, like doing work perspective inside of the Gemini app. You know, Canvas is another one of them, like vibe coding inside of Canvas, inside the Gemini app is a lot of fun. Someone needs to vibe code a Google reader. This would be the most viral thing ever. I don't know. Are you familiar with Google reader? Probably before my time. It was like an RSS reader and Google shut it down and like, it had like a thousand true fans and so they were like so upset about it i i was using it was fun but uh you know i understood that they didn't need it but uh anyway we'll see what we can do maybe we'll maybe we'll bring google reader back as a vibe coded that would be that you would destroy the internet that would be the really the greatest marketing for for everything that you're doing if you could bring back google reader um anyway sorry sorry you're i i want to i want to let him finish about his uh his his like most interesting like ai use cases and what's uh what's fun and what's in your what's in your everyday carry in the ai world yeah i think the only other one that i'll mention is in a this is something that's specific in ai studio right now and we'll hopefully have it in the gemini app and other places is we have this like live api or this real-time mode where you can go and screen share and share and like show this the model what you see on your screen i think this gets to like all the points that i was getting at before about like why is this such a magical experience?

1:57:33It's because right now, the product experience of using AI is I need to go and find all the context that's relevant for the model and get it into this text box somewhere or get it into this list of files somewhere. And the beautiful thing about screen sharing is all of the context that I need the model to do stuff with is already on my computer. I'm looking at it somewhere on my computer. I looked at it today or yesterday or right now. Just let me show the model what I see and have it do interesting stuff. So I've been playing around with a bunch of pair programming examples like that and just like generally like critiquing work that I've done and having the model sort of watching and with my permission able to see the things that I see and talk to it is a super cool like very futuristic feeling experience.

1:58:13That's awesome. Last question switching gears a little bit do you have any you know ignoring your work at Google and on Gemini do you have any takes on AI hardware? Do we need new hardware devices? Is it an area that you're excited about? Or how do you think about that generally? Yeah, I think on both sides from an enterprise and consumer side, I think there's a lot of opportunity. And I think in the platform shift, it makes sense to try to build something. I think, does it end up working? I don't have the crystal ball. But I think on the enterprise side, the opportunity is you go in and build hardware that makes LLMs a lot faster and more efficient.

1:58:54And I think that's warranted and somebody should do that work. And then on the even, yeah, consumer side, yeah, I'm curious. We had the founder of Cluely, which has been going viral on before you, and he was talking about how the end state is just being embedded into the brain directly. But yeah, I mean, You can feel free to pass on this question, but the founder of Cluely, he applied for a job at Amazon. He cheated on the LeetCode questions. If he were applying to be on your team, you caught him cheating, what's the punishment? Is he on the team or is he out forever? I think we've actually been looking at doing AI-assisted interviews and not AI-assisted interviews.

1:59:44I think the world needs both right now. It makes perfect sense to evaluate both. Yeah, yeah. Kind of interesting. I mean, he was doing it as like a publicity stunt, basically trolling. Obviously, I don't endorse actually cheating on real interviews. But it is interesting that like you will have to adapt just like the teacher will have to adapt that was previously assigning research papers that can be one shot by chat GPT. No, it's an interesting way to test how somebody works individually. Totally. Right? Single player. And thinking about AI assisted as a sort of like multiplayer experience. How do you collaborate with other people?

2:00:14Right? And that's like a really good eval, as you said. And in a world where the tools like actually make a difference in how much you can do, like, can you use the tools? I think that's like a fundamental question that I don't think a lot of people ask in these job interviews. Like, are you AI assisted in what you're doing today? And if you're not like, you know, it's a, there's, there's a Delta in your output. If you are AI assisted versus not across coding across every discipline right now. Totally. Awesome. This has been awesome. Come back. I'm sure you're going to have many big announcements this year.

2:00:43Always welcome to come on and jam with us. Thank you for making the time. Bye guys.

2:00:52We got Eric Torenberg coming in the studio in just a minute. uh he's uh injuries and horowitz's latest uh general partner gp uh good friend of us good friend of the show uh known him for years and uh excited to dig into his new role and uh do a little some personnel news segment we got a massive trade deal uh you know first round draft pick many times in his career max contract probably four-year deal you know one-year clip who knows Great to have you. Good to have you here. What's going on? Beautiful logo in the back. Welcome to the stream. How are you doing, Eric? Thank you. A long-time listener, first-time caller.

2:01:31Thanks for having me on, guys. Huge fan. Great to see you. Take us through the anatomy of the deal. Did they sit down with your parents, tell you, hey, we're taking them to the big leagues? How did it come together? I mean, I imagine you're close to a lot of folks over there for a long time. But is there anything you can share about how someone becomes a general partner at a storied venture capital firm? They brought a briefcase to my parents' house. Yeah. You know, just been working the family for years. Yeah. I love it. I love it. But yeah, I mean, seriously, I mean, you've known the team for a long time.

2:02:05What were the conversations like and how did it all come together? It's funny. part of me wants to ask like, why didn't you ask me this a few years ago? Why didn't they try to recruit me a few years ago? But it's kind of like when someone amazing asks you out, don't ask too many questions. Just say yes. But basically what happened was, so Mark and I have been close for a long time. And every year I basically, you know, sort of give him a state of the union on my career and how I'm thinking about things. And this year I was like, hey, I know I want to marry venture into what I'm doing. I have this media company.

2:02:44We've got sort of this founder social network thing. And everything around my career has been either investing or building communities, networks, media for founders. And so I was just like, I want to do it again, but I'm figuring out what was the right structure, how to do that. I can just start something, I can join something early. And then Mark's like, why don't you just do this at A, 16, Z? And I was like, I hadn't really considered it because such a big firm, you know, so specialized, I'm a generalist investor, you know, a more talent driven early stage. And he and I think like what I fit in and, you know, 700 % firm like, like, and he's like, let's, let's talk about it with Ben.

2:03:24So then me, Ben and Mark have a bunch of conversations, Catherine's involved to the other partners get involved. And basically, they had had a desire to go much bigger on content, like separate from me, and they kind of liked what I was up to at Terpeton with sort of this network approach. And they crafted the right dual role. That's a mix of investing and a mix of kind of media network stuff. And that's how over a few month period, we just kept falling in love with each other even more, in love with the opportunity. And so here I am. No, there's this thing in venture, specifically investors, they spend all their time working with founders and they, and they, and they think I need to be a founder.

2:04:06Like I have to start my own firm. And then you realize like, no, you can potentially have like a thousand times more impact if you join a platform and make it better and leverage all of those resources. So I think it makes a ton of sense. Um, I have, I want to go back like way, way, uh, in your early days, how did you, how did you get so good at networking? Uh, I think it's like, uh, uh, you're probably like, you know, the best, the best to ever do it. Michael Jordan of networking. Uh, and yeah, I don't, I don't say that lightly. It's, it's, you know, it's extremely impressive. Uh, and I'm curious, like, you know, if you identified early that you were just good at it or it just felt like riding a bike and you didn't, you know, it just was totally natural.

2:04:53Yeah. First of all, want to say i appreciate the compliment uh and no one who's good at it wants to be known as the thing that they want it to be like like delian's got a good network but it's like a byproduct of the stuff and so yeah we uh i would say hopefully it's similar to i mean i think the the way i realized it i was doing this music tech company out of college wrapped fm it was kind of a retarded idea it was like chat roulette for rap battles it yeah it made no sense as a as a business But it was just kind of a out of college project. We somehow raised the money for it This is when twitch twitch was getting off the ground people like hey twitch for music Maybe that'll work and live just the TV live video, etc.

2:05:34And I got People noticed that I had the skill for evangelism like getting people excited about things attracting talented people You know what a great founders do they attract capital and labor, right? And so I was I had some skill for that even though the idea was dumb and people were like, hey you have something here. And then, but I didn't put it together until product time, which is the thing I joined next where like, and you know, Ryan was CEO and he really got that off the ground, but I helped him sort of like bring that out of thin air, like, you know, create a strong community and network.

2:06:06And in that period, I was like, this is my superpower. I like up until that point, up until age 24, 25, I hadn't identified a superpower that I had. And I was just like, I'm just going to triple down on this. And so then starting on deck and village. And what is it? I think it's just like a desire to like an obsession with people. Like who is everybody? What are they doing? But then also desire to help them and think about sort of people dynamics. And but then also this, that's like highest level. But then the next level is like, build communities and networks that sort of help each other where people can help each other, like even when you're sleeping.

2:06:44And so That's what I spent a lot of time thinking about and doing. I want to talk about media, but first, let's talk about some of the industries that you're interested in on the investing side. Andreessen has a fund for every vertical now, bio, gaming, defense, crypto enterprise. What's interesting to you? Is there any crossover? We have this take that media can be fun and lucrative, but media is really hard to invest in. And if you want to invest in media, you've got to be on the platform side. You've got to be in Spotify, TikTok, Instagram, Twitter, Facebook. Like those were the power law outcomes in media.

2:07:23People don't really think about them as media. But if you were in YouTube early, you did very well. Very hard to invest in companies that are more in the middle. What are you seeing in terms of, are you excited about any investments on the media side? And then also, what are you excited about in the other sectors of the Andreessen portfolio? Leo? So I, I think Jessica Lesson really sort of helped popularize sort of the bootstrapped media company that, that Turpatine has followed. I think you guys are following. I think it just, it doesn't make sense for, for this kind of business, for the reasons that you're, you sort of indicated, but people can do very well.

2:08:00And so I, in general, you know, because I did a music company, a lot of people pitch me music. I'm like, I don't like this business because people, I do media, people pitch me media. I don't want to invest in, in media companies, but I do want to, I do want to build a lot of media at Andreessen Horowitz. Sure, sure, sure. Yeah. And I do want to partner with a lot of independent creators, like yourselves, like the Lenny Richeskis, like the Harry Stebbings, like, like, you know, all these guys who are building, you know, really awesome cashflow businesses, super lean, making a ton of profit. Sure.

2:08:33And sorry to interrupt. One thing I think is interesting is that venture outcomes require scale, yet media, and one of the things that makes media is amazing is anti-scale, right? For us, it's like having CNBC meets X, right? It's a niche thing, but it's great for a very core audience. And on that note, I think you've done this very well with Turpentine in terms of verticalizing out and being like, hey, there's this sort of show on a niche topic, and it may only be interesting to 50 ,000 people in the whole world, but those 50 ,000 people, like that's the show they're like waiting for it to drop.

2:09:13Right. Or it's, or it's a, a group chat that only has 50 people, but it has, you know, maybe like more activity for that, those 50 people than any other social platform that they use. So, um, can you talk about, um, how you think about like sizing sort of like projects, right? Because not everything that should be done should be massive, right? A hundred percent. It's interesting. There's this business industry dive that really taught me about something. It's a business that not a lot of people know about, but it's immensely successful. A lot of people, when they think about media businesses of the last 15 years, they think about like Buzzfeed and Box and it's kind of like, you know, upworthy.

2:09:56These businesses raised a bunch of money in the 2010s and totally crashed, right? There was, you know, there was a bunch of venture firms, you know, this one included doing sort of, you know, consumer media businesses. And people thought, you know, they were going to ride the platforms and they were big, you know, they were growing, but then sort of the, you know, as you alluded to the platforms, you know, took the value, it was where the value was. It was in the distribution, not in the, and these were just like content farms on top of them. But an under told part of that story is that business media has actually done pretty well.

2:10:25This company, Industry Dive, which is just a collection of niche trade publications. So things like utilities, weight, HR, pharma, just like valuable niche newsletters sold for$500 million, I believe, to Informa, which I believe is like a $10 billion trade show company. Wow. So business media business events, there's, there's quite a good business there in trade stuff on the B2B side. Because, you know, if you have the leading HR publication, and you know, you have 50 ,000, even 5000, you know, decision makers, reading or listening to your thing, that becomes a great channel for people who are trying to sell to that audience to, um, to reach them.

2:11:14And they're willing to, if, if, if, if, if, if making one sale can be tens of thousands of dollars for them, they're, they're willing to spend quite, quite a lot to, to, to reach that audience. So the, so the premiums, um, on the CPM are, are just fantastic. And so, and Sean Griffey, the CEO of industry dive has been like educating the market on this. And so I heard him and I, and I, then I looked at his, I think they're incredible at business. I think incredible at sales. I wasn't like blown away by the content itself you know they're a great business and they built something much bigger than I have but I was like wow I think more people could do this and the business work week also is is doing this and I wanted to do sort of a Silicon Valley take on that of like you know Lenny Wojcicki is is the is the goat of this for product managers the newsletter yeah I think he's almost had like a million subscribers and he's printing cash and I was like, hey, there's going to be a Lenny for HR.

2:12:09There's going to be a Lenny for CFOs. There's going to be a Lenny for every position in every sector. I want to create that. I want to create this roll up. And we started on the podcast side because that's what I know. I think that there was gap in the market for like, you know, the best of X. And we did that. I think we had, you know, reasonable success. And I was excited about adding additional businesses or business models on top of that, investing being one of them. And I think AC and Z is the best place to do it. Makes sense. What do you think about the different business models and media between ad sales, subscriptions, monetizing through trade shows or merch or products?

2:12:54I'm a big fan of Doug DeMuro. He value captured at the very end of building a massive scale with cars and bids, his online auction platform for cool cars made in the modern era. Yes, I'm a fan. I know this whole tagline. But in general, what business models do you think work in tech? What business models do you think work in other industries, in trade publications versus more traditional journalism? How are you thinking about the landscape of monetization of media? Well, yeah, it's a really fascinating question. It depends on what niche. If you're as big as someone like Mr. Beast, you want like a mass market thing.

2:13:35Even Mr. Beast, I do wonder if, and I'm friends with him and I think he's genius. But I do wonder if that's the right, if that's the most optimal business to be like, and I like he's also getting to gaming too, but like should Mr. Beast have like sort of a cash app competitor or like. I was pitching a VPN, beast vpn because he has a global audience low churn high margin it's not vpns are not suitable for venture backing typically it's all marketing that the entire expense is just getting people install the vpn so you know it doesn't matter where his audience is they can sign up for the vpn yeah yeah silly but i don't know probably not a good fit but you guys had the founder of honey on who's a very good friend of mine and he's building a really great business i just invested in um on the ad blocker side but like mr beast was saying that he brought a lot of honey's traffic And so should he have created a honey competitor back?

2:14:27Like there's gotta be something that's bigger, there's better margins than chocolate. That's more defensible. Well, there's an interesting thing with media, which is like this idea of value creation versus value capture. And it's possible with media to create a massive amount of value, but not fully be able to capture it. That being said, one of the best ways to capture value is to just put out amazing content and then invest in the companies that sort of come around from that. So I think that people have had this idea of media, of one of the best business models for media being investing. Yet that doesn't mean, you know, starting a solo GP fund and just having a podcast necessarily.

2:15:08Like it can look a lot of different ways. And that's why I'm I'm excited about Turpentine in the context of A16Z, because you can make the best podcast about biotech. And it's less about how do we scale this to millions of ad revenue and more so like how do we just have the most important conversations in biotech happen on this podcast and then invest in some of those companies and see the return from that. Yeah, I know you're still like second day on the job, but I'd love to know kind of how you are thinking about the current A16Z media landscape. People talk about like, oh, future was kind of, you know, didn't get didn't roll out the way they wanted.

2:15:48At the same time, it's like if the named partner can go on Joe Rogan, like you're kind of winning. And so, you know, I think about like the Joe Rogan appearance is like probably worth a thousand at bats that maybe, you know, are strikeouts or whatever. Uh, but, but, you know, how do you balance the, uh, let's build an intermediate, like a in-house media product versus let's empower the team to go and make a statement on other shows like what Mark did with, uh, with the, with the Rogan appearance. Totally. Um, yeah, it's a, it's a fascinating question. So I'm still, I'm still getting up to speed with the exact history, but my outsider version of it is that the future, well, let me step back.

2:16:33What I love about Andreessen, and it's not limited to Andreessen, but that they try a bunch of stuff. Like you guys, right? Like TBPN, you know, you guys had a few different iterations before coming to this one that worked. And people don't see that, you know, like you're making video content for years before this. And Jordi, you were making all sorts of like amazing branded content and doing all sorts of experiments. And and you guys tried it with different people first and you figured out the format that works like a couple of years in or a few months in. So, um, similarly, I think future was a good idea.

2:17:06Um, and I think we'll, it's possible there's some version of that, that, that comes back, you know, they already do a lot of long form content, but I love their everyday carry for founders where they would just ask the, they would just ask the founders in the portfolio, Hey, like tell us what apps you have on your phone right now. And that was it. And it was just like a listicle of like okay the founder of you know some amazing company like databricks like what's he using on it for his email client and it's like that's just interesting i don't know it's like yeah it's like total insider baseball it's totally like oh it's only andreason portfolio companies or whatever but like i don't care i just want to know like what these people are using it's interesting i don't know totally so they took a first stab i think there were a lot of great things about it i think it also coincided with sort of either the andreason uh decentralizing into different funds, right?

2:17:53Like they used to be main fund and now they're sort of different funds. And so there's some org complexity perhaps around that. I think there's much more buy-in and much more like maturity in the org to do something like it again. That's not immediate priority for me. I'm on the content side, separate from investing. I'm more focused on the podcasting to begin with. But to your question around own shows versus creators, I think we're gonna have a mix. So we have some flagship shows right now with the Ben and Mark show, with the ACES show, that we're going to have vertical shows relative to the funds.

2:18:26But I want to sort of really grow the Turpentine network. And, you know, the limitations of Turpentine were that we were a bootstrap business. So we have to make money from other shows. But, and Teresa doesn't have those same limitations. So we want to like do something like an affiliate network where we can partner with amazing shows, help them get distribution, help them make more money, and kind of make it a no-brainer for them to be part of this sort of broader podcast network that just sort of extends our influence and friends. And I think what's interesting is that if this works in podcasting, there isn't like a premier tech podcast network that works with like the best shows.

2:19:07Like, and there are some, you know, I love what Ashana is doing, but that's only like, that's like seven shows. Those shows are great um and i love um uh you know the hubspot's doing some good stuff but like you know lenny's not on a network dorkesh is not a network there's a bunch of people and because they don't need to be but but if you can make it a no-brainer um i think there's something interesting there and if that works in podcasting maybe you can work with newsletter writers maybe you can work with youtube creators more broadly um it's kind of like extended like i love what slow is doing they have a creator fund um there was a super super cool andresen games video that was produced by Secret Tape, which runs this video game documentary channel called Noclip.

2:19:47So they will do the full history, the most definitive documentary about Half-Life or Counter-Strike. And they'll go and interview all the people that worked on it. These really, really amazing things. And it was actually a viral video. It got almost 500 ,000 views. And this is why games like Baldur's Gate 3 are so rare. Just a deep dive on the history of Baldur's Gate, this great role-playing game. And I just saw this and I was like, oh, it's awesome that they found the right match of the team behind Noclip and these documentaries, Secret Tape. They're great at creating. And clearly, Andreessen Games, A16Z Games, was just like, go run with it.

2:20:25And it worked out really, really well. So I love the idea of that partnership model you've described. What do you think about the infrastructure around group chats? Nobody wants to download another app, but at the same time, like you do run into limitations, everybody that's built on, you know, you've built a number of communities, right. And I'm sure you've like hit the limits on Slack, hit the limits on, you know, run into issues on discord, you know, uh, signal has its limitations, WhatsApp, like, how do you, how do you think about, um, scaling these sort of like vertical, you know, group chat, like if you product a group chat, it's just like a, it's a curated social network, right?

2:21:06And how do you think about the infrastructure surrounding that? Do you think there's opportunities there? Yeah. So at first, zooming out, I think a lot of the last interesting, most interesting conversations over the last five years to me have been in group chats and I've been in some with you guys where that's happened. And I think it's just because, you know, Twitter, X, just social in general, became a little bit too noisy, a little bit too public and you needed to have these kind of like private spaces just to have conversations without a bunch of people sort of like getting in the getting in the replies um and you know there could have been a world where twitter that just happened in the dms but for whatever reason um that it moved out to different different group chats and so but the the platforms are not really built for you know signal whatsapp they're not really built for like hundreds of people in a chat, right?

2:21:56And they get notifications for every message. The identity is messed up. Sometimes John Coogan just adds a bunch of random friends on a Christmas Eve. You have to text me and say, Hey, are you going to stop or are you going to keep going, John? Because I made you an admin when this is a lot smaller and there's much more powerful people in the chat now. Why are you inviting all your friends? Yeah, that was back when I cared, when I tried to moderate the chat that since, you know, Bology just, Bology goes Bology and I can't. Yep, he goes full Bology all the time. It's great. That's what you live for.

2:22:29So I've thought about this for a long time and I worked with Shriram when he was at Andreessen because he built out their group chat operation. They basically have dozens of group chats, different verticals, different categories in WhatsApp. It's all separate chats. There's no like master directory. So, I mean, what it should be is like a real social network of like a thousand super interesting people and you can cross promote and it should be um sort of you know like uh interoperable in some way with certain permissions whatsapp's not built for it right now we've just optimized for ease of use um i i think that in the next five years like uh someone is gonna build a group chat social network where um every day well you guys cover the current thing every day um but twitter is too slow in the way that like like twitter um sort of obsoleted the new york the traditional media by just being way faster.

2:23:20And I think there's going to be something that's even way faster than Twitter. And by that, like Clubhouse got into it a little bit, where it's like, it's basically like the beef of the day. Like every day someone gets in a fight or there's some drama and there's going to be a platform where they just fight live. I feel like that's just going to happen. I remember Clubhouse with Mike Solana and Chesa Boudite, like infinite entertainment. And so group chat is just so much more live. Vlad Tenev and Elon Musk. That was iconic. So I was actually working on a product that was going to be like PubHouse for group chats, where you chat live with a public audience.

2:24:01You can sort of tier the permission gates and there's a speaker and an audience. That's fascinating. Because PubHouse really had some magic. But I actually think people said they got too big and they did. And they're pioneers, so kudos to them. But I think that also the audio format's a bit hard. You can't like scroll audio the way you could scroll text. And so I think group chat format is the right one. I was working on it. But then I, you know, spent time with Mark Adresin and decided to have other plans. And I also met another team that I'm not going to announce here yet, but I'll introduce you to them.

2:24:31And when they're ready now, they will. Who's building this? And I think it's a really exciting product. I think the, as you said, Jordy, the bar for a new app is really high. but there's a bunch of different ways that what's happened signal just aren't built for it and i think you could do a 10x better product there so yeah no i was asking in hopes that you were incubating or funding or he found something there he found you you found it you found the right one it sounds like but um but yeah it just seems like such an obvious opportunity when when when i hear about you running you know multiple groups i'm just like wow it sounds really exhausting the second that you get even more, even just one is a lot.

2:25:10Yeah. So one is a lot, uh, excited to see that roll out. I have a hot take. I want you to react to market maps are underrated. Oh, interesting. Um, flesh out the case. Uh, I just think that, uh, they got really like overhyped and they were kind of, so, I mean, the basic map, the basic mechanism of a market map is it's inherently viral because when Andreessen puts out a market map of like all the gaming startups, like every gaming founder is going to be on that list. They're all going to repost it. And so it's just naturally like bait for the algorithm that it's going to get a lot of attention, a lot of quote tweets.

2:25:43Oh, thanks so much for including me. Right. Um, but then it became like very like cringy and it was like, no, you should, you should be like thinking more deterministically about the future. You should be more aggressive. You should be thinking more about even like bigger problems, like the far future of the world. Uh, but, but, uh, you know, I've, I've, I've gone back to market maps and I think there's maybe something interesting there just to kind of take you on a tour of all the different approaches when there is uncertainty about something like AI at the application layer and how that's applying even to legal.

2:26:15Like everyone kind of knows Harvey, but there's probably seven different companies in the legal AI space, probably 50, 70, and they're all taking slightly different approaches. And it's kind of a nice like little window into this industry that you might not be immersed in. And so it creates some value for, you know, keeps the associates busy, keeps the partners busy. But it's also just maybe we can get past the criticism that like, oh, market maps are too basic and cringe and maybe just go to like, hey, actually, that makes sense. If you read a Goldman Sachs report or a Morgan Stanley report on Wall Street, you would often see something that looks like a market map.

2:26:52It's just instantiated in equity research, right? Or equity capital markets research. I'll go further. So I agree with you, but I'll one up you and say, I think lists are underrated. Like Forbes has trained us to think that, you know, the lists suck or they're obviously always gamed or everyone's going to go to jail if they're on a list, hypothetically speaking. But, you know, we've joked about sort of the Forbes 29 under 29. Yeah, I think you should do that. I mean, I don't know if it makes sense now. Maybe we do it. Maybe we do it. Or collaborate or something. But discovery, discovered, like knowing what people think about people is really important and it's what we do all the time we're like who's good you know um who's emerging who should we pay more attention to yeah and i think that functionality sort of like market maps we've like the execution has gotten sloppy and so we sort of thrown the baby out with the bath water but um those things they do perform best which is why they get gamed but if they're good they're they're really valuable totally yeah yeah we're actually going to do a midas list competitor it's going to be a hundred of the top vcs it's a hundred thousand dollars to be on the list a million dollars if you want to be in the top 10 uh but yeah just uh email jordy if you want to be on that list if you're listening uh we'd love to have you oh that's funny i love your whole corporate media like ran oh yeah independent media is dead corporate media is the future yeah you're either backed by a trillion dollar venture fund or you're shilling for ramp and bezel and ad quick and numeral and a sleep and wander uh you're one of the others on a more on a more serious note i'm curious if you're seeing from group chats the edge that you used to be able to get from the internet like internet used to be a place you could you know uh blake robbins talks about hanging out on the edge of the internet but like uh uh but but you could be in these sort of different sub communities and sort of discover somebody and you know we had this this uh kid uh this 21 year old named roy who like rocketed to like uh almost 80 000 followers in like a few months serious and historically that would have been like some you know maybe he would have had got gotten kicked out of columbia and like two people would have heard about it on x yeah no way but now it's like information moves so quickly is it you know are you seeing group chats as like the sort of where you get an edge right because it's like you know more niche down there's like less there's less noise there's less attention maybe if there's a thousand people in the group chat like you know it's there's less like almost like leakage like how do you where where do you think about so it's really interesting so a few years ago when people were more sort of centralized on on on twitter um there was more of a desire to um to you know it was critical to go to group chats because some someone uh would sort of was quicker to cancel you or call you out or there was just like more friction but social media has balkanized a bit, right?

2:29:40The whole blue sky, exodus, sort of, you know, people are threads, whatever, people leaving X. Like, it's a bit, and then Elon sort of, you know, with his sort of new impact on sort of the, like, free speech dynamic or even just like everything's about politics all the time. So people are more comfortable speaking up on X, but because they're more comfortable speaking up in public, they're also more comfortable speaking up in private. So I've seen what happens is these like eight people group chats, You know, me and John had a great one for like a couple of years. There's less of a need for them.

2:30:12It disbanded. Now we're in like multi-hundred people group chats because people are more public. And I think it's cool that, you know, we have people who disagree. Like we have this, you know, one chat where our good friend Mark Cuban gets in, you know, great debates, you know. I thought you weren't supposed to say people's names in that chat. I thought the whole reason was Chatham House Rules. So you're supposed to say a billionaire who's somewhat left wing. He was lampooned on Silicon Valley. Yeah, reality show. Pitches. He went on a podcast with Vivek. They had a debate. So this is public because they said their names.

2:30:53They did a debate on DEI in our chat. And then they took it to the podcast. Oh, that's cool. They mentioned the chat. So I'm comfortable mentioning their names as people who've had debates. and it's just fascinating like you know he doesn't want to go on twitter because he gets dunked on a billion times sure in the chat he'll get dunked on but with more thorough thoroughness sure sure and we all learn from it he gets some dunks in too and that that's the dynamic you just won't see anywhere else and i didn't even see it a few years ago because you wouldn't even think to put people who disagree in the same chat so um so yeah i'm still seeing the alpha well the most important group chat to me is the one that's just me and you eric our one-on-one group chat that's the best group chat in my DMs.

2:31:34Thank you so much for coming on. This is fantastic. Hey, Eric, little request. You're the first Andreessen partner to come on the show. Let's get the rest on. We need a lot more. Thank you. Only Nixon can go to China. Only Tornburgh can go to TBPN. I'm here to broker the relationship. Are you Kissinger or Nixon in this analogy? i think you're great to have you you sneak in through the back door and then nixon makes his arrival yeah we're we're very excited to uh to to watch you work and collaborate and congratulations again on the move we'll talk to you soon bye bye uh next up we got uh jack whitaker coming in to the studio uh the author of the fantastic sub stack bunny hopping i wonder if That's a reference to Counter-Strike.

2:32:25I'm a big fan of bunny hopping. I don't know if you ever did that or if that was before your time. Isn't that like a BMX reference too? Oh, yeah, yeah, yeah. You can bunny hop on a BMX bike. But he wrote a great post. Pre-training isn't dead. It's just resting. GPT 4.5, the value of reinforcement learning and the economics of frontier training. So welcome to the stream, Jack. How are you doing? Doing great. It's great to be on the show. Thanks for coming on. A little context for everyone. Jack is helping us with our distribution strategy. So thanks for all the support, Jack. It's great to have you on the show.

2:32:57Probably first of many, I guess. But I love the post and I'd love for you to take us through it. What inspired this? You kick it off with a couple of the reactions to GPT 4.5. The one that's popping out to me is from Jack Morris. It says, so GPT 4.5 is 10x bigger than 4.0 and only marginally better at most things. My read could be the beginning of the end for scaling laws. What happened here? Did we run out of data or do scaling laws not capture model behavior on tasks we really care about? What inspired you to write this post? Yeah, definitely. So a lot of me and my friends really like GPT 4.5.

2:33:32And we kind of had this high case tester mentality where we thought it came out really well. And we were kind of confused why people were underwhelmed by it. But a lot of the reason people were underwhelmed by it were the actual benchmarks. You can look and you can see it did worse than you'd expect. So me and Trevor were like, well, did it do worse than you would naively expect on the benchmark? have we actually graphed out a log linear law to this scale and seen what performance we'd expect on AI and E, on everything else, you know? And when we actually did this, we saw that it was about in line with benchmarks, you know, a little bit better, a little bit worse.

2:34:04And we thought this was a really important conclusion. I think what a lot of people don't realize is that these are log linear laws. If you double the amount of compute that goes into AI model, you're not going to get double the score on the math benchmark. You know, this is something that we expect that as compute gets larger, as buildouts get faster, we're going to get much better models, but not just because you can double model strength with doubling the amount of compute. So what is the implication for that? What model should people be defaulting to in the ChatGPT app? I think that's the thing I want to start with.

2:34:40Should people just trust you and say, I don't care what people say online, I got to pick 4.5 from the dropdown. Even if I can't tell, will be getting better results? Or is there more nuance there with some of the reinforcement learning that's happening on top and some of the reasoning models that might kind of take things to the next level, even if the underlying model is weaker and cheaper? Definitely, yeah. Well, I think the central claim here is that pre-training will continue to work as we're able to build it out, not that the pre-trained models are the best right now. I think OpenAI's O3 is the best model we've ever seen.

2:35:12It uses RL on every single type of tool use. It uses RL on chain of thought. It is quite a small model. Is O3 built on 4 or 4.5? I'm kind of confused at this point. Yeah, there's no public information on it. Me and Trevor did some estimates based on the token speed that comes out. And it seems like it's quite a small model, smaller than GPT-4, bigger than 4.0 Mini. Okay, interesting. So is this something that we're expecting them to optimize over time and eventually distill 4.5 or maybe just scale up the inference chips to the point where they can run an 03, 04 style model on top of 4.5? Yeah, as you scale up the inference and as you scale up compute buildouts, you could take something the size of 4.5 and you can do 03 style training on it.

2:36:01And I think that's going to be a really, really exceptional model. I think a lot of the point of the piece was that you have like a lot of axes on which you can improve AI and none of them are obviously showing diminishing returns. So there's just tremendous potential to make models better. And there really is no wall. But what about the economics? At a certain point, you need an order of magnitude. This is log linear, of course. So you need an order of magnitude, more compute. We're getting into the$500 billion data center. At a certain point, you get to$5 trillion. And you're talking about a meaningful portion of global GDP.

2:36:35and if the results are, oh, it goes from 128 IQ to 130, the economics don't really pencil out. What's your take on will we just see a pre-training winter just for purely economic considerations? Yeah, and I think that's what we're seeing right now. But another thing is that even as we have, it would be really expensive to train a model 10 times bigger than GPT 4.5 now and probably impossible to serve. But in a couple of years, we're going to see the type of algorithmic efficiency and we're going to see the type of chip improvements that make a model at this scale much more feasible. When you say algorithmic efficiency, are you talking about the type of optimizations that happen with DeepSeq on kind of memory and connect, the FPA, those types of inference optimizations at the actual inference level?

2:37:23Or are you talking about the actual design of the algorithm as it's trained or the design of the model? Yeah, basically on every level. I'm really referencing um dario's excellent piece about deep seek and export controls yep where he says that um deep seek is really good model you know but we do just see these continuous algorithmic improvements as we go you know so sort of the the functional ooms of compute are also scaling even if your compute isn't scaling uh talk about switching gears for a second what what is the uh what's the vibe on the stanford campus right now is is there a risk that everybody drops out to work on AI.

2:37:59I imagine it's like a constant conversation. Yeah, I wish it was honestly talked about a lot more. There's a lot of people, a lot of my friends around campus, like Mohit Agarwal and Jacob and Tamaki, you know, who are super aware of this stuff, you know, and are always kind of pushing things forward and thinking about these things all the time. But in terms of just like the populace getting used to AI and getting involved in it, it's just been a really slow and continuous crawl upward. I'm a CS major. I end up talking to a lot of CS majors and they're not fully internalizing what the model suite is going to look like.

2:38:35Everyone thinks that the models are going to get better around the same kind of paradigm and level. And no one is ever thinking, we have O3 now, what's GPT-5 going to look like in a few months? What's GPT-6 going to look like next year? What are you thinking GPT-5 will look like? Are we just talking about another order of magnitude on pre-training flops, or is there more to it than that? Yeah, so my naive guess is that GPT-5 is going to be a model that uses all the clever training techniques that OpenAI and Anthropic have worked out around RL, but it's scaled up significantly from these pretty small models.

2:39:10Sonnet 3.5 is pretty small. I think the O-series is quite small as well. um one thing that i think is the most interesting about o3 is that it has a lot of these agentic properties that we've been seeking so long for so long but it has them in this kind of narrow sense where it can agentically call tools it can agentically browse the internet but it's not necessarily like agentically going and doing a whole project you know i think that trajectory is one to watch as you look at gpt5 because you're going to see it being more and more agentic just very naturally as context expands and as the model expands.

2:39:41Yeah. Tyler Cowen said O3 is AGI. We had someone on the show yesterday who said we have 10 minute AGI, but that's not necessarily eight hour AGI or 24 hour AGI. How do you think about the length of reasoning chains and kind of that new frontier of optimization? Maybe we've hit the intelligence curve, but we need the agentic-ness to continue for a long time. Yeah, I basically think that's right. The 10-minute AGI take is exactly right. I mean, the AGI debate hinges a lot on what definition you use. Obviously, the Microsoft CEO came on DoorCash a few months ago and said, I'll only believe it's AGI when the GDP goes up by however much percent, you know.

2:40:25I think we're starting to see models that are very general and they're very intelligent, you know, and they can do a lot of the things that you might have naively described in AGI. to be able to do, but they still don't really have this like full capacity yet. I think a lot of this stuff just gets worked out as we kind of improve our current techniques though, and isn't necessarily like some barrier, some key problem that needs to be solved. Are undergrad CS majors bullish or bearish on like rappers or trying to go for the application layer, trying to do a startup or are the best and brightest saying, Hey, there's no way we're going to win.

2:41:01Let's go work for a lab. Yeah, I think a lot of the smartest people I know really want to work for labs, but I think they're almost too bearish on startups and on wrappers, you know, in that a smart wrapper can take a model and kind of scale up with it, you know, and I think this is what we've seen from places like cursor. You improve as the model improves. If you're making a wrapper that's a bet on models not getting better, you know? You're doing a lot of prompt engineering and you're really working to create the scaffolding to try and make the model better. Instead of just saying, okay, how do we get distribution?

2:41:34How do we get the user experience? And then the product gets better. Very light wrappers of the last few years. I think a lot of them relied on consumers not being aware of ChatGPT's capabilities. I mean, there's some apps in the app store that are literally like ChatAI. Yeah, there's still like billions of dollars of revenue out there. That's basically, they just happened to acquire a customer before OpenAI did or another lab. Yeah. Can you tell us any more about just the mood among CS majors around the opportunity in startups broadly? Yeah, I may worry about this pretty soon, but I do think the startup culture at Stanford is really not what it used to be.

2:42:18There was sort of this Halcyon era at Stanford where it seemed like there was so much energy around startups. And now it really feels more like people are doing their startups as summer projects and people aren't committing to them in the way that you really want to see. I think that if some Stanford student came to me and said, I want advice on what I should be doing this summer, I wouldn't tell them to go start a company. I would tell them to go work for Ramp. I would tell them to go work for Cursor. I would tell them to go see one of these incredible organizations, look at how they work, and then take this knowledge to go start a company.

2:42:48I think too many people are doing these things as like side projects and not fully committing to them. Yeah, we talked to somebody who is referencing the early Facebook days at Harvard where if you were an undergrad at Harvard or Stanford and you were interested in startups, like going to Facebook was the expression of that. Now, if you're interested in startups, like there's a$2 million seed round just waiting for you and you will be a founder, even though you could go join a 10 person or even 100 person startup, get a lot of that startup experience. So do you think venture capitalists are to blame here or is it something cultural or should we lay it on the university?

2:43:25who's the point? There's no actor who really made this bad, but the fundamental issue is that startups became too high prestige too fast. And doing real things and really building things didn't gain that same kind of prestige. This is a lot of actually what I valued most about my time at DoorCash Podcast, which is where I worked last summer, was that DoorCash felt like a startup. We had all of this energy, all of this attention that a startup had. But at the same time, obviously, we're doing a podcast. We're executing at a very high level, but it's a podcast and i think getting to work with someone like dorkesh who has this like agency and this knowledge and this drive to make things better and make his content incredible make his questions incredible was something you can learn a lot from you know so it becoming higher prestige to have like stealth start up in your bio than it is to have like i'm like working for nat friedman this summer is like pretty lame uh yeah i think that'll shift i mean working for nat friedman's pretty cool i i i think it will it will it will catch up yeah the benefit is we have a generation of people that are finding out just how hard startups are and there's not there's not land somewhere yeah they'll land somewhere but also realizing that if you're going to start a company once you know how hard it is you really really really need to pick yeah pick ideas carefully yep um totally uh can you give me your read on the last two episodes of door cash.

2:44:47He had AI 2027 and then AI 57, 2057 or something. Are you AGI pilled? Are you feeling the AGI? What's going on over in your world? Yeah, for sure. I actually, I mean, I think the Epoch people are fantastic. And I also think the AI 2027 people are very smart. And I've sent these to both sides here. You know, I think model capabilities will grow very quickly. And I think it's much less clear how this will translate into the economy growing really fast. I was in an interview recently and someone asked me, say, Jack, you keep telling us AI is going to be good. I think AI is already good. Why hasn't it changed the economy?

2:45:25I was like, I don't really know. It seems like the naive economic model says that diffusion of technologies takes a really long time. And it seems like intuitively that wouldn't be true for software, but practically it seems like it is, you know? So I think the AI 2027 people might be mostly right on how fast things are going to grow and how fast the model is going to get smart. And then the epoch people might have a really good sense of like, oh, but really getting this into the serious sector, getting this to disperse across the economy in a way that changes things fundamentally is a much longer issue.

2:45:55Yeah. I mean, maybe it's just like a human issue. You could kind of comp it to, I mean, you go back to like the PayPal days, like the internet got fast enough to transfer money very quickly. And then the percentage of money that was transferred digitally grew very slowly because it's human behavior. And it was only last year that Stripe became 1 % of World GDP transactions. That's crazy, right? And that's Stripe. And they're like the power law winner in the category. Yeah, it's crazy. Anyway, Jordy, you got any other questions? Not this second, but Jack, it's great to have you on. Every time you publish, hop back on, give us the breakdown.

2:46:31This was fantastic. Thanks so much. And Jack's coming to LA soon. Oh, fantastic. Looking forward to meeting you in person. Yeah, for sure. We'll see you at the new studio. See you at the Temple of Technology. Great coming. Thanks for coming on, Jack. Bye. Good to see you. Great to be on, guys. Thanks. See ya. Talk to you soon. Should we go through some timeline posts and then get out of here? Matt Wang says, I find, he's quoting, he posted a quote, I find that super subject matter experts can sometimes be very bad at predicting the things they're an expert in. That is because they're overweight their own expertise.

2:47:02interesting. I think we should deep dive this chat with Domer, the number one trader on Polymarket. Interesting. Kind of like an anti, it's not anti-wisdom of the crowds because that's Polymarket, but there's something there that, I mean, highly relevant to the AI discussion where everyone has a different prediction. They all have a different set of experiences and expertise and then also conflicts of interest and all sorts of things. But we should dig into that post. Joe Wiesenthal shares that markets are surging on this headline. Besant sees de-escalation of China, situation unsustainable. S &P now up 2.7%.

2:47:46Markets love when a situation is unsustainable. Yeah, we're de-escalating, but unsustainably. Who knows? Anyway, we got some big news in the media world. Evan Armstrong has walked away from his cushy writing job at Every to launch his own startup today. He's going founder mode. Founder mode. He says the leverage is his big swing. Starting a company while parenting a newborn feels a little insane, but I couldn't keep this idea inside me anymore. It had to exist. So congrats to Evan for taking the leap into the arena, out of one arena into the other. Maybe it'll be a public company one day. Maybe you'll be able to buy the stock on public.com, investing for those that take it seriously.

2:48:34They got multi-asset investing. He's got to be the first sub-stack to SPAC. Yeah, I'd love to see it. Anyway, we wanted to give a little congratulations to CeCe Gong. She shares a life update eight months after she went viral for supporting her YC boyfriend. I don't know if you saw that eight months ago. I missed it. She posted some photo saying, my boyfriend's in YC. I cooked him dinner, and it went very viral. But he proposed this last weekend by bringing together 100 of their friends and family to surprise her with a stand-up comedy show that she had to perform in. It's fantastic. She says, I posted the tweet in jest, but it sparked a global gender debate about women being invisible emotional labor sidekicks to men's visible professional success.

2:49:19In actuality, my fiance is my secret weapon. So congratulations to CeCe on the engagement. I hope the wedding planning is fun and enjoyable. And I hope his company starts ripping. You know, if it's not ripping already, he went through YC. Hopefully it's just up and to the right from here. Marriage is the greatest investment you'll ever make. It is. One of the most important. It is. Kevin says, never buying regular Zins again. five pound bucket of horse nicotine from tractor supply never underperforming or having a foggy mind ever again i am the hashtag boss lipping half a horse dose and i feel great so that so the next the next slide is is queued up for the mustang horse nicotine packets if you pull up the next slide do you know about this is is nicotine consumption in the horse community i no wait do you understand the riff on this so so basically someone got horse electrolytes yeah it was like This is horse aid, horse Gatorade went mega viral.

2:50:20And so now people are spinning off on that. This is clearly AI generated. Oh, OK, OK. So I saw this way too quickly. I saw it and I was like, John, is this size of an average horse? Maybe I should do this. We got to get Lucy to do some horse nicotine. There were other jokes about this. There was horse creatine. People were spinning off on it. Got it, got it. There's a variety of options. I got AGI's here. Yeah, AGI's here. Oh, this is big news. The star defense tech founder, Matt Grimm, was spotted soaking up some sun at his Costa Mesa headquarters. He is, of course, the founder of Anderle. We have eyes and ears everywhere.

2:50:59And he's stunned in a perfectly tailored three-piece suit. You can't step out of your office without one of our - We have paparazzi everywhere. We're bringing the paparazzi to technology. And Ben had a good idea. He said, need TBPN to start a weekly segment showcasing the best tech, business, and finance fits of the week, what league fits does for the NBA. We need for the high-class work fits. Maybe even throw in a yearly awards segment. Oh, we already did that. We already did that. We already did the best fitted in tech. Aiden Gomez was the runner-up. I believe Alex Karp won for his general overall style, but also his fantastic Patek Philippe Aquanaut with the gold strap.

2:51:39Well, great idea, Ben. Or the orange strap. We already talked about Clue Lee, but there's a lot of backlash. I thought it was a great interview. I thought it was interesting. It is a little... What's interesting is that I don't find his actual product that dystopian, the idea of having augmented reality glasses that I could have... I think he finessed the internet for a bunch of attention. Yeah, he finessed the internet. And I think he's going to leverage it very well. The most dystopian thing is that he's clearly a talented founder. He can't just say, I'm building augmented reality AI app. he has to go and put it in these provocative terms.

2:52:13That's actually the kind of more dystopian thing. And he wants to get to, you know, a implant effectively, a chip for your brain, but he's got to build enterprise sass to get there. Yes, I love that. And that's a good lesson for everyone. Well, speaking of a lesson, a Sam lesson said - I think we should, let's actually save this for tomorrow. Okay, yeah, let's bring him on. Go through it properly. Yeah, it'd be great to invite him on and talk through what to do as a seed investor and his TLDR. We will go to a post from Bezel. We already did the ad read, but they had a great post. Cartier said, brand everything.

2:52:50And they launched a horse headband that's branded. I love it. Camel drapery, branded. Grass, not yet, but they're working on it. Maybe we should do a TBPN horse headband or camel drapery for the folks in the audience who have horses and camels. I think it'd be great. Yeah. Put some ads on there, like the jacket. We should do a TBPN horse drapery or horse headband. Horse drapery. Yeah. Livery for your horse. I love it. Anyway, shout out to Michael in Hawaii. He says, doing my duty and TBPN pilling my Hawaii one big screen at a time. Size gong in full effect. Thank you, Michael, for putting us up on the largest TV I've ever seen.

2:53:33This must be a projector of some sort, but I thought this was an awesome post. Thank you so much for sharing us on the big screen. We love to see the show on TVs and in offices. We love when we're just passively on in the background. We'll get the subtitles going at some point. The live, like what you see in the gym. Live transcription. Yeah, live transcription. I'm sure we can do much better than what the standard is on TV. Yeah. That'd be great. Also, another post from Rahul. Been on the show. Big fan of him. He says, with intense focus. He's actually wearing the suit everywhere. It's fantastic.

2:54:08It's amazing. I'm so happy. Especially because he works in finance artificial intelligence. He does. So I'm glad he's dressing the part. Finance artificial. With intense focus, you can build a superior product. He's sharing a text message that he got from a friend, I believe, or a customer. Just tried Julius for my first real use case. I've always been annoyed that the 8Sleep app only gives you the last year of data. I found out that I could get a full export from them and ended up with like 10 megs of JSON. on. I tried visualizing the data with ChatGPT and Claude and they just couldn't manage it.

2:54:37I decided to try with Julius and immediately nailed everything I asked for. Keep up the great work, man. I love that. And promotion for Aidsleep. Use code TBPN. Use it. Fantastic. And then download Julius. He also was on a plane and he downloaded all my YouTube analytics and did Julius analytics on that and showed that I've been in the trenches. I've been in the trenches. Content trenches. Yeah, I think the first year I got like a couple hundred thousand views and then it was like 17 million three years later is great great great run um anyway this post from nat suki i don't know if it's it seems like it might have been photoshopped but it was yeah hey chat gpt look under there underwear uh lol made you say underwear haha well played say home home latitude 52.3974 this looks like an ad for gamdom or whatever why it's a gambling ad no no no this is a way so so this is a hack that betting companies are doing wait but why is this why is why is this ai poster posting out gambling i feel like that's no no so so basically what's going on here i bet natsuki just posts a lot of content like this sure and thought gambling companies i think are not allowed to watermark yeah they pay for watermark on a lot of posts weird weird weird.

2:55:57Anyway, last shout out and then we'll wrap up the show. Mass, you know him from the viral swing gate up in San Francisco. We covered on the show. There was a swing on a tree. It broke. He said you can just do things and built a new swing. And then that was taken down, went back and forth. I called it the most important political issue of our modern day. I said that unironically. but he just wanted to give a huge shout out to Daniel Strachman, Danielle Strachman and 1517 Fund. They are the goats because he's been looking for compute and he got some. He said, we were just approved for$100 ,000 in AWS credit.

2:56:39Thanks mass for working so hard on that. And huge thanks to Danielle for hooking us up. So he got some, he got some credits and he's going to be training a deca billion parameter time series digital twin of the global economy. The future is predicting the future. Are you ready? Anon. So congrats to Mass for working on stuff. I think he's working on the stock market prediction, hedge fund prediction. Very interesting area to apply artificial intelligence to. Obviously a lot of data. And I wanted to give him a shout out on the show. Anyway, we've had a fantastic show. Thank you for watching. Thank you for listening.

2:57:10And we will see you tomorrow. Have a great rest of your Tuesday. Enjoy your Tuesday. Goodbye. Chad Day.