New Tariff Implications, DeepSeek and Open Source Analysis, and Meta’s Big Bets w/ Byrne Hobart

Hey, Upstream listeners. Today we're releasing a conversation I had with Burn Hobart on Turpentine's show, The Riff. We discussed the recently instituted tariffs, DeepSeek, Meta's VR bat, fantasy stock picking, and more. Please enjoy.

Hello, Bernd.

How are you?

I'm doing great. I've just landed in Sydney where I've come to hand deliver the tariffs personally.

Have fun.

Yeah, yeah, exactly. Just kidding. But while we're on the topic, before we get into it, are you, do you, if you had to predict, do you think that some of the vibe shift will come to Europe in a similar way that some of sort of, you know, what's happened in the US might, you know, hit England, Australia? Or, or, or sort of, you know, these, these countries.

I think that in some ways the European vibe shift is more, more extreme in terms of voter behavior than it is in the U.S. There are, there are like, and some of it is just that if you have a system that is designed to have multiple parties, then you typically have room for a more extreme right and a more extreme left. But even within that, there's, there's been more, more reshuffling and you know, a lot of. Like a lot of what Trump does is rolling things back to a status quo that existed for a while. Like that is, that is what conservatives and reactionaries want to do generally is take something that they believe broke recently and put it back. And, you know, there's a debate on how far back you want to go. And there's also like within that debate, there's also the question of like, how literally do you take any of these rollbacks? Are you trying to do, are you trying to? Make it like the '90s again, where China is not in the WTO and you know immigration is a bit more restrictive, etc. Are you trying to do that? Are you trying to ask yourself, okay, what would like a centrist Clintonite from 1990s from the 1990s like you transported them forward in time? What would they be doing today? Because those those give you different answers, and I think that that does make some assessments of this kind of tricky. But like a lot of a lot of Trump's views on social issues are just things that were pretty pretty centrist in living memory. Or you know he's like drifted, but more slowly than the average person. And then on economic issues, he is more of a throwback there, particularly with respect to the tariffs. But I also can't tell how much of the tariff stuff is a Trump-specific vibe shift. How much of that is just general skepticism of globalization, skepticism of the economist establishment? I also think that within tariffs, and there there have been people who are clearly opponents of tariffs broadly, or you know not Trumpy people, but who have. Said that the tariff debate is a little bit more sophisticated than just talking about the first order consequence of there is deadweight loss. If you make it harder to import goods into a country, that loss is going to hit consumers. It's not just like everyone, everyone ignores the fact that there are tariffs and suddenly everyone who exports goods to the U.S. just accepts a, if they're from Canada, accepts a 25% lower margin on that. Like that's, that is just not a realistic description of human behavior. That said, within the, within that idea that, yeah, you could be more sophisticated on tariffs, there is a lot of nuance to exactly how you do that. It is not trivial to set up a tariff regime that is going to encourage industries where the U.S. could compete, but maybe U.S. can't, you know, there's, there's not a, the financial markets don't have a will to fund companies that will scale up to the point that they can compete, but the government can implicitly fund that by making consumers worse off. Like there are cases where that's a high ROI decision to make. And a lot of the East Asian countries. That had just enormous growth, particularly manufacturing-driven growth, one of the ways that they would do that is they did protect their home markets. So Japan was a great place to be an exporter and a pretty bad place relative to GDP to be just a consumer when they were at their peak growth period. But that, that works if there's, if it's an investment and if you can build up these export-driven industries such that they provide higher returns later on, such that you can, such that you turn those exports into importing things that people actually want. Like that is the end goal. And the U.S. is in a weird situation because we've sort of achieved that end goal where, yes, people are really, really happy to take dollars in exchange for things that American consumers want to buy. But the U.S. is also just not in a position to manufacture a lot of pretty, pretty straightforward intermediate goods. And some of that is also just when you have expensive labor markets, you can't, you can't cost, cost competitively manufacture some of this stuff. It is just going to go where the labor is cheaper. So there, I think there, there are a lot of problems that That are worth talking about, but blanket 25% tariffs on countries that are extremely economically integrated with the U.S., like that makes sense as a bluff. It does not make sense as a long-term strategy. And it also like, even as a bluff, it is such a, it is a sufficiently weird thing to do and has so many economic downsides. Like automotive, for example, if you have goods that are just constantly zipping back and forth between the borders where there's like Some component that is made in the U.S., it's put into something else in a factory in Canada that's sent back to the U.S. Ford puts it in a car. That car gets sold to an American. If you're putting a 25% tariff on every leg of that transaction, then a lot of complicated supply chains just can't be that complicated. And maybe it's worthwhile to have these supply chains all be equally complicated, but all be within the U.S. I don't think that is. I don't think that category of good like. intermediate goods made out of other intermediate goods that are put into final manufactured goods. Like, I don't know that that is the strategically important industry that the U.S. really ought to be protecting. So the tariff stuff is a, is a separate part of the vibe shift. But I think there is just a general, like, it is very old news that there are a lot of people who have defined an establishment that they are very skeptical of. And that used to be a very left-coded thing. We don't really talk about the man anymore, but it used to be a very left-coded thing that you sort of viewed the military and the Democrats and Republicans and big business and tech as all part of this big conglomeration of, you know, special interests that did not have your interests at heart. Like that you can find old, like, I think there used to be this, so on IBM punch cards, they used to have this warning, do not fold, spindle, or mutilate, which of course, if you, if you are, you know, folding the punch cards in order to stuff them in something, it's harder for the machine to read them, it'll jam, et cetera. And people They'd have to learn that. But that, that term, do not fold, spindle, or mutilate, people started using that as just a general political slogan, which was basically, I don't want my, I don't want my entire identity to be represented as a number that's run through a computer, especially if that computer is deciding who gets drafted and who doesn't. So now, now that whole vibe of there is an establishment, it is out to get you, that still exists. It is now right-coded instead of left-coded. But it's, I think it was directionally true in both cases that you can have an establishment, it can get kind of stagnant, it can end up pursuing its own interests at the expense of everybody else's, and that it's useful to shake things up. So that is, that is the vibe that is shifting. I am really, I think it's very hard to impute that kind of thinking to just everyone who is participating in this. And, um, There are just lots of different flavors of populism. Like that's, that's part of the nature of populism and part of, part of this vibe shift is that people are looking more at their country's particular interests rather than at universal goals, however laudable those goals might be. And when you're doing that, you just, you have a different set of interests if you are in France or in Hungary versus in the United States. So you, you should, you should expect populism to just have very different flavors in different parts of the world.

And for this reason, should we be bearish about the EU's sort of cohesiveness or sort of effectiveness over the next 20 years and, you know, anticipate more Brexit-like decisions?

I think the big downside to that is that people have been, for very good reasons, skeptical of the EU's cohesiveness and sustainability pretty much from the beginning. Like, it did grow, but it felt like every time it grew and every time it got more powerful, there was just a lot of local opposition and a lot of people felt like this can't go on forever. And it did just kind of keep And also, if you have a country that is aging, or if you have a region that's aging and its economy is growing more slowly, it's going to be more conservative in the more literal sense of trying not to shake things up because there are just a lot of ways to get more downside and they don't see as many ways to get that much upside. You obviously have these kind of hopeful glimmers of European startups that are able to compete globally and are able to produce high returns and provide excellent jobs. For people with technical skills, but there's there's a lot of the continent where it just feels like a lot of things can be routed around, and then you know there's a there's a question of does it happen sooner or later? But there are there are fewer cases where it's just an absolute linchpin in a broad sense. And then there are a lot of narrow industries where yeah there are European you know mid-sized European manufacturers who are absolutely indispensable to some supply chain. So they'll be around for a while, but I think when you have a case like that where you, you sell this one little component with a lot of pricing power, you sell one particular chemical that is used in some really high value process, you again, probably don't want to shake things up too much. You probably want to keep roughly the status quo because anything that changes that might mean that your product is going to be manufactured somewhere, somewhere else by somebody else. And there goes, there goes a little bit of your trade surplus.

Yeah. And going back to tariffs for a second, some people are on the bluff part. Some people are saying, hey, this is just needlessly hurting our relations with our allies and sort of negatively affecting our international position. And others say, no, this is finally us standing up for ourselves and flexing our muscles and peace through strength and projecting confidence and strengthens our international position. How do you evaluate those claims?

So at least in the case of Mexico, it sounded like the tariffs were really not about Mexican companies competing with American companies to create jobs and that Trump wanted those jobs in the U.S. and not in Mexico. It was, it was more about border-related things, about climate change and immigration and drugs. For Canada, there, it's, it's harder to make a case that immigration from Canada is a huge problem or that drug smuggling from Canada is a huge problem. There isn't, it's not that there's zero of it. There's just a whole lot less. So I'm actually less sure what concessions Canada can actually make to the U.S. that feel like they're worth whatever the hypothetical revenue of a 25% tariff would be. But I'm sure both sides can think of something. Maybe, maybe there's some, some line on the border that could squiggle its way over and maybe that's where things end up. But I, I don't know. It seems like the, the Trump administration's territorial ambitions are much more literal with respect to Greenland and the Panama Canal than, than Canada, where the 51st state thing just seems like more teasing than anything.

We'll continue our interview in a moment after a word from our sponsors. Even if you think it's a bit overhyped, AI is suddenly everywhere, from self-driving cars to molecular medicine to business efficiency. If it's not in your industry yet, it's coming fast. But AI needs a lot of speed and computing power. So how do you compete without costs spiraling out of control? Time to upgrade to the next generation of the cloud: Oracle Cloud Infrastructure, or OCI. OCI is a blazing fast and secure platform for your infrastructure, database, application development, plus all your AI and machine learning workloads. OCI costs 50% less for compute and 80% less for networking, so you're saving a pile of money. Thousands of businesses have already upgraded to OCI, Including Vodafone, Thomson Reuters, and Suno AI. Right now, Oracle is offering to cut your current cloud bill in half if you move to OCI for new US customers with minimum financial commitment. Offer ends March 31st. See if your company qualifies for this special offer at oracle.com/turpentine. That's oracle.com/turpentine. Most of you probably know how much the Turpentine team loves Notion, the sponsor of today's episode. We use it every day for editorial planning, transcript markups, project managing, knowledge sharing, and it's where I drop random 3 AM ideas that turn into episodes. Now it has the power of AI built right inside of Notion and woven across our entire workspace. It's been a total game changer for our productivity and focus. We just used it the other day to architect the content schedule for next month across our network. Notion combines your notes, docs, and projects into one space that's simple and beautifully designed. The fully integrated Notion AI helps you work faster, write better, and think bigger, doing tasks that normally take you hours in just seconds. Try Notion for free when you go to notion.com/upstreampod. That's all lowercase letters, notion.com/upstreampod, to try the powerful, easy-to-use Notion AI today. And when you use our link, you're supporting our show. That's notion.com/upstreampod.

Conversely, for China, is Belt and Road sort of— what is the status of that? Have they kind of retreated a little bit after early experiments haven't borne fruit? Or how do we evaluate sort of where they're at?

So my understanding of Belt and Road is that it was a tagline to describe a lot of things that China was already doing. And the way that I would look at it is that if you have a country that wants to exercise economic influence on other parts of the world, It is a lot easier to do that if you allow the free flow of currency and of people, et cetera. Like you're like the U.S. has a lot of influence on the rest of the world because it's really, it's relatively easy to move dollars around. And dollar is a good default currency to use for bilateral transactions between 2 countries who don't really want each other's currency. So it's also because the U.S. financial markets are so developed, borrowing in dollars or raising capital in dollars is generally more straightforward than other currencies. And there is actually just a valuation premium for equities that are listed in the US that you can't explain with any statistical tests that people try to do. There was a good piece from Verdotta Research on this recently. You control for everything you can control for, you know, control for size, control for growth, control for margins, and you still get a valuation premium purely for being US listed. So all of that makes the dollar just a really powerful tool for the US government because once everyone's using dollars, you can threaten to cut people off from the dollar system. It's a really big deal. Whereas You know, if I if I were cut off from the ruble system, there's just no point in my life at which I would notice or care. There's just nothing I want to do that requires rubles. But there are a lot of things that various powerful people in Russia want to do that are just much more convenient if you can handle and transfer dollars. So that's that's something that just naturally emerges if you have a large economy and you trade a lot with the world and you have multinationals. And China has a lot of those things, but it doesn't have a good way for Chinese currency. Currency to circulate outside of China and get back in. That, that process tends to be pretty controlled. So what, what China tends to do is take some of their trade surplus, use it to build up infrastructure, use that infrastructure to solidify economic relationships with other countries, and then kind of monetize it through importing more raw materials. And so they get some of that soft power and they also get some additional economic growth, at least if they make good investments. And they do that without losing that much control over their currency. But that's something that China had been doing for a long time. They had a huge infrastructure buildup when they needed infrastructure. And then they, when they responded to the financial crisis by just radically increasing state spending, pretty close to single-handedly pulling the global economy out of a recession just by using so many different kinds of raw materials and building so much stuff. I think towards the end of that process, they realized there's just, there just aren't a lot of railroads that are not yet built and ports that don't yet exist. You know highways that are like incremental highways needed. Like China's pretty much set in terms of that kind of physical infrastructure, but there are other parts of the world that still need some. And if China's economy is growing and China does have significant you know military and influence on global affairs, etc., they can they can cut some fairly lopsided deals. But right now, I think they they are they have to be just more focused on the domestic economy and in particular on the overhang from. A real estate bust, which had to happen at some point, but does mean that there are a lot of assets underwater, loans that really maybe could get paid, but probably should not get paid, probably should be restructured. And if they do get restructured, if they were marked to market, a lot of financial institutions would have very large holes in their balance sheet. So there, there are solutions to that. And countries have been able to recapitalize their banking system or just grow out of it. China did actually grow out of a similar problem in the late '90s where their banking sector was actually insolvent, or at least the big state banks were insolvent. But there wasn't going to be a run and the banks were structurally profitable if they weren't required to make politically motivated loans to state-owned enterprises that were not actually going to pay them back in a timely fashion. So there was room for China to just kind of roll with it. And there's less room now because growth is just less baked in and the creditworthiness of the marginal Chinese borrower is a lot worse if the country is just a whole lot more levered in general. So they, they do have fewer options there. And even fewer options if they don't want to concede that they made policy mistakes in the last round. And like, you can, you can do a thing where the central government says, well, everyone one level below us completely messed up their economic planning. And the thing is, if you do that in one province, it's probably pretty believable to say, you know, we set up these incentives to produce economic growth. We knew that they could be gamed. We just couldn't imagine that one of our own would do that. The problem is if it happened everywhere.

Yeah.

If everywhere people were following the incentives too literally and creating fake growth and doing a ton of off-balance sheet borrowing, at some point you have to say that actually the incentive structure was encouraging everybody to do this. If you don't have exceptions, if you don't have exceptional cases where people follow the rules, then obviously the implicit rule was break some of the explicit rules. So it is somewhat embarrassing for them. Maybe they're, you know, maybe in retrospect there was some time where they could look at Some high growth provinces and say, "Not only is this not only is this fake growth, but we're going to seriously punish the people responsible and let that be a lesson to everyone else." And they they occasionally did things like that. Like there were there was at least one province that just had fraudulently inflated its GDP by some huge huge amount years ago. So they they did have an accounting fraud scandal in the in the Chinese growth story. But I think that it's it's at the point where enough people have been following the actual incentive that there's just a lot of bad debt in the system, and it's it can't get worked out without. Conceding that there were some bad incentives that were put in place by the CCP.

Seguing a little bit to some of the DeepSeek sort of fallout or sort of response. We talked last week about how NVIDIA, you know, strong collapse or, you know, 18% or whatever drop after, after the news. But you also wrote about how a number of AI adjacent stocks moved in parallel at the same time. Why don't you share your reflections about this and what does it mean?

Yeah.

Yeah, it was.

So that was just a really funny, really fun time. Basically what, what often happens when there is bad news that affects some theme in the market is that you'll actually see the pretty high quality, high growth, large cap companies go down because they're all tied to that theme. And you'll also see a bunch of lower quality bets on the same thing go up because people will want to do this offsetting bet where You know, if we think that people are going to spend more time online and they're going to look at more online ads, but also that the small-scale ad tech companies just don't have the scale to monetize as well as the big platforms do, then you might be long Meta and Google and short a bunch of other mid-sized companies. And if you are running a totally factor-neutral book, you probably have other large-cap shorts, other small-cap longs, such that you've balanced your exposure to the size factor. But like, that's, That is generally roughly how it works. And so that's what I was expecting. I was expecting that at a lot of funds, the risk person would say, hey, you are paid to pick individual stocks. I noticed that most of your performance comes from things tied to AI. So it looks like we're actually paying you a lot of money to say AI is good and express that, you know, in 15 different ways in your portfolio. And we don't want to pay you very much money to do that. That's an easy one. But it turned out that actually people were getting paid to do that. And what I did see, which was really interesting, was I just went through the list of holdings for a large momentum-focused ETF. And I looked at the performance of everything that wasn't an AI bet and all of them were up. It was not up by a ton, but like Philip Morris was up and Progressive Insurance was up and there were pharma companies that were up. There was just a bunch of kind of low-ish risk stuff that had done well over the last year was way up. And so I realized that a lot of people were long AI. And since that means, since AI is working, that means they were long momentum. So they were shorting other momentum stuff to offset it. And then when they exited all those positions, the non-AI momentum stuff went up. So I thought that was, that was one interesting piece. And I'm still a little bit confused as to who, like, who was paying close enough attention, who had a sufficiently sophisticated view that they, they were able to update on the DeepSeek thing, but they also own some of the more fringe kind of garbage AI bets. And I won't, I won't name a bunch of them because people have talked about them in detail elsewhere. And also I'm short some of them and I, you know, I never, it's always, it was a dubious honor to be publicly identified as someone short a specific stock with a large retail following. You tend to get death threats. So I'm going to avoid that. But a lot of those stocks went down too. So that either means like a lot of retail investors are actually pretty keyed into day-to-day developments in AI or that a lot of, a lot of people, a lot of institutions were speculating on some, some fairly dodgy stuff that happened to be AI, AI-ish, AI-flavored. So yeah, it, I think that what's possible is that, and I'm sure some funds do that. I'm sure some funds come up with these qualitative ad hoc risk factors and tell their portfolio managers, don't bet on this risk factor and we won't pay you just to make bets on this factor. But I think that there's always room for refining how you think about what is idiosyncratic performance for one stock, and then what is just betting on a theme in a bunch of different ways. One of the other interesting highlights in terms of stock price performance on that day was that there were actually companies that did worse than Nvidia that were utilities or companies that build data centers, things like that. And that, like, one thing that occurred to me was I don't know that there's been a case where a US-listed utility stock dropped 20% in a day because of lower expected demand. I don't know that that's actually happened since the 1920s and 1930s. Like utilities, the demand picture has been very, very clear for a very long time. But when that demand picture changes, when you have a historically very low volatility company that has very, very predictable earnings power and you know roughly what the growth rate will be, any uptick in that growth rate has just an enormous magnitude of impact on the actual valuation. So I think that was a case where if you look in risk-adjusted returns, Sorry, in risk-adjusted terms, that maybe buying power companies was actually a smarter bet than buying Nvidia. But that also meant that the people who, like the marginal buyer, the price setter of that company was no longer someone who was a traditional utility analyst who is thinking about the things that utilities people think about. It was someone who was really gung-ho about AI, looking for an interesting bet that was going to capture some AI upside. And when those people went away and the utilities got valued like regular utilities again, that was a pretty big, very quick haircut to their valuations.

And for NVIDIA specifically, how much of it is really just about sort of the prediction as to whether, you know, the smaller models will sort of play a much bigger role than bigger models? Is that sort of the core sort of outcome that will determine NVIDIA's price, or is it more sophisticated than that?

So I think that's one piece, but there's also the software moat piece and that DeepSeek was able to, to not just use NVIDIA's own software for programming GPUs. They went a layer deeper and found some clever optimizations that NVIDIA's people apparently had not found or had not bothered to find for this particular use case for this set of chips. Like the set of chips that they used was pretty much, from what I understand, was basically there, there was an early round of sanctions and NVIDIA started shipping a chip that was just designed to be as powerful as it can be while still sanctions compliant. And That's probably just not, not a market that NVIDIA saw as a really great long-term market. It was more like, this is an opportunity to move some inventory and the rules will probably change again, especially if this product takes off. So exactly the kind of case where you wouldn't expect them to really focus on optimizing for that particular customer, that particular use case. So I think that that is like, it was still a surprise that they were able to, able to improve things as much as they were, at least surprise to me. But I think that is, that is part of what people were reacting to is that Nvidia, there can be like Nvidia has such a software moat that there could be a time where they are slightly behind on the hardware side, but nobody is going to take the hit from the sunk cost of all the CUDA-specific stuff they've done. So they just buy the next round of Nvidia chips and then hopefully Nvidia catches up again. It gives them a little bit of margin for error, makes it a slightly more durable, less hit-driven business. And if that goes away, then it is closer to a commodity business. Now it's a really good commodity to sell. That's the case because demand is booming and demand tends to has has over the last couple years persistently overshot anyone's willingness to supply enough chips. So still a high margin business, but maybe maybe one that deserves a lower terminal multiple. However you figure out Nvidia's terminal terminal metrics. So I think I think that was a big part of it. And I also think that I I still think that what people are underrating like there's so Jevons paradox instantly became a meme. It instantly became just like. A you know, like the the midwit explanation for why you should not update your views at all in response to this news. But I think if you flesh it out a little bit, there is actually a case for much more consumption as measured in tokens, even if total user minutes don't scale that much, and even if even if revenue takes a while to scale, like even if revenue doesn't scale proportionally, because a lot of the more interesting AI tools that are being launched by the major labs, they are they're agents, they or agent-like, they engage in. Multiple multi-step processes. And if you're trying to explore a, a problem space and you start out with, here's the premise, and then you come up with 10 different promising directions you could go and you follow each tree for a little bit until you find that, okay, this one is probably a dead end. We'll try something else, et cetera. Like you can just generate a lot of tokens to get to your answer. You can pretty much consume just arbitrary tokens. Like if that is how you scale the performance of that kind of model. Then you can just keep scaling. And if you, if the cost of a token drops by 90%, then the same O3 model that was answering graduate level math problems for $3,000 a problem, now it's $300 a problem. And if that happens again, it's $30 a problem. And at some point, just being able to have an on-demand grad student in any domain where the answers can be checked against something true, but the answers aren't trivial. Which is basically everything that is either math or uses lots of math. If you have that, there's just a much larger set of possibilities that you can look at. Like I would naturally turn towards investment research. And I would say that if you, if you're looking at some company that is more on the deep tech side of things and they've been developing a product for a while, or let's say they have a drug and it's going through trials, being able to, for $300, you know, that'd be 90% off what, what O3's cost was when Meta talked about its performance in late December. Being able to do that for $300, like for $300, I would be happy to just go through a list of biotech companies and, you know, dump in a bunch of data on what drugs they're testing and where they are in the, in the testing process and just ask, ask a really smart model, like, is it just, is it biologically feasible that this could actually work? And my guess is there will be cases where it's just not feasible. Like it is just clearly not going to work. It is not a mechanism that actually does what the drug is intended And that's, that's a pretty straightforward short. And then if you have an industry that, like most industries, has roughly similar risk-adjusted returns to the rest of the market, and then you can segment out some set of the companies that go to zero, then you can also just buy everything else. And now you have a nice factor-neutral, or at least industry factor-neutral portfolio where you're short all the biotech companies that will eventually blow up and long the ones with lower but non-zero blowup risk. And that's a good place to be. You could imagine doing this kind of due diligence for a lot of other technical questions. So, you know, maybe boom arrow, like just asking the model to start reasoning about things like how, how, like, how low does the cost of supersonic planes get under different scenarios and how many do they need to make before they start earning a profit on them, et cetera. These are, these are questions where you could spend a lot of time. You could hire someone, you could hire an aerospace engineer as a consultant and have them work on this for a couple of weeks and you probably get a better answer. My guess is that for a lot of these questions, there is actually just enough domain knowledge that you get a better answer. But if you can get an answer that is like 85% accurate and doesn't take very long and the cost is comparatively trivial, then there's just a lot more stuff you can do. And you can sort of, you can outsource certain parts that require a very deterministic kind of rigor and then focus on the things that are just more, more vibes based.

And you wrote also about how Meta is sort of making a bet on this too, right? In terms of—

Yeah.

Yeah.

And with Meta, like I was confused on Monday morning when Meta opened down because Meta is a consumer. They are a buyer of tokens in the sense that they are, they're monetizing generative AI by showing people more content, by prompting them to comment more, by pre-writing the comments for them, basically. And by being able to generate a very long tail of ads. So. So if like, you know, if it gets cheap enough, every ad just gets generated on the fly based on what is going to convince you personally right at this moment to make a purchase. And that, that opens up a lot of interesting possibilities. And it means that they, that there are even bigger scale advantages to being someone like Meta. Like if they have more data on your behaviors and your preferences, then they have a better return on making those ads. So they make more of them, which means they can afford to scale up their hardware more so than other people. I made a few people mad on Twitter recently by pointing out that as a percentage of sales based on Meta's current guidance, they will, they will perform, they will spend twice as much on capital expenditures as US Steel. So relative to sales. And that is like, I think that's a good thing. I think that they actually can get an incremental return on a lot of the hardware that they're buying and that they may be able to calculate that return more precisely than a lot of other participants in like a lot of other GPU buyers. But it was, it is kind of funny to me that this classically asset light business that literally starts as a group of teenagers in their dorm, they already own the computers. So like the only, the only expense is you need a server and you need a lot of Mountain Dew. Everything else is just, you know, it's stuff you were already spending money on. So, so now, now they're a very capital intensive company, but also a company that can probably project its returns on capital. Quite nicely. And if that's the case, then they should be really, really excited that there is a cheaper model. They said that they want to make the AI models as cheap as possible, and they want these open models to be ubiquitous. But also, my suspicion is that one of the reasons this is bad for Meta is that if you have a group that is building models that are open, that anyone can use, et cetera, that is really cool. Like you, you're able to build something people use. You're also able to talk about what you're doing in a way that someone working at a lot of the more closed labs just can't do. So it, it is uniquely attractive to some of the people that Meta would really want to hire. And in some cases, if they have those people and then they have some more prosaic task that happens to involve a lot of AI knowledge. So something like when they, when they spun up Reels, targeting Reels was much harder than targeting other kinds of content because they are not just looking at first and second degree connections. They're actually looking at the entire space of Reels content and Reels are videos. So you have to actually figure out what they're about. You have to actually analyze the videos and do that at scale, et cetera. So that was part of why Meta made this large initial investment in GPUs. And then they were able to repurpose that. But if they had AI researchers and they, they, so they have them on hand, they can, I suspect, tell some of them that, you know, as much as we like the pure research you're doing, as cool as we think it is, like for the next 6 months, we have this really important business goal and we think you can help us accomplish it. And then they could do that. And so Meta wants to have some reason for really good AI researchers to be on Meta's payroll and be available to occasionally do stuff that has more of a revenue or cost impact. And I think that's a harder case to make if there is another lab that's outperforming them and is using fewer resources. Although I guess the other important thing to point out is that we don't, I don't, as far as I know, we don't really have comparable numbers from other labs. That compare directly to the specific cost that DeepSeek's talking about. Because what they talked about was the cost of doing the final training run for the model, not the cost of all the experiments, not the cost of the data gathering, et cetera. And those costs are, are a non-trivial component of the overall cost. I don't know what the relative breakdown is. And I'm sure that breakdown evolves over time. But yeah, that the specific headline, like $5.5 million number is, is not directly comparable.

Yeah.

While we're on the topic of Meta, let's get into something you wrote about the other day and also deeper a few weeks ago, which is forward deployed engineers. I don't think we've sufficiently discussed that on this podcast. Why don't we talk about first where you first wrote about it and why you find it an important concept to understand?

Yeah. So I think efficient frontiers in general are really interesting to look at. They are. So that's the case where it is really hard to tell. Like there are 2 possible ways to solve a problem and it's actually just really hard. It's a judgment call as to which you use. And I think there are a lot of interesting companies that exist at different efficient frontiers. There are a lot of cases where like the YC advice of do things that don't scale. If you wanted to persuade that in a way that would really impress economists, it would be something more like try to explore the efficient frontier, like map out that frontier. Before you start pushing it outward, so you do a bunch of these non-scalable things. You figure out which of those non-scalable things you're actually doing over and over again, and then you figure out from there how do you automate them. And what forward deployed engineers? So that's it's a term that it's existed for a while. I don't know if Palantir actually coined it. They certainly made it a lot more popular. And in the Palantir context, Palantir has just this extremely elaborate set of data science tools, and they pitch customers on pretty broad solutions. That point to some specific cost that's going to go down or some revenue item that's going to go up or some risk that's going to go away. And then they have to actually make that happen. And what they could do, like the simpler the product is, the more you can just tell people, okay, sign up, you know, give us your email address, give us your payment information and click, click buy. And then done. We'll just dump you right into the product. And if it's Notion, you know, you can read a little tutorial and start making pages. If it's, you know, a forms product, then they just tell you, okay, make your first form, et cetera. But if it is some, suite of analytical tools, you have to figure out what connects to what. And often the person who is driving the purchase from a business decision perspective is not the person who knows which table in which database has the relevant numbers. And maybe, maybe that table does not actually exist in the form that the decision maker assumed that it would. So you want someone who can do a lot of the data pipelining work and who can also navigate the organization and figure out who actually cares about this. What specific benchmarks do they care about? And then you also do want to teach the rest of the team how to maintain what you've built and how to expand what you've built, et cetera. And then while you're doing that, what you're also doing is noticing that you did this for your last 3 engagements. And so maybe instead of handwriting a script from scratch to do whatever this thing you're going to do is, maybe there should be some part of the existing product that actually automates this for you. So that, that forward deployed engineer model, it works really well if you have a company that sells a complicated product, open-ended with high variance in how well customers could use it. And where you think that you are early enough in the existence of the product category that what it ends up looking like in the long term is very much unknown. But if you, if you do that, you're, you're constantly iterating towards what the final form of the product is. And I think FDEs. Maybe the downside is that just by necessity, they have to be less strategic. Like if you are, if you're working for a big oil company and you are monitoring methane leakages, like, and you have 6 months to figure out which pipes are leaking and how to fix them. If you're doing that, it's a really bad idea to take 3 months off and, you know, think about how to rearchitect the product or think about the value prop, et cetera. Like you want to actually get it done, but as you're getting it done, you can realize all the steps that you are repeating from previous engagements and start feeding those back to the product side where they can actually turn those into scalable parts of the overall product. So like, I think looking for forward deployed engineers is actually a really interesting way to look at where the frontiers are between what is just a trivially solved software problem, what's a pure human problem. And then what is that, that messy intermediate space where you could probably come up with a deterministic rule set that worked 99% of the time, but the 1% of the time where it fails, it fails really catastrophically. So you don't actually want that. But then figuring out how much of it can you automate and how do you have someone monitor a system that is mostly doing the right thing? These are, these are just not easy problems to solve.

Yeah. And do you think more companies are going to restart doing this?

Yeah, I think like the best bet on that is yes, because it is getting more straightforward to just build very general, very powerful products. And then that means it's actually a lot harder to figure out what those products should be used for. I'm sure there are people, like, I know that there are order of magnitude differences in how much people get out of an OpenAI subscription. And I'm sure that if it's enterprise software, there's an even wider range of what you can get out of it. Like there are plenty of stories of enterprise software products where someone at a company really, really pushed to buy it. And then it literally never got used after enormous expense and a lot of effort on both sides to, to get the contract signed. Sometimes that's politics. Sometimes that's just people overestimated how, overestimated their own willingness to actually adopt the thing that they they bought. Lots of reasons that that can happen. But it does happen often enough that having, having a job function where the goal is make the product do in reality what it hypothetically sounds like it should do is a really valuable function.

Yeah, that makes sense.

We'll continue our interview in a moment after a word from our sponsors. Every time I hop on the 101 freeway in SF, I look up at the various tech billboards and think, damn, them, Turpentine should definitely be up there, and I bet I'm not alone. In a world of endless scrolling and ad blockers, sometimes the most powerful way to reach people is in the real world. That's where AdQuick comes in— the easiest way to book out-of-home ads like billboards, vehicle wraps, and airport displays, the same way you would order an Uber. Started by Instacart alums, AdQuick was born from their own personal struggles to book and measure ad success. AdQuick's audience targeting campaign management, and analytics bring the precision and efficiency of digital to the real world. Ready to get your brand the attention it deserves? Visit adquick.com today to start reaching your customers in the real world. Again, that's adquick.com. Hey everyone, Erik here. In this environment, founders need to become profitable faster and do more with smaller teams, especially when it comes to engineering. That's why Sean Lanahan started Squad, a specialized global talent firm for top engineers that will seamlessly integrate with your org. Squad offers rigorously vetted top 1% talent that will actually work hard for you every day. Their engineers work in your time zone, follow your processes, and use your tools. Squad has frontend engineers excelling in TypeScript, React, and Next.js ready to onboard to your team today. Backend, Squad engineers are experts at Node.js, Python, Java, and a range of other languages and frameworks. While it may cost more than the freelancer on Upwork billing you for 40 hours but working only 2, Squad offers premium quality at a fraction of the typical cost without the headache of assessing for skills and culture fit. Squad takes care of sourcing, legal compliance, and local HR for global talent. Increase your velocity without amping up burn. Visit choosquad.com.

I want to get back to some of the sort of implications of the DeepSeek sort of topic, which are, do we update our thinking at all on how we think about open source and export controls? Because I see different camps who have the same goals of strength against China. Some say that A, C, and Z, You know, most sort of prominently says, hey, open source is critical, you know, in that fight against China. And others say, oh no, you know, you know, the Vinod Khosla and Josh Kushner, the world sort of OpenAI, you can't open source because, you know, sort of our sort of the China threat that'll just lead them to be stronger. What do both sort of parties believe that's different that's led them to come to different conclusions? I've heard, you know, someone summarize it as It depends as to whether you think AI is either nukes or code. If you think it's nukes, you don't want to open source. If you think it's code, you do want it because you would be stronger because of it.

So I think first, just on the terminology side, it frustrates me to no end that we chose the term open source to describe models that are open weight, because an open weight model is just a model you can run on your own hardware. And so it is as open source. I mean, it's slightly more open source than just a .exe file that you downloaded somewhere where, yeah, you can download it, you can run it on your own machine, but you need some very specific skills to be able to peek inside the code and understand what it does. It's my understanding is that you can make some inferences about roughly how the model was structured, but it's not like you can take an open weight model and just say, I am going to follow the same training approach and train it on different data. And I will have my own DeepSeek V3 that does the following things differently. Like, They didn't open source any of that stuff. They didn't open source what it was trained on, and we don't have the data pipelines code. So like, you can't just replicate it. You can run it on your machine, but you can't build it on your own machine. Although there is a project to do that. So what that, what the, what the open weight model does mean is that if there is some improvement in how you, in the efficiency of how you run the model, that, that does proliferate pretty widely. But if there's, but those, those, don't actually affect the, the ultimate skill of the model, except with, with scaling, scaling number of tokens that are used to generate a final answer, like doing just more chain of thought stuff. I guess, I guess in that sense, it does actually improve capabilities as well. Like I, what I suspect is that over time, there will be companies that are selling access to models. Those models are closed, proprietary, et cetera. There are companies that are, that are going to monetize the usage of models, and those companies are always going to subsidized open weight models. And that we probably shouldn't expect the open models, like we should expect that there should be a persistent gap in performance. If there isn't a persistent gap in performance, then you can just clone, like then at that point, the valuable thing about the, about the proprietary model companies is that they have a brand name and that they're associated with, with that brand name and people trust them. Like if you can build a product suite that has the same performance as OpenAI's and you pay less per token than they do, you know, one, congratulations, but two, you can just undercut them on price. Or offer, you know, more queries on the models that are query limited and just win that way. So it, like, there is an incentive from someone like Meta to commoditize these models as much as possible. But I think there's also an incentive to capture a lot of the value by having a proprietary model that is just not beatable by the public alternatives. So as long as there is just room for progress in model quality over time, as long as there's still headroom, you should expect there to be closed models that that long-term have a performance advantage over the open ones. And then if that ever peters out, then eventually everything the closed models have discovered, the open models catch up to them and there isn't a next discovery to make. And at that point, you know, we have really, really good models and that's how good they're going to be.

Yeah, I wanted to segue to another thing you wrote about, which was going back to our stock picking, you were sort of setting up almost this fantasy stock picking competition. Why don't you talk about sort of the inspiration behind it? Yeah.

So this is something that I, I've always been really interested in the question of talent and talent matching. How do people find the right careers and how do they get assessed? And through the newsletter, I've been doing recruiting for a long time. And a lot of that's been technical recruiting. And I think in some ways, technical recruiting is a lot more straightforward. Like you, you end up with a much more meaningful sample of someone's skills and abilities. And I I'm, you know, I've never been a professional software engineer. That has never been my job title, but I have written a fair amount of code. And so I could ask, you know, reasonable follow-up questions and assess someone. And then on the, on the finance side, like it is, it is harder to figure out how, how do you actually assess someone's skill and how do you, how do you understand whether they would be good at financial analysis? And really a work product is a great way to do that. But then there's another matching problem, which is there is a set of funds that are very high profile, spend a whole lot on recruiting. And pretty much anyone who is, anyone who has the skillset and temperament that would make them really, really good at coming up with intra-quarter reads on different companies and figuring out which companies could be long and short ahead of conferences based on how the, how the CEO typically talks at these conferences, or who's really, really good at inferring that when this company gave a range, they're rounding up and this other company's rounding down, et cetera. Like all of those skillsets, you know, they do produce repeatable alpha. It's very measurable. And it is very, very lucrative. But then there is a long tail of firms that are just looking at weirder, less liquid They're more qualitative. They do have a different process. And I think they're just, they're potentially just really hard to find. And for a lot of those firms, their turnover is also lower. So they just don't have this recruiting engine. It doesn't make sense for them to bring in a large class of analysts and expect for, you know, half of them to get fired and half of the remainder to quit within the next 3 years. And then whoever's left, both sides are pretty confident it's a good deal. Like they're not doing that model. So I think of it as a way to, to broaden the surface area and also just to, to talk to people who have some, some interesting investing ideas. So yeah, we're, we're reaching out to investing clubs right now. But I also think that it's a really bad idea to be locked into the university track. One, for like social reasons, I think that, that, that credential is getting weaker and, you know, it is probably a good idea to weaken it. There are other ways to assess people's skill at this point. But also, I, I do think that a lot of, a lot of hedge fund and asset management hiring does over-index on On pedigree. And often that is, you know, that is a very, it is a very good decision if there are a large number of people who can theoretically do the job. And if you really, really want to avoid mistakes and you also have a lot of money, like the value that these people can generate for their firm is also really high. Like it does make some sense to just pay 50% more than everybody else and only hire people from a very short list of schools. And within that, only people who've done very well in challenging majors, et cetera. But there are just a lot of people who, for whatever reason, their main priority in their teens was not getting into Yale, but who are capable of doing pretty extraordinary things. And so, you know, we were reaching out to these investing clouds, but we also want to find people who are just not part of that track at all and have just independently found, independently developed some kind of knack for investing ideas. We're also doing this pretty, pretty broadly, but we did set up kind of buckets of different investing categories. So people love Pitching small cap stocks. I do it too. I love reading the small cap stock pitches. And there is a universe of funds that basically spend all of their time looking at tiny companies and trying to figure out what, trying to figure out which of them are the very small fraction of these companies that are actually really good businesses and aren't being recognized by the market. And then there's more of the midsize category, which is more like, hey, you can take a directional view on a company. We'll, we'll like professionals will know about it. They will have opinions about it. They can push back, et cetera. And then we also wanted to look at large-cap pair trades, which is a lot closer to just what a modern pod shop fund would do. Now, what we, what they really do, as I've alluded to many times in this conversation before, is they try to identify a lot of the different factors that can drive price performance. And they want their portfolio managers to be pretty neutral with respect to those factors. So not big on momentum or size or industry or whatever. That is just, it doesn't make sense to ask someone for one idea that does that. You're actually what that is about is designing an entire portfolio. But you can conceive of that portfolio, like sometimes you get lucky and there is just a pair of stocks where they have very similar traits and you like one and you hate the other. Cruise lines are actually an interesting example of that because there are a few more now, but for a while there were basically 3 publicly traded cruise companies of any significant size. And so if you are running within one of those mandates, you're always long 1 or 2 of them and short 2 or Or short the other two or the other one, like always. So that's a case where if you have some view that here's what Carnival's doing right, that Royal is doing wrong, and therefore make the trade. I think that that fits a lot of that mandate. And then we also wanted to look at event-driven stuff because there's there's just a lot of alpha there. That is a case where part of what it rewards is a very legalistic mindset and a willingness to read lots of documents and read the footnotes too. But it's also it's a very human-focused kind of field. Like you are trying to figure out. Out who's bluffing and who's lying and things like that. And you're trying to figure it out by basically reading lots of paperwork, lots of regulatory filings. So there, there are, I think that that is a case where you can, you can potentially identify some, some outlier talent in that. And it's also a case where I think talking about someone's approach, like how did this get on their radar, et cetera, can be really interesting because that's, that is the problem that I always run into with event-driven stuff is that it's hard to track all of it. It takes a lot of time to try to track all of it. And then if you're not, don't track all of it, probably the thing that is not on your radar is the most interesting situation you could have been involved in. And then the last one is a macro pitch. Like, I, I think that there's a good argument to be made that macro is like discretionary macro is like the closest thing to pure alpha in this very Zen sense where it's really, really hard to reverse engineer someone's thought process if they are actually doing these kinds of macro bets. Like it's always betting on something that happens rarely or that has literally never happened before. And that always happens under different circumstances. Like you never have a sample size. And if they are doing some more repeatable strategy, like, okay, let's look at the 30 most valuable stock markets globally and let's always be short whichever one has had the best performance in the last year. I have no idea if that works or not, incidentally, but like that kind of thing, if that's your strategy, very trivial to replicate it. But if your strategy is some kind of of crazy thought, like, you know, crazy way of reframing a problem that you figured out because you read this Economist article a year ago, and then you were reading the FT today, and then you saw some weird price action in the yen, and you put all of these thoughts together and realize that Japan's going to revalue their currency or do something weird with rates or whatever. If it's, if it's something like that where you're putting together all these different facts and details and there is just just, you, you come to some conclusion where, you know, that conclusion fits the facts and you also know that's not what people are expecting. And that like basically any, what you really want is where any given subset of the facts you're looking at points to some different conclusion. And then the, the set of like the comprehensive understanding of what you're looking at points to only one conclusion. That's, that's really valuable and really hard to replicate. And those are also just some of the most fun pitches to read Because it is people just trying to deeply understand the state of the world right now and then ask, how is it going to change? How are people misperceiving it and what's going to change their minds?

Yeah, I'm surprised there aren't more competitions like that already, or that, you know, this kind of fantasy investing isn't more of a thing.

I mean, it definitely exists and it exists in a couple of places like Value Investors Club is still good. SumZero is still good. Seeking Alpha has the occasional really good write-up and is mostly not as good. And this like the, the exclusive sites tend to have a higher quality bar to get in. And so they tend to have a higher quality bar of what gets written on them. And actually those, like SumZero is also an interesting one because one of its founders was actually one, was the, the, one of the business partners of the Winklevoss twins when they were employing Mark Zuckerberg. So the, the team that started that, whatever that, that company was, whatever that company was like, that team actually had has some of the most successful alumni on a per capita basis of any startup. Everyone left and did other things. And 2 of those teams created billion, you know, multi-billion dollar or yeah, multi-billion dollar companies. And then one of them created a company that is like a fixture, an institution, and certainly like a lot of value passes, changes hands based on write-ups that first appear on SumZero. So very It's a, it's an interesting corporate family tree.

Before we get out of here, I also want to make sure that we get into the other piece that you wrote about, which was about sort of capital as legibility.

Yeah. So the idea on that was that what we're, so it's really about how when you have a business that's an asset-light business, that means that you are not constrained by capital when you grow. You may have other constraints, but like, it's not the kind of business where expanding means raising money from someone else and really getting permission from investors to expand. And it also means that it's a business that can start throwing off returns to shareholders much earlier. It's just a great place to be. And the problem is if it's an asset-light business where all the value is just in the heads of the people who work for you, it's really hard for you, the business owner, to actually capture all that much of that value. And so in the piece, I talked about a couple of different cases where mostly talent-based companies try to do that. So you have things like Like law firms where they do set up this weird Ponzi scheme-esque dynamic, or maybe you could charitably compare it to Social Security, where if you are a junior associate, you're working really hard, you're capturing a very small fraction of the revenue that you produce. As you achieve more seniority, you are capturing a larger fraction of the revenue that you produce. And you also have a different set of responsibilities, like you're supposed to be bringing in new business or at least maintaining business relationships. So that, that does tend to kind of work. And then ad agencies also, there are some scale benefits to just having every, every advertising medium in every country, you know, relationships with every, every publisher, et cetera, like having all of that inside of one company such that there's only one invoice to, you know, only one invoice to pay instead of 50 from all these different ad campaigns like that, that does give them some ability to capture value, but not a ton. And with, with someone like, like Meta, there's some of the value, some of their value capture comes from the fact that they have There are network effects, there are switching costs. If you recreate a subset of the Meta product suite, what you have is by default worthless. If you create something that Meta should do, then maybe it's worth a lot. Maybe, you know, you were wrong that they should do it, et cetera. But like, they, they still have this dynamic where people leave the company in order to, like, if you have a really good idea that is not, doesn't require a large user base to get started, it's often more economically rational and expected value terms to leave, start a company, have that company build the thing, and maybe get acquired by your previous employer or some other big tech company. And as, as companies find that capital expenditures are actually the main way that they grow, that, that labor bargaining power actually gets a bit weaker. Like, definitely there are still a lot of people who have a lot of negotiating leverage at those companies, but when, when more of the marginal value that is being created is If we had more GPUs to throw at the problem, we could do more generative ads. We know what the lift on those ads is. And so we're going to do it. At that point, more of what you would need, like more of the value creation is actually based on these physical assets, based on things that people don't actually walk out of the building with at the end of the workday. So I did have the joke that, you know, the classic line at any talent-centric business is our most valuable assets walk out of the building every night. And that if someone is saying that at Meta today, it probably means they have discovered discovered that somebody is stealing the GPUs. So you do, you like, in some ways, I think just like optically, just in accounting terms, it makes the businesses look worse. They are, they have to like each incremental dollar of revenue they get requires more, more capital expenditure dollars than the previous one. And that seems likely to continue for a long time. But another way to look at it is that these companies were generating tons and tons of cash flow and still are generating tons of cash And now they have found a way to reinvest that cash flow in a way that strengthens the asset-light part of the business and that gets a decent return on the assets that they're buying. And if you have a business that can continuously reinvest at a return that is greater than its cost of capital, when you adjust both for the risk that the company's taking, in theory, if they can do that indefinitely, the company is actually infinitely valuable. In practice, they can't do that indefinitely. And that's why companies don't ever reach infinite valuations. But if you can do that for a long time, then you actually create a lot of wealth for your shareholders. Even if creating that wealth means your return on equity is relentlessly moving down, as long as that, as long as the return on a marginal dollar that you retain into the business instead of return to the shareholders, as long as that return is higher than what shareholders expect from comparably risky investments, then you're golden. So I think it's, yeah, it's net good for Meta. Like they're actually in a better competitive position. With respect to their internal pricing power and the company's ability to capture profits, they're in a better position if they are a more capital-intensive business than they were before.

Right. And you as a shareholder, you're, you know, some people look at sort of their, sort of the amount of money they spend on things like the metaverse, is it sort of that bad? Or even things like Lawman are like, hey, are they in over their skis here? But you're bullish. You say, hey, this is an ambitious company that's betting on the future. And, you know, they're willing to take bets and they're not always going to be right, but when they're right, they're going to be right big.

Yeah. Yeah. Like I'm, I'm definitely, you know, Meta is a lot more expensive than like in terms of earnings multiple than it was when I started buying. So like, I think it is, you know, less of a pound the table thing right now. On the other hand, I wasn't pounding the table when I first started buying because there were actually a lot more risks. Like this was when they seemed to be losing attention share to TikTok and the ad targeting stuff meant that their revenue took a hit. So You had a lot of things and then the metaverse, they were also spending a ton of money on that. And so it was in that case, it was a bet that basically that Mark Zuckerberg is not going to spend all of the company's cash flow on the metaverse because that actually makes it hard to hire the people he'd want to hire to build the metaverse. Like if, if Facebook's equity compensation is questionable, then their ability to acquire the best people is also questionable. So it is like, yeah, it's a different cautionary note. Like now things seem to be going really well for the business. But the price mostly reflects that. So, but yeah, I still respect them a lot. And I think that is like part of what you have to do with a company like this is to say that at any given time, there, at any given time that you could have invested in one of the big tech companies, there was probably something that they were doing that was going to become a material contributor to revenue. And that would be the story behind why the stock was going up. It's something that either you don't know about today or that you would actually misunderstand at the time that they did it. Like, I remember when, when Netflix first started doing originals and I actually went to an idea dinner. I was at a hedge fund at the time and I went to an idea dinner where a bunch of people from a bunch of different funds were talking about this. And the consensus was, this is interesting, but producing movies is just not a great business, very lumpy. And that a lot of people, like, even if it's a big hit, a lot of people are going to sign up for Netflix and maybe pay for a month. They're going to watch all of House of Cards. They're going to be done and then they're going to be off to the next service. And I think that was somewhat true, but it becomes less true as it became less true as they were producing more shows. So by the time you're done with one show, they have another Netflix original for you ready to go. But that would have really freaked people out. Like if that had been Netflix response is, yeah, I think you're right that it's very risky to make make 3 originals. So we're going to do 30 this year and then next year we're going to do 300. I think the stock would have tanked if they had said that. So, and I wasn't, I wasn't super optimistic about Netflix. I was, I was looking at it when it was very, very cheap compared to today and was thinking there's probably a right time to buy this and I don't think it's now. So, so yeah, this stuff is, is really hard, but I do try to remember that a lot of people running these companies have seen a lot of technology shifts before I have, that some of these shifts are very expensive to jump And that even if their hit rate is not 100%, and even if they tried a few times and it's wrong, which happened with Meta and mobile, that their, their first mobile app, their early mobile experience was just not great. And they had to redo a lot of stuff. And for a while, mobile was actually a significant headwind to revenue because people were moving from the desktop, which was monetized with ads to mobile, which was not. And then once they turned on ads and once people saw how much that was contributing contributing, then it was a very different story. And now the desktop piece is just irrelevant to their business. So those, you know, I think that's, that's an important lesson, but it's also very easy to overfit. And you do have to acknowledge that as the companies get bigger, the bets also get bigger. So you are just risking more every time you invest in one of these companies. But there is like, there is a particular kind of, I guess, technology shift sensitive capital allocation. Which is a skill that I don't think you can really develop without just running a company, running a tech company for a long time and seeing everything that, everything that contributed to your original core business becoming completely irrelevant or even a liability. So if they can, if they can survive all of that, if they can run a business that is completely different from what they started and they're able to make those jumps, then I tend to infer that they will keep making the next jump.

Well, and this last question is going to let you go is, do you think it's too early to call their big metaverse bet? a poor use of capital? Or do you think, you know, years from now, how will we view that bet?

So I think the time, it's really tough to say, like my default is, yeah, that was actually a bad bet. But I think that there comes a point where the company should be able to say it was a bad bet. And they, you know, they've dialed back the emphasis substantially. Some of that is just, they found something else to invest a lot of money in. That has more direct returns for shareholders. So yeah, on net, I think it probably ends up being mostly a waste of money. I don't know if we ever make a transition to living in the metaverse. If we do, if we do end up doing most of our socializing while we're strapped into VR goggles, then it's still possible that they'll be, that they will have been early. But I think for, for you to make that case, what you have to say, there has to be some discrete innovation, whether it is a new VR headset that just gets massive distribution or some new interaction model or some advance in graphics or dealing with motion sickness, whatever. If there is some discrete advance that suddenly makes VR really, really appealing and we have another platform transition, then I think maybe it turns out to be a good bet, just early. But also, if you don't know what that advance would be, then you're always going to be early until you're too late. And so maybe, maybe you can view it as like, Like Meta is continuously buying call options, out of the money call options on the metaverse, and they have to spend a lot on those options. But if they keep doing that, then at some point the metaverse actually works and they exercise their option. It's deeply in the money and everyone's happy. So that, that return profile does kind of look like, at least the first part of that return profile where they're just constantly throwing money away. That is pretty much what we see. It remains to be seen what the second part of that return profile looks like, but actually, if I do want to steelman the just knee jerk skepticism of the metaverse. If I wanted to make the case today that they should just call it a day, shut it down. I think the actual case, like the, the reason that it has gotten easier to argue against the metaverse is that AI does function as an interface. Like there are AI tools, but AI as an interface is also a pretty powerful idea. And that might just be the new default interaction model. And that is very much an AR rather than VR vision. So that is not you strap on your goggles and you go to a different world. It is you put on your Meta-branded Ray-Bans. And now you're actually seeing the existing world in higher, higher resolution, more dimensions, however you want to think about it. And tools like that, where it's taking in all the real-world data and it's highlighting things you care about, it's answering questions based on what you see. That is just a really good way to interact with a computer. Like, it is a very helpful way to interact with a computer. It saves you a bunch of steps. Saving steps is always Really important for consumer applications. So, so maybe that works, but it also makes the metaverse just relatively less attractive. Like if you can actually have a more interesting, lively, dynamic experience with more distractions at hand all the time, wearing AR glasses instead of VR glasses, then maybe the VR glasses just gather dust. Then, so I, I will, I'll feel very vindicated if Meta shuts down or significantly curtails the metaverse. In order to focus more on AR glasses and specifically says it's basically taking like eyeball share. You're wearing one or you're wearing the other. So eyeball share actually matters here. If the eyeball share moves to enriched views of the real world rather than some other world, then, then I think metaverse was, you know, interesting experiment, probably spent too much money on it, but it doesn't work anymore.

Yeah. Yeah, that no that that that makes sense. Yeah, as an interface, that's a good sort of counter counterclaim. We're we're at time. As as always, Burn, this was a fantastic wide-ranging conversation, and until till next week.

Yes, indeed. Talk to you next week.

Upstream with Eric Thornberg is a show from Turpentine, the podcast network behind Moment of Zen and Cognitive Revolution. If you like the episode, please leave a review. Hey everyone, Erik here. At Turpentine, we're building the first media outlet for tech people by tech people. We're the network behind the show you're listening to right now. We have a slate of hit shows across a range of topics and industries, from our AI and investing cluster of podcasts to shows that drive the conversation in tech with the most interesting thinkers, founders, and executives. Founders, investors, and influencers like Econ 102 with Noah Smith. We're launching new shows every week and we're looking for industry-leading sponsors. If you think that might be you and your company, email me at erik@turpentine.co. That's erik@turpentine.co and let's partner together.