Linda Yaccarino Steps Down as CEO of X, F1's Red Bull Fires Christian Horner | Ben Thompson, Scott Belsky, David Marcus, Nathan Lambert, Richard Socher, Naeem Talukdar

29:41But so does Google. Yep. Well, let me tell you about Vanta. Automate compliance, manage risk, improve trust continuously. Vanta's trust management platform takes the manual work out of security and compliance process and replaces it with continuous automation, whether you're pursuing your first framework or managing a complex program. And in other news, we have Ben Thompson joining the stream today. So we're excited to talk to him. He just released a post about tech philosophy and AI opportunity and he has a very interesting breakdown of the, he creates an X and Y axis for how each company is thinking about technology and broadly AI, is it a tool or is it an agent?

30:25And he buckets the different companies into different, and then how important is it, what's the opportunity to grow versus what's the threat to your business. And so, you know, he shares a clip from Steve Jobs that I believe was like 40 years old. It's crazy how old that clip is, where he's talking about the bicycle for the mind and the idea that Apple is giving you tools that you can use. And then on the flip side, he highlights that Google and Meta are much more thinking about technology as basically agents, it goes back to this idea that the I'm feeling lucky button talking about the goal of the computer just doing something for you, and so that plays into what is the responsibility of the various companies.

31:17If it's a tool, the person using the tool is responsible for that, whether that's copyright infringement or doing something nefarious, but if it's an agent, than it's on the company, essentially, that's delivering that, to deliver a good experience. And so he kind of maps these all out and has, interestingly, he puts Anthropic all the way to the right on the agentic side and puts OpenAI on the left on the more of a tool side, which is interesting. And highlights the difference between, there's this tweet exchange between, about Cursor and Cloud Code saying like, why would I switch and this interaction is revealing that cloud code users are seeing it much more as an agentic tool in the sense that the UI is worse in most people's opinion but it's more likely to one shot the problem which is the agentic idea.

32:12And you can think of the I'm feeling lucky button as the original one shot the problem. Whereas Apple hasn't really had that many pieces of those products and anything that they've built software that's had that one shot. If we've seen anything today, you can have subpar UI, but if you have a truly great product, you can break through, right? Even OpenAI's experience of picking between different models still feels non-optimized and probably not the end state, but it hasn't really hurt adoption. Yep. Let me tell you about Linear. Linear.app. Linear is a purpose-built tool for planning and building products, meet the system for modern software development, streamline issues, projects, and product roadmaps.

32:57In other news, there's so much news today. Mark Gurman's been on an absolute tear, getting scoop after scoop at Bloomberg. He said, as I reported a year ago, hardware engineering chief John Turnus is primed to be the next CEO of Apple when Tim Cook eventually retires. This is very exciting. Very interesting. I don't know that much about their chief hardware engineering, the hardware engineering chief, John Ternus. We'll have to ideally have him on the show. But we'll have to read about him more and start to understand what the succession plan is. What's interesting is like there's been this rumble of like, oh, like Apple's not taking AI seriously.

33:36They're missing. But we've kind of had to take continuously that Tim Cook's actually doing a pretty great job. And if you look at the stock performance, the company's doing very well. They have so many different advantages. They can be a beneficiary of AI even just as a platform, remain the dominant phone. It's not as existential as Google. And even though they're behind, it feels like dealing with a trade war, dealing with tariffs, getting the exemptions there. Yeah, they've got a busy year. Feels like a bigger issue. But at some point, Tim Cook probably will retire. and succession planning is important for a company like this at this point.

34:15It will be interesting to see what the new CEO's comp package looks like because it will be another reminder to, like it'll be very telling to think of how Apple is thinking about compensation in the current era. A lot of people say, well, Tim Cook is paid$74-ish million because there's a lot of people that could run one of the best, you know, basically sell effectively at the end of the day, sell iPhones. Right. And, um, you know, I think there's, there's a lot you could do to debate that point. Um, but, uh, interested to see how it plays out. Yeah. I'm trying to think about like, you know, what, what made Tim Cook a great CEO for Apple when he took over, it was that Apple's supply chain was the number one thing holding it back.

35:06Like it seemed like Steve Jobs had laid down a like incredible vision for the products that were going to be delivering on Steve's near term vision. Yeah. And, and, and the work of Johnny on too. Um, and, and so, you know, you go back to, you know, the seventies and eighties and you find these videos of Steve Jobs talking about, you know, describing the iPad in, in the perfect detail, being like, you'll have this thing that's like a book, and it'll have a screen, and it'll be connected to a network, and you'll be able to do anything you want on it, and it'll talk to you. And Steve clearly saw the future, but then actually marshalling the manufacturing power to actually deliver that was the biggest challenge for the company over the last 20 years, and so Tim Cook was the perfect person for that.

35:57reading into the idea that like John Ternus is potentially taking over as CEO. It means, you know, he's running hardware engineering. So going deeper into hardware engineering is the read is like, let's continue there. They're not saying, hey, we're, you know, we're like, we're lining up to take AI even more seriously and push further into services and push further into software. Yeah, no, to me, it's exciting. The bear signal would be if the CMO was becoming the CEO. And then it's like, hey, we've hit peak iPhone. We're done. It's just about selling as many of these as possible, which in many ways, that is the game on the field today.

36:42But yeah, I think it'll be good to have engineering in the top seat. Yeah, I don't know if it is the game on the field today. How important is their marketing versus everything else that they have going on at the company? Because their marketing seems to be polished and well-run. They're getting impressions across things, and they're positioning the products as premium continually. But whenever they launch these ads, they have to take them down or apologize. And so the actual ads they're doing are not particularly moving the needle for them in a positive way. I'm not saying their marketing has been great.

37:18I'm just more so saying like the signal, like the difference of taking your most senior hardware engineer and saying you're going to run the company now is a dramatically different signal than taking somebody whose job is like the end selling of the goods and saying, you know, now you're going to take the top spot. Totally, totally. Well, speaking of hardware, OpenAI has, this is another scoop from Mark Gurman. OpenAI has completed its nearly$6.5 billion all-stock deal to buy an AI device startup co-founded by Apple's former design chief, Johnny Ive, cementing the chat GPT maker's push into the hardware market.

37:56So this was a deal that obviously had been announced. It was the intention to close here. And I guess as of the last 24 hours or so, it is actually closed. There are some more details here. So Johnny Ive is actually like on a contract where he will spend effectively like the majority of his time working at OpenAI. But he's still loved from his remaining a separate company with still has a couple marquee clients, Airbnb and Ferrari. Oh, sure. So he's, you know, and I think that can ultimately make sense for somebody in that creative, like effectively the role of creative director. Anything to put another node on the corporate org chart for sure.

38:43Yeah. Was there ever any doubt that this would go through? Like this doesn't feel like a crazy antitrust thing, but I guess it was trying to be blocked by that other company, I.O., right? Well, that was just more of a name. That was more of a naming. I don't think they tried to block the acquisition. They would have no grounds to do that. They were just forced to remove any mention of the I.O. branding. I mean, as crazy as it is, this is probably a beneficiary of not being public, right? Because the FTC antitrust regulations are probably a little bit lighter for a deal like this. Yeah, it would be hard to make the case as the FTC without a lot of clown makeup on to be like, this is bad for competition.

39:29Because you're taking one of the best hardware teams in the world and going into a market that effectively is the iPhone. It's duopoly. Duopoly. Yeah. You should love this. So this should be good. Yeah. And yeah, it would have just been hard to push back here. OpenAI had already owned 23 % of IO, going back to a 2024 investment. And so this is effectively buying the remainder and 55 various hardware engineers are joining the OpenAI team. So I'm very, very excited. This is a huge bet from OpenAI. Obviously, it was all stock. Yep. We covered it initially as paying a couple points to get Johnny Hive on the founding team of this new hardware effort.

40:11But I can't wait to see what they build. Well, the next, whatever they build, they're going to need to pay their sales tax. So they should get on Numeral. NumeralHQ.com. Spend less than five minutes per month on sales tax compliance. Go to NumeralHQ.com. Speaking of things that you sell online, need to pay sales tax on. Meta just is going deeper with Ray-Ban maker Eslior Luxottica. I cannot pronounce that first word, but people just call it Luxottica. And so Meta is taking a minority stake in Luxottica to accelerate its smart glasses ambitions, investing$3.5 billion in the iconic Ray-Ban manufacturer.

40:52We were talking to David Senter about the history of this company. It is fascinating. I'm very excited for him to break it down for us a little bit more. Hopefully he can come on the show and talk about it because it's a very, very soon. The founder has a crazy story group. And I think he grew up in an orphanage and, and, and it just, what are they, it wasn't, they didn't call him the pit bull. They call them something else, but he was an absolute savage. Apparently at one point he wanted to buy Oakley and the, and the founder CEO of Oakley didn't want to sell. and so the CEO of Luxottica acquired the largest retailer for Oakley's and just pulled them off the shelf and basically had started selling knockoff Oakley's even though they were trademarked and then eventually the Oakley CEO came around and said okay like I'll sell your cratering you know my revenue let's do a deal wow um so uh absolute dog and uh soon what do you have center to break it down what do you make of this idea that like you know Apple when they make a device, they redefine and very much standardize that particular market.

42:01So when they come out with watches, there are a number of styles of watch. There's the dress watch, the sports watch, the steel sports watch. There's the dive watch. There's the Casio style. There's a whole bunch of different styles, right? Apple comes in and just says, there's only one style, the Apple watch, and they become the number one Apple style. And they give you some variants in the band. I think the band, little stuff here and there. And they were doing partnerships, I think they did an Hermes band for a while, done a couple other things, but it's been mostly Apple's design language on your wrist.

42:34Whereas with the Meta Ray Bands, they're saying, and now the Meta Oakley's, they're saying, you like the look of Ray Bands, we're just putting our technology into the style you like. We're not going to try and create a new iconic style that says Meta like Apple says headphones. and they're just kind of like, they're very, very different strategies. And so it feels like - Well, so I think this is strategic. This doesn't mean that meta can't develop their own styles in time, but I think it's very smart to say, hey, we don't need to innovate on aesthetics and the sort of silhouettes, right? There's classic silhouettes.

43:13Ray-Ban silhouette is Lindy. These Oakley silhouettes are very Lindy. And they're different markets. The Ray-Ban wearer is different than - Luxottica has, I think, Garrett Leight and a bunch of other brands under it. So they're basically saying, through this, we can deliver. Luxottica has brands for every demo that Meta could possibly want as a$100 billion company. And so I think it's very smart. I think Apple, like you said, will probably take a drastically different approach in terms of standardizing around something. And that will say something. but accessories like I wear are just such a personal decision and such an expression of who somebody is that I think that you want to give people max amount of optionality.

44:00Yeah, it's just interesting because you could have said that about watches. Before the Apple Watch, you could have said that, well, somebody who wears a dress watch wants a dress watch. Somebody who wants a steel sports watch. Somebody who wants a G-Shock is a G-Shock. It's like the G-Shock, you say G-Shock and you just immediately think like, you know special operations guy or Jocko Willink listener like that that that it's like a durable rugged thing you say you know Rolex that's a different thing right and and Apple was able to standardize around it and it's interesting that that meta hasn't been trying to do that and instead they're they're focusing on partnership here it's just like a it's just an uncommon strategy but it seems to be working I there's another post in here I don't know if we have it here but I'm trying to think of a new like the key the key thing is Apple's great at innovating at multiple layers, but like generally it's very hard to try to deliver hits in like two specific areas, like aesthetics and design, and then simultaneously in something that's basically a fashion product and simultaneously deliver the technology.

45:04So I don't know. Yeah. Jack Ray here says after wearing Ray-Ban meta Wayfarer glasses for a few I feel kind of naked wearing regular sunglasses. I found three use cases that are hard to roll back. One, spontaneous photos of my kids when we're out and about. Any cool pose that has a half-life of three seconds I can now capture instead of pulling out your phone. Optionality of music or hands-free phone calls without digging around for earbuds. And three, knowledge-seeking chat when I'm walking around, usually for simple factual things. That's exactly what I experienced when I was demoing the Ray-Ban meta-Wayfarers.

45:42It turns out there's more questions I feel like asking when there's no friction. I'm very excited for multimodal and real-time translation use cases too. They're only going to get better. But I think those three are maybe enough. And I think with a lot of these products, just having one killer use case, like just replacing the headphones for hands-free phone calls or something. Like if you can just become someone's daily solution for music, like if that's enough to just sell the product and then sell them another one the next year when it upgrades a little bit, sell them another one, keep them as an active user and roll that out for a long time.

46:17And then if they can do the other stuff, that's great, too. But you just need to get these one really nail their single use case. And so, yeah, there's going to be cool stuff. But it's fascinating to see them roll this out. And it's also interesting how behind the ball it feels like everyone else is now. Like Google was was talking about getting into this this space. We saw some launches at IO. Haven't actually seen any of those in the wild. Haven't seen anyone really talking about those. Apple, it feels like this would be something that they could jump forward to with a stylish pair of eyeglasses with some basic functionality.

46:47Just take what's in the AirPods, take a camera. They could do something cool, but they're just much slower than the market. Yeah, the other thing with eyewear that's different, or that's gonna be a new challenge for manufacturers is that there's so many different situations where I might want to wear something like a Ray-Ban or a J.A.M. silhouette one day, and then I might want to... What's a J.A.M. silhouette? Jack Marie Maj. Okay. But the... And then that same afternoon, I'm wearing Oakley's when I'm playing tennis or something like that. And so there's a lot more like swapping than obviously something like mobile device.

47:29I mean, if they can keep the price low, you could maybe wind up selling to people multiple pairs and have indoor pair outdoor pair it's it's kind of inconvenient i feel like there's got to be a better solution to that but i don't know what it says what are this yeah the bifocals yeah yeah where they can flip down there's transition lenses but those never fully work all the way but then there's the flip down ones clip-ons there's all sorts of different solutions um well let me tell you about addio uh customer relationship magic addio is the ai native crm that builds scales and grows your company to the next level.

48:01You can get started for free at Adio. I am in Adio all day long. Love it. From the moment I wake up to the moment I go to bed. Well, more news, more, what do we call it, personnel news at Apple. Mark Gurman has a story in Bloomberg. Apple chief operating officer Jeff Williams is stepping down in a blockbuster changing of the guard as one former executive who reported to Williams told Mark Gurman, the band is dissolving. Details here, what it means and what happens next. It's interesting because like these personnel moves are so dramatic and yet Apple's such a juggernaut of a company that you don't see it show up in the stock price.

48:46You don't see it in like, you know, if this was a startup, like investors would be panicking and there'd be emergency board meeting. Maybe there are, maybe the board's talking. Secondary prices would be gapping up on the news. You're gapping down. But it seems like they are going through just a second act, a third act, a fourth act. I don't know what act they're on, but they are rethinking a lot of stuff over there. And it's been interesting to see. I think the biggest question keeps coming down to that salary question of if the market is going to sit at$100 million for someone who's two levels, three levels down from the CEO, like what does that mean for the upper ranks?

49:27I don't think it's going to sit there. Unfortunately for all the talented people in the world, I think it's a blip. I mean, I can see it. I just, I mean, private companies genuinely cannot afford to do that, and Apple doesn't have the appetite. I don't think companies like Amazon have the appetite. I don't think. What was Satya's total comp in 2024? That's a good question. I don't know. 79 million. 79. Pay these guys. It's like Mark picked the number to just like needle literally everyone else in the tech industry. It's like it's such a round number, such a viral number. And then such a perfect number to be just a little bit above every single one.

50:16Poor Andy Jassy barely cleared 40 million. Yeah. That is wild. But I don't know. It'll be interesting to see where these AI researchers sit in five years after they vest out and where the market sits and how much actual work there is to be done. It becomes more of like an implementation role, less research, less discovery of novel algorithms, novel concepts. Like maybe the salaries come down. But, you know, I don't know. Dworkhead made a good point when he was saying like just measure the value. And maybe the answer is higher pay for tech CEOs. I don't know. That could be a byproduct. But the main thing is that the inversion doesn't feel it will last.

51:00Meta spending billions to bring on a group of people that were previously at another company. He's effectively doing an indirect IP acquisition. He's effectively doing an acqui-hire. So if you think about it from that lens, it's a lot different than this is the market. the durable market rate for a group of people. Yeah, no, I agree. I think the number one takeaway is that it feels unlikely that AI researchers at Meta will be paid more than every other Mag7 CEO forever. The ratio will not hold. We might see big acquisition deals. We might see tech CEOs of the Mag7, the Mag7 CEO salaries go up.

51:50we might see the AI researchers' salaries go down. But I would not expect in four years or five years that we're seeing Mark Zuckerberg's direct reports, direct reports, making more than Andy Jassy and Satya Nadella and Tim Cook. That would be surprising to me. Anyway, let me talk about FinAI. The number one AI agent for customer service, number one in performance benchmarks, number one in competitive bake-offs, number one ranking on G2. Go check it out. The customer service agent of Anthropic. Oh, yeah. That's how you know it's good. Wow, that is actually a pretty glowing endorsement from Anthropic.

52:25Anyway, Ben Horowitz has some news. He's out of Delaware. Andreessen Horowitz has relocated to Nevada, and they think you should consider leaving Delaware as well. There's been a bunch of this. It's interesting that they landed in Nevada. We've got to dig into this or have somebody on to talk about what exactly they did. Yeah, Texas has been popular, at least for the Elon ecosystem. Yep, people have been talking to Texas. uh at the same time Horowitz has been living in Nevada at least part-time like for a long time totally so like there's definitely roots there that's where their LP conference was so yep they have roots so there's a post here it used to be a no-brainer start a company incorporated in Delaware that is no longer the case due to recent actions by the court of chancery which have injected an unprecedented level of subjectivity into judicial decisions undermining undermining the court's reputation for unbiased expertise this has introduced legal uncertainty into what was widely considered the gold standards of u.s corporate law in contrast nevada has taken significant steps in establishing a technical non-ideological forum for resolving business disputes we have therefore decided to move the state of incorporation of our primary business in ah capital management from delaware to nevada which has historically been a business-friendly state with a fair and balanced regulatory policies so again i think it's like something like 50 percent of Delaware's state revenue is from the C Corp.

53:53I'll confirm this. 50%. That's really high. That's really high. But it makes sense. I mean, there was never even a question when I came to Silicon Valley about, like, where would you incorporate? This was pre-Stripe Atlas, pre-Clerkey. But if you were in YC, it was like set up a Delaware C Corp, there's no question.

54:21But... Anyways, big move. Next time we have Mark or Ben on the show, we should break it down more with them. Well, if you're looking to invest in any state, in the country or any place in the world, go to public.com. Investing, for those who take it seriously, they have multi-asset investing, industry-leading yields, and they're trusted by millions. Joe Weisenthal had an interesting post. I missed this as well. Sorry, it just took me a second. It makes up about one third of the state's operating budget is from corporate license fees, franchise taxes, and entity formation fees. So very meaningful amount of their business, or sorry, of their state overall revenue.

54:59And I would expect them at some point to try to come out and basically try to resolve some of this tension, right? Because if you have Andreessen leaving, advising portfolio companies to leave as well, you have Elon, you're starting to get some, you know, very, very influential figures that are just broadly, you know, advising all of their different investments and new investments to get out. I also wonder the breakdown of that revenue, because if it's like millions of companies paying$100 a year, like a couple big companies leaving, like Andrews and Horowitz or Tesla, it's not gonna really move the budget.

55:43But if it's some sort of like tax-based percentage of revenue, percentage of earnings or something where the bills get really, really big, I feel like even though I've operated Delaware C-Corps at like significant scale, I've never run into a situation where it's like, oh wow, we're paying Delaware like tens of thousands of dollars. Yeah, it is. It is. So I think it's probably like a lot of small companies. And so we need to be like a real crazy tidal wave that just like continues forever and like a long time. Because like I imagine that the vast majority of the revenue comes from companies that have been incorporated for like over 10 years, are not going to move, don't care, are just completely fine.

56:19Because the real disadvantage to being in Delaware seems to come from when you're doing like crazy aggressive moves on the corporate side, like crazy stock based packages based on incentives. If somebody's spinning up like a design consultancy, they're not worried about, oh, the Delaware court of chancery is going to come after me when I try to do this reverse merger or this crazy stock compensation package. The benefit of being in Delaware was always that there's so much case law there that you can just rely on. Any lawyer can give you advice that holds very well because they're like, yeah, we've seen this exact situation dozens of times.

56:59There's nothing crazy about this. Like if you fill out this paperwork or use this form, everyone will understand what's going on. Yeah. So in 2023, there was 300 ,000 new formations, 220 ,000 LLCs, and 60 ,000 C corps. So pretty meaningful amount of C corps on top of over 2 million entities. And then the franchise tax starts at basically has like a minimum$175 to$400, but then it's capped at$200 to$250 ,000. Well, if you're trying to get the attention of Ben Horowitz, got to buy a billboard in Las Vegas. Every billboard in Vegas. Go to adquick.com. Out-of-home advertising made easy and measurable.

57:41Say goodbye to the headaches of out-of-home advertising. Only Adquick combines technology out-of-home expertise and data to enable efficient, seamless ad buying across the globe. And we have our first guest, David Marcus, in the studio. Welcome to the stream, David. How are you doing? Good to see you. Great, great. Thanks so much for stopping by. Would you mind kicking us off with an introduction on yourself and the company, just for those who might not be familiar? Of course. I'm David Marcus, co-founder and CEO of LightSpark. And basically, we're building modern payment infrastructure to replace correspondent banking with a fast open network built on top of Bitcoin.

58:20Can you talk to me about some of the history of your career and how that ties into bringing you to today, the decision to build it externally as a new company versus internally at some other company. Sure. So I've been building companies since I was 23 years old, came to America in 08, built a company that I ended up selling to PayPal through a series of twists and turns, ended up running PayPal for a number of years. Then Mark Zuckerberg convinced me to join him to take a little break from payments and regulated businesses, building messaging products at Facebook. and then started the Libra DM project there, which unfortunately failed because it was the wrong sponsor at the wrong time and too centralized.

59:11And the goal of Libra and DM was really to provide an interoperability layer for all of the banks and wallets that is real time, that looks like the internet, that is open, that is low cost, and that enables anyone to move money like you send a simple text message or an email. and sadly Janet Yellen the morning of June 2021 pulled the plug on this and so I left at the end of December 21 and in April of 22 started light spark with the benefits of all the learnings that I had collected with my team along the journey and and the one lesson was if you're trying to build something that truly looks like an internet for money it has to be built on top of something that's unassailable, decentralized enough.

59:55And Bitcoin happens to be the most neutral form of digital money ever invented. And so we're building technologies around that to ensure that you can move any currency at any given point in time from anywhere in the world to any other parts of the world at a fraction of the cost of the current system. Does it ever make sense? You mentioned like wrong sponsor at the wrong time. It feels like with all the news around stable coins and the Genius Act and the market structure bill, there's potentially this idea that it's the right time to build new stable coins and stable coin projects and just building crypto generally.

1:00:32But I'm interested to know, is there ever a situation where a big tech company or a bank or Visa or a different network might be the right sponsor? Or should it always be from an individual new company or should it always be should there be a particular structure to that where the actual project's decentralized or c-corp like like how do you think about the shape of the sponsor these days well i think you know stable coins are definitely booming right now and uh and everyone's building on it i think the real question is do you want something that's fully centralized again like the current payment systems and a current financial infrastructure or do you want something that looks and behaves a lot more like the internet, which is an open network that is permissionless, that enables developers and builders from all around the world to build applications to move money.

1:01:23And I'm definitely squarely in the latter camp. And I want to make sure that we have an open money network, an open money grid. And we believe that the only way to build that in a sufficiently decentralized way is to build it on top of Bitcoin and then build all the services and capabilities so that Bitcoin can actually move fast and cheaply, but also interconnects with all of the domestic real-time payment systems in the world and supports stablecoins. So that's what we're building at LightSpark. And that's why you're seeing now more and more digital banks around the world adopting this new standard that we open source that's called Universal Money Address that is basically like email for money.

1:02:01You would have like dollar sign your name at the bank or wallet, and then you can send whatever currency to whomever you want in the world, receiving another currency in real time on a weekend after 5 p.m. on a bank holiday anytime you want. And that's one part of the solution we're building. And then we're also building core infrastructure for Bitcoin to move faster, which has sparked this new Bitcoin L2. But we can talk about that later. Yeah, yeah. I'm definitely interested in that. Like walk me through some of the history and the current strategies to improve Bitcoin, the Lightning Network, and kind of bring us up to modern day in terms of just getting more out of what is undisputably the most successful crypto projects of all time, Bitcoin, but has had gaps where people have stepped up and created different L1s and different projects that haven't touched Bitcoin for a variety of reasons.

1:02:56Yeah, I mean, you're right. Bitcoin is by far the most successful digital asset ever created and network ever created. But the problem with Bitcoin for a long time is that it was too slow, not expressive enough in the sense that developers couldn't really build on it because there were no smart contracts and too expensive to move. So very secure, the most neutral form of digital money ever created, but slow, expensive, not programmable. And Lightning was a first attempt to that and we we spent the last three years building on lightning making it better and making it enterprise grade this is why coinbase and many of their larger largest exchanges in the world are using our technology to move bitcoin faster uh and cheaper on top of lightning but then lightning has a bunch of limitations like the first limitation is the channel-based payment system and so with that comes a lot of complexity in terms of liquidity efficiency routing and all kinds of fun other setbacks and issues.

1:03:58And so we built Spark, which is a brand new Bitcoin L2, which is taking off like crazy right now. And it's actually backward compatible with Lightning, but it enables all kinds of developers to build in a permissionless way, unhosted self-custody wallets, issue tokens, issue stable coins, create marketplaces and AMMs on top of Bitcoin for the first time. And that's really critical because we need to make Bitcoin the most efficient settlement layer in the world and also make it the most decentralized and trustless layer in the world to move value. And we think Spark is a massive step towards that.

1:04:38How are you thinking about adoption of Spark and incentivizing adoption? And I'm more interested in kind of like what are the pitfalls and kind of like traps or potholes along the road to adoption that you don't want to get caught up in. Because in some kind of counterintuitive way, people always talk about Bitcoin's volatility. It's been like the least volatile, really. and it feels like there's a lot of other projects that have kind of come out and created some viral sensation, some sort of massive incentive where a lot of people are making a lot of money very quickly, so it skyrockets, but then there's something that's missing fundamentally and then the projects or the trend kind of dies off.

1:05:21How are you thinking about engineering the incentive structure for long-term adoption? Well, look, the best incentive you can build is build a product that solves real problems. And I think, you know, unfortunately in crypto, sometimes that was replaced with, you know, a coin that some people call shit coins that are basically dumped on people to incentivize them to use the underlying technology. We don't have a shit coin like our coin and our unit of accounts for the network is Bitcoin that we don't control, thankfully. And so we have to work really, really hard to build utility that solves real world problems for many players.

1:06:03But in Bitcoin, there's a lot of liquidity because as you said, it's like the most successful digital assets ever created. And the most liquidity sits in Bitcoin, denominated in Bitcoin. And so you have a lot of traders out there. You have a lot of platforms that actually want to tap into that liquidity. And they couldn't until now because there wasn't a technology that enabled developers to build those marketplaces and platforms on top of Bitcoin to create these self-custody wallets that could move Bitcoin in real time at low cost and to do all of these things that Spark enables. And I think that's why we're seeing such an amazing early adoption and so much traction with a wide variety of developers, whether they're building payment apps, they're issuing stable coins, they're building marketplaces and AMMs and decentralized trading platforms.

1:06:54We're seeing all of that happen on Spark right now. I have this like kind of funny counterfactual that I like to run in my mind of the PayPal mafia and the PayPal diaspora is like so dominant in tech everywhere. And politics and everything and science and just literally everything. And I like to run this counterfactual of like what if the team stayed together forever? Like what would PayPal as an entity look like? Would it be just the biggest company in the world? If you had Elon and Peter Thiel and David Sachs and Keith Raboy and you and all these like the larger crew still there just like chopping down problem after problem after problem in the financial system and I'm wondering if you've ever played that that that you know mental exercise, but more precisely is was there something structural where PayPal was not able to jump headfirst into crypto as fast as possible like if If at the time everyone in the company had just been, as soon as the Bitcoin white paper comes out, like we are orienting the company around this, would it have been possible to actually lean in and be a leader and an innovator in that category?

1:08:10Or was it sort of like an innovator's dilemma problem where, I mean PayPal is doing great still, but where there was really no way for the structure of PayPal to play in the new paradigm? Yeah, I mean, look, I think it's an era thing. The era of all of the people you mentioned was predating all of this. And then the company sold to eBay and most of these talents were completely gone, building their own things. And so I think it's an era thing like, you know, the era of, you know, the Peter Thiel, Elon, Max Lefchen, PayPal was really, you know, the first really big push into, you know, consumer facing fintech.

1:09:00I think of like Visa as a fintech that's like basically built technology that enables money to move around like at an enterprise level serving banks, not consumers. And PayPal was kind of the first major fintech success. But it was really not the crypto era at all. So I think it's just a time thing. The next time PayPal goes through a leadership change, we need to have an all star game where we bring back the entire PayPal mafia for one quarter. Peter, Elon, everyone's full time for a full quarter. Just how hard can we go with PayPal? Let's make it great. It's fascinating. I'm curious what your conversations are like with people that are maybe just starting to build on crypto rails, excited for the first time or have been building for a long time.

1:09:50when talking with you and understanding your vision, it's very easy to see why Bitcoin is an obvious choice for a network to build on top of because it's decentralized, it's very global, but the default until now has been building on Ethereum or Solana or these other networks. How do those conversations go? Are people coming around to the idea that, yes, having a fully decentralized, network that no one individual or group has over influence on is a great place to build a business, right? Or bring your business. I mean, look, I think the only reason that all of this developer energy went to all of the other platforms is because you just couldn't build on top of Bitcoin.

1:10:38The tools weren't there. The technology wasn't there. And I think, you know, now, like, I feel like we're going to have a renaissance of developer energy on top of Bitcoin, because not only because of Spark, by the way, like there are many, many others that are building new capabilities on top of Bitcoin that are enabling those new use cases to happen without losing the true north of decentralization and trustlessness of Bitcoin. And I think it's very compelling for a lot of developers and a lot of people who want to build very successful companies because the again, like the depth of liquidity of Bitcoin, the desirability of Bitcoin in the world has just no parallel in the whole industry.

1:11:20So it was just a matter of removing the obstacles that were standing in the way of developers building really great products on top of Bitcoin. And I think it's happening right now. How do you guys solve the immediate, from what you've said, obviously you can have stable coins on top of the Bitcoin network, which solves one of the key issues with Bitcoin as a method of value transfer that also historically held it back is one, why do I want to buy a coffee? Anybody that bought a coffee 10 years ago with Bitcoin or pizza or whatever it was, probably regrets it now. But how are you, how do transact, That's terrible.

1:12:06Is it, in fact, you know, how are these, how are the, let's say I'm transacting with stables on Spark. What is the sort of like economic, what does that full economic exchange look like? What are fees paid in? Is it paid in the stable or is it network level? I'm so glad you asked this question because that's one of the killer selling point of Spark for stable coins, which is like, you know, if you issue a stablecoin on Ethereum or an EVM chain or any other chain, you have to pay gas fees for, you know, basically transaction fees in the assets that most people don't own, like whether it's ETH or Sol or whatever it is.

1:12:46In the case of Spark, you actually pay when you move stablecoins in the stablecoin. So it's very much focused on payments. And, you know, that's one of the advantages. So one, it's cheaper. Two, you pay with the asset you're transmitting, which is kind of the way that most payment systems at scale really work. And then you have the beauty of the trustlessness of knowing that even if you're transacting with a stablecoin, you can always have a unilateral exit to Bitcoin L1 with your stablecoin, and no one can prevent you from getting your money out. So it's the best of trustlessness, the lowest cost, the most efficient.

1:13:22And you don't have to complicate how you actually pay for the fees for most people who actually don't own the underlying asset needed to pay a fee. So it solves a lot of problems to make Bitcoin the absolute best platform for stablecoin payments. So, yeah, how did the actual dollars get custodied in that way? There's always this like hard interface between something that's truly decentralized and then like the U.S. Treasury at some point. And I feel like that's where a lot of the stablecoin companies kind of figured out how to, you know, bridge that gap. And they exist as this layer between the U.S.

1:13:59government effectively and the crypto community or like the programmable money world. And so it feels like we're on this trajectory of like, let's make this more programmable. But how close are we to something that's like fully programmable? Well, I think, you know, here's the issue, right? So, I mean, first of all, like stable coins are always going to be fully centralized. And so, you know, that's why it's so important for the network not to also be fully centralized, because then we're basically replicating the entire payment system that exists today with just new players. Circle can, is it true that Circle can just freeze all USDC?

1:14:37Yeah, sure. Yeah, of course. I mean, it's a company running, it's basically fully centralized. So like all of the stable coins are centralized. There are a bunch of people who attempted doing algorithmic stable coins that would algorithmically basically absorb like the, and it just doesn't work. It just doesn't work. Exactly. It doesn't work. And so stable coins are fully centralized. I think programmability always comes at the cost of trustlessness. So like the minute you can establish new conditions for how money can be moved with a smart contract, you lose the ability to have a full unilateral exit where no one can actually prevent you from exiting your fund from the network if you really want a trustless exit from the network.

1:15:23And I think that's the balance. And I think, you know, what we're focused on right now with Spark is really providing people with the right level of trust and the differentiating factor that Bitcoin can bring with trustlessness. But I think gradually what you'll see is different levels of trust for different levels of functionality. If you want more programmability, you'll have to actually relinquish a little bit of that expectation of trust to get more programmability. But those two things will always be in tension. How are you balancing go-to-market right now? I imagine you have this pretty intense tension between, you know, for example, like a developing country that's excited or, you know, companies in a developing country that's excited about the potential of Spark versus a Fortune 500 CEO that's saying, David, we want to do something in stables.

1:16:14Like, let's talk. Where are you splitting your time and where are you most excited? Well, I mean, right now, I feel like there are two parts of our business, right? It's like one part is like core infrastructure to make Bitcoin better, faster, more programmable, better for developers. That's all Spark. And then there's the mission of connecting all of the banks, all of the wallets, all of the payment networks in the world to Bitcoin with Universal Money Address or UMA to enable people to actually move money from their bank or from their wallet, the place they pay their bills from, the place they have a debit card or, you know, all kinds of different instruments attached to.

1:16:52to any other point in the world, making basically money flow in a completely open, unrestricted way, 24-7 at a very low cost. Both of these things are basically built on top of Bitcoin and serve different types of constituents that basically extends the reach of the network. But these are the two core focuses. And it's a very interesting time for us because on one side, we have permissionless building on top of Spark with developers building all kinds of different things like i i turn on my my computer in the morning or my phone and i look at like what people have built the night before i have no idea what's going on i don't onboard the business i don't have a contract with them like it's a wonderful thing and then on the other side of things i'm going through like diligence uh processes and compliance stuff with like the largest banks in the world that are coming onto the network so it's kind of a little schizophrenic on both sides of the business But both of these things accrue to the same thing, which is an open money network that enables both developers on one side of the spectrum and regulated entities like banks and wallets on the other side to move money in real time globally like never before.

1:17:58So it's kind of a fun two-sided part of the business right now. That makes sense. How are you thinking about corporate stablecoins? There's been some announcements, different PR stuff around companies saying, we're going to make our own stablecoin. And we were joking on the show a while back. Does that just turn into like a Kohl's Cash scenario? Do people want every big retailer that they interact with to have some native stablecoin or the issuers that we have today? It's a classic problem. There's too many standards. We need one standard. And then you have one more standard than you had before.

1:18:41I mean, look, when I think about these things, I always come back to one thing, which is what problems are we trying to solve? And I think, you know, it's very clear that like if you're trying to solve for dollarization in the world, like if you're in Argentina or in Venezuela or in Turkey or in parts of Africa, you'd much rather have a dollar denominated account with a U.S. bank. You can't have that. So a stablecoin, in this case, mostly Tether, is the solution to that. It's like it's the next best thing to having a U.S. dollar denominated bank account in the U.S. And it's great. And it increases the reach of the dollar.

1:19:18It's great for America. It's great for these people. It works. When it comes to domestic use cases for stablecoins, there's a bunch of really good problems to be solved in institutional capital movement. It's like, you know, you can't net settle trades on weekends or after hours. You can't move liquidity between institutions to make the market more efficient with the current system because it doesn't allow you to do that. Stablecoins help do that. But from a consumer standpoint, I'm kind of at a loss to understand what an American consumer would actually get from using a stablecoin. I don't get it.

1:19:52I don't think there's a massive problem to be solved. People can pay one another pretty easily with normal dollars. They're already digital dollars, basically. Like if you look at your Venmo balance or your Chase balance, it's already a kind of stable coin that you're seeing. It's like virtualized dollars that are controlled by the bank. It's like basically a stable coin. So, you know, I think this is a conversation worth having around like what is the consumer application using stable coins in the U.S. that is actually solving a problem for the vast majority of Americans. And, you know, I don't know what that is.

1:20:28Can you talk about open source and how that, like the current meta around open source in the crypto community and in just the role of decentralization? Was it ever an option not to have an open source project? It feels like kind of table stakes now, but do I have that kind of correctly in terms of the characterization? Yeah, no, I mean, it's super critical. And that's why almost everything we build is open sourced. And the reason for that is like, you know, people building in this industry are trying to make it like, you know, anti-fragile. And one of the ways that you make a technology anti-fragile is you don't concentrate all of the capabilities around one company that wins it all.

1:21:14And I often talk to my team here at LightSpark and basically tell them, look, we're going to be very successful the day we have a bunch of competitors building on Uma, building like all kinds of services to compete with us on the very technologies that we've helped build. And I think that's the way that we make ourselves kind of redundant and ensure that the network is actually going to exist, even if we were to disappear for whatever reason. And I think that's kind of an ethos of the entire industry that we care deeply about. Can you break down a little bit more how UMA works? I think anybody that's played around with crypto a little bit may have had the experience at some point of like sending bitcoin to a usdc address and realizing that it's just gone forever um so the idea of a universal address that can you receive and send a bunch of different uh currencies makes sense but um i'm curious how it how it works and how you're enabling other other companies to adopt it as a standard uh even if you guys aren't necessarily sounds like directly benefiting financially from that i mean we are for for the the companies that we serve we we are benefiting financially from that but like it's an open network but the way it works is that uh an institution like take new bank which is you know one of our early partners on uh on uma which has over 100 million bank customers in latin america uh and so they would assign you an address like dollar sign your name at new bank uh you your accounts is denominated in Brazilian reais.

1:22:47Let's say I'm here in the US and I'm with a bank. My UMA is going to be dollar sign David at bank.com. I'm sending dollars to you in Brazil. And what happens in the backend is basically UMA is a pre-transaction open messaging protocol. So So it enables me to go to the new bank server and basically say, hey, is this address a valid address? What is the currency that this person wants to receive? What is the exchange rate that you're going to charge me for this transaction? Then I can present a fee structure to the customer in the US sending dollars, show them exactly the amounts that the recipient is going to get in Brazilian reais.

1:23:32They click send. When they click send, basically the dollar gets converted into Bitcoin, gets pushed on Lightning to Brazil, gets to Brazil a second later, gets converted to Brazilian RIA, is deposited in the account. Works 24-7, super low cost and super tight spreads between all of these currencies because Bitcoin has so much depth of liquidity with all of these currencies because it's traded so much. And so it's like super cost efficient, real time, 24-7 and open. In some cases, like NewBank, they build the connectivity into Bitcoin themselves into Lightning because they can actually do the conversion on their side.

1:24:10In some other cases, we build the capabilities for, for instance, U.S. banks and European banks and Mexican banks and others to actually connect onto the network using their domestic payment system. So they would send, in this case, the dollars to us and we would convert to Bitcoin and push as a service to them so they don't have to deal with the Bitcoin portion of it. But Bitcoin is always the net neutral settlement asset between those currencies and allows to move liquidity across countries, across payment systems in real time, 24-7. That's the way it works. Very cool. Can I get an update from you on what's happening in Washington?

1:24:46can break down the different legislation that's going through the US government, kind of what your perception has been, your takeaways, status update, but also are you optimistic about where things are going on the regulatory side? Yeah, I mean, look, it's for a guy who's being shut down by the Treasury Department with, you know, the most public shutdown of the crypto industry, one could argue, it's quite a change, quite a vibe shift, right? And, you know, I was at the Digital Asset Summit at the White House and, you know, being welcomed at the White House in the East Wing in a very ceremonial way to actually promote the whole industry was the massive, massive whiplash of the best kind, right?

1:25:34And I think, you know, look, there's a lot of credit that goes to this administration, to David Sachs, Bo Hines, but also to people on the Hill who've been working on these important pieces of legislation that are going to actually make building the next set of technologies that will reinvent and rewire the world's financial system here in America, which I think was absolutely, absolutely direly needed. And so I'm super bullish. I think we have gone from an administration, a government that wanted to fully kill the entire industry to one that wants to promote it and ensure that American companies actually win at this and we win.

1:26:18And I think it's a vital interest for America that we continue to lead with financial services infrastructure. So I couldn't be more bullish of what's happening in DC right now around our industry. Is there anything that you're looking out for in the back half of this year? It's obviously, you know, we started strong with a meme coin out of the White House. We've got the new stablecoin regulations passed. Anything in the back half of the year that you're kind of looking at or anticipating? I think everyone's really anticipating market structure and like having a market structure bill that will clarify the rules of the road for that entire industry.

1:26:56I think that's as important, in my opinion, as the stablecoin legislation that is going through now. Yeah. That's great. Well, thank you so much for stopping by. This is fantastic. Thanks for having me. Great to be on the show. Always welcome. Talk to you soon. And next up, we have Ben Thompson from Stratechery coming into the studio. Very excited to talk to him. The moment we've been waiting for. Yeah. Welcome to the stream, Ben. Good to have you on the show. You've been a backbone of many analyses here on the show, and we're excited to welcome you to the show. How are you doing? I'm doing good.

1:27:34I put on a button-up shirt and a jacket just for you guys. Please feel honored. I am wearing shorts underneath. I wasn't with it. You didn't have to tell us that. People always ask if we wear shorts. We actually do wear the full suits. We got to stand up to hit the gong sometimes. There's a wide shot. I am the poser are here. So I'm happy to admit. Well, it's a great it's a great sign of respect in our culture to put on a suit for a TVP on appearance. And we're just we're so excited to talk to you. I, as you know, I've been lucky to read your work in my entire career. And I think it I think so many of the thoughts that I have are now like your your way of thinking about technology and markets is so embedded in my brain that ideas that I hold as true or just foundational beliefs are actually your beliefs that have just become so immersed.

1:28:25So it's great to talk. Well, thank you. I will attempt to implant new ones or maybe show you the error of your ways. I wanted to see what we're seeing. Sounds great. I do have a question on the nature of where you sit in the media world before we go into actual questions about tech companies. It's interesting that in some ways you're a journalist, but you don't really do the scoops and breaking news that much. But you also don't issue just straight up buy and sell recommendations. What was the thesis behind not just actually having a price target and not doing like this is a sell side bank, but independent?

1:29:08Well, when I started, I mean, it's funny to hear you talk about like my quote unquote place in the ecosystem. Sure. Because when I started, I had like 368 followers on Twitter. I was just some sort of random person on the Internet. In retrospect, sort of right place, right time, I think is certainly the case. But I did perceive there was a large gap between tech journalism, and I would include a lot of the bloggers there who were writing a lot about products. And then there was Wall Street that was very focused on sort of the financial results. And to my mind, there was a large space in the middle, which tied together the products to the financial results, but also the overall companies and strategies.

1:29:52And I'm very interested in culture and how that guides decision making. One of my sort of precepts is all these companies are filled with smart people. And a lot of people, when you ask them why they did something wrong, their only answer is that they're stupid. And I'm like, no, they're not stupid. It's actually much more interesting to assume they're smart and are doing stupid things and trying to unpack why they are doing that and what goes into that. And so that was sort of the thesis was that there is this space to explore these spaces. And then there's a business model aspect, which is I started Stripe2 years after Stripe started.

1:30:31I think they had just come out with their billing product. And the only alternative at the time was PayPal for subscriptions. and it was fairly sketchy and there was lots of like horror stories out there about you know stuff and just the stripe api was so great and the things you could potentially do with it and so on wall street you're putting a price on it you're also charging like a hundred thousand dollars a year or something like that uh and and so you get a small list of high arpu clients and my thought was i could go in the opposite direction and get a large list of low arpu clients thanks to things like Stripe and the ability to subscribe.

1:31:09And as part of that, I wasn't going to go through the rigmarole of getting registered and doing stock picks and all that sort of thing. I've always joked, if you want a stock pick from me, you're going to pay me a whole lot more than$15 a month. It was$10 when I started. And it's actually pretty great. Now, one of the critiques I do get, particularly from my friends on Wall Street is, you know, no skin in the game, XYZ. I think at this point I'm large enough that my reputation is significant skin in the game, but I do recognize the validity of that critique. Yeah, and you know if you make a bad call, you're going to have to circle back to it in two years and write about it yourself and admit that you got it wrong.

1:31:48Which hurts too. No, I had to write about this week. I was very optimistic about Apple's Apple Intelligence announcement last year and the theoretical power it would give them over the model makers and now I'm ready actually no they're gonna have to pay up and that's you know that was a bad call by me that you know I think was you know very well received at the time uh and might have gotten that one wrong and and so I do need to be straightforward about that and so I just this morning I was very crystal clear like I got that one wrong that was that was that was an issue what is nice is Stratechery kind of ended up being in this interesting place where I feel like I'm a little bit of like the Switzerland of tech and that no one pays anymore.

1:32:33If you're a CEO, you pay the same amount as Joe Blow down the street that is paying it. I don't invest directly, which I think made sense when I started because I didn't have any money. It's probably hurt me a lot over the years since then. I think this is a different West Coast, East Coast thing where it does feel like on the west coast everyone's talking their book sort of all the time and uh and you know that's why i generally as a rule don't have vcs on to do the trajectory interviews sure because it's kind of hard to get like a real take because because that is you know such a motivation um and so me coming in being like i have no book to talk i'm just here telling saying what i think I think has been good for the West Coast audience, which is my base audience, even if the East Coasters think that I'm being a big wimp.

1:33:27That's funny. Yeah, the talking your book challenge, we go through that a lot. You got sort of 12 VCs on a day. Well, yeah, and we just try to get a bunch of different opinions and triangulate what we think is real. I'm trying to come up. You have TPPN. I'm trying to come up with a P so I can get the talking book network in there. but um talking book production network yeah yeah for talking your book but yeah it is a real struggle to find somebody that for example has a deep understanding of every foundation model company but isn't massively conflicted at in some way or another extremely extremely yeah and so it's one of those things you just sort of you you end up like there's so much path dependency and all these sorts of things and like i mentioned like a big advantage i had was i started at a time when sharing good links was very high currency on Twitter and so you know I grew very very quickly much more quickly I sort of had a five-year plan to go independent I ended up doing it in less than a year in part because it just sort of spread really really rapidly and it was an ideal time to be someone sharing interesting links regularly and I wasn't showing them the beauty is my readers were showing them they were doing sort of the marketing for me and so I'm very cognizant of sort of the the luck I had in that regard and then just over time and it's been an interesting journey for me to grapple with my different position in in the ecosystem like so when I started the Struckery interviews that was sort of part of it which was I started out not knowing anyone I got to the point where I can talk to anyone that I want to and so how do I square that I can't be the guy with the chip on his shoulder trying to make a name for his self forever it sort of gets it's like the meme with the guy how are you doing kids like at some point you have to accept your part of the establishment how can I do that while still staying true to the idea that Shrekery is about the readers it's reader funded my loyalty is to them I'm very clear I have no loyalties to anybody else and so well I'll just I will talk to people in sort acknowledgement of what I can do, but it's going to be fully transcribed and published and sort of available to everyone.

1:35:46Have you ever dealt with or thought about the attack vector of a special interest, you know, buying a thousand plus, you know, thousands of seats to a single, you know, independent publication and saying like, yeah, like, you know, we're happy, you know, we got seats for all of our employees actually, because we really, you know, love the, and then, and then suddenly they're sitting over there and, you know, representing meaning, very meaningful 50 % of your revenue. 50 % of your revenue. I mean, I fortunately, I think of a scale that I don't have that problem. That's good. There we go. But no, I think audience capture for subscription sites is a potential issue for sure.

1:36:25And this is another thing, I was sort of right place, right time. I got big enough by the time that it doesn't matter. And if someone's really, like, I give refunds all the time. Actually if someone really upsets me, I will refund them and every dollar they paid me, I'm just like, go away. I don't, you know, I don't, you're being abusive or whatever it might be. And that, that is a beautiful thing about the relatively low price, high customer base model is no one has power over me. Like I have the burden of publishing, you know, as often as I do, I feel a heavy weight of duty to my customers. When I write something I'm not happy with, like I don't sleep well, but at the same time, there's no one customer or no, individual that can come in and be mad at me and impact my business.

1:37:13I'm seeing that there's maybe some sort of parallel between legacy media and independent media where independent media, it's not by default more pro-tech or anything, but there's just no salary cap. So if you're at a legacy institution and you're writing, it's probably some sort of rough, loose salary cap of a few hundred thousand dollars. Whereas you go independent, it's feast or famine, you might fail, but you might get really, really successful and have a huge income from that. And I'm wondering what we're seeing in the AI salary wars, where we're seeing more and more talent and Mark Zuckerberg potentially paying$100 million bonuses.

1:37:52Do you think that Apple will come around to spending more money on researchers? It feels like they kind of have an internal salary cap with Tim Cook making$75 million. There's now people that report two levels down from Mark Zuckerberg that are making more than Tim Cook. And you have this weird dynamic where even if there's no actual salary cap at Apple, you kind of have an implicit one from the CEO. Yeah, for sure. I mean, I think just to go back to the media observation you started out with is as you increase transparency in the market, as you decrease non-related barriers, which in the publishing world previously was really geography.

1:38:33And when everyone's on the Internet, you inevitably, just about all cases, you get a power law distribution. and a few people make a ton of money because they win most of the market and then some people make some and then there's a long tail that that sort of don't make any at all but it's it's very it's interesting it's it's fluid in a way but it can sort of become somewhat static as long as the people at the top sort of you know continue to do well but what's interesting about ai is for 40 years, you would have periods of time where you'd have tech companies going head-to-head in a product market.

1:39:14And I think one of the reasons, part of the software eating the world sort of idea is the way you get an apex predator is that that predator killed everyone else first. And so you had tech companies fighting each other for the first 20, 30 years of tech. The ones that emerged were lean, mean, killing machines. And they and the entire industry were sort of set loose on the rest of the world, and everyone was just like getting slaughtered sort of left and right. But what you also had over the past sort of 20 years or so is the big companies in particular sort of slotting into unique slots. So you have Facebook is social, Google is search, Apple is devices, Microsoft is business applications, Amazon, e-commerce, et cetera.

1:40:01And obviously these companies are very large and do lots of things, And there's some overlap in different places, but they've been fairly sort of distinct in their categories and they've been dominant in those categories. And so they've been in a place where like Hollywood is wanting to get to. Right. What is the dream in Hollywood? You want to have a franchise where the next Marvel movie matters more than who the star is. The reason that's so great is because you now have bargaining power over the stars. So you just sub someone else in. and whereas the old style like Tom Cruise makes the most money because Tom Cruise on a movie poster sells the poster and so in a negotiation he has massive bargaining power so he's going to get paid a lot of money.

1:40:45In tech it hasn't been that case. The companies themselves have been franchises and so the overall anyone who works in tech or probably works in any entity but you know there's a few people in each company that are critically important really make the whole thing go. Everyone else is fairly replaceable. Those people have probably always been somewhat underpaid for years and years and years, both just by the nature of companies and the cultural issues and your salary cap sort of analogy, but then also just like it's not a transparent market. It's not hard to price sort of what people are worth.

1:41:23With AI, everyone's trying to do the exact same thing. So you have multiple companies trying to do the same thing. The output is somewhat measurable. I mean, all the AI test stuff has issues, but by and large, everyone kind of knows who has the good models and who doesn't. The scalability questions, because all these companies are trying to do the same thing, we have a very unique situation where the bargaining power, you increase transparency, you increase sort of the liquidity or the ability of people to move around because they're doing the same thing, the bargaining power shifts to the people that are super valuable because suddenly it's much more clear who's valuable and their skills are much more transferable.

1:42:07So this is, I think, a very underrated bear case for tech in terms of AI, at least for this time period, is they've lost that murky bargaining power over employees that they enjoyed for decades. And currently, you're seeing what happens when you don't have that. You start paying employees what they're worth. And obviously, that's great. I'm not saying this is a business analyst. It's not a sort of a moral statement. But it is like what Mark Zuckerberg is doing, I think, is totally rational. I think it's a classic sort of Clayton Christensen from Facebook's perspective, AI is all upside. So of course they're going to invest what they need to do to win, but it's costing him a lot of money.

1:42:52And by extension, it's costing everyone else in the ecosystem a lot of money. Well, isn't it in some way, is the right way to think about the last couple of weeks, like more of like an unofficial acquihire in the sense that it's not just the people, but it is the know-how in terms of, hey, there's these things that we want to do that are important to our business in a lot of different ways. And we're basically, it's like the collective is actually more valuable than any one. Like the collective together, getting 10 researchers at the same time is meaningfully more valuable than just each individual researcher added up.

1:43:30There's probably something to that. But I think, again, like what is actually different between what Google is trying to do, what Anthropics trying to do, what OpenAI is trying to do and what Meta is trying to do. They're all trying to do the same thing. So I my suspicion, I'm not an AI researcher, so I don't want to overstate my my knowledge in this space. But my suspicion is skills are fairly highly transferable. And when that is the case, there is – in some situations, if lots of people can do those skills, that's terrible for the employees because then their bargaining power gets diminished because anyone can slot in.

1:44:06But we're in this space where the skills are transparent, knowable, transferable, and there's not very many people that can do them. And so it's a scarce resource that everyone's fighting over. And that's why you see this real shift in negotiating leverage as manifested through these dollar figures to AI researchers. Yeah. Do you think, I mean, Google seems like the most fragile and the most like paranoid about just disruption. It's not all upside. It could be very bad for them. The innovator's dilemma, you know, you had this back and forth where Sundar Pichai mentioned that he hadn't read the book.

1:44:42You said it doesn't matter because it's a structural issue. I think that's a good point. But if you play back the counterfactual, is it ever possible to disrupt yourself? And essentially, like, if the Gemini app had launched before ChatGPT and they had taken over that mindshare and maintained 90 % ownership in that, like it would be somewhat disruptive to their revenue and their profits as they transition over, but when I sum the revenues from OpenAI and LLMs and then Google search, I'm not seeing some massive drop off that actually would destroy Google in the short to medium term. But I'm wondering if you think it's like, is it entirely impossible to avoid the innovator's dilemma by disrupting yourself?

1:45:30Well, number one, you have to also look at margins, not just revenue. But number two, you actually answered your question. Google didn't launch Gemini as a chatbot. That's the answer. They were years ahead. They invented the transformer nearly a decade ago. And so in many respects, there's parts of this question that the counterfactual makes the point in that it is a counterfactual and it's not reality. now i do think i think google has done better than i expected over the last two years uh i i like what they're doing in search generally i i think they it does seem to be the one part of the company that still functions like they they can actually iterate and build products what we're seeing is reminiscent of what they did a decade at 50 or 12 years ago when everyone's like vertical search google's done all the never's gonna search in apps and google completely transformed the surf the search engine response page, whatever it is, the search engine results page to be local or to be shopping or whatever.

1:46:35And Yelp's been throwing a hissy fit sort of ever since. And so that's what they're doing with search, right, and with search overviews. And they have this new search labs or AI mode. They can sort of test stuff out once it's scalable, once they're confident about the monetization issues, they can sort of shift it over. I call it the search funnel, search AI funnel. I think it makes a lot of sense. And I think this is always actually kind of puzzled me where I think they're responding fairly well, even though this is seems to be a textbook case of disruption. And I went back to an article I wrote years ago called Microsoft's Monopoly Hangover.

1:47:14And I was I went through Lou Gerstner's autobiography and about how he turned around IBM. and his real insight with IBM was everyone wanted him to break it up into sort of different pieces and what he realized was IBM was so big and large from having downstream of been a monopoly that actually the only thing they were good at was being big and so breaking them up would actually just create a bunch of subscale low-performing companies that would all get wiped out. But as this behemoth, they could go to other big companies and solve all their problems at a very mediocre level, but it's still sort of an attractive proposition.

1:47:58And under Gerstner, they really rode the internet wave. They went to all these big companies and said, this internet thing's happening. You need help. We'll solve your problems for you and had a very sort of successful run, you know, kind of until cloud came along and which Gerstner, by the way, was, was, was a proponent of, but you know, by that time the IBM people were back in charge. And I was thinking about the context of Microsoft where business models are hard to change and disruption is ultimately about business models. And culture is even harder to change. But what can't really be changed is the nature of who you are.

1:48:37And I think there is, you know, in Microsoft, they were in a similar situation. They were a big monopoly and they weren't a product company and the attempts to become a product company with Windows 8 and all the things that went on around that time inevitably failed and Cyanidella to his great credit and you know sort of diminished Windows importance in the company broke it literally broke it into pieces spread it around and this was a multi-step process and got Microsoft back to a place of we're big and we'll do everything. We're not a Windows company. We'll go in there and we'll go solve all your problems.

1:49:16Very sort of reminiscent of the second version of IBM. And I go back to Google and I've always been intrigued by the I'm feeling lucky button, which doesn't exist anymore. But I always enjoyed that that button continued to exist long after it was impossible to click because the moment you started typing the search box, it would start auto searching immediately and jump right to a search page. But it was there in a – it's just so core to Google to give you the answer, to know everything, like to know everything about the world. And to – there's a bit where even though the core of their business model is 10 blue links, and it's not just the users choosing the search link, which gives them a data feedback loop so they know which results better, but also the users choose the winner of an auction Google puts on for ads.

1:50:10And it's an incredible business model. And there's something about that that's always been intention and counter to what Google was founded to be. And I feel like that germ of what Google was founded and meant to be is an AI answer engine. and it almost feels like even though Google is old and large and fat and slow moving, that core aspect of their nature is still in the culture. And that's why they're finding it in themselves, I think, to do better in AI than you would expect. Was it enough to launch a chat GPT before open AI? No. Was it enough to have any sort of cogent response for the first six to nine months?

1:50:57No. But it was enough that I think they've done better than I expected over the past year in particular and gives me, I think, more optimism than I expected I would have for the company when, you know, when ChatGPT first launched. AI overview from Google. If you search Google's mission, Google's mission is to organize the world's information and make it universally accessible and useful, which is exactly what language models do really, really well. Like the thing that's just undeniable, right? You can debate whether this is going to be the year of agents. It doesn't feel that way to me yet. But this is the year that most people have realized that, wow, LLMs are very good at organizing, surfacing, and making data valuable.

1:51:47You mentioned just the debate over breaking up IBM. I'm interested if you could take us through some of the - I was so glad you were talking about IBM today, did you? No, no, no. But I want to talk about Intel and kind of your, the history of some of your takeaways and what you think you've gotten right in the past, your perception of, you know, should they break up the foundry business, and what you think might be in the works with Lip Bhutan coming in there. Because I was listening to Dylan Patel talk about his conversation with the new CEO, Lip Bhutan, and it seems like they're doing lots of tightening up, lots of layoffs, but it's kind of, I don't even know what framework to apply to analyze, like is a breakup the correct thing?

1:52:35It feels like something people just say. Yeah. So Intel, it's funny. One of my very first articles was about Intel. And what I said at the time was, and this was 2013, and this was an art, like, you know, when you start a site like Strachekery, you're like a new band. And why does everyone think a new band's first album is the best? Because they've been working on these songs for years, right? And then the next album, they had a year to do it, and they all suck, right? So I'll let people decide if that applies to Strachekery or not. I won't be offended. The sophomore slump. But I'd have been on – Intel had been a thing I'd been wandering about for a long time, which was by 2013 when I started they had clearly missed mobile now it wasn't clear to them they were still trying to do the atom processor and just they're going to figure it out tomorrow and the the problem with missing mobile is the problem with Intel in general is Intel is always very biased towards high performance and this goes back to uh actually Pat Pat Gelsinger his first time through at Intel.

1:53:43Intel, you know, had the CISC, the way there's CISC versus RISC. It's like different ways of organizing bits or whatever. RISC is generally more efficient and actually even Intel processors today, even though x86 is CISC, the internal, it's retranslated internally to a RISC type language. None of that is really important other than to say in the 80s, there was a real push in Intel to switch away from x86 and to a risk type of architecture, but like for the processors. And Galsinger was a leading proponent that this is a terrible idea. And the reason it's a terrible idea is because there was already a huge ecosystem of software built around x86 and all this low level code and capabilities that no one ever, that was written once and no one ever wants to touch again because it's miserable work.

1:54:37And he's like, to rewrite all that stuff would take at least two years. And in that time, our ability to manufacture chips will improve so much that had we just stuck with CISC, our processors would be faster. And that was the right bet. and that's one of those foundational, that's why I like to think about companies and their history and what goes into that, which is Intel from the 80s on has solved its problems by having superior manufacturing and by moving faster. And yeah, our chips may be theoretically less efficient, but if our manufacturing is better and our transistors are smaller, it doesn't matter because that will swamp whatever theoretical sort of efficiency you might have.

1:55:21And this drove the entire computer industry. You would write, to write a program, just every second you spent optimizing your software in the 80s or 90s was a waste of time because whatever improvements you could get would be swamped by the next version of, if you went from 286 to 386 or 386 to 486. that jump was so large you were better off focusing on features even if it made your software sort of slow to use on the current hardware because the next generation of hardware would be so much faster it would solve your speed problems for you. Now this has generated a lot of bad habits amongst tech developers.

1:55:58That's why you get bloat and why you have like poor performing things and all those sort of things. But this was sort of super critical. And so Intel at its core has always been focused. They've always been manufacturing first and focused on better and better performance. What happened with mobile is in that calculation did not come efficiency. They were never focused on efficiency. And in mobile, efficiency was everything. So what happened with mobile is Apple went with an ARM processor made by Samsung, and they basically rewrote everything. All that stuff Intel didn't want to rewrite in the 80s or if they rewrote would just give other processor companies a chance to catch up with them had to be rewritten for mobile because efficiency was so much more important than performance.

1:56:44When that happened, Intel was screwed. Now, it took them a long, long time to realize they were screwed, but they were just fundamentally unsuited to be competitive. It was the whole Paul Ottolini turning down the iPhone contract is not true. Tony Fidel I said that once and I got a call from Tony Fidel actually this is when I had him on it for an interview and he's like this drives me up the wall Intel was not remotely competitive even though they had arm chips then even their arm chips then were focused on performance not on efficiency and and so the the problem for the problem for Intel is once you missed mobile you were going to lose your manufacturing lead at some point because volume matters so much and every time you move down the curve, your transistors get smaller, the cost increase massively.

1:57:31So you need volume to spread out the cost of building these fabs. Like back then when I wrote this article, fabs cost 500 million. Now they cost like 20 billion. And this is over a course of like 12 years. So it was clear Intel was going to be in big trouble back then. And so I wrote, they need to build a foundry business. They need to figure out a way to build chips for other people because in the long run, the cost of keeping up in manufacturing is not going to be tenable if you're not making mobile chips. And what obviously they didn't, TSMC made all the mobile chips for everyone. And guess what happened?

1:58:08TSMC took over the manufacturing. Now there's lots of other things that went into this, why Intel stumbled and sort of things, but at a structural level, what happened was actually inevitable. Once Intel missed mobile, unless they figured out a way to make mobile chips some other way. They didn't do that. What's interesting is what is the problem with that, it took so long to manifest. Part of mobile was you had an explosion in the cloud because cloud and mobile actually go hand in hand. Intel made all those cloud chips. Intel stock had an incredible run from the time I wrote that article for the next eight to nine years, and I felt like kind of a moron because I might say this company is screwed and they don't do what I say.

1:58:51They didn't do what I say, and their stock went to the moon. But the way it actually caught up to them has been in the past two to three years, where there's astronomical demand for AI chips. Only TSMC can beat it. Intel's not in the game. They're trying to shift to a foundry model, but they're so far behind. Being a foundry is being a customer service business. It's not being an Intel, we tell you what to do, or we tell our design teams how to change their chips to accommodate our manufacturing needs. It's totally different, and they needed a decade to learn how to do that. Had they changed in 2013, they would be ready today to capitalize on AI.

1:59:35And the counterexample here is Microsoft. Microsoft building Azure, yes, it got them somewhat in the game with mobile and things like that, But AWS dominates in that space. But by virtue of building up Azure, they were prepared when the AI opportunity came along. And now Azure is sort of a big AI player. And, you know, I wrote about these two examples a few weeks ago in the context of Apple. I think the concern for Apple isn't the short term. We're going to be using AI apps on our iPhones for quite a while. it's are they going to be prepared for what's next if they don't do some sort of sort of reset and pivot here oh sorry i didn't answer your question about intel anyhow yeah i mean it's i'm hearing like managed decline basically like just like you know just get as much cash flow out of this thing as you can while you wind down the business for intel yeah that's what i'm hearing yeah i mean it doesn't feel like oh yeah there's a silver bullet just split the business and they're good.

2:00:38Like, no, it's like, it's, it's all bad. The problem with the business is Intel needs volume and they get volume from Intel. Sure. Uh, and, uh, the, and AMD split their business a decade ago and it was really, they had a very hard time for many years and they had very tense and difficult negotiations between the global foundry side and the AMD side. Global foundries was AMD's, uh, manufacturing arm. Um, and it wasn't until really they got out of that and went to TSMC and then also completely rehauled their ship design business and all those, you know, that they got in the business they were. And then also that Intel stumbled.

2:01:13That certainly really helped them. Intel today, so if you split it up, like Intel itself is fabbing some of its stuff with TSMC. Yeah. Who wants to buy Intel's foundry services? The problem here is TSMC is located in a country called Taiwan, which you know what it is today, but five years ago, they'd been like, what, Thailand? Which, by the way, was probably much better for Taiwan security when the American side was Thailand. But, so there's a real national security element here. And it's just, it's a really tough situation because Intel is a failed company at this point. And the reason the failure is so total is because the aspects that drive their failure are the same things that drove their success.

2:02:05It was their arrogance. It was their sense that we're the best, that we will just win through manufacturing might and performance. And all those things work against becoming a good foundry, work against being a customer service organization, work against recognizing the fact that you're not going to make up for missing mobile through manufacturing, which was their bet for years and years. You had to accept that you lost. and that's a tough place for companies. It's not like someone made a mistake. It's not what they did what they did too well for too long. It's who they were. They continue being who they were.

2:02:44That's right. But who else are you going to get if you want an alternative to the SMC? It's a very tough situation. Last question, and I think we'll be forced to have you make a slightly shorter answer, unfortunately. I wish we had hours to keep talking. I wanted to get your updated thinking on XAIX, the combined entity. The last 24 hours have been very chaotic. When the initial merger was announced, it made sense for financial reasons for some of the different stakeholders. But I wasn't fully sold on this idea. You're going to force me to come up with takes that I generally just avoid writing about Elon Musk companies for self-sanity reasons, I think.

2:03:24I mean, I remember I wrote an article years ago about like when the Model Y was announced and I was talking about, you know, it's a Tesla and this aspect. What Elon Musk is very incredible at is sort of creating reality out of thin air. He's like the ultimate memer. And to create like – it's the way things used to work backwards. I remember I analogized it to like protests like a critique of modern protests is they spin up very quickly because social media makes it very possible. But there's no infrastructure under them, so they don't amount to anything. Whereas you go back to like the civil rights era.

2:04:03There was years of groundwork that went into like the million man march, you know, on Washington, D.C. And there was a structure in place that ultimately manifested in large crowds. But modern protests are the opposite. The largeness comes at the beginning and that all falls apart. There's nothing in place. and There's something that makes a challenge to write about anything Elon Musk related is the you have all the social aspects of creating reality. the stock was buttressed for years by these true believers even though the financial parts didn't make sense you famously had these wars with the short sellers and all that sort of thing and it worked it basically manifested a market for this Model Y and then the Model X not the Model X what's the other one?

2:04:51The Model 3 it was Model 3 when I wrote that article Model 3 and Model Y had this massively successful all the people that were true believers got very rich and congratulations to them. It's great. But it makes it almost impossible for someone, for what I do, who I want to look at structure and fundamentals. I can observe this effect happening, but you can't really say what's going to happen or the effects of it other than to say this is interesting. And so I wrote about that article and then the Solar City thing came out and he's like bailing out like his brother-in-law or something. And I'm like, I can't write about this.

2:05:26Like, what am I going to say? Like, there's, it just doesn't make sense. And so I think there's – to fast forward to XAI, yeah, there's a theoretical piece here. I think actually XAI would be an incredible acquisition target for a lot of companies if it wasn't saddled with X. So interesting enough – Yeah, it feels like the end state is like Twitter getting spun out again. Like that's my – that's kind of like my – it just ends up going back to Twitter and it becomes – The bluebird. No one actually wants to. Like, Twitter, there's never been a company in the history of the world, probably, where the impact of a company is completely and utterly divorced from its financial realities.

2:06:10Like, I think when Elon Musk bought it, and I assume that's continued through now, they'd have, like, one profitable quarter in their history. Like, it's an unbelievably terrible business. And so I think it's probably weighing XAI down. Yes, I get the theory that Twitter data helps XAI. Well, it helped yesterday. But you can contract for that data. You don't need to pay$43 million for Twitter, or$43 billion, I should say, to get it. Yeah, that was always my position, too. I don't think it helped yesterday when Mecca Hitler emerged on the timeline. It was a crazy day on the timeline yesterday. But anyways, I wish we had a lot more time here.

2:06:48Yeah, this has been really, really fun. Thank you so much for stopping by. Yeah, no worries. What was you guys doing? I actually had the idea of doing a daily podcast ages ago. Classic example of ideas don't count. Execution does. And you guys did it. I think it's great. Well, you're always welcome here. You're always welcome. Thanks so much. Thank you. We'll talk to you soon. Bye. We will jump straight into our next guest, Scott Belsky. I'll be right back. Hopefully, we haven't kept him waiting too long. And he is still in the waiting room. We'll bring him into the studio really quickly. Let me tell you about Wander.

2:07:17Find your happy place. Book a Wander with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning and 24-7 concierge service. It's a vacation home, but better, folks. And we will check in on Scott Belsky and see if he is available to catch up. How are you doing, Scott? Sorry for keeping you waiting. Ben Thompson, he knows so much about history. He runs, he puts out three hours of podcasts every day. I should have expected. I'm me, man. I mean, that's, you know. I just feel so bad, but I'm very excited to have you. Can you just give me a little update of what's going on in your world?

2:07:53I want to talk about the AI safety layers concept, and then we can talk about some of the current stuff that's going on in AI. It feels like, I mean, you published this post, what, two, three weeks ago, and it feels extremely relevant today, last week. So very excited to get the update from you. So kick us off. Yeah. Gosh, where do we begin? And while listening to our ongoing segment here on implications of the technology that's happening, that is happening faster and faster and faster, a few things top of mind. You know, we can start we can certainly start with AI safety layers. I think it's fascinating how much discussion there is of the dangers and the perils of AI without recognizing how it can operate as a layer to protect us.

2:08:36If you get some call from someone who proclaims to be your grandmother asking for money, that's clearly something that AI on the device, some form of local model can detect, given it's all happening on that device and warn you, this isn't your grandmother. When it comes down to all sorts of the creative and crazy scam phishing email type things that we get all the time, that's a perfect use case for AI, of course, and telling us that we need to be wary. But also, I mean, what about being polarized by algorithms and, you know, detecting an algorithm changing based on your engagement and, you know, an AI sort of saying, hey, Scott, you're getting on the fringe here.

2:09:16Like, watch out. You know, you're now like in this small 10 percent of society that now is going down this rabbit hole of some conspiracy theory. I just think there's so many there are so many use cases of AI as a safety layer that that the device needs to unlock. And of course, that means the operating systems need to figure this out. So I think that when most people talk about AI safety or safety layers or what you just described, solving the problems that will inevitably come from any new technology, they look at it through a technological lens. They see, you know, well, let's do more reinforcement learning.

2:09:50Let's align the model. There's a fixation at the model layer. For sure. Like we need to solve it. And I look at it almost entirely from an economic lens. And I just think about it as if the market cap of the company that's selling you the Skinner box is bigger than the company that's helping you get healthy. If the sugar company is bigger than the health food company, you're going to be fat. And if the health food company gets bigger, then you're going to be healthy. And so I think about it as like the doom scrolling. We have screen time apps. They're small. They don't monetize as well as doom scrolling, unfortunately.

2:10:28So there's some like economic considerations there. But at the same time, I come back to the scamming angle and I see this as like, you know, if the economic weight behind the good guys is bigger than the economic weight behind the bad guys, you get the good outcome. And that's why I'm not particularly worried about like super doom scenarios because I think that, you know, generally governments and people will align with like, hey, let's not get paper clipped. So let's build more systems to be safe in general. And then the bad guys, yeah, they might go try and build some really bad weapon, but they will be completely outnumbered.

2:11:05The question is like on the margin when we get into these pockets of, you know, the Skinner boxes, the doom scrolling, where the economic weight looks. So do you think about it in that same lens? And then the question is like, what business models can actually support this? Are we talking, you know, I need to have a subscription for like a Clue lead like app that's looking at everything I'm doing and then acting as that layer on top. How can we actually implement this? Or is this just like, we're hoping that Apple runs a great ad campaign around it and it becomes an Apple feature that they, you know, hold up at every chance they can.

2:11:42Well, I mean, first of all, I think that the operating systems of our life are the ultimate interface layers. And for many of us, it's either Android or iOS, but at work, it's many other companies that are operating systems at work. But those operating systems are trying to make us loyal. They're going to do so through remembering us, you know, personalization effects are the new network effects, I like to say. And I would imagine that protecting you from what's going to be a very comprehensive and very sophisticated set of social engineering and other sorts of long-form scams. I mean, you think about the most effective scams that are out there as when, you know, you really have this like very long, you know, experience or exposure to some entity to the point where you trust it and then suddenly gets your information and it's too late.

2:12:34And so that's that is a perfect use case for AI on the device to kind of monitor over time, you know, compare that data with any other scams that are reported. So in terms of the economic incentive, I mean, goodness, I feel like consumers will have high willingness to pay for that. if they don't get it free with their operating system. Yeah, I mean, I think you can already imagine the UI of you pick up a phone call and Apple, you know, you can think of the traditional layout of like the hang up button, the hold button, et cetera. And then there's just a, you should have a little bit of a tag there that says like AI voice, you know, detected or something to that effect.

2:13:14to just, I'm perfectly happy with talking with, you know, a model effectively on the other end, but I'd like to note, everyone should sort of know that it's a model. And I think we're in this weird period right now where people all the time are starting to talk with AI and not even fully realizing that it's a human on the other end. There's some ways that there's the counter-credentials movement, right, which I was involved with back in the days at Adobe, which is an effort to have models insert cryptographic metadata into anything that's generated, including live generation, like, you know, live audio that can be detected on the client.

2:13:51So there's some ways of going about this. But in terms of, you know, your point about the economic model is interesting. You know, my thoughts immediately went to alarm systems. You know, we all pay for these ring alarm systems, like all these alarm systems for our home that sometimes costs$60,$120 a month, you know, with monitoring and window motion detectors and everything else. But we're not really paying for an alarm system for our like, you know, for our devices in this new modern world where we're going. And maybe there is a market for an AI safety layer as a service. It's like new antivirus or something.

2:14:21We're going to have - Right, yeah. Memetic viruses and stuff. But it needs to screen record everything. Yeah, it needs to be at the operating system level. How are you thinking - Yeah, for mind virus detection. Yeah, yeah. How are you thinking about the evolution of those, like the mind viruses that come from just the accidental interaction with AI? I mean, people are, there's such a wide swath. When I talked to Jordy and we were having dinner last night with David Senra, and we were talking about how we use AI. And we're like, yeah, probably 30 minutes a day in chat GPT. Like, it's a lot. But these interactions are summarize this post, do this research, pull this things.

2:14:57It's like talking to a computer. I'm not saying, hey, how is your life? I never have that interaction, but there are a lot of people that do. And so what are you seeing? What anecdotes have you pulled from? How do you think that evolves? Are there any risks? Walk me through kind of the way humans are interacting with just language models broadly. That's interesting. I mean, I think one of the topics that is on my mind a lot lately is consumer AI. And I'm not just talking chat GPT, which is obviously a consumer product for many of us. But it's interesting. I was at a tech conference recently where all the trends that are popular now were being discussed.

2:15:34And I left asking myself, what's the one thing that no one talked about? And the one thing no one talked about were new consumer AI era social networks. And when mobile came around, there were a whole new variety of social networks. Like every time there's a platform shift, a lot of consumer mainstream applications or social networks, that sort of stuff is reimagined. right and so the question is why is why why is that not happening now and uh and then the whole saying in consumer investing is always around novelty preceding utility and so i'm trying to keep an eye out now for examples of um of consumer ai i mean there's this company called tolin which is sort of like a pet alien that you uh you you start having conversations with and they're doing really well i believe they raised around from you know some some of the top firms um you know i've been playing with this idea, a few ideas with friends, you know, one of a simulation representing our digital twins.

2:16:28So could you, could you kind of train a sort of an AI digital twin of you based on all your experiences in chat GPT or any other sources of data, and then deploy that in a simulation with mine and others, and we could start to actually just watch them interact with each other. And it's kind of, it plays with some fun ideas of plausible deniability, you know, oh my gosh, like, I'm so embarrassed. Like what my simulated twin said to yours. These are the types of things that are, you know, wingman as a service. Like, I don't know, is there an AI wingman that, you know, helps us when we're flirting with people on social, on, you know, on dating platforms.

2:17:04Yeah, yeah, the, what I want to see and you're kind of getting at this is just like more weirdness, right? It's easy to go build the neck, or not easy, but, you know, we were at YC and there's a lot of companies in the last batch building agentic infrastructure. And it's like that stuff needs to be built. But I also, at the next batch, I hope there's more people being like, yeah, a lot of people built all this infrastructure already, B2B, basically B2B SaaS. Why don't we just like take a crack at like, yeah, some dating simulation where it's like you create a digital twin and you just like throw it into the mix and it goes on a thousand speed dates with people in your city.

2:17:40Not even speed dates, but simulations of dates with people in your city. Yeah, we've talked about this before. this idea of like you have a whole bunch of people that are talking to a romantic AI partner and that feels super dystopian. But if Steve in Los Angeles is talking to the AI girlfriend and then Sarah in Boston is talking to an AI boyfriend and the two AI's realize on the back end that these people are super compatible because you have so much data from them, it's just introduce the two humans and say, hey, you know, yeah, you have to pay us to introduce you. We're going to pay a finder's fee and collect your LTV on this app for the next 10 years because you guys are going to probably live happily ever after.

2:18:20And that's kind of like the white pill scenario that I hope happens. And I hope the dating app companies explore. You have to break up with the AI companion though. Yeah, yeah, basically. Yeah, but the AI, it's the her scenario. You're still together, but your AI versions have broken up. Like, are you - Yeah, what happens then? Well, then you get a warning or it contacts like a divorce lawyer or something for you who takes a fee on that. Who knows? I think there's a lot of fun stuff to explore here. And one of the other random ideas I had was, I called it peanut gallery. And the idea was that the dirty little secret about why we go back to products like Instagram and others, oftentimes the traffic goes up after we have posted content because we want to see who else saw our content.

2:19:00And so playing off that idea, imagine a social platform called peanut gallery where you post your own content, but no humans are allowed to comment on it. It's all these like personas that are like, you know, tightly, tightly defined personas that are commenting and arguing with each other and discussing and you go back to see, like, how this AI is engaging with what you posted. And maybe that becomes the voyeurism of seeing how other people's, you know, contents were forming. I mean, these are the fun, crazy things that must be explored to find, you know, this edge that will become the center of social.

2:19:34Yeah, I've seen two things that are somewhat in that realm. One is just general YouTube thumbnail A-B testing services where you upload your thumbnail and it tries to predict based on all the data it has what the click-through rate will be. And then you can upload two and it'll say, hey, you should probably go with this one. And then the other, I saw I think Justine Moore at Andreessen posting some sort of app that you open your camera to the front-facing view and it gives you the sensation of live streaming with like hearts and comments and stuff and it's all fake but it's very odd but i don't know yeah i mean it feels inevitable that uh in many ways bots are a feature of it feels like bots are a feature of x now right they have not been eradicated yeah they're still here they're maybe hidden under some i mean that's the story of reddit right the early reddit days were you know it was all the reddit founders posting to seed this thing so if you think about a social network that needs to onboard you.

2:20:32There was that original Facebook thing where if you could get 50 friends, they'd keep you on the platform forever. If you show up and there's a couple bots that are just like, hey, good job. Hey, keep posting. Stick with it. Make some real friends, but we're here for you if you need a little encouragement, a little dopamine. I mean, on that note, the war between meta and open AI and the talent race and all the trade deals has been front page news for the last couple weeks. I'm, uh, Jordy's been saying that they, that the, the product that meta might wind up going after is less like a direct chat GPT knowledge engine and, uh, uh, app that feels more competitive with Google.

2:21:12And it might actually be something more like, uh, companionship and, and chat since that, that feels like the real threat of, of if there was an app outside of Tik TOK that was going to take user minutes from the enter, these sort of like entertainment, social minutes from the Meta ecosystem, it would be these sort of AI companionship, which function as entertainment, this sort of social experience, which is Meta at its core is effectively a social entertainment company. Yeah, and you think about all the rules of successful consumer products. They make us feel good about ourselves. They are sort of social lubricants in that they help us get connected to others in ways that we may not be able to do and be comfortable with in the physical world.

2:21:59And you kind of go through all the list of things and you realize like AI, there's an opportunity to really radically, you know, attack those vectors and make people have a really fun, engaging entertainment experience or a social experience. So it's not a surprise that Meta is going to innovate in that space. You know, I do also kind of wonder when I remember when we all remember when Facebook acquired Instagram and then, of course, when Facebook acquired WhatsApp, you know, they were acquiring network effects in essence, right, around messaging and images. And I wonder now, you know, now it's like it's a talent war.

2:22:35I mean, maybe AI is less about, AI is not really a network effect per se. It's more of like a talent driven differentiation. I wonder if that's also helping us understand the strategy of buying up all these different companies and people. Yeah, I think the framing that I've been thinking about is these are basically like unauthorized acquihires to some degree, where you're basically saying, yeah, these 10, if a company is doing an acquihire in general, there's like, we know this group of people is good at this thing and we want to do this thing and let's bring them over here. And so the premium on talent that we've seen in the last couple of weeks could just ultimately be that it's looked at as you're buying a team, which is more valuable than the individual parts.

2:23:22They just happen to all get chopped up. This has been the chatter around Alex Wang and Scale AI. People haven't been saying, oh, well, Scale AI is going to be this juggernaut in 30 years. But Alex Wang is a generational talent. He'll be around in 30 years. And so the nature of what scale does might change as we get to more data driven or just purely AI generated data and reinforcement learning with verifiable rewards. And scale AI has been through a couple different things with self-driving cars and then RLHF for light language models. And that business, it's not the same as Instagram where it's like, okay, there's a network here.

2:24:02And so there's like this asset value in this, but it's, yes, much more talent driven. And so that's why you see all these people coming together. But it's fascinating. It is interesting to see if Meta is focused more on just let's make Llama great so we can use the best in class AI effectively for free all over our products. Or if they're trying to aim for something that's like an entirely new experience that will be vended to their billions of customers. Probably both, honestly. Why not? And make ads more efficient while they're at it. Yeah, for sure. I mean, that's the crazy thing about this is like$100 million doesn't take that much to generate$100 million if you make the ads 0.1 % more efficient.

2:24:42So it's all economically rational, but we just haven't seen it in tech yet. And that's why these big numbers feel like, oh, we got to talk about this. A little bit of a tangent, but have you thought about how LLMs now are immediately and I'm assuming pretty aggressively shaping actual human communication? Like right now we're in this period of M-dash, gotcha. You wrote that with ChatGPT and people that love the M-dash before are disappointed. But at the same time, it's not like you see that people calling out the M-dash other places on the Internet outside of basically teapot. And I just wonder, we're in this dynamic now where we have the most prolific, like prolific writers throughout history have shaped communication.

2:25:33And now we have LLMs, which are effectively the most prolific writers in history, producing more written word than any one human could do in a lifetime in minutes. Right. Yeah. And it just feels like we're potentially in this interesting flywheel that's just going to keep spinning. A couple of thoughts. First, I feel that LLMs are going to start fine-tuning more towards how we want them to talk to us. So if you want your LLM to be straight to the point, no BS and all lowercase and short sentences, Like that's what your experience of any information retrieval and conversation will be. And that might be different from mine.

2:26:18So I do think they'll all become more personalized for us in these like dramatic ways. I also, though, wonder just like when music becomes generic and then some band or some star like just does something entirely new and creates this new genre. I mean, similarly with writing, like, well, what will human writing be like as a result of LLMs in five to ten years from now? When you pick up a novel that actually captivates your attention and keeps you engaged, you know, what sort of writing will be necessary to do that in this age where, yeah, your LLM can spit out poetry or write a, you know, a short novelette, you know, upon command.

2:26:57So it's fascinating. I mean, technology has always had this impact on us and culture. It's just never been easy to chronicle because it's always happened over such long periods of time. And it feels like those windows of culture change are happening more quickly. and so it's something I'm looking at as well. It's an interesting question. I'm generally still long tool. This is a tool. Creativity is still undervalued or it's not going away and I keep coming back to the idea that there should be nothing easier for an LLM than to write a great tweet. It's 280 characters. It doesn't need to really maintain some long context to get it and yet we haven't really seen anyone break out with an account that people are following and entertained by that's fully AI generated.

2:27:44And there's been some experiments, but usually it's like you're following it. Even like that, do you remember horse eBooks back in the day? I don't know if you remember this account, but it was like said to be just randomly, algorithmically generated from these eBooks, but it turned out that there was actually a human writing it. And there's been a few examples of that where, or the stuff that does go viral that's like AI generated is like, oh, it's hallucinations. And so the fascinating part about it is not the underlying product. It's the fact that it's generated by AI. Yeah, it is interesting that we have this band, the Velvet Sundown, I think they're called.

2:28:20They have a million listeners on Spotify a month right now that claims to be fully AI generated. And it's funny that we got that before a prolific poster that is fully AI. That has 100 ,000 followers and is popular. Here's the thing. I mean, we talk about, of course, like taste being more important than skill. And I think you're tuning into the fact that Ken LLM's like output tweets that are compelling and therefore have taste. And I think one of the questions is, is taste not just about each tweet, but also like consistency of good judgment and great, you know, great content. It's just like they say, a brand is like the hardest thing to build, the easiest thing to lose.

2:28:56I wonder if taste is a similar way. You know, if you have AI pumping out tweets in an account, but if 20 % of them are like, you're like, wait, what? That wasn't clever. you know, do you just lose, does that, is the credibility of that account gone? So I think humans are good at, you know, humans with good taste are good at knowing, you know, yes and no, yes and no, like what should and shouldn't be shared or said or written more consistently. And I wonder if, I wonder if, you know, LMs can do that. It's also a memory. It's a context thing. To be a good poster, you need to really understand the fullness of the zeitgeist and where in the current thing and and and all these meta trends and yeah yeah and it feels like even the longer context windows are still losing focus because there's some sort of fundamental limitation of the transformer uh we talked to dorkesh a little bit about this and uh the continual learning breakthrough is maybe still a few years away but uh certainly will be interesting to see uh how it develops i'm i'm optimistic.

2:29:57I'm still looking for that. I keep coming back to that idea of the Lisa Dahl match against AlphaGo, where AlphaGo dropped Move 37, this very unconventional play. Everyone thought it was a hallucination, a mistake, and it turned out to be kind of a genius new move. And I feel like we haven't had our Move 37 moment for LLMs yet, but it's probably coming at some point. Well, I'll tell you, each time we have these conversations, the whole world will be different. I guess that's the thing we're learning these days in terms of the pace of change. But good to see you there. That's great. As we accelerate it, it goes from monthly to every couple weeks.

2:30:32Weekly, daily. Weekly, daily. And then every hour that we'll be feeling the acceleration. But great having you on as always. Looking forward to the next one. Sounds good. Until next time. We'll talk to you soon. Bye. Next up, we have Nathan Lambert coming in to talk about an American DeepSeek project. But first, let me tell you about Bezel. Go to getbezel.com. your bezel concierge is available now to source you any watch on the planet. Seriously, any watch. Go check them out. And I'm very excited to bring in Nathan and talk about Deep Seek Llama. How you doing, Nathan? Boom. Good to meet you. What's going on?

2:31:05Good. Thanks so much for doing it. Happy to be on this format. You guys got a loaded lineup today. I was like, wow, I got on the same day as Ben. Ben is like the motivation for why I started writing about AI. That's amazing. Somebody has to do this for AI because there's so much to talk about, But all he does now is AI anyways. So yeah, he's a competitor. He's awesome. The Ben Thompson for AI is definitely Ben Thompson. But I mean, it is a little bit different. There's so much different space in terms of whether you're going after the business models or the actual infrastructure or what you were writing about earlier with kind of the open source geopolitical angle.

2:31:39So take us through the recent piece, the thesis. And then I have a bunch of questions about both DeepSeq, Lama and kind of how this could come together. We were talking to the CEO of Grok yesterday. He's obviously extremely long, open source, and it's very interesting to dig into a million different threads here. So just kick us off with an overview. Thankfully, we were talking to the CEO of Grok with a Q yesterday. Because at that very moment, there was a different Grok. Yes, hallucinating and at scale. Yeah, there should be more Grok news later. The tweets are to be believed. But the American DeepSeek thing is largely a forcing function to make the AI research ecosystem in the United States catch up.

2:32:20I think we were talking about niches and lanes and like Ben Thompson, it's the biggest one. A lot of mine is on the research side and kind of understanding the emerging trends on research that are getting picked up by companies. And one of the clearest ones that we do is I lead this with a couple other people as we keep track of all the open models and data sets and make sure of research and startups are releasing. And there's been a huge shift in the last three to five months, and pretty much everyone builds on Qun. And there's a long tail of kind of business or geopolitical reasons. Some of them are sensitive and some of them are kind of obvious, where American companies don't want to build on Chinese models.

2:32:54And that's one thing. And then also, America should just take pride in what has been a great research ecosystem. And we want to have that and own it whole stack, which is otherwise most of the leading AI research is going to start coming out of China. And I think so politically it should be an easy win. And in terms of cost to maintain the open ecosystem, it's so much less than what these companies are pouring into their AI models. So it's just kind of getting a bit of the tractor beam of AI onto this open source and open. It's like just building models that have all the data and code released so more people can start building on them.

2:33:30So we can go to the details from there. DeepSeek versus Quen. I feel like DeepSeek had this like crazy viral moment. But now you're saying that Quen's been kind of on a compounding growth for a while. What's the dynamic between those two companies and what's driving Quen's adoption over DeepSeek? Yeah, so this is a great example of one I started using and we'll loop it into Llama. Essentially DeepSeek has these frontier class models that are extremely good and they drop the weights on Hugging Face and they have been switching to permissive licenses. These are models that anyone could pick up and use and dump into a product that they want to ship.

2:34:05A lot of startups use these things. We see all the clouds hosting them. That's one thing. Not a lot of people actually fine tune DeepSeq because it's huge. It's already so good. Like, what are you going to get from this? And then what Quen is doing is they're releasing, honestly, tens of models at different formats, both base models and post-terrain, and from size scales as small as like 500 million parameters up through these bigger MOE models. So for a researcher, it's pretty much somebody trying to build a really niche product and something at the cutting edge, it's like, QN will have the model at a certain size that you need in order to fine tune it or figure out if you can build a certain thing at a certain cost profile.

2:34:43And especially for researchers that have some sort of limited compute, like training and working on DeepSeq is super hard. And we've seen this with LAMA2 and LAMA3, we're much closer to this QN approach when it was seen both in the data that we have and on the ground is like LAMA was the open standard for research. I used to joke around that like hugging face is just going to be rebranded Llama because you see Llama everywhere, especially around Llama 3. And with Llama 4, Meta started to go, like, release Llama aside, they've started going more towards their bespoke solutions, and they're also releasing the models, which has made this big opportunity for Quen with Quen 3, which, honestly, earlier Quen models were already starting to fulfill this.

2:35:22But that's kind of been a big mass shift and attention shift in the last few months. So what do you expect out of Meta with the new talent acquisitions? And it seems like a redoubling of the efforts on AI broadly, super intelligence. But maybe, you know, the strategy of LLAMA could be shifting. Are they the lot? I've often thought that, you know, right now with the dominance of DeepSeek and Quen internationally and kind of these like jump ball, half ally countries, frenemy countries, Mark Zuckerberg should be like a national champion and we should be pushing Lama at a national level all over the world.

2:36:06But what do you think is going to happen there? Yeah, so there's two things. Mostly, I'm going to gossip, as anyone will, on what will happen with Lama. It's very 50-50. I think with the leadership they've brought in that there's less attention and value behind the open thing. So Zuckerberg historically has been very pro-open. And if more of it is shifting to other people, they kind of loosen that vision if other people are at the lead of leadership. Like that's what people are saying. Like they need more leadership to build this AI org. And so a best case scenario is Lama invests more in AI and the national champion becomes even better and they just crush this.

2:36:47I don't think that's the outcome that I expect to happen, which is I'm saying there's a cheap off ramp. If you take the COMP packages for a couple of those researchers, like that's the cost to get a whole research ecosystem built around a fully open U.S. model where we have all this, like the data is released and a nonprofit and stuff can handle data releases a bit better than a big tech company that has all these eyes on their back. And then just have research happen on these things. So what are you advocating for? Are you advocating for nonprofit taxpayer funding? You know, we've seen a nonprofit before and it turned into a for-profit.

2:37:25How are we going to keep this in a for-profit? And then how are we actually mustering the will? Because yeah, it sounds like, yeah, just put$100 million into a nonprofit. That's a lot of money. This is where we get to the nitty gritty. I work at the Allen Institute for AI, AI2, which historically is even more academic than OpenAI was. So I think culturally there's not that type of feeling the AGI like supervision that Ilya had on the scaling deep learning. That was kind of the thing that I think drove them from the start. They were like, we need so much money. So AI2 is set up, it's like you can do digging, but it's so different in that culturally that it could never happen.

2:38:04So if the leadership here tried to do that, the company would just implode going for a for-profit. Because, I mean, most of the leadership has co-appointments with professorships at UW and things like that. So it is already half embedded in the ecosystem. And then practically speaking, moving AI talent around is so hard that it's like the government exfiltrating researchers to fund like a open source government lab doing this is so hard that it's like you have to find the money or the partners to do this where there are people. So sitting on the ground where I am, we're trying to train our next model, like it's pre-training now.

2:38:39It's like we just need more compute. So if we double or triple our compute, America will have X or like these open models will be just X percent better. And it's it seems tractable. It's just hard to get the right. It's a lot of politics to get these things in place. Yeah. I mean, I guess to to dig in there, I'm wondering if like, you know, the the initial like economic model for Lama was always in a little bit of debate. Is it a recruiting effort for Meta? Is it their desire to decouple and not be dependent on Gemini or OpenAI or Anthropic and just save costs there? There were a whole bunch of different economic motivations.

2:39:28But I'm wondering if in the long term we don't see something like a Red Hat Linux where it It is a for-profit company maintaining a non-profit or an open source software package. Or even, I mean, you can run Linux on Azure now. And so is there a world where you just have every different piece of the stack, whether it's a consumer app that people pay monthly for or an API that people pay on a per token basis or an open source model that you're paying for, you know, Red Hat style consulting services on top of all within one company? Like, is there no hope that OpenAI's open source model or, you know, AWS or Microsoft open source something that actually competes significantly with DeepSeq and Quen, but is still within the typical corporate structure?

2:40:24I think it'll come from the biggest motivators have to be NVIDIA and AMD. So the long tail of if Quen is going to keep doing this, and then it's something like Huawei, they start working with Huawei, this like Huawei libraries, they want to support them. And then US researchers are starting to dig down into those levels because they want to understand how these models were trained. So those are the people that have the most direct exposure. It would be NVIDIA and AMD. It's cheap for them to do. They get the benefit of the researchers keep working on their hardware and software ecosystems. There are outlandish stories that you could tell about open source AI where that type of thing could emerge, but mostly they revolve technical breakthroughs that you can't plan on.

2:41:06If you could do weird model splicing where you train a bunch of MOEs and then you cut an MOE out of one model because it's really good at healthcare, and there's kind of this open marketplace for model parts, and then that's all in the open. And the person that kind of writes the software by which those pieces of models are combined and standards by which those happen, that could exist. So there's kind of wacky ideas, but I think that that's a lower probability outcome. And it's just like, let's get good big transformer models trained that anyone can download and poke around at and just get more people involved in AI research that we have complete control over.

2:41:43Last question for me. What are you expecting out of OpenAI's open model? Yeah, so I think one of the core things about OpenAI culture is that they really like to deliver extremely cutting edge and good artifacts and research. So I expect it to be one model that fills a niche that they're either hearing from customers or the community that isn't quite filled, whether it's an extremely, like, super long context or low latency for agents or a certain size of reasoning model. It's going to be this type of thing where it's a certain niche and it works really well for it, where it could be like a DeepSeq style release where it's just super strong model that people can plug into real world products and applications.

2:42:26but it's not going to be this Qen or Lama suite of models that researchers look at for all sorts of things. And OpenAI has been saying that they hear the license critiques of Lama and stuff, and they're going to commit to the actual permissive license, which are things like Apache or MIT, that these Chinese orgs have started using again, which I think is a nice thing to kind of make all of that simpler. You release a model, it doesn't have terms and conditions on it. like Meta's not trying to say like your legal department has to talk to us or avoid these use cases. It's just get people using the model that your company released and take a simpler approach.

2:43:04That makes a lot of sense. Anything else, Jordy? That's it for now. Thank you so much for stopping by. This was fantastic. Thank you for working on all this. Yeah, we will talk soon. Looking forward to the release. Cheers, Nathan. Have a good one. Bye. Up next, we have Richard from you.com coming into the studio. do we have any more ads? Go to AIDSleep.com get a pod 5 5 year warranty 30 night risk free trial free returns free shipping at AIDSleep.com I'm basically you're in a hole my household is in sleep shambles well let's bring in Richard talk to him putting up the fact that we're doing this on 5 hours of sleep consistently impressive how'd you sleep last night Richard?

2:43:47how'd you sleep? good to meet you hey guys Nice to be here. I slept all right. That's good. I'm complaining. You would have slept better on an eight sleep. So we'll work on that offline. But great. I did buy it. I retired it. I think it's better if my body sets its own temperature. Oh, wow. Interesting. Would you mind kicking us off in the introduction on yourself and the company? Happy to, yeah. Richard did my PhD at Stanford, brought neural networks into the field of natural language processing, laid a lot of the groundwork for what now is Chad Chuby T. I was a professor also on the side at Stanford for a couple of years because no one was teaching neural nets like transformers and so on to students back then.

2:44:28This was 2014 to 2018. But my main job was starting Metamind. I started and made it very easy to train neural networks for other companies. We got acquired by Salesforce, became the chief scientist, and after two years, executive vice president, running most of the AI efforts, starting the Einstein suite of things and so on. build out the research team there. In that research team, we ended prompt engineering, a paper cited by the early GPT papers from OpenAI and others. And in 2020, I decided to start U.com to bring better answers to the world, to change what I thought initially was search, but I think now I think it's something else.

2:45:07And also started AIX Ventures. It's a relatively small venture firm, about half a billion AUM that invests in early stage AI companies. It's not that small. Half a million AUM is pretty solid. Congratulations. A humble 500 million of AUM. Humble, yes. On you.com, where is the business today? What are the biggest challenges? I feel like we've been hearing more and more about the data wars and how high are the walls and the beautiful gardens that we have all tended to with our Slack installations and our Google Drives. And this feels like the logical thing that, you know, yes, I own my data until I want to give it to you or literally you.com.

2:45:51That's right. Actually, you know, as a startup founder, you have to remind yourself every price is an opportunity. And the opportunity here is actually that a lot of data is in silos and those companies don't want to give it out, but they do need to make it useful. And so one of the many things that we've learned over our changes and focus deeper, deeper on enterprise is that doing good internal search is actually quite hard and quite useful. And so we're partnering with a lot of companies and do search over their entire archives of decades and also really up to date things for publishers and insurance companies, like pretty gnarly, complex problems.

2:46:29and combining that with web data, which also has its own complexities. And a lot of folks are in some ways trying to pay people in news, which makes a lot of sense, but also potentially threatens the entire free open internet. If you have to pay for everything you read or crawl, then only Google can really afford that and then you have an even bigger monopoly. So I do think we need to keep the web open and free for that. And so merging all of that together to make companies more productive is what we now focus on. What's kind of the best practice in the modern enterprise these days? Is it like try and be really diligent about sucking data out of every platform you use into some sort of data lake or like Snowflake installation and then dropping u.com on top?

2:47:17or is it figuring out a way to actually deal with the sharp elbows of direct API integrations into the databases that are managed by the other companies that I'm you know purchasing SAS from that is a great question I wish there was a simple silver bullet always do X and just have it all in data lake in one place but it's kind of messy and usually there's some data that's just so large you don't want to have another copy somewhere else. And then there's some data where we have the whole internet, like we have an internet index, right? And you can't bring that into your virtual private cloud and so on.

2:47:52It's just too expensive for each company. But then in some cases, it does make sense if you want really deep understanding, reasoning, over complex, structured and unstructured data inside an enterprise, then you have to often copy it over and bring it into a new setup. So one of the big things we just announced actually is a big partnership with Databricks, where we are sitting on top of Databricks and we can actually answer questions over data that is in Databricks. And it's been a very exciting partnership already. Yeah, that makes sense. They also enable all their LLMs to have web index access.

2:48:27Oh, sure. Okay. Yeah, that makes sense. How are you feeling on acceleration, deceleration? It feels like the vibes have shifted most recently. We had Dwarkesh Patel on the show on Monday. He's kind of pushed out his AGI timelines. Folks are talking about reinforcement learning not scaling as fast as people thought it would, the problem of continual learning. Just if you take a step back and maybe put on more of your academic hat, how are you feeling about the current state of AI? Obviously, even if the, there's also the take that like, I don't know if you agree with this, but like even if the models plateau, there's still so much enterprise value and so many problems to go self, but I'll let you answer.

2:49:09Where do you stand? Yeah, I'll try to keep it short because I can talk about that for hours. I think it's true that there are so many simple jobs that can actually be done already with the technology that's there, assuming you have good data access and you had all the recent information and you know the company context and all of that stuff. So there is already a lot of low-hanging fruit. At the same time, it's actually quite exciting for the researcher in me that hasn't fully died yet as an entrepreneur and CEO for many years, like that it's time again for research. Like in many ways, we've known that you need large neural nets with a lot of data on GPUs and highly paralyzable training.

2:49:51And you want the whole thing to be ideally end-to-end trainable in some fashion. It's sort of been known ingredients for over a decade now. And indeed, we've crossed thresholds by scaling all three data compute and and model size we scaled those three things up and it worked better and better and better and it created these emergent properties similar to like a smaller brain of a monkey just not doing certain epic things even though the brain kind of looks similar that has neurons too but then you cross a certain threshold you get these emergent intelligent properties and so what that means is now is the time again for actual research, not just engineering and scaling things up and throwing more data and better data at it and bigger GPUs and all of that, but to actually go back and say like, okay, what is true intelligence?

2:50:40How can we get to superintelligence? What does it mean for intelligence to increase exponentially for a certain amount of time? I think there are actually different dimensions of intelligence too that you have to look at separately. And some do have upper bounds. and some of the times these upper bounds are astronomically far away, like grounded in physics. And in other cases, the bounds are not that hard, like classify every object on the planet in computer vision. It's actually not that hard in comparison to have all knowledge about the universe. And intelligence should have a lot of knowledge and it'll take us a while to get to as much knowledge and the bounds of how much knowledge you can collect are rooted in physics and the speed of light cones around all the sensors you can have.

2:51:21So there's basically a time again for research. And that's exciting. So yeah, it feels like we're kind of paradoxically in an AI bull market summer in the Stripe dashboard or in like the ARR sense. Like we've never been adding more EV, never been doing bigger contracts. Everything's good on the business side. But maybe we're kind of counterintuitively in a little bit of an AI winter on the academic side. My question is, if you agree with that or not, but also, do you think that with the AI researchers, it feels like the top AI researchers are getting poached into meta and they're going into open AI, do we think that the next Transformers coming from, or the next major research breakthrough comes from a foundation lab or a big tech company, or is there a role for academia to step up and do some like kind of longer timeline, unbounded research to try and go explore even without even an economic model in mind?

2:52:18I do think you cannot build another OpenAI by just building an LM. The LM was the one thing that worked out for OpenAI after they spent hundreds of millions on robotic hands and on Dota, like computer games and reinforcement learning and all these other projects. And one of them actually worked. And so if you want to replicate that kind of success and do research again, which I think is that the time is now, it does make sense to have that kind of entity. And it can, I think, be done. And so I do think academia has a role to play in that. Thanks to open source models, academia can be relevant again, because they have access to them.

2:52:55They just couldn't have afforded it before top open source models were available. And I do think I think a lot of folks are chasing sort of the latest employees in the top labs like OpenAI, Anthropic, and so on. But you can also go a step further and look at who's actually, who's trained those people. How did they learn how to do research? And then you get to folks like Chris Manning, who's one of my PhD advisors too, who just recently joined AIX, our venture fund, like in a much larger capacity as a GP. And so those are the kinds of folks that I think we'll need to rely more on again. And many of those are also moving out of academia into kind of labs that push frontier research forward.

2:53:39What I have to ask because it's so current. What do you what's your thesis on what happened yesterday with Grok? How do you have that big of a general, I don't know, alignment? Oopsie. Oopsie daisies. In prod. You know, I think when you ship very fast, these things are about to happen, right? Like people can push them in the conversations into certain directions. You sample from the same model multiple times, you get different answers. And I think if you try to be sort of a like free speech maximalist, which on many levels makes sense, but turns out there's a lot of funky speech out there you know with no guardrails whatsoever it will go into those very dark places yeah what are you expecting uh since it's in six-ish hours i i believe i hopefully they're still announcing and launching grok 4 tonight what are you expecting out of grok 4 in terms of um maybe benchmarks aren't aren't the right um just even even uh way of thinking about it but but in terms of progress?

2:54:47My hunch is it'll be more of everything, but nothing like, wow, like binary, like a novel thing, right? It'll be more multimodal and deal with better understanding of images and videos and maybe sound. It'll be larger memory. It'll be slightly better reasoning. It'll be probably, I mean, we won't know for most of them, but more parameters and so on. but maybe not like completely novel research that has a capability that no one else has. Sure. Yeah, that makes sense. Last question and we'll let you go. What are you finding most exciting on the investing side? For me,$500 million fund, it feels like the foundation model labs, the big training runs, the multi-billion dollar rounds ship has kind of sailed on that front.

2:55:35So maybe the time is for application layer investing, but what are you seeing? What are you excited about? Yeah, we've been actually very fortunate at AIX Ventures to avoid some of these massive rounds. The way I describe that is that not every company that raises a ton of money in a very early stage is bound to fail. At the same time, you basically combine a seed stage risk with a late stage return in many cases. And that expected value just doesn't work out very well. And so there are a handful of foundational companies like Hugging Face that we invested in at a seed round and Windsurf. Congratulations.

2:56:22Those are great companies. Yeah, these are all like companies invested in this seed round, invest in Perplexity and Flow and Whisper, Tobit, Ambience, like a bunch of really amazing companies. And so I think there are only one or two dozen foundational model companies that the world needs. and then there are thousands of application companies. So short answer is yes, you're right. I think you will see a lot more in the applications. I also believe that all the stars are aligning for biology and hence medicine and hence health to have a major moment thanks to AI. Now, software is faster than hardware and hardware is faster than wetware people and biology, right?

2:57:01And so it takes longer, the cycles are longer, but you have enough data, you have the right compute, and we can eventually simulate more and more of biology and then make it into not just, oh, memorize what nature has kind of hackily evolved towards, but make it an engineering discipline. Think about how we can actually change a specific antibody and target a specific gene and like change our epigenetics and improve aging and cure cancer. All of these things, I think, are within our grasp and reach over the next few decades. And I think that will be a massive group of applications. Amazing. Well, thank you so much for stopping by.

2:57:40This is a really great conversation. yeah we'll have to have you back later I love you back on again soon cheers Richard thanks for joining good chatting next up we have the founder of Moon Valley coming on the show talking about generative imagery we're gonna pull up very very cool stuff so welcome to the stream hopefully we can pull up this website because we have fully passed the uncanny valley don't you agree I mean And this, are you on the website? Can we pull up the website really quickly and show the video? So everything on your website, this is AI generated. Is that correct? That's right.

2:58:20There's some like After Effects stuff on top of it. But yeah, it's basically all AI Mary generated. It's remarkable. And I feel like it's under discussed. Anyway, please introduce yourself and the company. Because this is fantastic. Yeah, absolutely. Thanks for having me. Great to meet you guys. Great to meet you. So I'm Naeem. I'm one of the founders of Moon Valley. You know, at a very high level, we're a team of kind of a unique structure. You know, a good chunk of our team are world-renowned researchers in visual intelligence primarily. But, you know, our focus in a specific level is we're building the biggest and the most capable production-grade generative video models.

2:59:01And on the flip side is we also have a large chunk of our company are filmmakers. So we have a movie studio in L.A. It's one of the oldest, most well-preserved sound stages in the world. Some of the first Charlie Chaplin movies were shot there. And so it's like kind of a ground zero. And we have, you know, folks that have won Emmys that have been nominated for Oscars on the team. And so we've just kind of brought both worlds together to figure out how do you take this tech from being, you know, kind of interesting research and, you know, cool things that you can share on X and actually become things that sort of push the boundaries of, you know what we sort of we think of it as like movies and stuff but these are sort of limited abstractions more of just like visual media broadly um and and kind of the artistry around it so how do you think about the trade-off between training new models being a foundational model company hiring researchers huge training runs and then the application layer the distribution actually working with filmmakers that feels like most companies have kind of split between one or the other?

3:00:05Are you doing both right now? How would you describe the shape of the business? Yeah. You know, it feels a little bit, maybe it's like a bit of a post-chat GPT phenomenon, but I do think that increasingly foundational companies, quote unquote, are having to think a lot more about the application layer. And I think that that's only going to continue to be the case. Like I think a foundational research, which, you know, and you kind of alluded to it in your chat with Richard, but like there's an element of commoditization that's happening, right? There's an element of you'll get to there's a point, and this applies to visual media as well, where the the output of the model, you kind of hit this point of diminishing returns from a consumer perspective, from like a user perspective.

3:00:52So there might be really interesting kind of research thing that's happening and you'll continue to invest in that, but to drive the same amount of like business value as like, you know, a GPT-4 style leap that requires kind of thinking about other other areas. So and the other piece for us is I think it's different with things like LLMs, but with visual and video in particular, I think one of the issues is that like research labs and technology companies that have been in this space, they have been largely divorced from like the end practitioner in a lot of ways. And for LLMs that are so such general technology, I think that makes sense.

3:01:33But in our case, you know, we're building tooling that filmmakers will ultimately use that like creators will ultimately use. So we have to understand that inside and out. And that needs to help guide the research rather than the research happening somewhat in a vacuum and then, you know, kind of trickling down to the target user. Yeah, a lot of the crazy technology vision of the future is kind of just like, you're going to, with one prompt, just be like, make me a new Top Gun movie and it'll just one shot it. Clearly going to be a while until we get there. Where are you actually seeing value or demand from Hollywood, from filmmakers?

3:02:10Because, you know, AI is so broad, it could mean just like pull a green screen key better. Do some rotoscoping, do some camera stabilization. there's been AI tools in filmmaking for a long time. They're obviously ramping up set extensions. There's so much that you could do in the 3D pipeline, the VFX pipeline and 2D pipeline. Where are you seeing actual adoption? Where are you excited for there to be adoption in the next year or the year after? Yeah, for sure. I think, you know, it's a good point where, especially in video, I think more than in other kind of fields, there has been this like, you know, when it first started, there was this sense of like, well, you know, kind of like the holodeck, right?

3:02:50Like that's the world that we're going to move towards. And to an extent it is. I do think that there's kind of like a misunderstanding though of where the value of the end content comes from. And it's sort of like, we're now at the place where you could relatively credibly write a book with an LLM. Like you could have Chad GPT, you know, publish literature. Problem is nobody's going to read that literature. And like, that's the, that's kind of the missing piece you know people have done it people have and if you go on the kindle store apparently it's like swamped with it's like 90 yeah and but like and every once in a while these things break out but it's more of like a novelty like oh wow like somebody actually did this thing like let me leaf through it wow yeah it's you know they hit the periods and there's tons of i think it might just end up reflecting human creation where it's ultimately creativity creative products or hits business where there's a lot of ai songs right now but the only ai artists that i I can think of is that the Velvet Sundown or whatever that's gotten popular in the last month.

3:03:50And so it's, yeah, I just think it's like this ultra power law, potentially. A lot of media properties also, and art generally is also very story driven. Like the story behind it. Like part of the reason why I like to go see a Tom Cruise movie is because I've heard the story that he's doing the stunts. And even if - Cause we know who he is. Yeah, we know who he is. We know the story behind the story. And that's what drives a lot of value in art is like, oh, this painter really spent years doing this thing and he cut off his ear. And so that this painting has a crazy story behind it. So it's valuable, even if like my kid could do it.

3:04:25It's like my LLM can do it. But yeah, your LLM didn't. Yeah. But I think it's like in AI music, I think you're kind of starting to grapple with that a little bit where it's, you know, you can you can use a lot of these tools. I think that there's like kind of lowest common denominator content. Like what, you know, what you don't see as much anymore is like blog spam. That was just like crazy in the 2010s. Right. Nowadays, that's just replaced with AI. Yeah, exactly. I think that with AI music, it's like there's certain strands of like top 40 kind of radio, you know, where it's like it's already a very commoditized, you know, sort of form.

3:05:04That's what it does well. But I just don't see a world where a transformer model replaces, you know, Kendrick Lamar. Like that's it's so, you know, there's such a big gap there. So we think about it like internally, it's the same way we, you know, like John Carmack calls it like power tools. And that's that's largely the model we use. I think in film as well, it's very acute because to your point, like unlike other spaces, it's actually it's a continuation. Like there are a lot of the things that AI video enables. It's not novel. Like, you know, you'll hear studios that we talk to, they'll say, you know, a research company came to us and they said, hey, you can do all these new things with with these video models.

3:05:42And they have to remind them that, no, we can do all of these things. Right. Like what we're talking about here is potentially doing them in an easier, more flexible, more powerful way. But with VFX, like there's nothing you can do today with in terms of like the output that you're creating that you couldn't do with VFX. It's it's a workflow thing. Right. We're just making that process easier, that process more affordable and that kind of thing. So it's yeah, we look at it. We're selling we're selling SaaS. That's what the market needs. What about what you know, it's been interesting to see X is its own kind of universe in terms of AI content.

3:06:21What gets picked up? A lot of it ends up being stuff that that's getting made shared on TikTok or other platforms. and there's been this idea of like an AI content creator, which is like a new personality that is just being generated by a human that's creating, you know, using a video or an image model to generate this person, you know, going about their life. And then there's this sort of like the, what's considered the slop, like the Italian brain rot, all that stuff. um what about like do you expect to see like an entirely new class of filmmakers in like you know like basically like net new youtube channels things like that of people that are just a you know could be a teenager or just somebody sitting in a room by themselves focused on that storytelling focused on basically creating um you know the internet was beautiful because it lowered the cost of distribution to zero and you know mobile devices lowered the cost of content creation to zero or effectively zero and now the the cost of of producing films is not going to go to zero but but may as well on a long enough time horizon and so i'm curious when you think that moment will be where we start to see kind of the the velvet sundown equivalent of of of film filmmaking Yeah, I would say that like, we're a lot closer than people think.

3:07:55So, you know, we like Moon Valley models, because we have, you know, clean models or models that have been trained on licensed data, there have been the first models that studio, you know, legal teams have really like allowed them to use. And so that's been part of why, you know, when you heard about Sora like a year and a half ago, since then, there hasn't actually been that much AI adoption in the industry. Now it's starting to pick up a little bit. I would say that like the the the point where the technology is capable of doing that kind of thing, we've already reached. reached. And there's places.

3:08:33It's there's pockets of the world where especially like smaller film communities in different parts of the world, where you're starting to see like real world productions, productions, often people aren't necessarily aware of it. But very large parts of not just the pre and post production, but even the production process itself are being driven by these models. Now, I think that there's like, beyond, especially in this industry, beyond the technology, there's a lot of other barriers to adoption here, right? Like you have to, it's a whole new, you know, it's a whole new set of tools that you have to figure out.

3:09:06It's a whole, you know, the adoption curve is very different. And of course, you know, for better or for worse, there's like, there's a big kind of dialogue around AI in movies and what that means and, you know, how we feel about it. But to your point, like what we get excited about internally is there's this dialogue of like, you've got on one hand, just like day-to-day people like myself, that maybe, oh, now I'll be able to create a movie. On the other hand, you have the far end of, well, now studios will be able to make the same movies for much cheaper. And that seems to be where a lot of the focus is.

3:09:38To us, the area that we're the most excited about is that middle, where you essentially have this band of millions of creative people in the world, like artists, people who have real taste and talent and ability, but they don't have necessarily the access and the infrastructure to do that, right? And that's not necessarily like, you know, we have one of our alpha users is this. He's a filmmaker in Senegal. And like he's been he's been a filmmaker for over a decade. His sting, his thing or, you know, his focus is on doing like music videos for local artists. And it's like this kind of funky, like Afro beat style, you know, West African kind of flavor.

3:10:19And now he suddenly started making those music videos using generative AI. And like the production quality of these have soared. And he's had a number, like he went from being just like a very obscure, you know, person in his local community to he's had some videos that have gone up like 10 million views on YouTube now. And nobody has any idea. They just think, oh, this is just like a sick artist that like I haven't heard before and, you know, cool visuals. So it's that middle layer. It's like the independent filmmakers who today, unless you're friends with one of the top five studios in the world, you have no capability of making a big budget production.

3:10:55Now you'll be able to do that. Right. And that's not just, you know, obviously this individual is one example, but for instance, we have like somebody like Natasha Lyonne, who we work really closely with. She's, you know, an industry insider. She's a, you know, one of the kind of top people in the industry today, but she's working on this new movie. That's like, she's been leveraging AI to help do it because this has been something that she's wanted this is a movie she's wanted to make for well over a decade but she just couldn't she would like she talked to the major studios and it's like hey it's going to cost us 30 million dollars to do this right and and we just can't we can't do that now it's like well if i can suddenly potentially do it for 15 million dollars now i can you know i can make this thing happen right and and so it's there's there's there's this like idea that you'll be able to do, you know, movies for cheaper.

3:11:46But really what we're seeing in the studio is a existing movies. You're now just doing more than you expected to do before. You're now having to compromise less like directors on saying, hey, I had a$75 million budget. Now I'm going to do it for 50 million. They're saying, hey, with my$75 million and my team, I'm going to go and now do all the things that I couldn't do before unless I had a hundred million dollar budget. That's like one thing we're seeing. And then the flip side of that is you walk into any studio for every one good production that's live there's 10 that weren't greenlit because they just couldn't get the budget right now suddenly you'll have a lot more of that so that's like high level how you know i think that the way that this stuff is actually getting implemented it's happening in a different way than i think where there's been a lot of fear um and and you know i think it's totally justifiable fear but i do think that ultimately it's the artistry that wins out more than like you know budget requirements or budget constraints yeah and the i mean the exciting thing from my point of view is if movie studios keep budgets relatively the same because there is this incredible demand for content but then you can take a hundred million dollar budget and it can now go to 10 different films you get 10 more shots on goal it's 10 more teams and maybe the underlying teams even can create you know better margins themselves so very exciting yeah yeah this is awesome thank you so much for stopping by hope you have a great rest of your day yeah come back on when you talk to you soon awesome thanks for having me guys later um and that is our show for today jordy do you have any other breaking news you want to share breaking news brandon jacoby trade deal well i guess now's a good time as any for some personal news after wild chapter at x i've officially wrapped up my time there i'm extremely proud of the work we did and you'll see more of it soon more to come on what's next dot dot dot uh brandon has been a dear friend for a long time and uh i'm excited for his next chapter me too and uh i'm gonna miss being able to text him when i have bugs uh but uh tyler you're up next expect some bug reports but um yeah fun show i'm trying to think if there's anything else that we missed.

3:13:57I don't think so. We hit all of our ads. Chamath was saying that Meta switches to Sonnet for coding instead of using Llama, that they had fine-tuned on Meta's own code base. Now they're just one-shotting everything with Sonnet. So very interesting. Since the change code suggestions are generally better, engineers can change back to Llama, and occasionally do when the fine-tuning makes a difference. Internally, this is a big change, given how big, how heavily Meta has invested in the Lama project or product. This move officially acknowledges that Anthropics models are currently far ahead of Meta's own LLMs, even when fine-tuned.

3:14:32I suspect Meta will double down and try and make their next Lama versions more capable for coding, but until then, it doesn't want to hold back its engineers. So very interesting that they're using Anthropic and just kind of letting it rip. Anthropic is cooking. last post to end it from at Jarvis best he says he's sharing a screenshot of Linda Yaccarino saying after two incredible years I've decided to step down as CEO of X obviously has a much longer post which we covered earlier and Elon just comments thank you for your contributions Jarvis says LMAO cold as dry ice anyways at least they were cordial Yeah, at least they record your own.

3:15:19Leave us five stars on Apple Podcasts and Spotify, and stay tuned for our stream tomorrow. Can't wait. And hope you're having a great summer. Have a great evening. Talk to you soon. Cheers.