OpenAI Responds to Meta Offers, Meta's Secret List, Apple Scores Big With 'F1' | Will Brown, Will O'Brien, Kylie Robison, Joshua Meier, John Glasgow, Tom Lee

28:07If Zuck were able to gain access to OpenAI's linear, it would be illegal, but it would also tell them a lot. Yeah, you don't want to do anything illegal, Zuck. You want to be on Vanta. Automate compliance, manage risk, prove trust continuously. Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation, whether you're pursuing your first framework or managing a complex program. So the news comes as competition for top AI researchers heating up in Silicon Valley. Zuckerberg has been particularly aggressive in his approach, offering$100 million signing bonuses to some OpenAI staffers, according to comments Altman made on a podcast with his brother, Jack Altman.

28:48So this$100 million signing bonus, this just got baked into the lexicon. But I was listening to Dylan Patel on Jordan Schneider's China Talk podcast. It's actually a transistor radio, but it goes out on China Talk anyway. And Dylan Patel was saying that that might be a kind of a game of telephone, a little bit, in the sense that it's possible that you could maybe be making$100 million over a number of years in stock based on appreciation. but he was very skeptical that anyone was getting 100 million on day one without any sort of strings attached for like a true signing bonus. So I don't know how real that number is.

29:32And remember that the Zurich team came out and said, the Zurich team explicitly said they didn't get it. But the way that he phrased it, they clearly were getting something. No, no, no. I do think that the round numbers are easy to latch on to. Like why is mom Donnie talking about billionaire specifically? He doesn't have a problem with 999 millionaires because like that that's not easy to grasp like people latch on to round numbers 100 million is so much a billionaire is so much and it's easy to it's easy to quantify And so the hundred million if it was 80 million it wouldn't be going nearly as viral as hundred million And so so I I don't know I don't know how real that is Maybe that was thrown out as a total package, and then it got kind of telephoned into, well, you know, it's basically a signing bonus because they got the deal as soon as they signed, but they do have some sort of earn out or the shares are locked up in some way.

30:29But there's a bunch of creative accounting that can go into basically delivering someone what feels like$100 million of value, which is not far off from, okay, what would it take? You know, Johnny, I've got, you know, a huge offer to go to OpenAI. And it was structured in an interesting way with like an acquisition and you know, but he got what we got a few points Yeah, we were it was a few it was a few percentage points and when you think about The scale of open AI and the scale of Johnny Ives legacy like yeah Like what would it take to get someone like that? He's a founding designer basically, so it's a couple couple percentage points So multiple sources of open AI with direct knowledge of the offers confirmed the number to wired so I don't know because it kind of benefits like both sides to have that number out there.

31:18It's kind of unclear who would really want to confirm that and then what the nature of these things were. The other thing that's interesting is if you're working at OpenAI, and let's say their top researchers have on average $100 million in OpenAI shares that they're like the truly top people, And then you get an offer that's like a$20 million signing bonus, but it's kind of paid out over time or they can claw it back if you don't stay past a certain date. And you're kind of looking at it and be like, okay, I'm already a Senti and I'm getting an offer to go to a new team that I don't know if I'm going to like.

32:00And it's unclear. The vision is not quite as clear at meta. Right. I'm joining the super intelligence team. It's going to be an incredible group of people with a ton of resources. That's exciting. But it is a big switch if you're already happily investing out tens of millions of dollars. So it would make sense that these 20 million dollars, you know, if it's 20 million dollars, I can see that. If it's 100 million dollars for the for the top, top, top people, it makes total sense. Yeah, I mean, the deal to bring on Alex Wang was probably way over$100 million to him personally based on his ownership in that company, which was acquired.

32:44And so that number doesn't feel impossible. I would just be surprised if there's not some structure around it based on what Dylan was talking about. There's just enough precedent here that, again, going back to Zuck raiding Google in 2010, 2011, he was offering tens of millions of dollars in stock packages to mid-level engineers just because he was like, I need a really good ad product. I want to make my ad product great. Yeah. I mean, have you seen the, why am I even asking this? Have you seen the Blackberry movie? Of course not. But in the Blackberry movie, there's this whole story about they go and they try and poach someone.

33:23and then it was controversial because they backdated some stock options for someone to basically give them more cash, but without having the initial impact of that cash. It's kind of interesting. But yeah, I mean, to poach top talent, it's expensive. It kind of benefits both, maybe. I mean, I don't know. Like, on the meta side, do they want, does Zuck want the idea that he's willing to pay up to$100 million out there? like maybe because that that certainly leads to more people taking the phone call of like oh wow mark is really taking this seriously i should i should talk to meta's recruiting team i should actually hear them out and then if they wind up coming back to me with a 50 million dollar offer or a 10 million dollar offer like you know we can discuss that and i can see it and they can evaluate how much i'm worth they can make an offer but um at least it got the conversation started but I don't know there was also the theory that Sam Altman put that 100 million dollar number out in order to kind of like poison the well because if if Mark comes to you and says hey you're a top AI researcher come over here and then the top AI the AI researcher says like yeah great I hear I hear 100 million dollars at the going rate and Mark's like no no no that's like fake news that that Sam was spreading like we're actually paying 15 then you're like oh no like I was I was hoping over 100.

34:45Sam is one of the most Machiavellian people. Yeah, it could be like 4D chess on both sides. I don't know. Over the past month, Meta has been aggressively building out their new AI effort and has repeatedly and most unsuccessfully tried to recruit some of our strongest talent with comp-focused packages, Chen wrote on Slack. A source close to the efforts at Meta confirmed the company has been significantly ramping up its research recruiting with a particular eye toward talent from OpenAI and Google. Anthropic, while also a top rival, is thought to be less of a culture fit at Meta. One source tells Wired, they haven't necessarily expanded the band, but for top talent, the sky's the limit.

35:25Which is funny because, yeah, of course, Anthropic's so missionary-driven. It's going to be very hard to go to Anthropic research scientists and be like, ads. Better ads. They're like, we're building God here? Anyway. 100 million dollars is 100 million dollars not for not for anthropic guys anthropic guys are true believers they don't care what's 100 million dollars in a post-scarcity world it doesn't matter you want to be on the you want to be on this on the on the post-scarcity train uh chen's no it is so it is so telling though like how how you like sort of like try to you know pinpoint like zuck's views on ai he's like i'm not even going to bother with those people not a culture pit.

36:14Two AGI pilled. Yeah, that's great. Senator, we sell ads. Anyway, if you're looking for sales tax super intelligence, go to numeralhq.com. Spend less than five minutes per month on sales tax compliance with super intelligence for sales tax. The official sales tax automation provider for Lucy. Yes, that's true. Chen's note included messages from seven other research leaders at the company where they wrote notes to staffers in an apparent effort to encourage them to stay. One leader on the research team encouraged staff to reach out if they received an offer for Meta. If they pressure you or make ridiculous, exploding offers, just tell them to back off.

36:51It's not nice to pressure people in potentially the most important decision. Wired is not naming the leader as they are not a C-suite executive. This is weird. Where's the end of this quote? That's odd. I'd like to be able to talk you through it, and I know all about their offers. The remarks come as OpenAI staff grapple with an intense workload that has many staffers grinding 80 hours a week. OpenAI is largely shutting down next week as the company tries to give employees time to recharge. Let's give it up for OpenAI's grind set. It's really fantastic. Also, extremely American to give a full week off right around July 4th.

37:28Like, I'm normally pretty anti-taking time off, but if you're going to go really hard on a particular holiday, make it July 4th. You know, give a week off around July 4th. I love it. I was posting, you know, the over, I need a poly market on whether or not we get a Zuck American flag video on July 4th. Do you think it'll happen? I actually think there 100 % should be a poly market on that. Right? It's kind of 50-50, right? Like he is definitely locked in, very busy, probably not a lot of time off. He sees that, he knows from this that OpenAI is down this week. So that's probably like really encouraging to be like, I want to go harder this week.

38:06They're off. Yeah. My advice for founders this week is one of the greatest pieces of content he puts out. Think about how hard Zuck is going right now. Go harder. The week of July 4th. And just don't get outworked by him. Yes. Just don't get outworked by Zuck this week. It's going to be tough. He's in year 20. What excuses do you have? You're three months into your startup. Yeah. Come on. Go harder. Go harder. I love it. Yeah, Meta knows we're taking this week to recharge and will take advantage of it to try and pressure you to make decisions fast and in isolation, another leader at the company wrote.

38:43If you're feeling that pressure, don't be afraid to reach out. I and Mark are around and want to support you. While OpenAI's leadership is taking Meta's efforts seriously, Chen also said that the company is getting too caught up in the cadence of regular product launches and in short-term comparison with the competition. The sentiment is backed by a former OpenAI staffer who worked closely with Altman and said the CEO wanted to see buzzy announcements every few months. This was a key to their strategy. Always have a buzzy announcement ready if the competition does anything. Just drop an absolute.

39:15Oh, Google I.O. is happening? Wouldn't it be crazy if we acquired a company called I.O. the day before or something like that? We're like, oh, Google's announcing the new Gemini. Let's just steamroll them in the timeline 24 hours before. I'd love to introduce you Deep Research. I mean, it's masterful. But that's the game you got to play. That's the game on the field. I think he's being aggressive, but he's not doing anything wrong. He is up there with Elon in terms of creating and riding hype and constantly delivering enough real product value to justify the hype that was being sold six months ago, basically.

39:58So like staying on the bleeding edge. And this is why the anti-tech people hate him and Elon so much. Because they're like, it's overhyped. It's not useful. And then like a year later, it's like there's Tesla cars everywhere. The rockets are landing. And like 90 % of people are eating Chachipiti every day. Tens of billions of visits monthly. Okay, so it was real. But also he was overpromising. But he delivered at a higher level than anyone in human history. And that's just like those two things are in conflict. Like someone said this, like Elon is both Thomas Edison and Barnum and Bailey or whatever.

40:33P.T. Barnum. He's a circus man who can put on a show and put a robot in a human in a suit acting like a robot. But then he can actually go and deliver the thing, which is just it just breaks people's brains. Anyway, we need to remain focused on the real prize of finding ways to to compute. A lot more supercomputers are coming on later this year into intelligence. Chen wrote, this is the main quest and it's important to remember that skirmishes with meta are the side quest. Last but not least, I'll be around this week, recharged and ready to go pound for pound. DM me anytime. I love it. He's ready to go.

41:08He's fired up. Get him in the ring. I love talking with Mark. Yeah. That was a fun conversation. Also, I mean, guys in shape. Let's get him in the octagon with Zuck. Mark on Mark violence. Let's see it. I mean, if there was a pound for pound list for the top AI researcher in the world, he's in the conversation. He's in the conversation. It's amazing. It's really been amazing to watch Mark's leadership and integrity through this process, especially when he has had to make tough decisions Altman wrote on Slack in response to Chen's message. Very grateful we have him as our leader. I mean, it can't be overstated making these decisions, right?

41:46Yeah. One of your best people comes to you and says, hey, I'm really happy here at OpenAI. I have$80 million of stock. And I was just offered$150 million over the next year. And it's highly liquid, or over the next few years. It's highly liquid, meaning that I can market sell it quarterly forever. And meanwhile, you're going through this for-profit conversion. Do you think the liquidity thing is real at all? I mean, obviously, what are you I mean, like if you want to let's say you want to buy a 50 million dollar house and you have 100 million dollars in like illiquid open AI. Like, is that going to be hard at all?

42:29I feel like Goldman Sachs and Morgan Stanley, private wealth management are going to be beating down your doors to give you a loan that's backed by the shares. Sure. You're only going to have to put down 10 percent. Sure. But but meanwhile, there's so many selling to go through a for profit conversion. Yeah. But I mean, so that is. Satya's taken 20 % off the top. I mean, that applies. No, I'm just saying. That applies like a small haircut, I would say. I'm just saying there is always going to be a tension of somebody having massive wealth but not having greenbacks. Yeah. There's always going to be that tension.

43:04Yeah. And people seeing that Meta is one of the most liquid stocks in the world. Yeah. Can constantly trade in and out of it. And it's ultimately just, it's stable. I mean, if you're interested in liquid stocks, go to public.com. Investing for those to take it seriously. Multi-asset investing. Industry-leading yields. Good, old-fashioned stocks. Trusted by millions. And who knows? Maybe they'll have open AI soon. Anything's possible. Yeah, we should. Robinhood has the tokens around private companies. Maybe we should tokenize nonprofits. I want to be able to trade, you know, PETA. You know, let me go long PETA.

43:52Let me get some leverage on that. I want some levered options on PETA. Look at Tyler right now. Let's get on the Tyler cam. How are we doing on timing? I think I'm in the right mission. Oh, you did it? No way. 15 minutes left. Let's see it. Okay. Okay. I can't believe you got it set up so fast. That's amazing. Okay. He's sitting at 26 minutes so far. All right, well, have fun in there, Tyler. I just look over and I didn't realize he was on VR. It's amazing. I think he's actually playing it. This is crazy because my number one complaint with the Quest was that it just takes a long time to set up.

44:31And I think that it could cause a lot of dust. You intentionally did this challenge. It made it seem like, oh, I'm going to happily give away my new VR Xbox. thinking that there's no way he could finish it in 40 minutes. But now he's speedrunning. Pillar of Autumn, it's happening. Anyway, Swix has the story. He's breaking it down. Yeah, some good highlights here. The Mark V. Mark war has resulted in going from, oh, the good people don't leave, to someone has broken into our home and stolen something. One week, companies shut down, question mark, question mark. I mean, I was going back and forth here.

45:13I wouldn't be surprised if the shutdown was somewhat strategic in that, hey, people already need a break. We shouldn't have them walking around the office. The way they were writing it, it felt like this had been on the table for months because the company has been working really hard launching a ton of stuff. and it seems like if it just seems like an odd choice to if you're in the middle of this mark v mark war that that would be what you choose to do would be a one week company shutdown but I don't know maybe I'm not 100 % that like the AI talent war has resulted in the shutdown it's more just like the shutdown is happening at a weird time anyway this third point is very interesting pivoting product launch calendar to AGI race especially when Stargate comes online.

46:01Dylan Patel says December is when Stargate's coming online, and that should be the biggest cluster ever and unlock even more, maybe pre-training scale, but there's a question about how important that is and maybe you can do more reinforcement learning in a more distributed way, so I don't know. Also, it's not exactly like Zuck is GPU poor, So I don't know how big of a deal it is, the Stargate thing. We'll see. But number four, for some reason, Meta is ignoring Anthropic. Don't care about safety. If this is true, of course. The Anthropic point was very funny. Anyway, there's other posts in here that we should cover.

46:43Satwik Singh says, what Sam never understood was that Zuckerberg torched 20 billion over VR headsets with Nintendo Wii graphics and didn't even flinch. Didn't even flinch. You know what's so funny? like those the the nintendo we graphics thing that was like that was like three months and then after that it was he was doing that that lex interview in the metaverse did you ever see that where it's like it is like photo real immediately and we got a wild demo at meta hq yeah of all the new products and we were blown away i'm not even sure i don't think we can say much talk about it in like the abstract but but yeah I mean none of that was like Nintendo Wii level graphics but but but I mean that that was that one picture was like iconic and kind of like it felt like oh this is kind of a step back in terms of well many more people saw that picture than actually use the product yeah and so yeah because there were plenty of there were plenty of VR games that didn't have Nintendo Wii graphics like the hero game for the for the quest when it came out was Robo recall which looked super sci-fi and just looked awesome and And it was great.

47:46And there were a bunch of other games. But yes, I mean, the point is that like Mark is totally ready to spend$20 billion on a big project if he thinks it's valuable. And here he clearly does. And so he's going after it. Anyway, similar founder-led company that's building amazing AI stuff. Finn, the number one AI agent for customer service. We've had Owen on the show before. He's been on a generational run. Number one in performance benchmarks. Number one in bank offs. Number one on G2. I want you to. Let's go. So if you're looking to do customer service, head over to Finn. I'm saying this, they're not saying this, but I'm gonna say this is the official

48:30AI CX tool of both AI safety and being AGI-pilled. Okay. Because Anthropic uses. Oh really, no way. There we go, okay, awesome. Tope. Yeah, I'm sure Anthropic will love that. They're like, we said you could put our logo on your website. We didn't say these technologies. Two podcasters could go way over their skis on this endorsement. It's the official CX agent. Anyway, Rune had an interesting take here, diving into why these trade deals are so, they're not really trade deals, but these poachings can be so damaging. And he's trying to put it in broader consumer terms. We're gonna debate whether or not his point is well taken.

49:13He, of course, works at OpenAI and has for a number of years. Rune says, intellectual property by default is a market failure. Single, well-informed, talented defectors can walk away from organizations with billions of dollars of tacit value of knowing what works and what doesn't. The athlete metaphor is somewhat wrong. In industries where patents and such aren't viable, market participants, aren't viable, market participants will drastically under-invest in R &D efforts due to IP seepage, just like the I drink your milkshake effect of oil seepage and land rights. There is a compelling pro-consumer case for non-competes.

49:53Interesting. Very rare that you hear a tech person argue in favor of non-competes. Interesting. I mean, the immediate impact of non-competes in this context would be dramatically lower wages for AI researchers like Rune, which is interesting. The other question is— Yeah, and it's interesting. You'd have to move AI out of California. Oh, people would then. Maybe. Or they would choose to be there. I'm just saying if you really want to use non-competes, go where they can be enforced, right? Yes. California has been historically impossible. Yeah, but then as an AI researcher, you'd say, I definitely don't want to leave California because I'm going to get a much better deal here.

50:37I've got kids in school here. Yeah, my question is like, I've always been interested in like, why can, you know, like the consumer packaged goods industry like patent like a tab on a can for 25 years and like Walt Disney can own like the shape of mouse ears for like 80 years, but like Google can't own the transformer for like even a minute. Like, I get that it's like some sort of fundamental discovery, but I've always been interested in how little IP works in technology. So that's interesting. It just is a byproduct of hacker culture and sort of like positive sum culture of Silicon Valley. Yeah, but you'd think that would change if there were tens of billions of dollars on the table, right?

51:20Yeah, I mean, there's a culture of, you know, competition doesn't kill businesses. So I agree with you, and I'll give you an example. So pull to refresh on the iPhone was created by a Twitter employee, like a Twitter UX UI developer, built that user interaction, and then Twitter had a patent on it. Lauren Brichter. Yes. American software developer, best known for creating Tweedy and the pull to refresh interaction. So he created that design pattern, and then he had a patent on it. Twitter owned the patent. and Jack Dorsey and the Twitter crew were libertarian and said, we're not gonna enforce this.

52:01And then it got baked into Instagram and it's now natively in iOS and it's in all of, it's basically like one button to like - If you're truly terminally online, you're using it hundreds of times a day. True, but now we've actually moved past that. We've moved into endless scroll where you actually don't need, like if you're on TikTok, you're never pulling to refresh. You just keep scrolling to the bottom. So we're actually post pull to refresh, and now we're in the endless scroll. But endless scroll, I believe, was developed by Pinterest, and they had a patent on it for a while, I think. I might be wrong here, but endless scroll was also something that was not able to be owned, which is just odd to me.

52:42And then there's the other question of like, Coca-Cola's been able to keep the formula as a trade secret for like 100 years or something, And it's just odd that tech companies can't seem to do either. They can't keep the secrets. You can have trade secrets. It's not like an open AI researcher can go over and say, here's a bunch of code. Let's roll it out. It's basically the sort of way of doing things that they've internalized from being at a company for a certain amount of time. Tacit value of knowing what works and what doesn't. And so my naive read on this is that if you had had a top tier OpenAI researcher at Meta during the LAMA 4 build out, that researcher would have said, hey, don't focus on pre-training scale as much.

53:33We need to get a reasoning model out and we need to focus on RL and post-training much more, much more. And when you talk to Dylan Patel, I was asking him about DeepSeek and how DeepSeek was doing all these innovations and like moving to floating point eight, like smaller, basically smaller numbers storing the weights. And he was like, oh, yeah, but OpenAI was doing that like two years ago. DeepSeek just open sourced it. And so it's like the world found out, if you found out the best practices from DeepSeek because they open sourced it, like you were not on the cutting edge and you should have been coaching AI researchers from the top labs earlier.

54:08So two things could be happening here. One, Zuck is right to just go and overpay and try to recreate some of the magic and get a leading model or set of models out of it that he can use in a bunch of different ways that we've discussed. Or none of this matters and Rune is baiting Zuck into just overpaying. This is like the 4D chess is like, oh yeah, this is, they're getting everything. They're taking everything. Yeah, yeah, yeah. No, I don't necessarily believe that because I did talk to Rune like years ago and I was like, but like, is this going to commoditize? Like, is the foundation model layer going to commoditize?

54:52And he was like, they're like, they're like the secret to open AI in many ways is like the Thelian secrets. Like they understand things at a lower level that other, that other people just don't. And so that will compound and stay at advantage. And he hadn't even predicted the dominance of the consumer and like the aggregation theory that happens when you're the front end to consumer AI. That's incredibly valuable. But just on the research side, just on the research side, like they can stay ahead just by understanding like the secrets of what works and what doesn't. But it's interesting. Anyway, let's tell you about Adio.

55:24Customer relationship magic. Adio is the AI native CRM that builds, scales, and grows your company to the next level. Next level. It well, I mean, we you know we have a light guest booking today But you know we're Spending an hour talking about one story because it's fun stuff. We can continue we can we can do the go over the Wall Street Journal because in the exchange section there's a fantastic article related to this says it's known as the list and it's a secret file of geniuses and I'm looking at this it says Mishka Belenko, Yu Zhong, Mark Zuckerberg's been reviewing this list, Lucas Byers on this, the new guy.

56:07It says Jordy Hayes, John Coogan, Tyler Cosgrove, Ben Kohler, a couple other folks, but you know, it's just the usual people that we know. Anyway, it's known as the list, and it's a secret file of geniuses. Only select AI researchers have the skills for the hottest area in tech. Mark Zuckerberg and his rivals want to hire them, even if it costs ungodly sums. We love ungodly We should be ringing the gong more this episode. All over Silicon Valley, the brightest minds in AI are buzzing about The List, a compilation of the most talented researchers and engineers in artificial intelligence that Mark Zuckerberg has spent months putting together.

56:46Lucas Beyer works in multimodal vision research, which Tyler broke down for us. Can you imagine if Zuck spun up basically like a hot or not style tool that just pins two researchers together. He makes other researchers just do the test and then just puts out a list of, this is the top 50 people. Yep. This is so good. So Lucas Beyer describes himself as a scientist dedicated to the creation of awesomeness. Very Reddit-coded. But he's a poster on X as well. Yuzhong specializes in automatic speech recognition and barely has an online presence besides his influential papers. Misha Belenko is an expert in large-scale machine learning who also enjoys hiking and skiing, or as he puts it on his website, applying hill climbing search and gradient descent algorithms to real-world domains.

57:49That's very cute. The recruits on the list typically have PhDs from elite schools like Berkeley and Carnegie Mellon. They also have experience at places like OpenAI in San Francisco or Google DeepMind in London. They're usually in their 20s and 30s and they all know each other. They spend their days staring at screens to solve the kinds of inscrutable problems that require spectacular amounts of computing power. And their previously obscure talents have never been so highly valued. The chief executives of Tech Goliaths and heavyweight venture capitalists are cozying up with a few dozen nerdy researchers because their specialized knowledge is the key to cashing in on the artificial intelligence revolution.

58:25It is funny. Some of the losers out of this talent war are just VCs because how do you compete when historically you could have gone to some of these and really overpaid to invest? Yes. You could have come to them with a ridiculous, what would be very ridiculous for a VC, which is I will spin up a new company. Not ridiculous. It happens, but it is kind of ridiculous in some ways. Spin up a new company. I'll give you 200 on a billion and you can take 10 million personally and secondary immediately. And that is no longer even that compelling of an offer if the odds of actually winning are low and you just want the most access to compute and be traded.

59:14Absolute, absolute banger ripping. This was Tyler's handy. Oh, Tyler, how are we doing? I'm in like some kind of maze right now. Oh, no. You lost, didn't you? No, I lost already? I think he lost. It's past noon. it's past noon wait Ben what you got on the timer I've got four minutes still on the timer four minutes still okay let's give him another four minutes you better finish this I'm surprised this game is so hard but I bet you if we roll the tape it took him 30 minutes to set up and he's actually on track I bet he's going to finish the pillar of autumn the level in normal time like 20 to 30 minutes because I only gave him 10 minutes for the setup Mogged anyway This is the reality of pledging.

1:00:02Podcasting. Pledging to be a technology brother. Nobody in this escalating arms race is chasing the prize recruits quite like Zuckerberg, who has tried to raid Silicon Valley's top research labs, dangling$100 million pay packages. We're hearing it again. To a select few superstars in hopes of poaching them. The billionaire CEO of Meta wants them to join his company's new lab focused on super intelligence. One recruit who has spoken with Zuckerberg, who is personally courting his dream team of potential hires, describes the company's goal as nothing less than a transfusion from the country's top AI labs.

1:00:37That's all it is. It's a transfusion. It's a blood transfusion. It's a blood transfusion. It is wild. Yeah, we got to talk to some more VCs about how they're coping. We got to get some coping VCs on here to hear them cope about it. Yeah, yeah, it's actually a fine time to invest in AI. Hey, what do we got here? Tyler. I'm done. I'm done. You're done. No way. What's the clock? What's the clock? Two minutes, 45 seconds. Wow. No way. He did it. Let's go. Oh, that's amazing. Let's go. Congratulations, Tyler. Did you even get to play your new? No, no. He had to unbox it. I did. I wanted the unboxing to be part of the experience.

1:01:16Wow. You just ate your lunch, John. He did. Congratulations, Tyler. Very good. That's amazing. Okay. So break it down. How long did it actually take you to get set up and get the game running? Yeah, well, okay. For the first five minutes, there was just a software update. That's what I was worried about. Yeah, and then I had to connect to my phone. There's the Meta Horizon app. Yep. And then it was probably another couple minutes. I had to create a new Microsoft account and then get Xbox Game Pass. Yeah. And then I wasn't sure which Halo it was. It took me a minute or two there. Yeah. And then the first probably five or six minutes of playing the game was just cutscenes.

1:01:51It wouldn't let me skip. It's like the tutorial of how to play the game. And then finally, and then I just crushed through it. Okay. So how long do you think it actually took you to go from like, I gave you the box to you were playing the game. I actually have no idea. I mean I was pretty locked in. I don't know how long I was playing. Okay, we should have had splits because my big pushback on VR right now is that if you get an N64 or you get like the old, if you got an Xbox for Christmas with Halo the DVD, like you would put that thing in, plug it into the TV and you would be playing in two minutes.

1:02:30Like it was so fast to get up and running. And I've always been upset with how many things you have to, you have to link it to your phone, take it off, log in, create a password. Like all of that I've always been really annoyed with. I think that they should make you do the account creation after the fact. After they let you get hooked on playing and be like, this is fun. Okay, cool. Yeah, now I want to create a social profile because I'm having fun. Anyway. I mean, the app seems very unnecessary. I don't know what the app is for. Odd. Because it wasn't like to sign in. I still have to sign in on the, using the joysticks.

1:02:59I mean at one point I had I had four I counted I had like six different passwords that were relevant to play Oculus games because I had I had a pin code that was like you draw it. I had a Oculus password, but then Oculus got rolled into meta and so I had a meta password But then I also had a Facebook password that was different And then I had like a pin for paying with things and it was just like so many different things and and it's not that hard when you're on your phone and you get one password and you just kind of like auto fill passwords all day long and two factor codes. But when you're in VR, it's super hard to type everything out.

1:03:36So, wow, I'm very, very impressed that you pulled that off. I thought that that was going to be a failure. So congratulations. Enjoy it. Wait, so break down the experience of actually playing. Were you playing on a big 2D screen and you were sitting there in like a theater basically? Yeah, yeah. So the screen was probably like six feet in front of me. It was like in front of this light. It was like great experience. I didn't find it to be like, like the refresh rate was like totally fine. I'm not motion sick at all. No screen door effect or anything? No, it was totally fine. And then also breakdown.

1:04:08So were you technically streaming the game over Wi-Fi? I believe so. I'll look, but I think it was like cloud gaming. Yeah. So I think the actual Xbox like frame rendering is happening in the cloud and they're streaming that to you and then rendering in VR and then the headset's just pinning that to the world, which is really, really cool. So I have this thesis and you'll have to play with the headset more and let me know if it plays out that I think that Palmer Luckey has that take that's like soldiers will be wearing VR headsets on a daily basis before consumers are because the military can underwrite like a$5 ,000 VR headset.

1:04:51And so he's betting on augmented reality glasses for the warfighter being the first real use case. I have this thesis that before people were using the iPhone for dedicated iPhone apps like TikTok and Uber, people were using it to make phone calls. It was a phone first. And I think that that will be basically a TV first, and people will be playing traditional video games and watching traditional movies before we get VR GTA 6, people will just be playing GTA 5. And so I think all the VR companies should focus on making the movie watching experience and the traditional video game experience like super, super smooth so that you just get in, you can actually chill in that thing for two hours, actually beat Halo.

1:05:37Maybe, you know, the whole game only takes three hours to beat. Maybe you got to beat on the screen now. Okay, gamers. Anyway. Let's get back to the news. Let's get Back to the sleep scores. How'd you sleep last night, Jordy? Okay, before we... I was absolutely mogged. No. What happened? Get a pod five. Get a pod five today. 69. Rough. I got a 92, baby. Incredible. Generational run. Hit me with the Ashton Hall. You know the rule. If I beat you, I get that sound effect. Ashton Hall. Boom. 92. Get... Go to aidsleep.com. Slash tbpn. I'm worried to think about how I would have slept if I didn't have an eight sleep.

1:06:18And, you know, the only thing that I got penalized on was I went to bed too early, 9.30. And they're like, you don't usually go to bed at 9.30. And so I pulled it off. I was penalized by screaming children in the middle of the night. Brutal. Inside the secretive world. Yeah, we have some breaking news. I'm going to dive into this. So this morning, Zuckerberg announced his super intelligence team. 4 ,000 likes on this traded post? No way, Tyler. Oh, you're on a generational run today. That is amazing. Okay, wow. 4K likes. Okay. It's awesome. Back to the breaking news. Yeah, yeah, please. So Zuckerberg sent a message this morning in a memo introducing the new super intelligence team.

1:07:08So we were just covering his list. Yep. He got seemingly a lot of people from his list. So Zuckerberg introduced Wang, who will be the company's chief AI officer and leader of MSL, as well as former GitHub CEO Nat Friedman. Friedman will co-lead the new lab with Wang with a focus on AI products and applied research. Here's the list of all the new hires as seen in Zuckerberg's memo. It notably doesn't include the employees who joined from OpenAI Zurich office. So this is Kylie Robinson's reporting over at Wired. So Trapit Bonsal pioneered RL on Chain of Thought and co-creator of O-Series Models at OpenAI.

1:07:50Correct. Xu Chao B, co-creator of GPT-4 Voice Mode and O4 Mini, previously led multimodal post-training at OpenAI. Hua Wen Chang, co-creator of GPT-4 O's Image Generation and previously invented Maskit and Muse text-to-image architectures at Google Research. G. Lin helped build 0304 mini GPT40 GPT40 GPT40 GPT40 GPT4.5 GPT40 image gen and operator reasoning stack Joel Pobar inference at Anthropic previously at Meta for 11 years at HHVM hack flow redex performance tooling and machine learning Jack Ray pre-training tech lead for Gemini and reasoning for Gemini 2.5 led gopher and chinchilla early LLM efforts at DeepMind Hong Yu Ren, co-creator of GPT-40.

1:08:39And this just goes on and on and on. But the big takeaway here is that these seem to be some important people and not what Sam had originally messaged as none of our best people. Wow. That is a big... But obviously things have changed since that original. I think we need some more trade deal posters, except it's not much of a trade deal. Just full poaching. It's pretty crazy. No, this, there will, whenever, whoever writes the book on this era of AI. Ashley Vance. Obviously. Will have at least a chapter dedicated to this talent rate. Yep. This is huge. Pretty unprecedented. But yeah. I think the difference when Facebook was raiding Google in 2010, it was just spread out over a much longer period of time.

1:09:38And it wasn't, I mean, this is unique because in that situation, Zuck was poaching from Google. Zuck poaching from a company that still is like many people view as a startup, right, even though it's worth hundreds of billions. So. Yeah, this is pretty remarkable. Turner Novak said is it true they poached Matthew McConaughey from Salesforce I like that one anyways so we could go back to this list but clearly he made a list he checked it twice and he got a very meaningful it's time to head over to bezel and get him something get himself something to remember this moment by because this bezel concierge is available to source him any watch on the planet seriously any watch he could get a new richard mill he could get a new vacheron or a new patek he has a cubitus maybe he needs to fill out the fill out the stable a little bit after this uh pretty remarkable yeah the other thing that's that's interesting here daniel gross is not anywhere in this memo oh interesting who was rumored is the full memo listed or is it just like they just didn't include this because it's just wired reporting and they don't realize how big of a deal DG is.

1:11:01I think they would have included it. Yeah, maybe like the negotiation is still ongoing. Like it is a more complex situation because it's like a CEO leaving an active company and so it's not like a full free agent or acquisition or going from W2 to W2. It's a little bit more complicated than that. So fascinating. Well, the AI talent wars are the current thing, and we have been enjoying following them. It's a great time to be an AI researcher. Even better time to be a podcaster. Yeah, incredible overnight success for all these researchers. It's not like they spent 10 to 15 to 20 years grinding. Also in a field where people were like, oh, that will never happen.

1:11:46Like this is crazy sci-fi. This is not going to ever be real. Like you're just, you're doing like complex math. Cute chat bot. Does it work? No way. Yeah. Cute picture that makes a morphed hand. Yeah. Look at that hand with six fingers. Yeah. That'll never be commercial. You bozo. Oh, you're working on image, Jen? Yeah. Why does that person have? Oh, deep dream. The thing that puts dogs in every picture. Oh yeah. That'll work. Yeah. Right. Yeah. cat detector hot hot what is it hot dog not hot dog yeah now it's the most valuable technology in the world yeah i i'm i'm super i'm super interested in seeing like what they build obviously like it's going to be like a good model like we know that they're going to try and build like a a big good model it's gonna be open source maybe it might not be maybe not there's like Mark hasn't guaranteed that he'll be open source forever.

1:12:48Actually, the model is really good now. We're going to charge for it. No, but he could. He could. I mean, like open source, yeah. I mean, it could be closed source. It could be open source variations. It could be, you know, a chat-driven product that directly competes with ChatGPT. It could be something that just lives behind the scenes. He's going to, yeah, yeah. It could be a tab. It could be a new tab. You know what? I actually like, I would like to own a piece of search too. I got a bunch of DAUs. They're searching stuff all the time. I kind of like that market. I think people might want to discover products that way.

1:13:21I think you could build a pretty good ad product on top of that. It is interesting, though. I mean, it feels like LLMs right now are so single player and Meta products are so social and communication and entertainment oriented. Launching a consumer. And the social thing is like it's right around the corner. It's right around the corner. Already, I've done this with you where I will go and generate an O3 Pro response, let it cook for 10 minutes, have 4O generate some charts or graphs or images off of it, and I will just share the chat with you. And then you can trace through the whole thing as you like.

1:14:02And that workflow back and forth is amazing. Ben Thompson wrote about it, how he has a research assistant who isn't at the level of him in terms of analysis, but can ask a bunch of questions. and then he can go back and see, okay, he asked this question slightly the wrong way, so he got slightly the wrong answer. Let me just continue chatting with the same interaction in GPT-4 or 4.0 and actually improve and get the answer that he wants. So it's fantastic. Are you about to hit the gong for something? I mean, there's like 20 researchers that just signed a meta. Do it. And I'm just going to do it.

1:14:36What is that? Why did you throw it? We're just experimenting here. Okay. Yeah, it's funny. Dan Ratcliffe in the chat says, every researcher hire I see, I think of Jordy's tweet on companies doing NBA style trade deals, open AI trading researchers with Anthropic, throwing in a couple 2026 Waterloo grads. The only thing here is it's very one way. So I called it, but I kind of didn't call it in that Zach's just absorbing. He's not saying. he's not turns out in tech everyone's in free agency yeah at all times at all times yes there are no binding contracts that can keep someone at a company at a company for five years it would be yeah it would be very interesting if sports shifted to that model where there was no long-term contracts and you could just mid-season be like oh i guess tom brady's going to the eagles like Like, say, Quan Barclay is being traded mid-game at halftime.

1:15:37He switches teams because there's a team that paid more. That would be possible, I guess. And that's what we're going through here. I don't know. We'll have to get Tyler to give us a breakdown of these folks on tomorrow's show. Maybe we could go through these folks and talk about some of their history, what they've actually done, who's the most legit. I wonder how important the, oh, what's that called? The Dunbar, is it Dunbar's number? No, I forget. There's a quantitative way to assess how powerful, how, why can't I, Jikester's number? I forget. Benchpress? Yes, Benchpress. No, basically you look at the papers that they authored, and then you look at how many citations that paper has.

1:16:27And so if really great researchers are citing your research, then you are a great researcher, that type of thing. It's kind of like a six degrees of separation, like the Kevin Bacon thing. Anyway, should we go to Apple F1? It's also in the Wall Street Journal. Yeah, let's cover this journal piece. Can Brad Pitt's F1 movie finally deliver Apple a big screen hit? It did. It did. This is old news. Also, it's funny because the paper version says, Apple is now trying to make movies people actually watch. Did you see how brutal the written headline is compared to the online one? So the online one is, can Brad Pitt's F1 movie finally deliver Apple a big screen hit?

1:17:15Like, okay, that's like kind of shade at Apple. Like they haven't really had a big screen hit. But then this one is just like, oh yeah, they haven't even been trying before. And now they're trying to make movies people actually watch. Brutal. Well, if you're familiar with F1, you're probably familiar with it from some billboard ads you saw. So if you're trying to run billboard ads, go to adquick.com. Out-of-home advertising made easy and measurable. Say goodbye to the headaches of out-of-home advertising. Only adquick combines technology, out-of-home expertise, and data to enable seamless, efficient ad buying across the globe.

1:17:46It's crazy how – so the – I was off earlier. this is Apple's first box off it hit. It's a$55.6 million debut. And the poly market had it. There was under$43 million,$43 to$47,$47 to$51,$51 to$55, and then over$55 million. So it had it at 71 % just about a week ago. And that ended up being true by half a million. So close, but accurate. Oh, interesting. I thought it was like 140 million. What was that number? That's different? I think I - Okay, so they actually clocked it pretty well with the - Okay. Anyway, we'll get the details here. So in the Wall Street Journal - It was - So it was domestic - Oh, the domestic box office versus international?

1:18:39Yeah. Those numbers? Okay. Well, Ben Fritz and Joe Flint over at the Wall Street Journal have the story. They say, when the team behind this weekend's Brad Pitt movie, F1, first pitched their idea around Hollywood, nearly every major entertainment company wanted in and a bidding war quickly erupted. I like that. Then Apple Bigfooted them all. Tech companies, they love Bigfooting. The tech giant agreed to spend nearly$250 million, more than two AI researchers, more than any other studio. thought the drama about an aging driver's last shot at glory was worth, said people with knowledge of the matter.

1:19:21Pitt was paid more than the$20 million standard for A-list movie stars, and will get a cut of the film's revenue if it's a hit. F1 is one of Apple's biggest entertainment bets since it leapt into Hollywood in 2019 and embodies the unusually lavish and meticulous approach to the business, which has brought little commercial success. People who have worked at or with the company say Apple executives crave hits, but only want to produce content they believe reflects the high-end aspirational qualities of the iPhone and MacBook. There's little chance, in other words, that Apple would make a critically reviled blockbuster like a Minecraft movie.

1:20:04Yeah, I talked to a very successful Hollywood producer about this dynamic with Apple that so much of being a movie studio is it's a hits driven business. You got to slop it up sometimes. You got to do the kids movie like Shrek. You got to hallucinate a little bit. Yeah, Shrek is not an Oscar winning film, but it's a hugely valuable franchise. It's an Oscar in my heart. And so he was basically saying that Apple had one hand tied behind its back because of the brand standard. And that it would only want to do things that had a polish and a prestige. And that were expensive. Expensive. And if you're only doing like half of the surface area of what a movie studio can do, you're just going to get eaten by Netflix.

1:20:50That's going to say like, yeah, let's put the reality show on this show. Let's do this. Let's try this. Let's try some prestige. Yeah, let's do Love Island. Yeah, let's do Love Island, whatever. And Apple just wouldn't do that because it's not the brand. But it seems like they're figuring it out, and F1's a good example of that. So with sleek production values and strong reviews, F1 is exactly the kind of movie Apple wants. It was made by Top Gun Maverick director Joseph Kosinski and producer Jerry Bruckheimer, and Apple executives have privately said they hope it will be the land-based version of the airborne blockbuster, which grossed$1.5 billion in 2022.

1:21:28Wow, you think it could get up there? That would be fantastic. But pre-release surveys indicate F1 is struggling to generate interest among audiences beyond older men. Its success or failure will be a referendum on Apple's ability to meld carefully curated content with broad appeal after six years in which it hasn't released a single box office hit. To date, worldwide box office of Apple original films. Killers of the Flower Moon, did not see that one. 2023, 160 million. Napoleon, that was an Apple movie? Oh, 221 million, I saw that, did you see that? People didn't like it, I thought it was pretty good.

1:22:12It was long, but it was fun. Argyle, 96 million, that one didn't do too well. Fly Me to the Moon, 2024 did 42 million. You have not seen any of those, correct? Nope. Fantastic. F1 will be available. It's okay. I haven't seen UFC 724. Who cares? To each their own. F1 will be available on Apple TV, the company's streaming service, after its run on the big screen. Apple TV Plus. I knew this movie would be somewhat significant. Yep. I felt inclined to watch it. Yes. And so knowing that it was releasing, or I thought it was releasing, I was like, okay, great. I'll watch it on Sunday. There's no HBO.

1:22:56I'll watch it on Sunday. Okay. And what happened? And it wasn't on Apple TV. It was at the movie theaters. Yep. And I didn't want to go to the movie theaters. You're allergic to movie theaters. When was the last time you were in a movie theater physically? When we saw Dune 2 with an absolute squad. Yeah, we did. This is the movie to take the boys to. We got to do it. It's hard to schedule, but we got to do it. This is a call to action for all the fans. Many podcasters are descending on Malibu for the month of July. This is the thing. So maybe we'll get it done together. So yeah, and Ben Thompson's been beating the drum on this, this idea of like Taylor Swift understands windowing better than like Netflix executives.

1:23:41Have you seen this thing? No. So basically like windowing is the concept that when you have a valuable media property, you should monetize it at increasing rates of cost based on time. And so when Taylor Swift comes out with a new tour, you can first go and see it live in person at SoFi Stadium and tickets are like$100, maybe more, like a couple hundred bucks. It's expensive to go see it. Then months later, you can see the movie version of her tour in movie theaters, but it's not on streaming. And so then you're paying$20 instead of$200. Then after a couple months of it being exclusively in theaters, then you can stream it for free if you're subscribed.

1:24:28So it's like, let's call it like an average cost of like$2. So you went from like 200 bucks to see it live in person, 20 bucks to see it in the theaters,$2 equivalent to like stream it on Netflix for like, you know, 10 % of your Netflix spend that month or whatever. And then eventually it'll be like free on TV or whatever. It'll be even freer, maybe she'll put it on YouTube. And so it's this very good way to price discriminate and get the maximum amount of value out of a piece of content. And Hollywood had been very good at this for a very long time. Like you would come out in the movie theaters, then there'd be a big gap for like three months, you just couldn't see the movie at all, no matter what.

1:25:03It was like, you gotta go see it in the theater because if you don't see it now, everyone's, you won't be able to see it for six months. And then and then the DVD comes out or the blu-ray comes out and it's like exclusive just and you have to buy it And then and then after that it goes to blockbuster You can rent it and then it goes on TV or HBO later and like it might not be It might be a year or two until you can actually go and see the movie for free and so Hollywood was really good at that and they were printing money and then then you know The streamers came and disrupted all this and like the binge era started and you can just watch everything for free constantly but Ben Thompson's point is that they kind of went too far and they started giving away things that people would actually go see in theaters for free on streaming and so even this the I think there's going to be a lot of people who are like okay it's an Apple TV plus exclusive like how long do I have to wait to just watch it at home oh like two weeks like okay I'm not going to go see it in theaters but if they window it hardcore and they say look yeah it'll be on Apple TV plus in seven months then people will be like well it's worth it for me to go see it now anyway uh well we have some notes in the chat one what's up michael ginsburg says community note shrek won two oscars owned owned okay okay apologies apologies not only a cult classic but it's an award-winning one it won wait did it win best animated no way because that's actually hard like that cgi or I don't know.

1:26:30Best music or something? And then Blake Dreyer says, just got back from F1 IMAX. Incredible. As a fan of the sport. Very well done. I think you got to see it. I think it's also the type of movie that you want to see. Jason Turner says, it wasn't a bad film. Cool. Nick Kotov says, it was just too long. F1? Too long? How long is it? Over two hours? Got to be 90 minutes. Get me in, get me out. Two hours, 36 minutes. That's so long. Anyone that would put out content more than two hours a day. is just completely out of line and unhinged. We would never put out three to four hours a day. This is a pot call in the kettle black situation, but we're doing it.

1:27:10Yeah, so Shrek, according to Michael, says best animated feature and best adapted screenplay. Okay, okay, I stand corrected. Let's give it up for Shrek. Do we really think Shrek meets the brand standard of Apple? Like, no way, right? My point holds. I don't think Apple - Tim Cook would kill to be able to put out a film of that Caliban. You think so? Well, the Wall Street Journal disagrees with you. They say that they wouldn't put out the Minecraft movie. Anyway, Apple TV Plus has earned a reputation for content that excites critics more than audiences, including the Leonardo DiCaprio, Robert De Niro.

1:27:47Content that excites critics. Yeah. Apple TV. There's a lot of examples of that in media. There's a lot of tech media that's just for the insider, for the media folks. It's journo on journo content. Killers of the Flower Moon, Oscar's Best Picture winner, Coda, literary adaptation, Pachinko, and the spy drama Slow Horses. Yeah, I tried to watch Slow Horses. It was pretty good. Its few mainstream successes include sports comedy Ted Lasso, sci-fi workplace drama Severance, and straight-to-streaming film The George or The Gorge. We're into entertainment to tell great stories, and we want it to be a great business as well, says Tim Cook.

1:28:26He told Variety, an Apple TV Plus spokeswoman. and declined to comment. Analysts have questioned why Apple spends billions annually on a business so far afield from its core consumer electronics operation, even if it's barely a blip for the company worth more than three trillion. The strategic value it brings is sufficiently mysterious for people not to talk about it very much, says Craig Moffitt at Moffitt Nathanson, good buddy of Ben Thompson's. Apple TV Plus costs$10 a month. And yeah, so a lot of people think that the Apple TV Plus production team is to take pressure off of the 30 % Apple tax discussion.

1:29:00Anyway, we have our first guest of the show. Will is in the studio. Welcome back to TVPN. How are you doing? How's it going? Congratulations on the trade deal. Thank you for having me back. What supercar did you buy? Did you go Bugatti Chiron? Or you go F80? I mean, we hear that the numbers are huge right now for folks like you. I mean, I'm taking a plane more. so like okay that was yeah yeah yeah yeah yeah people say you know fly you know we wanted flying cars we got the fact that you would imply that will drives himself yes it's offensive um i mean yeah a lot of waymo for the first time recently i'd like i'd been to like sf but like i haven't been around sf proper that much since the the waymo craze took off and so like that's been like the the car experience really just like sure self-driving cars damn yeah so are you in sf full-time now uh still back and forth new york i'm like kind of slowly starting to spend more time in sf it's on brand you're decentralized you're decentralizing ai it's it's it's perfect fit perfect fit uh anyway did you touch grass at all this weekend or were you locked into the timeline a lot of stuff was happening on the timeline i did i mean i'm in new york right now.

1:30:20I did some New York stuff on Saturday. But yeah, I mean, it's been it's been good. Lots of drama, intrigue as usual. The trade deals. I know I've been tuning into earlier. I know you guys have gone over all the major roster updates. But yeah, it's exciting times. Open AI, open model soon, hopefully, but not this week. Yeah. So, yeah, I mean, the first place I want to start with is the actual trade deals. Do any of these people jump out to you as particularly interesting? Have they been on your radar? Have you met any of these folks? Can you give us more context on how academic are they? How product oriented are they?

1:31:08It feels like this is purely research, you know, beefing up of the research team. And Meta kind of has products solved. Is that the right model to think about this? How can you characterize the shape of the team that Zuck's building? Even there's thousands of really great software engineers that work in AI. Sure, sure. But even kind of getting a sense of how many people was Zuck really going after? Was it 250? Was it 100? The list that came out today is somewhere, I don't know, there's like 20 people on it. something right right yeah yeah i mean it's a great list it's a lot of like very senior senior researchers who know who have like been through some like real product cycles who like i think the goal is people who can bridge the gap and research your product um because like like one thing meta has a strength of but i think doesn't translate in like this era is they have a lot of really good like academic researchers like jan lecun's whole org has a lot of very talented very capable researchers who write important papers but it's the kind of stuff that doesn't really turn into exciting products it's like more betting five or ten years out how do we want to do things eventually versus like going through a pre-training cycle for a gemini or a quad or an opening eye model like that has a different set of considerations and so this to me looks like they want like 10, 15 great people who have been through those research to product cycles in terms of the model being the product.

1:32:49And it seems like they probably, like they got Alex and then supposedly Nat and or Daniel to kind of handle a little more of the, not the business side, but producty business stuff. Yeah, it's a management role. Yeah, yeah. Yeah. What is your take on Jan LeCun right now? How should we think about his position in the AI world? I see these memes of like, he's a non-believer. He doesn't believe in super intelligence and AI God. He's a little more practical. It seems like he's been right about some things in the sense that we're not feeling a fast takeoff right now. So I think you got to give him some credit for that at the same time like there's Ben Thompson's been criticizing on Lacuna a little bit for not driving the Product side of the business.

1:33:47It's like it's one thing to say that hey, we're not gonna go straight shot to AGI God but then the the the Downstream implication of that is that if you believe that we're gonna be living in like B2B SaaS world of AI implementation and productization, then go do that and make sure that the models work really well for business. And so I don't know, what is your take? How would you describe Jan LeCun in his history and kind of where he sits in the organization to like a layman? Sure. Yeah. So I think people over index on thinking of Jan as the meta guy like too much. I think it's a shame that they didn't necessarily have someone in the Lama org who was as visible as him because he was not involved with llama at all like fair fundamental AI research is like a whole other group that they mostly write papers and do very academic work Jan LeCun's work is very academic um and it's really like it's I think he's like I think on one hand he like is right in the sense of like we don't have the full picture yet like we don't have full-on AGI that can like do a human's job forever and I think that's kind of the drum he's beating is like there's more work to do there's more science to do it's not just productize the current set of techniques um in terms of like the end of ai forever um but his goal is not to do the productization like he like the org he's in at meta is really isolated from product um he's there to kind of boost the uh reputation of the org academically um as well as to be able to potentially advise on other stuff that's more productive but his job is not to like go train a business model train a model that is going to be used by businesses and like i think people like that org fair the friends i have who are there like are there because they essentially wanted to be a professor but they want tech company resources and not to not have to teach and that is essentially what that org is yeah does that lead to better recruiting on campus essentially like you can go get researchers because Jan LeCun can come do a really powerful lecture?

1:36:03Maybe, but I think of it more as like the same way that Zuck did metaverse stuff. Zuck is very willing to make 10-year bets. Yeah. And so that org for meta is not about what product they're releasing next quarter. It's what are our 10-year bets. Yeah. Okay. I wonder... What about... Sorry. Yeah, you can go. So MetaShares just hit an all-time high. Really? So the stock's performing. I'm curious from your time at Morgan State. Thank you for that, John. It's incredible. That is incredible. But generally up 5 % over the last week. Wow. How did Wall Street generally sort of like process the like meta AI story?

1:36:49Because now, obviously, at least the market seems generally excited. it may get to the point where we did with VR where people are like, hey, whoa, whoa, whoa, cool it on the metaverse, you know? Well, it's funny because like, you add all this up, even if the$100 million offer number is real, it's like, okay, so now we're at like$1 billion in spend. It's 5 % of the metaverse hole that was burning. And it's like, you know, from a Wall Street perspective, you're like, I don't care about$1 billion. I care about$20 billion Because that was what was weighing on the stock during the metaverse buildup.

1:37:24But this is like AI is so much more productizable. You can make so much more money from it upfront. And it seems like even though it's a big number,$100 million offer, it feels like the spend is maybe less. But I don't know. What do you think, Will? Right. Yeah. I mean, I think the, I don't know the context fully of the$100 million. I know people have all these theories, but like the scale AI numbers we do know. Yeah. And like that, that's a big chunk of their like, I don't know, they don't do those every day. Yeah. But they've done a handful in the past at that scale. Like WhatsApp was 20 billion or something like that.

1:38:03And I mean, it's hard to read into like the charts. Like I think people are generally like expecting that meta will take AI seriously and are kind of happy to see change. whether that is like justified or not. I mean, we'll see.

1:38:23Yeah, it was interesting. The product side is going to be interesting because like they're not a B2B SaaS company. Well, so I want to push back on that. They're a content, they're an entertainment company. And I think that I don't think that people are fully understand the potential of Gen AI around entertainment. Like it gets talked a lot around, oh, you're going to be able to generate an entire movie or generate video games or things like that. And I think that we haven't seen, we've seen some fantastic examples so far, but nothing that is, I think like George Hots was on our show and he was talking about how like basically AI is going to be like having five CIA agents follow you around all the time, convincing you to buy products.

1:39:11and like that is like one kind of dark bowl case for dark bowl case dark bowl case for meta in the context of ai because it's like it's possible that that facebook is already the best ad advertising product in human history like period hands down there's nothing better and then could you make it like two three times better it's very possible yeah i want to push back on the b2b thing because yes, yes, they don't sell B2B software, but I was thinking about it in this terms. Like if you, if you were running a company where your entire, you only had one client and that client was meta and you sold them CRM and infrastructure and LLMs to improve the back office and do, you know, censorship and, you know, and reality checking and looking for bad words and looking for improper posts and recommendation algorithms.

1:40:09Like how much would that company be worth and how much revenue would they be making just from that one client? And I think it's in the billions because it's just such a large organization that it's potentially like just the B2B applications of AI inside of Meta are like, it's a multi-billion dollar like cost center or revenue driver or something like that. I don't know. I think a company at Metascale has already rewritten all that stuff themselves for the most like they probably have some services yeah that they're using but like they like they could they rebuild their own cloud like because they use aws for a lot yeah yeah um can they're not a compute company though and in terms of like platform stuff yeah i feel like those companies once you hit a certain size and you just like have so many engineers that you want everything like google rewrites all their own stuff yeah um microsoft rewrites all their own stuff um Um, and can they, is there another, like a tale of end of that that has not been done yet that they can squeeze some more money out of her?

1:41:07Um, entertainment will be interesting. So one thing that just, I was just reminded of, um, have you seen the Italian brain rot videos? Oh yeah. Love them. Yeah. Yeah. So like, I know people, it's like silly, it's stupid, but there is something there in terms of this like communal narrative storytelling with a level of vibrancy that we haven't seen so far. It's like Looney Tunes. but people just like come up with these where you could imagine meta kind of eating a good chunk of the like video gen market um if they have an answer to like vo3 um where that becomes part of the platform it ties into the whole metaverse thing of like creating stories and sharing these stories with your friends um but instead of like instagram stories it's like okay what's what's the evolution of story yeah have you have you seen higgs field no field ai i think it's an x snapchat team yeah their new image model is basically already at a point where it's creating images that are photorealistic but in the scent photorealistic it looks like instagram photos not like full-time influencer generated it after spending hours trying to create it and you can't tell that it's not real and so it feels like um it feels like meta integrating that kind of thing where it's like yep it'll be interesting to see how this actually plays out because everybody will be able to be a super tasteful creator or like generate these sort of unique styles and i'm sure at that point it'll switch everybody will be like well i only follow organic farm to table influencers who knows yeah yeah i feel like they're uh ai integration so far have all been kind of weird like there's these like instagram accounts you can chat with and then there was the whole uh meta ai homepage disaster where old people were like leaking personal info without realizing it but yeah i guess facebook's always had that um yeah i think it'll they'll have to thread the needle on product in a way that will have to get creative because they haven't like it's been a while since they had a product innovation that was homegrown that really stuck like yeah it is interesting that they haven't tried to create a studio ghibli moment by you're like going i was saying this they should just pre-render everyone's profile picture as a ghibli and just when you open the Instagram app, it's just like, here, we did this for you.

1:43:42It would be incredibly expensive from an inference cost, but then it's just like you could share it and it's just more virality around it. But I don't know. And like even baking that down into a filter. No, it can't be so like OpenAI branded now. They have to go with their new version. I'm saying like an entirely new. But yeah, basically like the filters in the Instagram app should be like style transfer or like fully generative instead of just like color grades. That's clearly the next thing. And they should just do that. But what do you think of Sam Altman's ability to get through this? Like he's had a series of exoduses and seemed to continue to march on the anthropic departures, then SSI and thinking machines like XAI.

1:44:30Like this is not the first time. It's not his first rodeo. It's not his first rodeo. And so there's this whole narrative like never bet against Zuck. He's been down on the metaverse came back He's been down on mobile with the HTML5 thing and he came back from that went native and Super dominant and bought Instagram whatsapp super dominant and mobile And so never been against Zuck, but then also maybe never get it better get bet against Sam Maybe they both win and maybe like the real loser is like I don't know some some other company or something. I don't know I think as long as opening I doesn't totally drop the ball on product or research like they have the center of gravity for the AI world like in the same way that no matter what Android ships, people are not going to switch from their iPhones.

1:45:11Like it's something that I have to go very wrong for people to not see open AI as the winner, I think, or there's the kind of the default. Yeah. I mean, the other thing is the average chat GPT users not waking up today being like, Oh, I can't believe OpenAI lost a handful of its top researchers, right? They're just still using ChatGPT as a Google replacement or they're talking with it as a companion. So, yeah, I think the lead is still very, very real. And I just, and again, it's going to create opportunities for people to say, hey, I want to go work on this product that hundreds of millions and soon billions of people use every single day.

1:46:02Right. And I do think it'll get to a point, especially with 2025 into 26, where there's a lot of product stuff that people seem... I think we're early on products and we'll keep getting better versions of the same things in a way that's kind of predictable. The image models will keep getting better. The agents will mess up less. But we can already use those things and we can start building a group of concepts around that. And the question is like, okay, what are the winning apps? Like, I imagine that like the real-time, clearly kind of thing, OpenAI is going to do their version of that at some point.

1:46:42Other people probably will too. Like, does that become a modality that people really want like live overview on their screen, on their phone? Like, what's the way that people are going to be interacting with AI generally? um because a lot of those use cases like it's they're already smart enough it's not about making the models like way smarter at whatever it's like how do you have the model be useful or fun engaging there's a bunch of paths here going back to the clearly thing where one roy is like right about the ux and like wins the market and the other option is like he's right about the UX but doesn't win the market, and then it's just not the right form factor or whatever.

1:47:29But at least the first two outcomes are hilarious in that Roy Lee ushers in the new paradigm for engaging with language models. I'm curious, is all the stuff around every B2B SaaS player descending on this sort of like single interface like a chat interface that generates software was that predictable to you do you think that's do you think that that's like part of a multi-year trend or is that just fomo i mean like copilot was early it was like 2020 21 like the first gpt3 copilot came out and like that was already one of the early like lm applications that people were interested in at all and then it took until cursor for it to really like i think cursor plus like three five sonnet was when it became a thing that was good enough that people were excited about it um and it really ushered in the trend because people were starting to find it more useful than a toy and like a thing that actually want to use day to day um and so i think like that's one path is like and then the more background agent kind of things are like starting to take off now which i imagine like those will get reliable enough that they're like useful for cranking stuff out um they already are kind of depending on what you're doing yeah like i think yeah we're about halfway through the year what what are what are kind of like the big moments that you're kind of tracking the the open ai open source model could be one whatever meta launches next right i'm assuming they're going to be dark until this new team can can really cook and bring something great it maybe they don't do anything this year but i would assume they they come with with something what else are you are you tracking yeah i mean for me the o3 release in chat cpt was like a pretty like game changer kind of thing where it was like we saw with like deep research that like okay they kind of figured out how to make agents work but it was also just like this one version of an agent and oh three you can kind of get it to be a pretty general agent where it can like do some pretty complex stuff that was kind of new to see from like the geoguessor thing was crazy um and having that as a like vision of like what agi starts to look like i think it's pretty cool of course from the like research open source world there was a deep seek as like the rl craze taking off but like i mean i'm i work on rl so i like obsess over it and like think about it a lot but i do think we're really starting to see these recipes, at least in the broad strokes of like, okay, here's how the LLM thing can go.

1:50:15We figure out what we want it to do. We give it some tools. We set up these environments. We figure out how to evaluate it. And then we can just kind of like let it go. And these things get better at doing those things via trial and error. And so like, I think that is one way to kind of forecast where things are going. It's just like, what are the plausible use cases that people want to use all elements for they want an agent to do XYZ and then how do you make this a thing that you can help climb and I think like you see a lot like there's been a lot of news about opening I trying to like sell their RFT service there's these startups popping up there was a was the RFT service okay yeah and so like they're pushing this pretty hard you're spending over ten million dollars like Palantir opening I will also customize a model for you?

1:51:04So they have like a serious heavy, like a heavy paying customer one, but they also have like a more like in the thousands of dollars service where you are getting to essentially do fine tuning on OPPOR Mini. Yep. And that is a little more self-serve. But they're also like doing consulting around that, the forward deployed thing around that. Is that an important strategy for open AI to kind of stay in the game against open source models like LLAMA? Yeah, but I think also against other thinking machines. There was some report that this is a version of the strategy they're taking. It's like go to enterprises, talk to them about their problems, turn these problems into things where you can create really good customized agent models.

1:51:56and if the that's one potential roadmap for having more AI everywhere having services whose job is to come in whether or not it's fine tuning or not make them into agent shaped tasks where you can then optimize the model have someone craft the model experience for you because I think a lot of enterprises are still like in the spot where they don't really like there's there's not as you see by the like the talent or there's just not that many people in the world or in the market who really understand this stuff at a level where they can go make it happen and so like the open source thing on one hand it's cool that you in theory can go do it but it's also like there's not that many hands who can who are equipped to like go make the thing happen actually fine-tune llama 4 or whatever exactly Yeah, we actually talked to a startup that's doing basically that, like small models for specific business use cases.

1:53:01Like, okay, you just have a ton of CSVs that need to be turned into JSON, and we build you an LLM that just does that or whatever. Interesting. How do you, after the scale news, and then there's a bunch of players popping up, or everyone from Mercore to Labelbox to Handshake, shake a lot of different people competing for that market how do you see that market evolving over the next one to two years i think there are certain people saying scale got out at the perfect time in many ways but clearly there's a lot of at least gross revenue up for grabs right now in the near term yeah i mean i think like the tap the broader sphere of like creating stuff to train models on with humans in the loop to do that curation is going to be important especially for these uh domain specific applications um i think it's going to be less like oh we just need more tokens and it's more about we need uh curation of goals and objectives um because like tokens you kind of hit diminishing returns pretty quick in terms of just like more internet text or more like human written math solution examples um and it doesn't scale super well but the thing but nice thing about the past specification is you can kind of like pour in more compute without necessarily needing more data um it scales much better with compute um whereas we don't really have like a great way of like spending 100x more compute per free chain token other than make a giant model yeah do you think uh llama will stay open source for the foreseeable future we'll see i wouldn't surprise me if they go kind of the google route where like they are still they still do open source they still have the llama brand but it isn't like their flagship thing where they start doing whether it's for internal products or it's they really want to show that they're the winner um releasing closed models uh via api or other services where do they have the infrastructure to serve a closed model?

1:55:20I mean you mentioned that they use AWS so it would be kind of awkward right? Yeah I mean I don't know that they really want to that much I think that it'd be something that comes in a product sure yeah that makes sense or maybe they partner with like you could see them partnering with someone like a CoreWeave or NVIDIA directly NVIDIA is hoping to get into the inference serving game it sounds like Yeah. Do you think so one of SWIX's takeaways from the the Wired article was that like one big development that's coming is when Stargate comes online, OpenAI will have the largest single cluster for pre-training.

1:56:04but it feels like we might be at the end of that game and we might be doing more compute intensive work in a distributed fashion. And so maybe having it all in one place is a little bit less relevant to just, you know, oh, trump card, Stargate, boom, I win. I have the best model by far. How do you think about the impact of Stargate on like the AGI race? I mean, I think the biggest experiment that they're definitely going to do is take a pretty big model. I don't know if it'll be like quite as big as like a 4.5, but like a big model and do way more RL on it than anyone's on the app. Like what does like an 05 level of RL look like?

1:56:47Okay. See how good that is. See what new task that unlock. Do you need Stargate to do that though? Or could you do that just across a bunch of data centers? Because I've heard like a lot of RL is like you're generating, you're doing verifiable rewards, you're generating in a bunch of different data centers you could do it completely distributed and you don't necessarily need a stargate to do that you can do it distributed like it is very inference heavy there is still like a lot of weight updating you have to sync the model across you have to keep sending the training model to your inference workers okay um and so having it co-located certainly makes it easier you sure it's easier to do it in a distributed way if you want to versus pre-training um but it's not like trivial um But I think like a lot of this is just data people are building the data centers without necessarily knowing exactly what experiments they're going to run they just know more stuff is coming but also a lot of this is going to be inference where like they want to be able to do like meta has a ton of gpus they just use for instagram ads um recommendations um a lot of stargate will probably be serving ghibli like i mean that makes a ton of sense like i'm i'm running into rate limits i haven't seen one literally everything like like google it's like how how How do they run out of TPUs?

1:58:01They spend$60 billion on CapEx every year at least. And an O3 Pro too. I mean, I get timeouts on this stuff. Last question. I'm guessing you weren't surprised to see OpenAI leveraging the TPU for inference or reporting to be starting to use it? Was that kind of - I mean, it makes sense that they are considering all options given the friction with Microsoft. like they don't have a core cloud that they're super close with. They're not like there's, they need to both try to lower their costs for inference as well. Like that maybe was a factor behind the, because the O3 had a big price drop that might've been TPU related.

1:58:44I don't know your speculation, but like Google has TPUs and they can serve Gemini really cheap and really well, really fast. if OpenAI wants to compete at that level they kind of have to consider all options makes sense well great day to have you pop on yeah thanks so much for helping out knowing how things are unfolding we'll probably ping you to jump on again later this week stay ready great to have you Will cheers I'll be right back and we got another Will coming in the studio it's Will Day on TVPN And we got Will from Ulysses coming in, not talking about AI, talking about the ocean, talking about boats, talking about the gundo, military.

1:59:30Who knows what we'll talk about? But we will talk about the origin of his company. How are you doing, Will? I'm good, John. How are you getting on? I'm great. Good to have you on the show, finally. Would you mind kicking us off with just an introduction on yourself and your company? Yeah, absolutely. My name is Will O 'Brien. I'm one of the founders of Ulysses. We are building autonomous vehicles to solve the most critical tasks in the oceans. So in practicality, what that looks like, we build an autonomous surface vehicle. It's a 30-foot vessel that launches and recovers swarms of autonomous underwater vehicles.

2:00:08These underwater vehicles can read as well as write. They can sense. They can take maps. They can take images and collect data. But they can also do physical manipulation as well, robotics. They can do kind of whatever you want, whatever you want to task them to. And we make both these vehicles probably about 10 times cheaper than anyone else in the world. So while some people are building maybe humanoid robots as a general purpose robotics platform for land, we're building the maritime equivalent of that. Humans obviously are not evolved for land, for the ocean, so we need something for that.

2:00:40So that's a bit about us. We're based in San Francisco, as you can tell I'm not an American. I'm Irish and some of our founders founding team are Irish and Scottish. And yeah, that's a bit about us. How did you settle? Yeah, yeah. What was the inciting problem set? How did you land on this industry to work on? What got you in the industry? Yeah, good question. So two years ago, one of my co-founders, a crazy Scottish man called Jamie Wedderburn was on a surf trip. and uh he got talking to a biologist working on this plant called seagrass he'd never heard about seagrass before i had never heard about seagrass really before despite growing up by the seaside and uh he became captivated by it and then he like uh you know like any good uh autistic engineer does he goes down a rabbit hole learning about it and telling all his friends about it and telling us about it um seagrass is this wonder plant in the oceans it supports 20 of the world's fish stocks it holds about 20 % of the carbon in the ocean as well.

2:01:44Underneath it, it's 35 times better than rainforest or moving carbon, but it wasn't in Finding Nemo, so nobody knows about it really. So we kind of discovered this, and in learning about it, he realized it's dying off all around the world, and it's basically very time-intensive and expensive to bring it back to restore it. And basically, the methods people are doing to restore it are manual. They're divers going down with their hands, and they're manually planting seeds, and takes a lot of time, takes a lot of money, and it's not really working. And we kind of had this insight then that, like, well, we've mechanized, you know, nearly every kind of form of large-scale outdoor work on land.

2:02:20That really hasn't happened in the oceans yet. Like, nobody's built the tractor equivalent for the oceans. And that really, like, a lot of the kind of tasks in the ocean, whether it be in this, you know, narrow, weird, wonderful domain of seagrass restoration or in the world of defense or in the commercial industry, if you're building infrastructure, we're still sending down tons of divers to do manual tasks that really just make absolutely no sense. They're dangerous. They're expensive. So we kind of started in this world. Seagrass got very obsessed with that and had a very successful kind of first year just doing autonomous Seagrass restoration services for different governments around the world, working with governments in Australia and the US.

2:03:01We did a million dollars in revenue last year and we're on track for over two this year. Congratulations. When I met Will and I heard this initial wedge, I was like, this is so fascinating and interesting. Find the most weird, obscure market that no one would ever think of and just absolutely dominate it and then use that to earn the right to do many other things. And I think I introduced Will to like 10 seed funds in like 30 minutes at one point without getting the opt-in from Will. It was like, people need to meet you. and then within a few days people were committing so that's awesome yeah i learned to scuba dive off the coast of catalina island and when i was i was like 15 when i was doing it and it was um had all this amazing tall california kelp and i went back yeah a couple years ago and you've heard the story there's like an invasive kelp species species that came over from japan i believe on like a shipping container shipping shipping tanker and it completely wiped it out and it looks completely different.

2:04:02And I don't know if it's particularly important to restore the original California kelp. I imagine it probably is in some abstract sense, but it was just a bummer as a scuba diver to be like, it's not what I remember, which was like swimming through this beautiful forest. It's just kind of these tumbleweeds now. Yeah, man, it's those fucking purple urchins. That's right. That's it. What happens, a big way of like that invasive species are transmitted in the ocean is like a ship will like um you know be in in the in the in tokyo and in the port and then it'll fill up its it's like its ballast so it'll pull in water and when it's pulling in water it'll take in fish or urchins in this instance and fill up and then we'll go right to sea and then it'll land in san diego or san francisco in the port and then it'll release them all so this happened as well with lionfish um in the caribbean um in miami and uh uh yeah those lionfish are nerdy they're just like chewing up the coral everywhere um i i went i went calling them once you get i was in in the in the cayman islands uh working there one summer and we were getting paid to like spearfish yes i've heard this yeah i mean that is great yeah that is great that is the good side of the lion fish is that you can just hunt them relentlessly because no one cares about them because they're so invasive and they're like pretty fun to shoot and you shoot them and you can make leather wallets with them yeah yeah i'm making leather water and you're like i'm actually i'm hunting but i'm saving the world yeah yeah but these are like pigs in texas yeah it's like pigs in texas yeah i mean it's actually a problem and and it needs to be like really dealt with but as a as a consumer of hunting it's fantastic oh it's great but yeah these are these are good examples of tasks again that are just like still today it's like our stage of spears going down just like yeah like they're really realistically there aren't that many people that want to go spearfishing lion fishing all day every day forever because like the like yeah a couple tourists doing it once the weekend is not going to do the job so yeah you need automation 100 yeah and like we've yeah so that's the kind of these are the types of this is like you know the same platform that we've built for seagrass it's a service vehicle it's underwater vehicle with a different robotic payload and a different model on board fine-tuned on finding urchins or finding uh lionfish you could you know we could have like a service that just like solving this problem for someone.

2:06:20What's the origin of the name? Yeah, it was, there was, there was kind of a few things. I mean, as most great ideas, it was conceived in the midst of drinking a record points. And, but we were, we were, me and my co-founders were, we were just like thinking about different names for the business. And then we were thinking about that 1997 animated movie, the, called Treasure Planet where these people go on this quest in the submarine to kind of save the world and we looked it up and the submarine in that movie was called Ulysses. And then, interestingly when we got our first office in Dublin, just after we raised the pre-seed in October 24 or 23 and we got our first office and to the left of us was the house that James Joyce grew up in I was sorry, Ulysses has the Irish connection as well which is another reason we chose it but to the left of us there was a house that James Joyce grew up in to the right of us was the House or the Museum of Literary Ireland where the first copy of Ulysses is kept and in the park opposite our street there's a big huge statue of James Joyce with a quote from Ulysses there and that also a bit a week later my mom found a photo of me at home next to a picture an extra James Joyce quote and there's a Ulysses quote underneath it all this kind of happened in the same week so it's kind of like okay we're keeping it now yeah because the nominative determinism here is that i mean the story of ulysses odysseus is like a 10-year grind and it's like absolutely miserable like facing the cyclops and they're getting enchanted by cersei and the sirens and like i imagine that if things if nominative determinism holds the sirens for you or You won't be getting a quick buck in an aqua hire at Meta anytime soon.

2:08:12You're going to be at this for a very long time. I'm rooting for you, and I think you will ultimately be successful. But buckle up. You're going to go on an entrepreneurial journey if the name holds. Who are your biggest Irish tech heroes? What makes the Irish so infatuated with technology? um i mean what makes irish people so infatuated i mean you guys are well aware of the hibernian conspiracy of course as uh as to uh i i don't know what you guys have to say about your uh the the accusations that you know that mr coogan and mr hayes are at the core of this uh cabal of uh irish people in um technology or irish never beating the allegations never beating the allegations what about paki mccormick what about molly o'shea what about all these people with like Irish surnames, what are they doing?

2:09:04Why do they all suddenly have media programs? Why do they all follow each other? Why does Apple domicile all of their cash there? It's a good question. Yeah, you think it's the Rothschilds, you know, running the money splite, actually, Collinsons. Exactly. Satoshi Nakamoto, Nakamoto is actually an Irish name. There's the old father on Nakamoto. So there's all these connections here. It goes deeper. it goes deeper my favorite irish technologist would probably be george bool the inventor of boolean algebra oh there you go foundation for modern computing yeah i didn't realize he was that deep cut uh yeah and then modern modern ones owen mccabe is a good friend as well big fan of one everything they're doing it's very very inspiring um to see to see someone like that and obviously the collisons as well um but i feel i feel the colson's everyone appreciates the calls us i don't think i know people appreciate the great business that the guys at intercom are building and i know you had him on the show recently and we just partner with him yeah he's a sponsor now ring the bell ring the bell yeah yeah you can have in your tinfoil hat uh conspiracy there's definitely a line between us and own and then us and intercom and own appeared on the show He just happens to pay us through Stripe.

2:10:18You know, I had a random question for you that I'm sure you have strong opinions on. Every now and then I'll see a video that's off the coast of Chile. And there's just like just fleets of fishing vehicles as far as the eye can see what's going on down there. Is there any solution to it? I'm sure you've gotten asked that. is is your guys's tech something that could help yeah address that it is yeah and we're speaking to some partners about about kind of supporting on this problem right now i think there's like two aspects to the problem so yeah like characterizing the problem um the chinese are um you know destroying um many oceans around the world particularly in countries where they basically just don't have the money to monitor um their their seas and enforce marine protect areas and these sorts of things so up and down the coastlines of south america this is like the single biggest thing that like say the american navies are spending time on right now west africa as well um is just being um completely the fisheries they're being completely decimated in and around the antarctic as well there's massive krill populations and krill populations are they're this kind of foundational species for a lot of other um marine life in the ocean you take away the krill which is the kind of bottom of the food chain you know it knocks all the way up to the to the whales and and then you lose the whales as well and nobody wants to lose the whales um so that's kind of like characterizing the problem okay the reason the problem exists is okay two reasons naties that basically just can't cover enough area and get maritime domain awareness they just can't they don't have enough assets and you know that's part of the problem and there's like a governance aspect to the problem as well right so okay how does china get away with it well so you know the the people responsible for kind of looking after um you know law on the high quite often is what you'd call the flag state so where the boat is registered all these boats are registered chinese chinese kind of like knows this is their kind of this they use this as what you know you'd call like gray warfare and it's kind of similar with the critical undersea infrastructure you know it's not really officially the chinese state but it may as well be because they're not like prosecuting on it um so uh on the on the technical side i suppose like the problem here is like the again the paradigm of maritime operations is kind of you know is is stuck and it's it's broken right it's you know you could probably characterize it we call it this the heavy tech bottleneck right so basically any mission you want to do in the ocean kind of has like a few um features around it like it generally uh you know relies on these expensive manned assets so these could be airplanes like p3 orions or they could be um you know expensive of national security cutters these big coast guard cutters these 100 foot 200 foot long vehicles and these costs like tens of millions um to you know in capex and then the running cost for them on a day is like tens of thousands if not hundreds of thousands of dollars to run daily and then you use these things as like a platform on which you put sensors um uh or you use it like an unmanned system and that's kind of like what it relies on so it's like small numbers of very expensive assets whereas the way we're seeing autonomy going around the world whether it be like self-driving transfer networking like self-driving cars or drones in ukraine is like we're moving away from these small networks of exquisite vehicles to large networks of small autonomous vehicles right so it's like going from buses and trains to like a network of waymos or tesla robo taxis uh you know on the battlefield it's moving away from fighter jets to like a bajillion near us drones um you know it's like it's like this is like just the way we're going but that hasn't really happened yet in the ocean and the limiting factor we think is that um basically drones are still like crazy expensive in the ocean on my systems they're crazy expensive even if you want to buy you know the the sexiest on my surface vehicle you're still paying hundreds of thousands of dollars underwater vehicles you're spending hundreds of thousands if not millions of dollars and but we haven't had like a dgi yet for the ocean right we haven't had someone that really completely collapsed costs to the degree that you can buy fleets instead of single units so So, again, that's kind of like what we're trying to position ourselves as is like this person who you can buy if you're like the Argentinian Navy and your fisheries just getting obliterated.

2:14:32You can buy like a swarm of surface and subsea vehicles to kind of solve the problem for the fraction of a cost you purchase a mand acid for that works 24-7 with zero opex. How much is the problem that, you know, these governments don't know something is happening or they know? but it's 30 boats that have showed up somewhere and they're just going to fish, you know, and basically decimate all life until they're maybe pushed away by a boat. Like what is I can see the technology angle of getting just like more coverage and knowing more about what's happening in these different areas. But it feels like there needs to be more will from the government as well to just say, like, we're not going to let this happen.

2:15:25Yeah, I mean, the will is like certainly there. It's just they don't they don't have the budget on current assets like it like, you know, like some of these like if you want to buy more of these, like um like boats that they would need or these planes that they need to like survey or or get like awareness on the area like these are like tens of millions if not hundreds of millions for some of these assets and the argentinian navy just doesn't have that kind of money um so they know that in general there's a problem there but they just don't have enough assets out there um i mean it's crazy that like most of the argentinian air force's time is now spent on looking for chinese ships over their waters it's it's um so it's like they just yeah they don't have the money so yeah you really do need cheaper solutions for these types of folks.

2:16:05What other stories are you tracking or kind of problems globally that you think you're I want to hear about like this big shipbuilding bill. Like how does the new budget fit in? What are you tracking in DC? I know you're still an early stage company, but I imagine that the long term has to be figure out a way to get to program of record, work with the DOD, but break it down. How are you actually thinking about that? because it feels like there's a lot of different applications you got to stay dual use you got to stay agile while you're building the underlying tech but what are you what are you looking at in dc yeah like for us on the kind of um government bd side there's there's kind of three plates that we're always spinning right there's like the the dc plate there's the like operator or warfighter end user state and then there's like the acquisition uh teams it's met play so we're always just like in constant dialogue and with like you know different members of these these three and these bodies so yeah the the the most recent um you know appropriation for for defense is very interesting um there's uh many many billions uh going to uh maritime unmanned systems particularly surface vehicles and underwater vehicles i think there's a general recognition um on the hill right now that like you know you know the chinese are producing like hundreds of ships per per year and you know us is producing one ship per year and that like trying to beat them on ship building is like probably not feasible we're not going to ramp up that production that quickly we need to just play a different game and i think they're seeing like unmanned systems is like the the way to that and so um yeah just this week it was announced again and another um nine billion um you know a few months ago as well there was another four billion allocated to this line item so um yeah they're the things that that that we're tracking there's some some existing um program or records that are already set up that um that we're you know in the kind of running to um to kind of to kind of challenge for and uh get a get a position on and uh there's also talk of some new ones being set up for these some of these newer categories of vehicles that haven't been developed yet but yeah out of the three plates that we're spinning right now it's you know generally that you know you you do want to have like um a presence in dc kind of like at this early stage it's not the main focus the main focus is refining our product with the end-user understanding what their needs are and building the link to spec and so that's kind of like our main focus right now and spoken to speak to some some folks in Navy about that and and members of DHS and Department of Homeland Security and other other groups as well on this and so they're kind of our main focuses right now in that front probability P Atlantis what's the probability that you guys find it I think very high.

2:18:51I think it's probably... Very high. You think that Atlantis exists? Yes. I think there's two very good candidates for Atlantis. There is one structure in modern-day Mauritania called the RICAT structure. So the RICAT structure is basically three concentric rings in the middle of the desert with salt deposits between each of them. And it looks man-made. You can see it from space. You can look up the RECAD structure, R-I-C-H-A-T. And the only description we have of Atlantis is actually from Plato. Plato describes it as being this, like, walled city full of gold built with the kind of city hall being at the middle of three concentric rings.

2:19:36The rings were each filled with seawater. and um modern day mauritania is also where um the richest uh the richest uh man that ever lived uh existed so there was this mauritanian king who had um more gold he would have been the i think the you know the first trillionaire um and uh yeah that's here and uh and he uh and and so that that's the gold part of it the recap structure the three rings of that part of it and then there's also these all these like salt striations all the way from the coastline in west africa to this location so it's kind of the theory that there was this um there was this uh apocalyptic event about 11 500 years ago called the related to the comet hitting the earth and atlantis was there there was a great flood and uh this this city was um lost so um that's that's one location the other one is in the is in um the atlantic ocean um which then would lead to some of the first survivors potentially leaving the Atlantis and going to Ireland to set up the next modern day civilization.

2:20:37Okay, we're bringing it back. This is fantastic. Next time we're getting to the bottom of the Bermuda Triangle. But thanks so much for stopping by. This is fantastic. We'll talk to you soon, Will. Great having you on, Will. Thanks for having us, Lads. Cheers. Bye. Really quickly, when we find Atlantis, you know what we're doing? The cure for male loneliness being called Lads. Lads, yeah. What are we doing? What's the first thing we're going to do? The first thing, set up a wander. That's right. You got to find your happy place. Find your happy place. Book of Wander with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, and 24-7 concierge service.

2:21:09It's a vacation, but better. We have a video dropping with Wander this week. We do. I cannot wait to release. And we have a new guest dropping on the stream. We are surrounded by journalists. We're going to bring in Kylie. Kylie did the reporting that we covered for Wired, covering the leaked memo. Very cool. Um, so let's bring her on in. There she is. What's happening? Welcome to the stream. Hello. Hello. My first time. Thanks for having me. First of many, hopefully. Uh, how's the day going? You've been scooping, scooping up a storm? Yeah, it's been busy. Um, just that this morning and tweeting nonstop.

2:21:50And now you guys. Fantastic. Awesome. Uh, so what did your, what did your weekend look like? What's the, what's like break down like a a weekend of of scoops it's pushing a signal messenger to the max um well it was pride weekend here in san francisco so that wasn't uh fun to combine i helped a friend move and then i got the first scoop and ran to pride and dolores drank a bunch of white claws ran home slept for 12 hours and um my editor started scooping again then we're scooping again And Monday, first thing in the morning, scoop it in. Wait, so what, everything's been moving quickly. What was the first scoop?

2:22:30And maybe like break that down. Yeah, yeah. So I think that was Friday. I reported on a research paper that's not being released at OpenAI because it talks about the definition of AGI. And sources have said that allegedly managers were saying that the Microsoft deal and those negotiations were a reason for not releasing that paper. And it's really, really tense as the journal and Reuters have reported those negotiations. Then Saturday, four OpenAI researchers went to Meta's new lab. Sunday, Mark Chen, an executive at OpenAI, sent out a memo saying they're basically, Meta's basically robbing our home.

2:23:09My editor published that one. And this morning, Mark Zuckerberg sent out a list of all his new hires. It's been fun. Huge. So we got to ask, who's your source? No, I'm kidding. Not you guys. You're not going on the record as you guys? Yeah, yeah, yeah. So, yeah, I'm curious kind of what you expect next in this saga. Do you expect, I mean, clearly OpenAI is not seeing Zuck announced his current list and being like, okay, we're safe now, we don't have to worry. But at the same time, it's very possible that the worst is behind them. Yeah, my question, I talked to Will about this, was just like, were you surprised that the memo didn't lean more on the fact that this has happened like three or four times at OpenAI and they've gotten through it every time?

2:24:01Like, I would be leaning on the fact that like, yeah, Dario left. Elon started XAI. Thinking machines exists. SSI exists. We still have 90 % penetration in consumer AI. Walk down the street, ask someone what they use for AI. They're not even going to say OpenAI or ChatGPT. They're just going to say chat. And just leaning into that, hey, we are this compounding dominant force in consumer AI. This is the place that you want to stay and be, as opposed to getting, I think, a little bit more in the weeds of the nature of the deal. It felt like an appeal, maybe not restating the missionary element. Yeah, it was just very defensive and came across.

2:24:42I mean, they were clearly scared. Yeah, yeah. It seemed very focused on like the actual dynamics of like how one of these offers comes together as opposed to just reiterating the strengths. That like here you get to work on the consumer product in AI. Yeah, it was like a lot of messaging around, please talk to us if you're getting offers versus let's rally, let's ride. in there. Let's ride. Yeah, I think it's, you know, a big part of it came over the weekends, like these people are just these executives are scrambling to tell staff, like, for the love of God, do not talk to anyone at Meta and sign any offers.

2:25:23And they're going on this like week long break next week. So I think they were just trying to get ahead of that. But I think that your guys's argument, that's right. I think it's just like a very emotional, high stress time. They're trying to ship two very large models. And Mark Zuckerberg is coming in calling employees directly. And offering millions and millions and millions of dollars. So I think like their response is emotional and defensive because it's like just so high stress, but that would have been a good tactic. And I don't doubt it's coming eventually. How solid do you think you are on the$100 million offer?

2:25:55That's been floated multiple times. And we were kind of digging through the game theory of it. Like in one way, it benefits Sam to get that number out because if someone, if Mark tries to poach someone for 50, they might be like, I've been hearing that everyone's getting 100. Like, come on. I was one of your top guys. I was near the top of your list. And then, you know, in some ways it benefits Mark at Meta because if you're hearing about these$100 million offers, you might be more willing to take that preliminary phone call because you know that they are ready to put money where their mouth is.

2:26:27And so it kind of benefits both sides. Dylan Patel was on a podcast a couple of days ago talking about how they're not really sending$100 million wire with cash on day one, that there's earnouts to these structures and they can claw it back if you don't stay. And so how much nuance do you think there is on that number specifically? Because it's certainly headline grabbing. The Wall Street Journal, I mean, they put it in the special infographic right here. Because$100 million is a nice round number and it reads very well and it works in a headline. But I'm just wondering if you think if you think it's 100 % real or 50-50 or there's more nuance to it?

2:27:06Yeah. So my take is the Wall Street Journal and the New York Times both printed that number, and they have really, really rigorous editorial standards to get to print. So they must have seen something, some high-level sources that proved to them that that number is real. I, for one, have not seen that number. And for me, I would have to see an offer letter with that number. But knowing how tech salaries work, I don't think it's just like a wire transfer. It's you know, going to be golden handcuffs. It's going to be over a certain amount of time. And, you know, the information had a really good story breaking down, you know, researcher salaries.

2:27:38And just imagine if you came to OpenAI a couple years ago at 80 billion, 40 billion valuation and what that is now, they don't really need, you know, all this crazy money from Meta. But it is attention grabbing, attention grabbing for sure. It's also just funny when you when you say 100 million dollars for one person, it sounds like a lot. But if you were to say, we're going to spend five billion dollars to put together a team that is super core to our entire business and win, it doesn't feel nearly as significant for a company that is a multi-trillion dollar company. So I'm looking at the WhatsApp acquisition price and number of employees.

2:28:21I'm pretty sure they paid more than$100 million per person. Because it was$19 billion,$4 billion in cash, $14 billion in Facebook shares, which are probably worth like$120 billion now because this was in 2014. And at the time, WhatsApp was a remarkably lean team with around 50 employees. So, I mean, not far off. I mean, what if they had... Slightly different comp, though, just given that they were buying a product with... Totally. Yeah, yeah, the product is super valuable. But, you know, the researchers that can build the next model and bring Lama to the frontier makes a lot of sense. What else should we dig into here?

2:28:59I think that was it. I don't know if you touched on this specifically, but any specific predictions about what comes next or are you just waiting for one of your little birdies to send you a note? Well, if you're watching birdies, my signal is at Kylie.01. I will be impressed if they get someone of Ilya Suskiver's capabilities. I think that the people who want to build superintelligence are true believers. And you're giving them a value proposition of, we will give you so much money, but you're making superintelligence for Mark Zuckerberg. And I think that some of the best brains that really believe in AGI is this prosperous future or perhaps the end of humanity don't necessarily want to do that for Mark Zuckerberg.

2:29:45So I will be impressed if he gets one of those minds. And he still has about 50 people he wants to hire for this team. So it's just going to keep coming. I think a lot of them are going to continue coming from the top labs. But what the vision is, what Mark's vision is, I'm waiting to see. What about, last question, what about Daniel Gross? It had been initially reported, but he wasn't on the list that you released earlier. Yeah, I noticed that. Neither was the team from Switzerland. I'm going to guess this is pure speculation for Switzerland. They have EU laws that make it strange to employ them right away.

2:30:22But for Daniel Gross, I'm not sure what happened. I saw that was reported by the information with Nat Friedman. I've tried to be in touch with anyone who could tell me if Daniel's going, but he wasn't on the list. So I guess we'll see. Yeah. Very cool. Well, I wouldn't be surprised if you start getting some major offers. I could see the AI researcher space is heating up. the AI journalist, it might be the next domino to fall. Oh, God. We're here to help talk you through any sort of offers that you might get. Perfect. You guys can take a percent. We'll work it out. Fantastic. Awesome. Well, great having you on.

2:31:00And, yeah, excited to follow the rest of your reporting this week. Thanks, y 'all. Cheers. Bye. Next, we have Joshua. From Chai Discovery with a big launch this week. It's always tough to launch when the biggest trade deal in tech history is going down. Not even a trade deal, the biggest poaching. We keep calling it a trade deal. It was not much on the other side of the trade. It was a non-consensual trade deal. It's more of like a draft. Suck is just drafting from the entire league. I take 12 of your best people for nothing. Yeah, he's kind of just building an all-star roster. It's like all-star weekend over at Meta this week.

2:31:41Anyway, Joshua, welcome to the stream. How are you doing? I'm doing great. Thanks so much for having me. Excited to be here. Of course. It's great to have you. Great work launching on such a wild day in tech history. Is the time run really getting completely steamrolled by this story? I don't know. It feels like that. I feel like there's also people on vacation and stuff. But break it down for us. What is the big news? Well, the big news is we've just announced CHI2, which is our latest foundation model at CHI Discovery. CHI2 is actually allowing us to now generate antibody sequences with a success rate that's high enough for us to skip the usual high throughput lab testing.

2:32:24This means that you can go take a target, put that into our model, go generate a set of potential protein sequences, and then test them in the lab. And they tend to work with extremely high hit rates. Previous methods were like 0.1 % or so, so you'd have to screen hundreds of thousands to potentially find something. But here the hit rates are actually upwards of 15%, approaching 20 % of success. Amazing. Break down some immediate applications of that tech across maybe different types of categories. You want to ask about making a more potent testosterone? Maybe we'll get to that next. Yeah, maybe that's the right spin out.

2:33:06TBP on incubation. We said we would never incubate it. I think one thing that we're really passionate about is going after areas that are really hard to do today. So, you know, like this allows us to do things faster and cheaper potentially. But I think what's really going to be game changing is just opening up new kinds of biology that wasn't possible before. So we've got scientists that are like, you know, maybe working for years trying to enable a certain class of drug program. and if we can now use machine learning to jump to potential solutions, that could really just kind of push up the bar of what people need to do.

2:33:41I think it's take a step back. Like, look at what happened with LLMs. Everyone has to be an AI company now to compete, right? What's that going to look like in biotech where if you don't have access to a technology like this and then you start to have these certain areas that are closed out to you, like, what is that going to look like? So I think that that's something that we are really excited to figure out. To benchmark it to the LLM market, do you see yourself taking more of an anthropic path or an open AI path? And what I mean by that is, or do you develop the drugs? Do you wind up being the front end to this and kind of being full stack?

2:34:14Or do you go more like B2B and you have a bunch of pharma companies or traditional biotech companies that are buying a product or service from you? Yeah, it's a great question. You know, I think if we take any lesson from the LLM market, it's that the winners are going to be the ones who are most agile. The technology is evolving at a faster clip than we ever expected. Our company goal for this entire year was to get to a 1 % success rate on this task. And, you know, we're halfway into the year and we're already past 15%. When you have progress like that, it is really hard to predict like how the market is going to evolve.

2:34:47And that's why I think just like being flexible, making sure that we have the best technology, making sure that we have the best way to use that technology as well. You know, it's not just throwing a model over the wall and you get a drug. We really have to make the models work well for the specific applications where they're most important. And we also have to make them much easier to use. We have to get them into the right people's hands. There's a lot of scientists that if we can bring their creativity and ingenuity together with these new models, that's where the magic is going to happen. What's the state of the art in the actual impact of antibodies?

2:35:18The last time I heard antibodies in the news was when Joe Rogan was taking monoclonal antibodies for COVID infection. And that was one of those weird things where it felt like a very modern piece of science. And the narrative around him was that he was not adopting the modern science, but he was kind of on the cutting edge of that fork of the tech tree. And so is our monoclonal antibodies like the most like the latest and greatest or are there other applications that are particularly interesting or just driving value, even if they're even if they're there are indeed like the traditional way? Yeah, so about half of drug approvals these days are actually biologics, protein drugs, so antibodies.

2:36:00Monoclonal antibodies are actually just one type. That's closer to what, you know, we've been exploring in this paper. But I actually think what's going to happen is that people have now started to push beyond that. They're basic like by specifics are all the rage right now. That's basically you can think of it as two antibodies that are kind of stuck together. So you can actually go after multiple targets at the same time or you can actually get like a more potent behavior for single target. You know, it's kind of just like two things in one. So if we can make, you know, the if we can make each one of those chains like easier to design, then, you know, bringing them together is going to be going to be even easier as well.

2:36:31So I think it just like raises the bar for the kinds of therapies that you have to work on because it becomes easier to make those drugs. So that's a monoclonal antibodies are. they were really important in COVID, right? They can neutralize viruses. The way that people found them during COVID was actually went to patients who had COVID. They just like, you know, you would go and just like take someone's blood, for example, and try to find antibodies that are in there and then try to develop that into a therapy. But now we can just make it on the computer in 24 hours and then go test it in the lab and have some results in a few weeks.

2:37:00So there's a lot more work to do to make sure that these molecules will actually work in the clinic and will work well in patients. We have early signs that they have drug-like properties, which is really nice to see. But we've got a lot of work ahead of us in order to really bring that to production, if you will. There's still some more engineering that you might want to do on the molecules that are coming out of the platform today. Yeah. What is your sci-fi vision for the state of drug development 50 years from now? And how would the FDA have to adapt regulations to enable that? Because I can imagine my vision is like you come in to some type of doctor clinic setting, you have some set of conditions, and they're just sort of like auto generating drugs on the fly, specifically to your current conditions, but that obviously wouldn't work in the current paradigm.

2:37:52Yeah, I mean, that would be the dream. The North Star that we have for this space is something that we think of as like zero shot drug candidates. Right now we have these zero shot proteins, right? So we specify target, we get some protein that binds it. But the fact that this is possible makes us believe that someday we'll be able to actually just like have the models think about all the other properties that we need for a drug. And it's possible that the first sequences that come out of the model will actually be good enough to eventually move into patients. And yeah, if you can do that, then, you know, maybe the next stretch is like, can we actually do hyper-personalized medicine, right?

2:38:28Maybe there's something where we take some samples from a patient. We figure out what exact target we need to go after. Then we go and generate the drug. I mean, there's multiple leaps that need to happen. But I can't predict what's going to happen 50 years from now. Like, honestly, I failed in predicting what's going to happen in six months. I thought you're going to be at 1%. Do you use AlphaFold or is it just a separate tech that is kind of complementary but not in your direct supply chain? Yeah, the foundation models are built in-house here. So we don't use AlphaFold. We actually can't use AlphaFold.

2:38:59The latest one isn't available to companies. So we built our own. We have a state of the art model for predicting protein, protein structure and its interactions with molecules. We actually open source that. So we started the company a little over a year ago and then six months in. That was the first thing we did, because you really need like an atomic level microscope if you're going to design drugs. That was our thesis when we got started. But we realized that, you know, that kind of level of the technology we thought would get out there anyway. So we open source that. It's being used in most of the major pharmaceuticals today.

2:39:30I think the community has gotten a lot of engagement around that. But now the models are becoming even more complicated. So that's why we're starting to invest in not just the model layer, but how we actually make those models useful. They're really starting to look more like pipelines. It's kind of like ChatGPT 03. It's not just a single inference. There's actually a bunch of complex tricks that you need to run if you're going to get the best results. so that's kind of how we're thinking about it right now very cool well congrats on the launch we'll have to have you back soon this is fantastic yeah with the rate that you guys are moving I'm sure you'll be back on in a couple weeks so congrats to you and the whole team we'll talk to you soon cheers up next we have John from Campfire coming on the show welcome to the stream John how are you doing

2:40:21i missed this you haven't been hitting the soundboard enough i love it boom welcome to the stream how you doing john can you hear us yes i can thank you so much for having me on today fantastic uh do you want to kick it off with an introduction on yourself and the company absolutely i'm the founder ceo at campfire we're based in san francisco and I was a finance exec for about 15 years and just lived the pain. Thank you. Don't get the unsung heroes. The unsung heroes. Never featured in the all hands, but we're working on changing that here and lived in spreadsheets my whole life and just helping.

2:41:02I don't think they're going anywhere anytime soon, but helping folks take all the transactional accounting, all the financial reporting and get it into software. And so Campfire is the modern ERP for high growth companies to just help them move to an AI powered platform. And so that's how young were you when you realized you wanted to build the modern ERP? I would say it's not something you think of growing up. It's not like, I'm sure some ambitious kids these days think I'm going to build, I'm going to build the next gen ERP, but you'd be surprised. So going into white Combinator, there's about 4000 companies that have gone through it, but they've had hundreds of thousands apply.

2:41:46They're like, you know, no one came to us and said, we want to build a new ERP. They're like, you're, you're like the only one excited about the category. So I would say this was in 2020, 2021, when I really was feeling the pain at a midsize company that I'm like, this is the largest category in software why is nobody disrupting it and that's that was a that was a critical moment for me well it's interesting because it's it's not going to fit the pro like you need to have lived and breathed the problem to know what to build and i think that yc has like the the sort of default type of founder that does that maybe they have a few years as an engineer palantir or they were a product manager at coinbase or something like that it's very different than being in the C-suite, actually living and breathing and having that ability to go and figure out what to build.

2:42:41Give us the latest on the fundraising side. Yeah. We announced today Excel is leading our$35 million series.

2:42:51Here we go. I love it. Thank you. I want to talk to you. You said high growth. I remember I was running a company and there was a big debate over like, oh, thank goodness we don't need to implement Oracle, we get to implement NetSuite. It was still a lot of work. It was a big, big ERP. And so it seems like the playbook is start with high growth startups, be the first ERP and then scale into that and then eventually you'll be able to service the really, really large Fortune 500 companies. But it's fair to say that you're not ready to go into a Fortune 500 company today, is that right? and how do you think about actually growing the business long term?

2:43:33You said it well. We're hiring on the sales team, so we would love to have you. But in all seriousness, we are not ready for the Fortune 500 yet. We do have customers with hundreds of millions in ARR that have ripped out SAP. They've ripped out NetSuite. They've ripped out everything below that. So we are filling the pull-up market quickly. And one of the reasons for the raise is to go make us enterprise ready. And our customers are saying, I don't want to be on an ERP for a few years. right i want to be on it for 10 plus i talked to the cfo at a 10 000 employee company still on net suite they're really scaled into a hundred billion dollar plus market cap on net suite and they bought it at 100 employees yeah and so it's like how do we go how do we get them at 100 and scale with them to 10 000 and and that's really where we're going yeah i mean pretty sharp guy running the company over at oracle so it is a competitive market um but it does seem like there's there's lots of pockets of opportunity because the, I mean, I remember just like the implementation bill for NetSuite was going to be like half a million dollars in consulting or something like that.

2:44:38It was pretty staggering. How do you think about pricing and getting a growth stage company to actually match the value that you're delivering with the money that you make? It's obviously great if you get a big contract up front and they're locked in for a long time and they're paying this massive implementation bill. There's real costs to you if you have to implement stuff. How are you thinking about ramping a customer into a great profit center for you long term? Yeah. First off, we rethought the whole implementation process for the reasons you've stated. You tell your accounting team, we bought an ERP.

2:45:15Everyone's ready to quit. Nobody wants to go through a one-year implementation. And so we've made it where you give us a login to your old system, we go do all the work because everybody says, I just don't have time. Our team is way too lean to do that. So a lot of the switching costs, to your point, a lot of it is just like dollars, but a lot of it is just time. So we have solved for both. But to your point on like, how do we get them into a profit center? We're just very focused on getting them in the door and making people super happy because we know like customers drive customers, like referenceable and one of your sponsors ramp has done a great job at driving referenceability and customer love has really been i think the virality within the finance and accounting community it's been a huge driver of their growth and so we're focused on the same how are you guys uh leveraging ai internally and then in the in the product itself never heard of it irrelevant don't need it don't need it what it stands for i'm not familiar yeah yeah don't need it built different yeah we're doing it all everything's if statements over here yeah i know you know i found i forgot what it was but excel's if statement limit i like over you know you can that's the final boss of erp you have to build an erp after you hit the limit on excel you just unlock a like satya nadella just comes through the computer and says good you've broken out of the matrix welcome to the big leagues writes you a check anyway our very first version of campfire was I built a three statement model that was iframed into like a login.

2:46:51And it was me on the back end running a three statement Google Sheets for customers. Wow. You got to start somewhere. But yeah, we're super excited about AI. So obviously one thing about us relative to the incumbents is like we are investing and putting AI everywhere. But whether it's meta winning the AI talent war or open AI, like we're always going to be on the latest models and we're always going to be multi-model. And so we're going to be inheriting all of the amazing work of all of these amazing AI engineers. And we're of course building out our own proprietary AI models. But the gap between us and the incumbents is growing every time a new model ships, irrespective of a vendor, because we're always putting in the latest models.

2:47:36A few examples, customers are uploading their board minutes in the last meeting and saying, take all of our every row and campfire of every financial data and come up with a story narrative for the next board meeting. And it's it's reading the board minutes it's going through and it's like, here's the five things you should focus on based on all of it. On the accounting side, it's automating a lot of the accounting. But we've built like a chat GPT for your finance and accounting data. It'll write to the ledger. It'll report for you. Customers are saying like, kind of feels like we got a promotion.

2:48:12I'm not in there doing formatting my Y and X access anymore. You know, the charts are coming out through AI. And that's been exciting to me because I live that pain. Guys, I had a diligence report. I had to build the day of my wedding. Wow. That's brutal. Because we were in a bad ERP. So don't be mean. Enjoy your wedding day. Enjoy your wedding day. Last question for me in the news. Xero acquired the American B2B payments company Melio for$3 billion. I think this was announced last week. Two and a half. Who's counting at this point? Can you explain to me what the rationale you think was behind that?

2:48:52And Xero feels like a different tier of product, not full ERP, more accounting suite for small businesses. Why would they need a payments product? Are you thinking about payments as a potential revenue generator in the future, whether it's through Stripe or Stablecoins or maybe you're going to buy a payments firm at one point? I don't know. Yeah. So I can go really deep on this one. So as deep as you guys want to go. But at Invoice2Go, we were an AR company, 250 ,000 businesses. We were powering their invoicing globally in 150 countries. We competed with Melio. We sold to Build.com. Obviously, they're a titan in AP, a billion in AP revenue.

2:49:30And so I know Xero and Melio well. I'll say Xero was really looking to be stronger in the US market. They're very strong in Australia, New Zealand, very strong in Western Europe and other parts of the world. QuickBooks, unsurprisingly, given their headquarters is here, is dominant in the US. Melio brings a strong footprint. QuickBooks also launched their own AP products and Xero needed a response to that. And so getting into the payments game, getting into APAR in that solo entrepreneur category where I came from at Invoice2Go is where Xero and Melio have a lot of strength. We don't compete with either at all.

2:50:10So we're really like a cut of bob. So no competitive angle for Campfire. But I think one that actually no one has really covered is Melio is very strong in indirect. So Fiserv and others, and this is public, Fiserv and others are selling Melio through their channel. So Clover and other systems are out. And so this is a new go-to-market that's becoming really popular. We saw Brex launch embedded. Bill is embedded. Embedded go-to-market is really a way to solve for really efficient go-to-markets. And I think this is a new one for Xero that is a great play. That's really helpful. Thank you for that context.

2:50:51I appreciate it. I always like when people break down other stuff. And congratulations. Yeah, congratulations. We'll talk to you soon. Have a good one. Bye. Cheers, Sean. Up next, we have Tom Lee from BitMine, the first new crypto treasury company, this time on the New York Stock Exchange, not the NASDAQ. We're going to talk to him about crypto. Really an iconic stock exchange. It is. It's the New York of stock exchanges in many ways. Totally. The city that never sleeps, the Big Apple. Totally. If it was a stock exchange, it would be the New York Stock Exchange. Very excited to talk to him. Yeah, so today they announced a$250 million private placement to implement buying Ethereum for their ETH treasury.

2:51:39Yes. The stock. ETH has been one of these. It's up 48 % in the pre-market. Yes. So welcome to the stream. How are you doing? Good. Good to meet you. Would you mind kicking us off with an introduction on yourself and the company and some of the news today? Glad to. My name is Tom Lee. I'm the founder of Fundstrat, which is a research firm. And many of those might know me from that role where I periodically appear on CNBC and talk about markets. I'm also the chief investment officer of Fundstrat Capital, which is the asset manager, and we have an ETF called Granny Shots. That's$1.4 billion of assets now.

2:52:27And Fundstrat as a research firm has 300 hedge fund clients and 10 ,000 RIAs. And today it was announced that I'm taking a role separately from Fundstret as chairman of Bitmine Immersion Technologies. And the ticker is BMNR. And that's to help the company transform its existing treasury strategy to be one that is focused on Ethereum. What was the decision to go with Ethereum? We've seen a lot of Bitcoin treasury stories happening over the past year, I would say. I mean, even going back further with MicroStrategy. But Ethereum is one that we haven't seen that much yet. I just would highlight that BitMine is up 679 % today.

2:53:16So the market is excited about what you're doing. Congratulations. Back to John's question.

2:53:25Well, to me, one of the things that caught my eye is that stablecoins are the chat GPT of fintech right now because it's seen rapid adoption by consumers, merchants, and banks, and now Visa and credit card providers. It's because it really does have a log function benefit over traditional dollars or over the credit card system. and uh it's 250 billion dollars today treasury secretary besant estimated or said it's reasonable it could go to two trillion that's a 10x market you know addressable market change which means it's an exponential opportunity stable coins have to transact on a layer one blockchain and the the blockchain that is considered the biggest and the most secure and the most commonly used for stable coins is Ethereum.

2:54:19So we saw an opportunity to say, listen, let's think about more tokenization.

2:54:28Stable coins is the chat GBT, but it's opening doors to other ways to tokenize assets. And so we thought it would make sense to do a treasury strategy around Ethereum. And it plays a role that the bigger we become, the more important we are to the Ethereum ecosystem. So I think we're actually playing a role in stabilizing Ethereum because we essentially act as a staking entity. And if we sort of think forward a few years, sorry if I'm rambling. Goldman and JP Morgan, as they get into the stablecoin business and Walmart and Amazon, they're going to realize that they need to really secure the layer one that they're operating on.

2:55:07So they want it to be a big blockchain, a lot of validators. Ethereum is the biggest And then the opportunity is that we're kind of like the plumbing of what banks will look like in the future, because they'll be staking Ethereum on their balance sheet or crypto in order to fund and operate their stable coin. Can you talk about the differences between the New York Stock Exchange and the NASDAQ and what it means to be listed on one or the other? Yes. If you look at our press release, what we announced and we put prominently in headlines was New York Stock Exchange listed. And we listed all of our partners.

2:55:50The reason for that is that most people know me as a traditional finance world person that also is trafficking in crypto. We've written about Bitcoin for eight years. We first recommended it when it was under$1 ,000 to our clients, but it's been a slow road. This project, the BitMine Immersions Treasury Strategy, is really a convergence between traditional finance and crypto, right? That's where stablecoins are intersecting. And the partners we brought on, Founders Fund, Pantera, Kraken, Falcon X, Galaxy Digital, are companies that are really straddling the crossover world. And we want it to be on a prestigious exchange.

2:56:33Most of these treasury strategies are trading on the NASDAQ CM, which is like a sort of a junior league of NASDAQ. We want it to be on something prestigious because we want our traditional clients to come in and say, look, Tom, if we're going to be interested in this treasury and learning about why this might be a new bank model, we don't want to be buying stocks that are on a bulletin board. So it was really important to be on a prestigious exchange like the New York Stock Exchange. So I think that that's a real, you know, one of the advantages of Bitmine Immersion is that they invested the time to be uplisted and so that they are on the New York Stock Exchange.

2:57:17What is liquidity like different on the stock exchange versus like, I mean, as someone who's like younger and more, I guess, internet native, like the idea of needing to buy a stock on a stock exchange is a little bit foreign compared to just like self-custodied crypto directly. And then you You have access to buy the major cryptocurrencies on pretty much any app in the app store these days. Like, it feels very accessible. And yet, in the past, we've seen the ETF as a catalyst for price movements in the major L1s. So talk to me about, like, what's actually going on in the market and characterize some of the different buyers and market participants and why they choose one way to get exposure over another?

2:58:10Yes. Yeah. And so let's say I'll take it in two parts. You know, the first is like, in terms of a product to buy, as New York Stock Exchange listed stock, which is what BMNR is, has market makers. It's almost like you have this invisible hand protecting price and the exchange will halt the stock. If it feels one-sided, that's really important because it protects investors. And these treasury strategies need to get institutional investors to sort of back this and become essentially create a two-way market so people can combine and sell these with little price erosion. Now, the second part though is someone say, well, Tom, why didn't I just buy Ethereum ETF or I'll just buy player one?

2:59:03And I think the Ethereum treasury strategy. Now there's another one, which is Sharplink, which was led by Joe Lubin and consensus and that tickers SBET. I think the treasury strategies, especially for Ethereum, can actually deliver better returns for holders because let's say the metric for a treasury strategy is Ethereum per share. Now, if you buy an ETF, your Ethereum per share is going to always be fixed because you're just buying a unit of Ethereum. So you only benefit if Ethereum price goes up. Whereas a treasury strategy, let's say has an Ethereum per share value, then there's three ways that that price of that treasury stock goes up.

2:59:57One is that the company generates staking yield. So you're getting a little income. The second way that you benefit is if the company can capitalize on volatility and Ethereum has twice the volatility of Bitcoin. which means the stock's twice as volatile. That means you're going to be able to fund transactions like convertible instruments or sell stock in the open market, and people are buying it for the liquidity, and then you can increase the net asset value of Ethereum per share. So you're kind of like growing. Someone's actually, if you own a share of a treasury stock, someone's working to get you more Ethereum per share for every share you hold.

3:00:40So they're actually like growing it. And the third, of course, is the price of Ethereum goes up. And that's where I think it's really interesting because like Circle, which is the second largest stablecoin issuer, and it's probably the most successful IPO in like several years. It trades at around 100 times enterprise value to EBITDA. So for every dollar of EBITDA, it's like$100 of market cap. Ethereum has a network yield of four. So in other words, like Ethereum's four times cheaper than Circle, even though Circle runs on Ethereum. So usually, like the more you get into the infrastructure layer, the higher the multiple is.

3:01:26That's why software has a higher multiple than a retailer. So that means like, but if you just said Ethereum should be a 1 % yield, then Ethereum should be like 12 ,000 per unit. And it's only like, you know, 2500 now. So I think that there's the third way is, of course, Ethereum could go up a lot. What do you see as kind of the important milestones in Ethereum over the next few years? We went through kind of the NFT boom. We're now in the stablecoin era. People have been still talking about store value. They're also starting to talk more about prediction markets. What is interesting to you about maybe crypto broadly over the next few years?

3:02:17Yeah, Galaxia. Well, you know, I'd say the story arcs that we see is one, the regulatory landscape has become very friendly. for crypto. Huge deal. The second is that institutional fiat money, traditional, is getting into crypto. 95 % of institutions don't own crypto. So that's a huge addressable market. But then there's technological innovation, as you're pointing out, and there's a lot of great projects because of that. So Bitcoin, it's still our favorite layer one. But that to us is digital gold. It's really the best way to measure trust. And if it ever traded at the value of gold above ground value, it's over a million dollars.

3:03:03So we still see a lot of upside. Ethereum, the unique feature is the smart contract. And we know that the majority of real world assets, tokenized things, including tokenized dollars and tokenized real estate and equities, etc. They're actually being minted on the Ethereum or created on the Ethereum blockchain. So I think as the world starts to tokenize and experiment, remember, there's a lot more equities and real world assets to tokenize onto layer ones. Then the other way is the traditional financial world is trying to equitize crypto through ETFs. There's more to move into tokenizing on layer ones.

3:03:47And the third, of course, is things like solving some of the risks of AI, the whole idea of a decentralized blockchain. I mean, look, let's face it. There is a non-insignificant risk that AI is going to be malevolent. Like there'll be a catastrophe because every innovation has always led to a catastrophe in the short term. Like consumer credit was a technical innovation for consumer access to money. And nearly every country that rolled out a credit card, whether it was Asia or Latin America, had a credit crisis. And then in 2008, we had a mortgage innovation with the securitization of mortgages.

3:04:28And it boomed for a while. But then we had the GFC. So I think AI is obviously going to be innovative. And there's a lot of opportunities. but it's inevitably going to lead to a catastrophe. And I think that's when decentralized blockchains are going to be one answer to sort of managing the risk around AI. And so, I mean, I don't know if BitTensor is a solution, but I can see something running on Ethereum and validators, you know, human validators is one way too. Interesting. Yeah. Talk to you all around about that. Any reactions to, there's been a couple announcements in the last week. Republic came out creating their mirror tokens products, which are bringing private company stocks on chain.

3:05:11And then Robinhood had an announcement, I believe it was this morning, bringing OpenAI and SpaceX on chain. Any kind of predictions around private company assets coming on chain? I love all these experiments because they're trying to provide a hybrid form of liquidity to things that are illiquid, you know, and so it makes sense. And, you know, the answer to that is it is a big deal that Republic and then Robinhood are doing this because you're going to create the probability of a more healthy market. This has to become a two-way market because if someone's just tokenizing assets, but then very few people use it.

3:06:01And what I mean by use it, someone may want to own it, but someone might want to hedge something. So like an OpenAI token isn't just someone's long bet on OpenAI. They might be invested in Grok or something, and then they need to hedge it by shorting the token. Like if you can create a two-way market, then it's definitely going to work. But if the market is purely on like, let's just tokenize it, you buy it and hodl it, then it's going to have the same liquidity problem that it originally had. And then they might even trade it at discount. And then you create other risk because now you're putting it on a blockchain.

3:06:39So now you have the actual risk of the physical asset being missing and then the blockchain being broken. So I think it'll be important for two-way markets to develop around it. Very cool. Well, congratulations. Yeah, congratulations. On the announcement. And we're excited to follow along. Hopefully have you on again soon. Yeah, fantastic. Thanks for having me. We'll talk to you soon. Cheers. Bye. Well, Drew, is there anything else? What else is going on? There's all of our ads. We are out of ads. We need more sponsors. Put more on that soon because we have a new one coming. There's a story on Apple Wang from Mark Gurman.

3:07:18Wang using Anthropic or OpenAI to power Siri in a major reversal. But let's save it for tomorrow. Save it for tomorrow. We do have some other news. Cursor is now available on your phone and on the web. Spin off dozens of agents and review them later in your editor. Andrew says, I joined Cursor to build the future of coding. And yes, as of today, you can code from anywhere with a coding assistant on every web browser and phone. Very, very cool. We'll have to see how. I'll have to give Tyler over there in the VR headset a challenge. Oh, he's back in. He's back in. There he is. What are you doing over there, Tyler?

3:07:53I'm just chilling in the metaverse. Oh, yeah? Yeah. But what specifically are you doing? I'm not playing video games. No? You're grinding? No, yeah. Yeah, yeah. How is the, now that you've had the headset for a few hours, how? It's sick. Well, it's way better. So I have the Quest 2 at home. And I think it's just all the Quest 3s, but they have this, like, back strap. I don't know if you can see it, but it's way more comfortable. It's way more comfortable. It's like totally comfortable to wear this right now. That's awesome. So are you doing pass through? Are you using your computer? I can see you guys right now.

3:08:23You can see me. How many fingers? Four. No way. He's got it. How many fingers? Two. Oh, wow. You can see right through. It's amazing. I didn't realize the pass through was that good. New challenge. Don't take it off for 24 hours. Starts now. Go on a date with it. Use Cluely with it. Yes. Yeah, I mean, this is the Cluely device right here. We need a Cluely app in there ASAP, ASAP for sure. Yeah, you show up there. Why don't you take off your MetaQuest 3S for the interview, actually? I'd prefer to leave it on. Anyway, there's some other good posts we should go through. Apparently, Xfinity is allowing people to use Wi-Fi signals to detect motion in your home at no extra cost and this five seconds account says, I apologize to all my local schizophrenic, schizophrenics, I owe you an apology.

3:09:24I wasn't really familiar with your game. Because it's like - You remember there was a super, there was a video that came out, it was probably a year ago where somebody was showing how you can use wifi signals to like get it. To detect people. Yeah, to get a very accurate map, like a 3D render of what's happening in space. Yeah. and this is 160 ,000 likes. It was almost hard to believe and then now Xfinity for free. No extra cost. Crazy, crazy. The concerning thing here is introducing this technology to more people. They're going to realize, oh, I can, if they're in an apartment building, they're like, oh, I can see when my neighbor's home or a thief could set up a Wi-Fi router outside of somebody's home and start mapping it.

3:10:08See, what's this person's schedule? When are they going to put this on an assault rifle and give the special forces wall hacks? I bet they are. No, no. Somebody was saying the CIA and other law enforcement groups use this already to understand when there's like an active shooter scenario, understand where they are. The positioning of everything. We've got to fly on the set. Yeah. This is going after you. But anyways. In other news, Christoph Thulin Futurist says, I am developing an LLM benchmark test. that I believe is truly the first of its kind. Very excited to launch this soon and hopefully partner with some of the Frontier Labs on it.

3:10:44Congratulations. We'd love to check it out when it goes live, Christoph. So thanks for the hard work. You're going up against ArcGi. So, you know, it's not an easy, easy, the benchmark battle is hot, but good luck to you. And Delian reflects on the progress of AI. He says, imagine gathering Alan Turing of 1942, Marvin Minsky of 1992, and Nick Bostrom of 2012 and explaining to them that we have passed the Turing test several years ago and yes, there are some changes to society, but largely day-to-day life is the same. It would be a funny conversation. It is crazy. The soft singularity is real. Everyone's feeling the soft acceleration.

3:11:28Anyway, any other, I mean, it's just like, yeah, there's just like, yeah, we beat the Turing test and now we're on to super intelligence and then eventually giga intelligence. But that's a great place to end the show. Thanks for watching. We will see you tomorrow. Leave us five stars on Apple Podcasts and Spotify. And have a great Monday. See you soon. See ya. Bye. Have a great evening.