Elon Musk’s Network Effects and Checking in on Chinese AI w/ Byrne Hobart

Hey, Upstream listeners. Today we're releasing a conversation I had on Turpentine's show, The Riff, with my friend Bern Hobart. We discussed the benefits Elon Musk gains from having so many companies, the chaotic nature of Doge's work, Chinese AI development updates, and more. Please enjoy the conversation.

Hey, how are you? Hey, Burn. I'm doing great. How about you?

I am good.

Shall we, shall we get into it?

Yeah, let's do it.

Awesome. Burn, there's a lot I want to cover this week. One is you had a piece on, on Don Jr. and how he's leveraging the, the Trump brand. I sort of want to have you unpack that a little bit, but also just ask the question, if Don Jr. was focused on making as much money as possible, by leveraging the Trump brand, and he was saying, hey, Byrne, what advice would you have for me? What might you say?

Yeah, I'm not sure I would tell him to do that much differently at this point. I guess, you know, there are ways to get a little bit more of the upside, and certainly there is stuff that could be done but may already be being done in terms of trading crypto in particular ahead of different announcements. But, you know, there is this long-end kind of Sordid political history of major politicians, less known relatives getting into various business dealings. It's like, it is one of the sort of kind of comic relief side stories in any presidency that there is some nephew, cousin, half-brother, whatever, who is constantly saying that they're very close to the president and that of course You know, of course, he can't ever be directly involved in this, but you know he's he's really on our side. And that I had heard this story, this theory about Hunter Biden too, which I thought actually had some explanatory power. That yeah, all the emails are real, and Hunter really did tell people it's ten percent for the big guy and so on. But he just made sure to not actually tell the tell the actual president any of the things he was doing trading on the name. And meanwhile, you can't really collect on that. You know, if you are a corrupt oligarch and You hire someone for this completely fake, very well-paid job, and he says, yeah, you know, it's the org chart is there's me, and then under me is the Secretary of State or whatever. It's not like you can sue him for misrepresenting himself if that turns out not to be true. So it is kind of a good, a good grift to run. And I think in the Don Jr. case, he, he is at least among Republicans, a celebrity in his own right. And Does does operate in that capacity. Like he gives speeches and things. He has his own media presence. So he is also part of that orbit. But yeah, he seems to seems to be doing a lot of the a lot of the kind of smaller scale business dealings. But it's also it is a feature of the market right now that retail investors are pretty big. They tend to get pretty excited about things like that, about you know brand name board members and. Especially for companies where either A, you can't really evaluate the business at all. It's too early for that. Or B, you can, and you can compare it to other companies and say, this is just not as good a business as the other ones, but it is my kind of not as good a business. So what we've sort of done is democratize the idea of subsidizing a, you know, nepotistically subsidizing a family member who really wants to run a coffee shop or used bookstore or something high status, but not very lucrative. It used to be that you had to be wealthy enough to actually do that, to actually fully subsidize that business. But now you can actually do the nepotistic business subsidy thing through your Robinhood account. And it's a smaller amount, but you still get the same fuzzy feeling. So yeah, maybe this is, this is all about bringing, taking parts of capitalism that used to be available only to the 1% and making them something anybody could do.

Yeah, that's a good explanation on the Don Trump phenomenon. I want to also segue to another version of kind of like, what would you do if you were this person? You also mentioned, wrote about how Dustin Moskovitz left Asana and had an interesting take as it relates to sort of how he's thinking about what to do with his time. There is this sort of question of, you know, if you have that level of resources, what's the highest leverage thing that you could be doing, especially if you're worried about Sort of runaway AI.

Yeah, and I think that is basically the story. So it's part of what I was writing about is just anytime you have a founder-led company and investors like the founder and the founder clearly has high conviction in the business, it is very troubling if that founder steps down. And, you know, it often for good reasons drives the stock down a little bit. You know, you wonder if all the stock is going to get liquidated, but you also in many cases just want this person who is one understands the business very deeply and who's respected by the employees and so on. Like, it's better for them to be in charge than someone else, even if that other person is also pretty good. But I do think in this case, if, if there is a strong explanation for why that person would want to leave, it's a weaker signal because sometimes people do just leave because they can see the business is actually not going to go all that well in the future. And so that's what keeps them unexcited. But in other cases, yeah, it's— There, there are other higher value uses of time and usually those uses of time are just Either giving away lots of money generally, or just spending more time with your possessions. And there has to, you have to be pretty motivated to run a business if you could also just enjoy being a billionaire indefinitely and never think about money again, except thinking about how nice it is to have so much. But in, in his case, he's been alarmed about AI for a long time and has been donating to a lot of EA effective altruist causes, including causes that are about AI and existential risk and things like that. So I think it's actually, it's pretty aligned in that case to say that as, as AI accelerates and as it gets better at doing human-level performance, that if you're worried about it achieving superhuman performance in a very negative way for society, that, that should probably be your full-time thing. So yeah, some, I, so I guess it is kind of analogous to if someone steps down because they really want to provide full-time medical care to a relative with a terminal illness. It's just in his case, the, the person with the terminal illness is all of us. And the illness is that there's this AI thing that's ramping really fast and could, could kill us all.

Yeah, it is. And so what do you expect him to do?

I mean, I don't know. It's, it's always hard to come up with expectations for someone who is clearly smart and effective and has a lot of financial resources and has been paying attention to this rapidly changing field for long enough time to have some sense of what's coming next. So I think it will be very high signal, whatever, whatever he does next. I think that the, you know, even, even the fairly straightforward things like, okay, we will, someone will start another lab that is even more safety focused than the other labs. I think that that probably just doesn't work because you probably, if you do view, if you view safety as actually complementary to building a more powerful AI system, then maybe you'd say you don't need as much money as OpenAI or as as Google because you can, you can actually get more out of the dollars you spend. But I think that's, that's not necessarily true. It is probably true in some sense over long periods that if you have this smarter reasoning model, it may reason its way into fairly pro-social behavior, but it's, it's hard to say for sure. Anyway, so I, I really don't know, but I think it'd be very interesting to see, you know, if someone does have this fortune and they now have time to devote to these kinds of causes. What is it that they see as the highest leverage thing to do? But yeah, no idea what that would be. And, you know, I could be wrong, but like the, the, the, the, the fact that that is a good explanation does not mean it's the only explanation. And so it is possible that either A, it's like quitting to do nonprofit stuff, but it's a totally different nonprofit thing. It's shrimp welfare or something. Or B, that it is just quitting because it is really relaxing to not be running a large company and the money, the money will still be there.

Yeah. But what is your sense on sort of, you know, Dustin's broader project or his sort of contributions to sort of, you know, or what he's doing with his, his work and his foundations and just the broader sort of field of existential risk? Do you, do you think they're doing a pretty good job? Do you think it's sort of like impossible effort on its face? Or what was sort of your analysis of our sort of existential risk, you know, apparatus?

I think it's good to think about existential risks, but I think that you really want to avoid prematurely defining risks of a particular nature and saying, this is what counts as an existential risk, because there, there is this just ongoing existential risk of we end up reaching, we end up creating a really complicated society. It has high fixed costs, not in terms of it takes a lot of dollars to keep things running, but in terms of it takes some very functional institutions and the ability to put competent people in charge of solving the right problems. And if you lose that, it's very hard to replace it. And so you can have these radical decreases in the complexity of your society. And in the case of modern human civilization, that just has to mean not just a large decrease in standard of living, but actual starvation. So that's the kind of thing that I worry about more is sort of cultural decay and risk aversion and just people. Not having kids and thus not thinking about the distant future. I think that is over long periods an existential risk that will always be with us and is worth addressing. And then there are other more near-term existential risks, whether it is the Genghis Khan or nuclear war or synthetic bioweapons or rogue superintelligence. Like these are, these are all things that could actually wipe out humanity or functionally wipe us out. But they, they're always, they're always in that mix. And it does, it does feel like the, the risk of, you, you could sort of think of like a hot apocalypse and cold apocalypse where hot apocalypse is nukes start flying. And by the time they stop, civilization is over. And then cold apocalypse is things are just a little bit less functional over time. And people are a little bit less willing, a little bit lower trust and a little bit less willing to invest in long-term projects and take big risks. And. Smart people do things; they opt out of making challenging, high-stakes decisions in favor of other stuff that they can use their intelligence to do, and the whole thing just gradually falls apart. But it would—I suspect it would fall apart in a pretty invisible way. Like if you read about people who are in late-stage empires, they don't—they don't usually feel like they're at the end of the empire. In fact, it's—it seems more common for people to have this really paranoid feeling that. This could collapse at any moment for them to feel like that on the way up. This is one like Bill Gates apparently in the like throughout the '80s where it was you know throughout the '70s where it definitely made sense and then continuing into the '80s but then into the '90s where it was really hard to justify Gates would always talk about how Microsoft is just a couple bad releases from being a completely irrelevant company that no longer exists and I'm sure a lot of people who had those conversations with him in the '90s were thinking to themselves Bill no one has ever made as much money as you have at your age and you're you're a monopoly. in this incredibly valuable product category and everyone, all of the smart people, you are very close to the top of the list of every smart person's ideal place to work. Like, how could you possibly think that there's some risk of all this going wrong? And I think his answer would be, yeah, you could have said that about a lot of other people who, who, if I name them, you won't actually recognize the names because it is true that a couple of mistakes are enough to kill a company. And then it doesn't feel like, it feels like in the, in the mid-2000s when Microsoft was kind of in this lull where Yes, they were actually growing, like revenue was actually pretty good. Margins were going up, but people were just less scared of them. Like this was something Paul Graham wrote about was that he no longer really had to warn startups to think about what Microsoft could do to kill them. They had to warn them about Google, but Microsoft was just not a threat to startups. Like in that period, it does feel like people weren't especially worried that Microsoft would die. They were, they sort of felt like it was. impregnable, but was just not really going anywhere. And that is, again, a more meaningful risk. Part of what got Microsoft out of that was a series of pretty bold bets, some of which were done under the Ballmer era. And then Satya was able to spearhead one of those, which did well enough that he's now in charge of the company. And he does seem to have the right kind of paranoia. And maybe it is helpful that it's, it's interesting actually that Microsoft Has a lot of categories where they are just number one. They're synonymous with the category. They've had other categories in the past where they were synonymous with that. And then they're run by someone who previously worked on their cloud business, which was not number one. It was the company that was trying, the business that was trying to catch up to AWS. It actually did a really impressive job, but that is, you know, at a company where you, where they are number one at a bunch of different fields, you actually do have to kind of go out of your way to pick someone who is not from. a business that was the best in its category, but maybe that's what you actually need is for them to always think of themselves as underdogs who are not necessarily everyone's first choice and who need to be pretty scrappy to survive.

Yeah, that, that, that, those words are clear to me.

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. Every time I hop on the 101 freeway in SF, I look up at the various tech billboards and think, damn, 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.

Speaking of sort of high-leverage people who are thinking about existential risk, I want to segue to Elon because you both covered that sort of an update with Twitter and how we're impressed with how Twitter has operated or sort of the ability of Elon's financial engineering prowess. But then also on the other side, Doge has had a lot, you know, been less in favor over the past few weeks. And I'm curious for your— Your assessment of what's going on there.

So the Twitter thing is partly financial engineering that Twitter has this relationship with xAI. And it is just one of the perks of being Elon or just being someone who is in control of a bunch of different entities in which his ownership stake, his economic ownership varies, is that he does often have the option to move cash flow, talent, infrastructure, et cetera, to whichever company gets the highest return from. And that kind of works really, really well unless you just run out of cash flow to reallocate or all of your Tesla engineers are so busy working on something for SpaceX that they don't also have time to fix some problem at Twitter or whatever. But he, he was able to technically get Twitter's equity value or Twitter's enterprise value back up to where it was when he bought it. But it's just, it's a very different entity right now. And it does feel like Some of the capitalized value of Twitter circa 2022, when he did raise external capital, including on the equity side, some of that was people who wanted to do him a favor, who figured it was much better to buy something that Elon is buying that makes him feel really good about himself at 150 cents on the dollar than to keep that money and not invest it somewhere else. So basically a fee to get on Elon's good side. And then now it feels like there's a, there's a kind of different math behind being on his good side, which is that Twitter remains a really important place for ideas to rapidly disseminate. It is where the discussion happens when there's a breaking news story. That doesn't seem likely to change, even though they have had some attrition and lost some of their contributors. And they've, they have the same kind of boiling off effect where if everyone, if people who leave are more likely to be progressive, then the site gets more right-wing as they leave. And so it ends up being too right-wing for the next set of more progressive people. But then some of the people who stick around either just really are open to a broad spectrum of views, and they know that they have plenty of progressive friends and can get the progressive side of something or can just generate it on the fly. And, and so they want to hear what the right-wingers are saying. And then the other piece is just people who like scrapping and who absolutely want to go to a site that is now more conservative because they get to have arguments with people and dunk on them all day. And that's, that's quite fun some of the time. So So I think Twitter, it is, it is not quite the town square, but you also can't really run the town square if your town spans multiple countries with different legal jurisdictions and your owner, the owner of this town square has business interests, including interests in different countries that have very different views from the U.S. So yeah, you, you just can't have the town square if the town square is owned by someone who does a lot of business in China. There are just going to be limits to the things that he's able to say. And then if he's also Part of the Trump administration, there are further limits to what kinds of discourse he wants to happen. But yeah, I think, so I think it is partly just a story about how if you, if you have a bunch of different entities and you just have a bunch of different trade-offs you can make, you have a lot of opportunities other people don't have. And that is a lot of what negotiation is in general, is figuring out, okay, we're, we're both negotiating on this one dimension and it's clear that we both care about the outcome. And we can't quite get to the middle. Is there some other trade-off that I care about a lot and my counterparty doesn't care about that much or vice versa, where we can actually get closer to an agreement? And so Elon just has more degrees of freedom than just about anyone, and then more hard constraints on how much he can use at each of those things than just about anyone.

Are you, what are your sort of, what is your outlook on Doge going forward? And if it's perceived to be, you know, its ability.

Yeah. So I think if, if I were part of it, what I would hopefully be thinking is the shock and awe phase is over. We have convinced everyone who gets a government contract that these contracts will be scrutinized. In fact, they'll be scrutinized in a pretty hostile way. And sometimes you'll be doing something good that sounds not so good. And so you still lose your funding. And so That, what that forces is some level of transparency and really a focus on KPIs. So if you, if you have some grant that is actually serving U.S. interests and it's doing so in a kind of opaque way, just making that more legible to someone who's new to the system is probably a good thing. And I think you can, you can look at a lot of the mistakes Doge has made and say, this is why we need people who are, who have much more expertise in how, how U.S. foreign aid works and how the U.S. government allocates its operating expense budget and things like that. But I think the, the other side of that is if you're deferring to people who do have that expertise, you're often deferring to people who also have a vested interest in how things currently work. And you don't really, really know what the balance of that is unless you actually start cutting and see what goes wrong and what doesn't. But once that shock and awe phase is over, I would love to see them just go after, I think there, there are areas like Medicare where there's just a, there, there are a lot of opportunities for people in general service providers to take advantage of the system in a way that was not intended that probably shows up pretty straightforwardly if you are doing statistical analysis on things like I don't know, like you could look at some category of surgery and look at what demographic factors predict, you know, things like age predict how common that surgery is. And then just look at, okay, per capita, is this happening a whole lot in very specific places? Maybe there is someone who is really, really interested in getting subsidized to do this surgery and is willing to do it for people who could probably, could probably do without it. I think something like that. It's, Maybe, maybe less exciting and certainly less politically polarizing, but is actually a useful thing to do. The problem with that though is if they had tried to do that early on, there would just be a long period where Doge is announced and then nothing happens. And then when they do have an impact, it's these tiny incremental changes that no one really cares about. And so Doge just becomes this minor sideshow. And maybe in the next State of the Union, Trump gets to say, here's how many billions of dollars this organization has saved and everyone Well, just would just be asking that happened at all. Everyone would be asking, what's Doge? Whereas if in the next State of the Union, he quotes a number, it's an accurate number. It's a few billion less than if Doge had been really rigorous from the very beginning. At least in this case, everyone knows what Doge is. Everyone involved with Doge gets to be proud. Like people know that they have actually achieved something. And given the size of the team and given the scope of US spending, It's probably a pretty high ROI decision, even if it's not, not quite the magnitude that Musk was hoping for.

That makes sense. The, yeah, it is interesting. What is your sort of assessment of sort of the first, you know, couple months of the, of the Trump White House more broadly? Do you think this is what we expected? You know, sort of the, the, the chaos, sort of the ups and downs. Do we think things are going to normalize? Is it just this, this status quo going forward? How should we think about it?

So I would say the, it is very nice of you to ask if it's as much chaos as expected, because yes, yes, it is. It is as, as chaotic as I expected, because I remember the Trump won days very clearly. There was a lot of chaos, but it was chaos that didn't actually lead to that many outcomes, or there was, there was a very high ratio chaos, discourse, news flow, et cetera, to actual changes. And it feels like they've, they've had that same level of discourse and uncertainty and things, but have actually implemented a bunch of changes. And, you know, I think some of those, like some of the cost cuts that Doge did were just bad, or at least were very ill-considered, if only from a political standpoint. And then others seem actually pretty reasonable. And it also, like the U.S. government, from what I understand, in many cases, there are pretty generous work-from-home policies, but they're real estate footprint has not dropped that much. So that is the kind of realignment where I think a lot of, a lot of companies in the private sector would have very quickly adjusted their, their spending to make sure they're spending for offices that actually have people in them. And there's less of an incentive to do that within the federal government, but it's, it's probably helpful to, to do that kind of downsizing. On the tariff stuff, it is, you know, it's almost, you almost don't want to say this is a really effective bluffing strategy. In fact, it should, if you don't like tariffs, you should make it somewhat taboo to say this is an effective bluffing strategy. What you should do instead is pretend that you take these completely seriously. This is a complete reshaping of the economic order, and no country will be able to sell anything to America Americans without paying a toll to the U.S. government. Like if you say that, then tariffs are actually a more credible threat. If everyone understands that a 25% tariff is the start of a negotiation, the actual tariff will be much lower in the end. Then the the ratio of starting bluff to final outcome gets a lot more extreme in favor of less important final outcomes. One thing that has been interesting is to see just how, how other countries are reacting in terms of doing their own defense spending. And then there's, there's the story that came out, I think over the weekend about how the UK had a digital, has a digital service tax. They are considering preemptively removing that tax because the tax is basically a tariff. It is not on goods, but definitely on services. And the way it's structured is that it's a tax on social networks and search and a couple of services like that. But these are all categories that are just overwhelmingly dominated by American companies. So it does in effect mean that there was a tariff on the U.S. and now it's going away. And as I've pointed out a couple of times, the tech companies are in a really interesting Part of the crossfire, which I think they probably did not anticipate. I suspect that some people would have been a little bit less Trumpy had they thought that the tariff talk was serious than it actually is. Sorry, my dog just jumped off the bed. But what but these companies, they're they're kind of swing voters or swing swing influences where they used to be very pro Democrat, and I think if you ask them. They just wouldn't have been able to credibly say that they were centrist. They would have said that they, yeah, that many of their employees are, you know, they want to give each side a fair hearing. But yeah, it's much, much more likely that a Democratic candidate than a Republican candidate would be able to do a fundraiser and get a lot of these people to, to attend and donate. And now big tech is still, still probably tilts left, but there are plenty of people who are vocally right. And that means that if, if you are running trade policy in some country and the U.S. is threatening that country with tariffs, one of the things you think about is who are some swing voters who we can really punish this politician through. And if everyone who, everyone in tech who supported Trump associates Trump with lots of bad things that happened to American tech companies, then it's probably much, much harder for Vance to raise money from these people in 4 years. So yeah, it's, it is a form of trade policy that is very heavily informed by U.S. domestic political considerations. But that's just how the world works. But like it is, I think from from from a big tech perspective, if the UK decides to preemptively be nicer to big American tech companies, then that's actually enough of a tariff victory that they feel good about it. And I presume Trump feels really good about the idea that he didn't even have to directly threaten a tariff to get this. So so maybe that does end up leading to an outcome where there there are probably going to be slightly more barriers for trade, but. It's still a recognizable economic order. And then another thing is that if the U.S. is protecting global seaborne trade routes by shooting rockets at people who try to attack ships, that is economically equivalent to a lower tariff. It it is it it's eliminating deadweight loss instead of eliminating attacks, but it still does reduce the cost of trade. And part of the story of globalization, part of it was changing. Legal and economic infrastructure such that trade was more viable. But a lot of it was just that container ships are a really, really cheap way to move things arbitrarily long distances. And they tend to create much more complicated supply chains that are much more sprawling and very hard to unwind. So it is doing things like that is in some economic sense equivalent to making slightly better container ships or just slightly reducing tariffs.

Yeah, that, that, that is a good articulation.

We'll continue our interview in a moment after a word from our sponsors. 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. For 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 free 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 and mention Turpentine to skip the waitlist.

One of the things that people have been thinking about, of course, is what's the future for the Democratic Party and Ezra Klein and Derek Thompson just came out with a new book called Abundance, and you gave it a brief blurb in your long reads. Antonio Garcia Martinez joked on Twitter or on X that, is this just helping Democrats appreciate supply and demand a little bit? Is this just a basic, hey, markets aren't that bad? I just tried to phrase it in a non-right-wing way, or kind of what's your assessment of the, of the, of the, of sort of the ideas there and how it relates to the Democrat Party?

No, I like it a lot. I think that abundance, it maybe has ended up, you could definitely tell a, tell a conservative story about abundance, but it's already kind of the conservative bet, which is basically on, on economics. The, the traditional conservative thing has been that if we aim for more economic efficiency in, in the econ sense, that we'll all be much better off and that, that actually solves a lot of the problems that redistribution is supposed to solve, that basically you, if you have a sufficiently rich society, it is very cheap to keep people out of abject poverty. And there are just a lot of job opportunities for people. So you don't have to have as many transfer payments to deal with that. And that is, that is always one of those situational things. And there can definitely be times where for a particular voter, it is just way better for them to have more transfer payments coming in rather than for GDP growth to be half a percent higher next year, but over really long periods, the GDP growth thing is really all that matters, and you'd much rather live in a country like the U.S. at pretty much any part of the economic stratum than a country that has perfect redistributive strategies, has had them for so long that GDP growth has undershot the world average for multiple generations. Like that country, you will you'll be a lot closer to your neighbors in terms of standard of living, but none of you will have a great standard of living. And if you live in a more globalized world where people can stream videos of shows that are set in the United States and they can talk to Americans about what their day-to-day life is like and they can buy a plane ticket and they can potentially get a visa of some kind. Then you'll, in, in those more egalitarian countries, they will tend to bleed off talent and that talent will go to higher variance places. So yeah, it's in, in the long run, you do need some level of growth and you need some growth-focused policies just for your country to stay competitive. And then there's another debate in that over how you, how you handle people who are mostly experiencing the downsides of that kind of growth. But I, there's also a, an interesting kind of political entrepreneurship here where the Republican Party is just a lot less growth focused than they were a decade ago. And they are a lot more interested in other issues. Some of it is that some social issues have gotten more salient. Some of it is that immigration is more salient. But some of it is just Republicans are a lot less jazzed to talk about things like how tax cuts will affect incentives and cause people to start more companies and things. And that is true. Like it is true that on the margin, the tax rate you pay does affect the decisions that you make. And in particular, if you have a very high marginal tax rate and someone, if someone is considering a very high variance opportunity, the effective marginal tax rate is actually even higher because In many outcomes, they fail. And then in the outcome where they succeed, they pay very, very steep taxes on on that success. So it does truncate the the tail of the bell curve where you actually the tail of the the distribution where a lot of the a lot of the companies and other institutions that actually create long-term growth that's that's where they are is in exactly that part of the distribution. So it is good to be very careful about that kind of thing. And there's a lot of just nonpartisan low-hanging fruit. It feels like if If there, if politics, if foreign policy got more boring, if social issues were less lively, and if there were actually just space on front pages to talk about things like permitting reform, that we'd probably be in a much better spot. And also just as the parties reshuffle themselves, there is the question of where some people go. So I'm sure there are plenty of people who were libertarians and held their nose and vote Republican for a while and have switched Democrat and then vice versa. There are plenty of libertarians who made the opposite switch, often much more vocally. But things haven't really settled and which party gets which constituency is always an open question. One of the other pieces I linked in the same newsletter where I was talking about abundance was this breakdown of all the detail, the detailed demographics we now have on which voters were swinging in which direction in the 2024 election. And one of the things that struck me was that when I first heard about just how different voter voting demographics work, the, there were a lot of cases where if you're in the middle of some distribution, especially something like income or education, you're Republican. And then if you're at either tail, either very poor or very rich, you're more likely to be a Democrat. Then similarly, yeah. And then for education, it was, if you are a college graduate, you're very Republican. If you have a PhD, you probably vote Democrat. If you have a GED, you probably vote Democrat. And now those are, neither of those are really true. Like the Republican Party does skew more towards lower educated voters and the Democrats and, and they also skew more towards lower income voters. So it is just a different coalition than it was when Mitt Romney was running. And so there are a lot of people who just, they registered when there was one set of assumptions about not just What do Republicans believe, but who is a typical Republican? And now they feel a little bit more isolated. But when you think about something like that, it is also important to realize that we're not talking about something where 80% of the people who earn more than $200K a year vote Democrat. And, you know, 80% of the people earning less than $30K a year vote Republican or whatever, whatever the ranges would be like, there are still, there's still a lot of political diversity within those cohorts. And so plenty of people, they may not, plenty of high income people or well educated people, they may not know it, but they do have Republican friends. So there, it's, it's not as if there's this totally, totally income and class-based distinction, but it is, it is different from what it used to be.

Yeah, it's, it's really interesting. It's, it seems like whatever's going to be popular is just going to be whatever Donald Trump hates or doesn't want because of the sort of the negative polarization. And so the more that he is sort of anti-free market or anti-trade or anti-immigration, the more that Democrats You know, will be, will be in favor of that.

Well, I think for, you know, in particular for Abundance, I, I think there's, there's some level of opportunism in that now is a really, really good time to tell Democrats that free markets are a good idea because you can actually run as the free market person. And America, you know, free, the term free market probably polls reasonably well among Americans. I'm not, not 100% sure on that, but I'm sure there are plenty of terms that indicate that and that are things that a politician wants to say a whole lot. So I don't think there's that much, I don't think much of this is driven by people just saying, I'm going to be inverse Trump in every respect. But I think definitely you want to highlight that message at a time when it's also very politically viable.

That makes sense. Speaking of sort of Republicans and the economy, you know, there's been concern over the last few weeks about, you know, are we going to enter a recession? You wrote a great Capital Gains piece about sort of the different ways that That, that economy could fall into recession. Why don't you sort of unpack that a bit?

Yeah. So a lot of what I was thinking about in that piece was that you, when you look at older recessions, they're, well, the further back in history you go, just the less comparable things are. And financial crises can happen in many different environments, but they are just a much worse thing to happen in a more financialized economy. And so they, they do become in many ways a bigger factor when just there are a lot more, a lot more borrowers, a lot more lenders, and a lot more interconnections within the system. But also, yeah, recessions, sometimes they are just kind of a more random outcome. Economies don't grow forever. And there are feedback loops where if there's a growth shortfall in some sector, for whatever reason, that can ripple out into, into other ones and In some ways, I think you can worry a little bit less about recessions than you used to have to worry about them because the rates are not zero. And while inflation is not running, you know, it hasn't been completely conquered. It's not nearly as salient as it was a while ago. And I think one of the indicators of that is actually that people are specifically talking about the price of eggs because eggs have been idiosyncratically very expensive recently. And if inflation were a broader problem, there would just be more product categories that we would talk about than, than just eggs. So in some ways there is more room for the government to react, but you can have a recession that, that stems in part from businesses just being really uncertain about whether or not the capital investments they make right now are worth it. And I think that that's one of the negative outcomes from, from tariffs is that over a long period, there would be adjustments. There'd be more factories in the U.S. building goods that were previously built outside of the U.S. and shipped here. And so labor costs would be probably mostly higher in the U.S., but your transportation costs are lower. And depending, depending on the, the overall mix of what moves here, what doesn't, you, you maybe end up with a similar or only slightly higher cost structure. And you're also stimulating demand if you have more of those goods built by American factory workers who are going to spend most of their income. So you can have that kind of adjustment. Things can end up being kind of close to normal. And yes, we pay higher prices, but we've secured important parts of our supply chain and that's just A worthy price to pay. You can also have a case where the cost of importing goods is higher because of tariffs. Companies are considering relocating, but they also ask themselves, well, if these tariffs cause a recession and the recession means that Republicans are unelectable in 2028, and the first thing a Democrat does when they're in office is eliminate all of these tariffs, then I just have this big factory in a place with more expensive labor, and I will really wish that I'd kept it in Malaysia or Vietnam or Mexico. Or wherever, so that that kind of uncertainty can can mean that the the inflationary impact of tariffs isn't very self-correcting, and it's hard to say exactly what you do about that other than you can make your tariffs more expensive and more inflationary if you also couple them with subsidies credit or otherwise for people who build in the U.S. And and maybe if if it is just much harder for environmental review to indefinitely delay building a factory, then. The more of those factories get built, because it is probably the case that if you, if you decide today that you're going to build the next Toyota factory in the United States rather than in some other country, it's not necessarily the case that you actually get that thing done and up and running by the time Trump is out of office. So there are a lot of incentives to wait things out and streamlining the process by which projects get approval might actually be a way to make the economy a little bit more adaptable to that.

Yeah. Interesting. You know, there was this sort of line of thinking in the past few weeks around like some people were trying to spin this into a good thing or this meaning the sort of stock market going down. I think they're trying to say things like, hey, this will help us with our debt somehow, or this will sort of like clean the sewage out the system. I heard people say things like, or like, you know, there's a difference between Wall Street and Main Street and this is just affecting Wall Street, but sort of Main Street's not, not affected or, or, you know, the stock market's not the economy or so like, what, what is, what is the right way of thinking about it? Is there a way? I assume that that's all cope basically. And that everything that affects sort of Wall Street also affects Main Street. But let's disentangle this.

Yeah. Yeah. Like if you, if you get fired from your job and your first reaction is, well, the office chairs were really uncomfortable. So this is actually a good thing. Like, it's obviously cope. It is like all else being equal, it is probably better for the market to be up than down. It probably means that the U.S. is producing more things and that these are things that people want such that they are willing to pay a markup for them. So these companies are profitable and people have confidence that the growth will continue. So it's not just that earnings are high, but multiples are rising, et cetera, et cetera. But it's also the case that, yeah, sometimes there. There are political decisions you can make that do redistribute some upside from the, from consumption to investment or from labor to asset holders. So yeah, a rising stock market is not always a good sign and it's always going to be out of sync with the rest of the economy too. So if you look at a chart of the S&P 500 versus unemployment rates, you'll definitely find a lot of cases where they're moving in opposite directions, but it's because S&P is going down in anticipation of a recession. Unemployment still low because the recession hasn't hit yet. And then unemployment starts going up, market's still going down. And then at some point market starts rallying and unemployment is still rising. And so you look at this and you're like, what are people doing? But what they were doing in March of 2009 was saying unemployment used to go up by 0.4 to 0.5 percentage points per month. And now it's only 0.3 to 0.4. So we know that things are actually getting worse more slowly than before. And, you know, housing prices, They're dropping more slowly than they were a couple months ago. And so we can kind of see that at some point the second derivative means that there will be growth again. And you just, you don't want to wait until unemployment is back below 5% and say, okay, now's the time to get into equities. You miss a lot of the upside there. If you see that there's a path towards that, then that's when you buy. So yeah, there's, there is a lot of cope here because it's, it's not as if What's going on is that we have these policies that are just redistributing a lot of what would have been money going towards buybacks into consumers' pockets. That is just not the case. What we see instead is that there's a set of policies that do cause a lot of uncertainty. They cause some deadweight loss. They do have some upsides over time, especially if implemented well, and it's very unclear what those policies will ultimately be. But in the meantime, I think pretty much anyone who is analyzing analyzing stocks in the S&P 500 prefers the policy mix of a couple months ago to the policy mix of today, at least in terms of this year's expected earnings per share. Yeah. So there's, there's a lot of cope there. Yeah. You really, you really, really don't want to hear people in a policymaking capacity working for the US government saying, well, the upside is we'll be able to refinance our debt really cheaply because the, The lower, you know, rates go to zero when there's a depression. So yes, it's true that you can refinance your debt really cheaply, but it's also true that it's not as if that's going to be the main thing people are talking about. Like the headlines in, in September of 2008 were not Treasury, you know, interest as a percentage of total federal spending likely to drop over the next several years. That's, that's not what people were really fixated on. In fact, when when that happens, what the government does, which is a reasonable thing to do in a variety but not every economic model, and it's no one that I really subscribe to. The the reasonable thing to do is okay, money is free. We can borrow a lot of it. Let's borrow a lot of it, and let's spend it very quickly. We're basically taking if you think of the marginal Treasury bond buyer as someone who is saving rather than spending, you want to take the money that they put into Treasuries and turn that into spending rather than saving. So. You're recycling that money towards people who are lower income, have a higher marginal propensity to spend. And all this stuff, it does just, redistribution just happens as a matter of policy. And it is, it is kind of a social norm that we don't want a, we don't want a strictly, no one wants a strictly fair set of outcomes for every situation. But it is also true that if we're, if we're thinking about that, those issues of redistribution versus growth and the trade-off. Between the two, it's just a whole lot better to be thinking about those at a time when the economy is going reasonably well and it could be going a little bit better versus at a time where you just don't know which critical institution will fail next. And you don't know which major industry in the U.S. is going to be going through the bankruptcy courts. That's just, that's a time when you have a lot of really narrow questions around this many tens of thousands of people are not getting the next paycheck that they thought they were getting. And what do we do about that? And, you know, those, those questions have answers that are sometimes, sometimes involve Lots of financial policy changes, and sometimes it's like let's make sure that we know how many police officers are available to to do some overtime if things get really bad. So yeah, you just you don't want to be making economic policy when that's one of the considerations is like how much tear gas do we actually have right now? And I you know we're we're not anywhere near there, and there there are a lot of economic missteps that that can be made that that switch growth from high growth to less high growth. And even if there is a recession, the recession is partly the economy calling Trump's bluff on. How high a tariff should be, and I don't know exactly what Trump would do in a situation like that. But it's entirely possible that he would find some way to gracefully save face. And this is actually a case where the rest of the world would really, really, really want to tell Trump that he won, and he that he got everything he wanted, and you know that they surrender, etc. Because they really the rest of the world needs U.S. demand during a global recession. The U.S. is a large enough share of Global GDP that a recession in the U.S. is probably a global recession unless China deliberately intervenes in order to cause growth, which has its own set of problems. So you you what you'd probably see in that case is that Trump would be getting a lot of credit from world leaders, which very few really informed observers would think was all that credible. But in the end, it would be you know we they would be telling Trump you got it right. We were exploiting you in terms of global trade now. Please lower these tariffs somewhat and start spending more money so that the world economy can grow again.

Yeah. But is it your view in a world where we spend less on Ukraine and Europe spends more and in general kind of like reindustrializes a little bit, is that not in the US foreign policy interest to have Europe sort of sort of reinvesting in their own sort of capabilities there.

What are your thoughts? An empire is an asset; it's also a liability, and sometimes the balance between those two can slightly shift such that you suddenly want to be way less of an empire. And I think the U.S. view for a while was there is no other country that can credibly defend the free world, and so the U.S. has to do it. And there are advantages to ultimately calling the shots on all matters military in. In the U.S. sphere of influence, and then at other times, yeah, that's just there's there's less upside to that, and the countries in Europe are able to take care of themselves. Some of them do actually spend more heavily on military, but most of them don't, and it kind of shows. And so I think it is it is probably it does weaken the U.S. in the sense that it is the U.S. can can really throw its weight around more if it is the only way that Military force can be deployed in a meaningful scale outside of just defending those countries, and even then it's kind of iffy. But maybe that's just not not really worth it. And I also think it's just kind of weird for there to be a handful of real countries, and then a lot of countries that are on this weird continuum where it doesn't. I'm just not entirely clear if a hundred years from now, when people when historians are drawing world maps of what the world looks like and what the borders were in, say, the 1990s, do they actually— I would assume that the right thing to do is you color the U.S. one color and then you use a slightly different shade of that color to color in most of Western Europe and Japan and Australia and a few other places because they don't really have the scale to, to make a lot of independent foreign policy decisions. Like a lot of their foreign policy is implicitly through the U.S. and within the bounds that the U.S. sets. But I'm not sure that America actually is all that good at that. And it's certainly not something like, I don't think America thinks of itself as being this imperial core with a bunch of vassal states of varying degrees of independence under it. So I think if you, if you have an empire that's run by a country that doesn't think of itself as running an empire, then that's probably going to lead to some dumb mistakes. So if nothing else, it just aligns perception with reality by shifting reality more towards, yes, Germany is a real country. It has a military, you know, not just a flag, but a military too. And in the event that Germany's interests are best served through the use of force, their first decision, you know, they're, they will call the U.S. president to tell him what they're going to do and not call the U.S. president and ask him what he's going to do. Yeah. And.

Okay. Okay. So that, that's Europe. And by the way, you probably saw the, the JD Vance group chats that just leaked. It was very funny where they basically let the editor of The Atlantic in by mistake.

It was pretty wild.

Yeah. And he's saying how he hates you.

Yeah. I was thinking of just dropping random messages into some of my group chats saying things like, yeah, we'll probably be launching the missiles in exactly 2 hours. Oops, wrong chat. Sorry.

Yeah, that's so funny.

But Signal, Signal does make it slightly easy to send a message to the wrong person. So I do try to start one-on-one Signal chats with just idle chitchat that makes me comfortable with the fact that I'm actually talking to the person I think I am, because there are just a lot of people with the same first and last initial, and you don't, especially, especially if you have like an urgent thing that you need to tell this person because they specifically would find it extremely funny. That is often the case where you really want to make sure you have the right person first.

Yeah, it is. That's like every journalist's dream. You just get accidentally added to the most powerful group chat in the world. It's hilarious. The, we just spoke about Europe. How about China and particularly as it relates to AI policy? It, you know, the Sam Altman interview with Ben Thompson, Stratechery, kind of hinted that OpenAI might be doing a bit more open sourcing or going back to its roots in some ways. And so maybe, you know, seeing that because China was able to sort of, you know, create DeepSeek, it seems like there's less of a desire to not be open source at the moment. I'm curious what you think about that debate, or also just the export controls, or what's your view basically on, on sort of the merits and trade-offs of, you know, China AI policy options.

Yeah. So I think if you hold the models constant, the questions are the advantages you could have are one, distribution, 2, you have access to the best hardware, and 3, you're applying these models to data that you can actually use to make decisions you really care about. And China is obviously pretty far behind on the hardware question. They do seem to be not, not necessarily catching up, but getting closer. And I think that is an important thing to watch. It's been historically quite a struggle for China to turn dollars of semiconductor industry subsidies into working chips that are close to what TSMC can do, but they've been getting closer on that front. And then there, there are still apparently many ways for them to get access to some Western hardware. So they, they're still able to do some of that, but then on the, on the distribution side, there was a story, I think in Rest of World recently about how Many different Chinese companies and government entities are trying to use DeepSeek's models everywhere they can. And I found that just because it is one of the, one of the obstacles to AI having a big economic impact is that people need to actually decide to use it and they need to rethink a lot of different workflows based on the fact that they have these tools. Like there are just a lot of things where if you, you thought about doing something and you thought to yourself, well, this would be worth, this might be worth doing. I would have to read, comprehensively analyze 10,000 pages of PDFs in order to decide if it was actually worth taking the next step. And that upfront cost is so high that I'm just not going to bother. Well, analyzing large amounts of unstructured text and pulling out specific insights is something LLMs are quite good at. So it does expand the scope of things you can do, but if you're not thinking about it from that perspective, then you may not adopt it because the things it can do are things you weren't going to do and you don't realize that you weren't going to do them because you didn't have the technical capabilities that you, you now have available. So there's, so the distribution side is interesting. And then the data side, China, just the Chinese government collects a lot of data and they have a lot of questions that other governments don't have about what people are saying and who they're associating with and some of the things that they might do, et cetera. So there, there is a lot of room for China to be just a more effective authoritarian country. With access to even models at the, at the current status. And I think one possibility is just they get the models good enough that dissent is not really an option. CCP is pretty sure that it will be able to celebrate its 100-year, the 100-year anniversary of winning the Chinese Civil War in style. And at that point, it becomes less urgent for them to compete. So that, that is one possibility. And beyond that, it does seem like there's, there's this brief window where the, the last generation or 2 of hardware is still good enough. And if you are constrained on that, and you also have just a really cracked team of people who are, who love doing low-level optimizations, that you can make up for some of that lack of hardware access. But then the problem that you run into after that is, okay, you've identified this set of low-level optimizations. You can get X% more out of a given chip. You've probably given a lot of your American peers some ideas for what to do with their much larger supply of now much faster chips. So I don't, I, I could see like if in general, if you, if you open source models, hardware does become a more hardware access becomes a more important competitive factor because that's, that's the difference. If everyone's running the same model, it's just who's running it faster and who needs fewer watts. So. At that point, it actually becomes in the U.S. interest to have really good open source models. And if you do build models that are open source, or really it's open weight, open weight models that are trained on a corpus that does tend to serve U.S. interests and that will censor the things that you want censored basically, and not censor the things that you'd actually rather the average Chinese netizen have a little bit more familiarity with, then, then maybe that actual open source arc is a good thing. And it's certainly something where it gives Sam Altman some political cover if he can say that there's a model that is better than anything DeepSeek has. And yes, it knows who Tank Guy is. And yes, it knows which beloved children's character Xi Jinping has an uncanny resemblance to. So this model will be running on every, every smartphone. Smartphones tend to move across borders pretty frequently. And therefore this is actually U.S. soft power being projected into China. And then that, that thesis actually runs into the problem of, okay, if, if China just wants to get models, widely deploy models that are good enough for keeping an authoritarian state in power, well, if those, if there are better models out there, does that mean that the state is actually getting weaker unless it continues to try to keep up and continue to advance the state of the art? And that, again, very interesting open question. And it will be fascinating to see what, what beliefs are implied by things China does in the AI policy realm.

Well, Bernd, that's, that sounds like a great place to, great place to wrap. As always, this has been a wide-ranging conversation and until next time.

Yes, indeed. Great chatting.

Thank you. Upstream with Erik Thornberg is a show from Turpentine, the podcast network behind Moment of Zen and Cognitive Revolution.

If you liked the episode, please leave a review in the Apple Store.

Store. 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, 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 E-R-I-K at turpentine.co, and let's partner together.