Samo Burja - Live Players and Institutional Reform

Samo is an institutional theorist, the founder of Bismarck Analysis, and the editor of Palladium Magazine. Highlights of our discussion include why society’s political agreements are 80 years out of date, the pressing need for procurement reform, and how Estonia in the aftermath of the Soviet Union provided better public services with fewer people.

The audio is available on Spotify and Apple Podcasts. I’ve recently added an RSS feed, meaning the episode is now also available on the likes of Pocket Casts, and all other major podcast apps. Just search for ‘Public Service Podcast’.

You can find a transcript of this discussion below [with notes, set out in this format].

This is the penultimate episode of my ‘Emergent Ventures’ series. I’ve one more talk in the series, which I’m very excited to release. Thanks again to the Wallis Family for their unwavering enthusiasm and support, Sammy Cottrell for the superlative recording studio, and my editor for putting this together.

Timestamps

(00:00) Intro
(02:18) Live Players
(09:14) Effective Bureaucracies
(15:45) Obsolete Political Bargins
(22:17) Procurement Reform
(34:25) Wasteful Processes
(41:13) Economic Challenges
(51:12) Demographic Challenges
(55:38) Ways Forward
(1:04:48) The Failure of DOGE
(1:12:26) Industrial Policy

Transcript

Adam: Today I'm speaking with Samo Burja. He's a sociologist and political philosopher, the founder of Bismarck Analysis, an institutional research firm that consults with governments and major companies alike. They publish most of the work online as the Bismark brief newsletter which is available to the public via subscription. Samo is also the editor of Palladium magazine which covers topics related to the future of governance. Samo, thanks so much for coming on.

Samo: It's great to be here.

Adam: You've been very prolific - publishing online in various forms both through the institutions that I mentioned, and personally. I think you've been really useful to people interested in governance and introducing various ideas and in helping condense quite complex systems into you terms that can be utilised help smooth the wheels of discussion on these topics. I think you deserve a lot of credit for that. One of the ideas that comes up again and again is this idea of ‘live and dead players’. Would you be able to give us a brief overview of what your theory is there?

Live Players

Samo: Yes, I think it's useful terminology I first introduced in a 2017 essay on the topic. Essentially, a live player is an agent or a group of uh agents that are capable of adapting to new circumstances, not just operating off a script. It came out of the analogy of ‘imagine yourself playing a chess game’. So if you're playing a chess game that's against a preset game when someone is just taking moves out of a playbook where a previous, even masterful chess games, have been executed, eventually you'll be able to beat them, right? Because no matter how good that pre-written, pre-recorded, chess game is, it can't respond to someone actively studying it and adjusting to it. I think the terminology is useful in the world because most people don't pursue action from first principles. There's a whole lot of imitation in every field of complex practice, of course medicine obviously, law, politics, economics right? Because of that, I think it's better to think of most people as operating off of scripts that they didn't come up with themselves. Then, when someone actually does undertake effective action that's not on the basis of those scripts, it becomes very surprising. It becomes very interesting, right?

It is in fact not the case that, before the 2010s, starting an electric car company seemed like a good idea. In fact, it seemed like an absolutely terrible idea. It was only Elon Musk's Tesla that was actually pursuing it as a business, where customers might really desire those cars and so on. Starting an electric car company, even one with self-driving, now in 2025 in China? Well, that's not a live player move. Even if they might be very successful companies. One thing that I do have to emphasise here is that following these pre-computed scripts things like general professionalism, and so on, it can still lead to good outcomes. It's just that in a competitive context it won't tend to suffice. So, if you can identify if someone has the cognitive capabilities, or if we're talking about an organisation has the organisational coherence, to execute things that are non-standard for their industry, or completely jump domains, then they will almost certainly, in a way that's quite unpredictable, eventually out compete any dead players in their domain, no matter how significant the seeming incumbent advantage of those dead players would be.

Adam: Sure. Thinking about it a public institutional context here. Really love what you're saying there. It's not so much that… this isn't a totalising judgment, but rather an assessment of whether or not it's able to respond to competition. There's a sort of fractal nature here as well?

Samo: Well, not even just that.

Adam: And uncertainty?

Samo: Yes, respond to competition, but more importantly be capable of breaking whatever script or recipe is followed by default, as either the industry standard as the social norm or standard, as the political convention, and to do so in a way that is still effective action. So in a competitive context, this is most relevant. In a non-competitive context, well, it's it's hard to see what the counterfactual is, right? The organisation, say a major company, like Boeing, right? They might not be innovating in aerospace because there's no innovation left to do, or they might not be innovating in aerospace because they're somewhat dysfunctional organization. At the end of the day, who are they even competing with? Basically, just their own products or I don't know, maybe Airbus.

Adam: Other lobbyists, maybe.

Samo: Exactly. So, it's more that the competition reveals who were dead players, and who were live players. I feel this is a surprisingly fruitful binary. It's surprisingly useful to start an evaluation of an organisation with the question of ‘are they a live player?’ Before you even figure out, ‘are they profitable, are they well connected, do they have the best experts?’, just first ascertain whether they have the capability to carry out new things well.

Adam: Right, so just moving away from the private sector into the public I would imagine here that the distinction is most often felt in cases of uncertainty or unprecedented circumstances right? So like might look at Covid as such an example where you know countries responded quite differently. Even within different countries there were organisations and institutions that performed extraordinarily well. I just was just, earlier today, talking about the example of the UK's special forces that within something like 24 hours were able to send logistics equipment and organise a heap of protective equipment to be brought back to the UK at short notice. Meanwhile the Ministry of Health was completely…

Samo: In shambles.

Adam: Right, unable to really understand what even the nature of the problem was. There was a literally a playbook at the Ministry of Health for pandemics, that was generalized to meet all pandemics, and the play was herd immunity. There can be a debate about whether or not that that was a good thing or not, but certainly not a live player in this case. Just to fully explore this idea or explore it a little more, you could have potentially a live player within an organisation that is itself, at least temporarily, not a live player, right? So somebody is working within reforming doing something unusual, they don't necessarily fit in in some respects, but maybe rise to the top, take over the institution, and reform it. Maybe a Deng Xiaoping type character might be such an example.

Effective Bureaucracies

Samo: Yes. First off, it has to be said, historically unprecedented things tend to happen every 2 to 3 years. So maybe you're lucky and you get through a typical Western political four-year cycle without one, but that's being lucky. It's likelier than not something unprecedented happens multiple times in the political lifespan of any administration, let alone in the lifespan of an effective civil service bureaucracy.

The question of ‘Okay can we have an institution be piloted by live players that is not in ordinary times a piloted institution in this sense?’ It can happen, but this is likelier to happen if the organisation was already fairly well-run. So even if it didn't have the capability to do something new, it was a functional institution very much aligned with its own function; in service of the explicit function of the organisation. This is likelier because you can always appeal to that mission as a strong moral or even legal argument for reorganisation.

The less effective an institution was to begin with, even at executing some sort of script on the basis of some politically legitimate legally mandated rationale, the less weight that argument will carry because well ‘we've already not really been fulfilling our legal mandate. No one's really held us into account. No one here is really that motivated to do something in this space’. So there's nuances even among dead player organizations, right? You can have a functional or a nonfunctional bureaucracy, right?

I think that in the political world… Well, you mentioned Deng Xiaoping, but he really only had a shot at power when sort of Mao's immediate successors in the Gang of Four completely burned their credibility, and when political infighting had destroyed many of the other many competitors for power in that stage of the political evolution of China's system. You can have such things as productive conflict, but it's not always just conflict that brings that to fore, right? I think in functional organisations rapid promotion happens for this reason. I bet there were people at the relevant Ministries of Health [globally] who knew exactly what needed to be done, but had no capability of bringing others along with them. So it was a case where the whole was less than the sum of its parts. I think that the key challenge in a bureaucratic context is always to have the whole be at least as much as the sum of its parts—‘more than’ is very tricky—but at least as smart as the people making it up. That's a very smart bureaucracy indeed.

Adam: Indeed. I struggle to try to make a coherent picture of all this, given just the complexity involved. In many cases, you know… international organisations, large bureaucracies that have many different functions, that many different things could be said about [positive and negative], that was all true. How do we orient ourselves in this space? I feel quite strongly that there's a group of society that is increasingly thinking to themselves, looking at public institutions thinking, ‘they're not working for us. Let's burn it all down’. We see that a little bit at the moment. There are riots in Northern Ireland. There are reactionary forces at work. The Trump administration is doing this quite explicitly and openly. Yet there's value in stability and the rule of law, these principles. How do you how do you even begin to talk about these things coherently? How do you judge action? I'm very interested in your views on this.

Samo: I think it is a mistake to equate inaction with stability because you know at every point in the short run doing nothing is the least risky thing. In the long run doing nothing is the most risky thing. At some point the lines cross and whatever was the default has become thoroughly obsoleted by several factors. For example our social systems were set up in a time and place where society was about 30 to 40% poorer, but also younger by about 20 to 25 years. A lot of these countries, like Japan and Germany, the systems of social transfer or bureaucratic functioning they were designed for much younger societies. They were also designed for societies that did not have instantaneous mass text communication, social media, and so on. They were designed for societies where agricultural labor was still an important component. Now I know that, say, New Zealand is a little bit different, but in most countries around the world it's only like 1-2% of the population…

Adam: [New Zealand is] about three and a half. So it's not all that different. A lot of mechanisation has occurred.

[There are annual fluctuations in employment here, but around 5.2%-6% of New Zealand jobs are in the Agriculture, Forestry, and Fisheries category, so around 3.5% once you adjust for the fact that only 60% of the total living population is in paid employment]

Samo: It's not all that different by global standards. In the early 20th century we're talking about populations like 20-30%, even in places like the United States. People talk about job automation, but as far as farming is concerned the robots are already here. Yet despite all the automation, and despite only 1% of the population being sufficient to produce all the food and calories we need generally, we still talk about labour shortages in farming. Isn't that interesting?

So, anyway, that's a small a small tangent, but I think a telling one, right? Because laws that were made to protect 30% of the population might today only protect 3% from say economic competition or something like that. So what I'm saying is a lot of the political bargains that underwrote laws, policies, institutions, bureaucracies..

Obsolete Political Bargins

Adam: The New Deal, for example.

Samo: All of those, all those balances of power, distributions of economic value, strategic considerations, 80 years out of date, right? 80 years out of date at this point. You need to update those, but to update those entails some amount of political fighting. But I really do think that a lack of belief in our ability to make new settlements, new predictable equilibria, that are just as good in terms of stability as the previous ones were… I think there's there's a crisis of our confidence in our ability to do that, right? Let's remember, these institutions themselves came as rapid, at some points even breakneck-rapid, adaptation to new economic and political circumstances, both domestically in various countries dealing with the Great Depression, but also internationally with a with a system that had been shattered and not really rebuilt effectively by World War I and then of course maybe, let's say, somewhat stabilised and rebuilt into the bipolar world order with World War II, or the aftermath of World War II.

So the origin of these systems of stability that we have today was in periods of instability, and was was itself a great gamble. So, I think that we are thoroughly approaching the end of the lifespan of these institutions. This does not mean that the lessons behind them are obsolete. Not at all.

Adam: The principles, perhaps.

Samo: The principles. Nor does it mean that this is needs to be like a total overhaul of Western states or anything like that. It just has to be at least as an ambitious adaptation to the 21st century, which is now in full swing, as in 19__. You know, today in 2025 we should be at least as ambitious as adapting to a much changed 21st century as we were in 1925 to adapt to a much different 20th century. A refusal to adapt, that's actually what leads to instability. That's actually what leads to the equivalent of the Bolshevik revolution. That's actually what leads to the German unpleasantness of the 1930s, it was the inability of these new young constitutional republics at the time to resolve these problems. The countries that did have that ability like the United Kingdom, or the United States… it's not that they weren't risky things about their politics. There certainly were. There were questions of high state discussed, and decided on, most of them seemingly well enough. They didn't undergo such crisis, and that's not because they were conservative (I mean in small-c conservative sense) it's not because they were afraid of change. No, they just took principal change as their guiding principle. So that would be sort of my way of squaring that. I feel like stasis offers more the illusion of stability rather than actual stability. Actual stability is like a healthy living organism renewing itself, right? You know, it's not mummification, it's life.

Adam: Right. This is possibly is a bit of a tangent, but there's parallels there to the stock market. The liquidity provided by short sellers and high frequency trading which is often maligned in various ways—‘you're making money and out of other people's misery’, these sorts of things—but the outcome of this sort of engagement is far more kind of smoothness and stability, rather than like some kind of short sharp discontinuity—suddenly the pound collapses, and everything's in turmoil—allowing instead new equilibria to be found more organically.

Samo: Yes, because if you insist on appearances, eventually the underlying balance— be it of trade, or power, or technology that underwrote that appearance—you're just letting those two be disconnected more and more, and you wake up bankrupt metaphorically or literally in one day.

Adam: ‘Nature cannot be fooled’ as Feynman said of the [Challenger] disaster of the Space Shuttle.

It's fascinating hypothetical, and I think from a theoretical point of view you can see the parallels you can see that that is obviously so. Any system that's not changing and and adapting, any life form that's not growing is busy dying.

But how should we be thinking about this in the 21st century given that I think the Ministries, and the things that we've created, are by and large perhaps, roughly speaking the sorts of things that people want? There are, after all, elections, and change from time to time, mergers [of bureaucracies], and so forth. I don't see a massive change in the role of the state necessarily being something that people want. Yet I agree with you entirely. There is a need to ‘refound’ but without necessarily… So yeah, can you walk us through like a particular example perhaps?

Samo: Of a desired reform in a 21st century, or a country that is [already] well reformed in some way?

Adam: Sure, or take a place like the UK, US, New Zealand, just walk us through like an institution that's not performing well, or seen not to be performing well, and how you might look at approaching like that reformation. Like how do we identify whether something actually is poorly performing and in need of change, versus just simply carrying some degree of dead weight loss, for example. Through to what you would be looking at doing in terms of like setting it up for the 21st century.

Procurement Reform

Samo: Well, one of the biggest questions is the relationship with government contractors in most of the major countries. Most of the major countries in the Western world, as all modern states do, have strategic needs that have a big, difficult to deliver, technological component. Now whether we're talking about building nuclear reactors, or we're talking about aerospace, or whether we're talking about secure data centers for say the NSA, or something like this, these are complex technical projects, and many of the companies that work with major Western governments have been companies that have been primarily government contractors for decades. The problem with a market where there's only one customer is that it's not a real market, right? It's a monopsony. Conceptually the opposite of a monopoly, but it produces some of the same problems, right?

I think many of these companies at first were very lean, hungry, effective companies. Companies like Northrup Grumman when it was still Northrup in the 80s would take interesting risks in developing new technology like the B-2 Spirit stealth bomber. Of course, the price point might have been too high, but the technology worked. Run this forward a few decades, these same companies will have become quite used to their own unique, and at times irreplaceable, role. So I think Western governments have do something that is politically very uncomfortable. I think they should bet proactively on new companies developing replacement products.

In fact, we have one such success story in the western world. You might hear that the Falcon engine, used in the Falcon 9 rocket, that it’s named after the Millennium Falcon or something like that? It's not. It's named for a particular project that was sponsored by DARPA, the Falcon project, which was this small scale launch. So there was a small scale launch that caused the development of the Falcon rocket. It allowed SpaceX to become a private space company. Then roll the film forward 10 years. You now have a government contractor that can deliver payloads into orbit much more cheaply than Boeing can.

I think this should be done in a whole variety of fields. So, I'm not calling for the privatization or the nationalization of anything that has not been privatized or nationalized. I'm saying that we should basically reset almost every public-private partnership that is 20 or 30 years old. Merely open something small, allow completely new and unrelated company to compete on the bidding. If it does a good job, take the radical, and politically unpopular, step of going with the disconnected company that doesn't have lobbyists in DC or in London. Nurture that company for a few years and then perhaps you'll get a much much better product for the same cost or a lower cost. Right?

There's another point here. You know, Ukraine is at this point often declining US direct military aid. They would much prefer $2 million, or $20 million, or $200 million, that they can use to buy drones manufactured in Turkey, or build their own drones, than to receive one or 20 or 300 American drones [respectively].

Adam: These legacy systems…

Samo: Exactly. For $10 million, they can sure build a lot of drones. The US can, I don't know, maybe build five drones for $10 million?

Adam: You're talking about these, is it the Reaper [MQ-9] or something, these enormous single systems that are 20 years out of date.

Samo: It almost doesn’t matter. Even if we’re talking about missiles, right? What I'm saying is that almost all Western military hardware is: it's not that it's obsolete, it's just obsolete if you try to use it cost effectively. In other words, it's useless against a power that's technologically advanced. Even useless versus Russia. If it takes like, you know, a $200,000 missile to take out Russian equipment…

Adam: Better off just using a truck…

Samo: …and letting the drones fly. Well, you know, people noted the destruction of the so-called strategic bombers, which just to be clear, I think those strategic bombers in terms of delivering nuclear munition, those were already obsolete, and had been obsolete for many decades.

Adam: Turboprops?

Samo: No, not quite turboprop. They're fast enough, it's just that they're very visible to sensor systems and radars. You know, the B2 bomber, the American one from the 80s is much better than they are, but even the B2 bomber is like probably somewhat obsolete, and could be shot down by with modern sensors. Still, more importantly though, you know, the US parks its bombers and its aircraft outdoors as well. They don't have them in the hanger, right? So the US would be just as vulnerable to the same drone attack. Note the Russians, yes, lost a significant part of their Black Sea Fleet to Ukrainian drone attacks. Let's say, I would believe American and Western ships are about twice as good, or maybe even five times as good, at shooting down incoming drones as Russian ships are. I don't believe they're 100 times as good. So, in other words, hey, maybe boats are obsolete. How would we deal with that strategically? What should the Navy be procuring in that case?

So, there's two related problems. One is over time we've had sort of feature creep. We've had organisations that have not been under fiscal discipline because they've been serving the same customer, and the customer just has to believe you, maybe occasionally throws a tantrum, when you say how much something costs. So, we're buying what we want and it's too expensive. The second thing is, the problem I'm now pointing to, is we're buying or wanting the wrong thing. We keep trying to rebuild a late 20th Century military.

Adam: The Gulf War.

Samo: The Gulf War was the last paradigm of warfare that was really comfortable in the Western world, partially because it produces the idea of a war without many casualties. It's almost unthinkable to say that the United States might find itself in a war where it loses 100-200,000 people. That's almost unimaginable to propose, but I think is the reality of where the balance of power is, in an era of war like what we're seeing in Ukraine, and we would also see that elsewhere.

Now this is not even a specific point about war fighting. I think this generalizes to 20th century infrastructure. Energy infrastructure. Data infrastructure. We keep wanting to buy the wrong things, because we're still in a very 20th-Century mindset. If 10 years ago you proposed: ‘we need to build many many more power plants because data centers will consume a lot of energy’. I don't think this would have been a compelling argument. Of course, today it sounds more like a compelling argument, doesn't it? But to start building power plants today rather than 10 years ago, well, that's almost too late. So we were planning for past needs and not anticipating future needs.

Adam: So just thinking about this idea of, essentially you’re proposing [widespread] procurement reform. I think anybody paying attention would very clearly note that defense procurement just needs to radically change. That's essentially unquestionable at this stage, but I wonder if you could talk about something a little bit more… give us more mundane example like in terms of software, or like road maintenance contracting. Are we are we talking about that kind of thing as well?

Samo: I think that there are interesting and difficult problems in the domain of construction. Worldwide I think in every society around the world, the construction business and the corruption business are awfully close to each other. For example you know Erdogan, before becoming the president of Turkey, was the mayor of Istanbul. There were many high-profile construction projects, many of which that greatly enriched Erdogan's political allies who then later backed his campaign for president. That's not unique to Turkey. That's just something that could have happened in 1970s Italy, honestly 1980s Chicago, and so on. Construction is just somewhat illegible. There are always ways to sort of siphon off money. So I have no concrete prescription as to how to make construction more cost effective, which by the way is clearly needed in the western world. We're talking about extreme situations like a mile of subway costing like almost half a billion dollars or more in New York City. And in China, you know…

Adam: Or even forget China, even, like, Spain. We're not talking… I think this is something that's often underappreciated. This idea of like, ‘ah, well, yes, China can do things at scale, but that's only because they can, at the stroke of the pen, just annihilate entire ecosystems…

Samo: But even in Spain it's much cheaper that New York.

Adam: Right? It’s like a tenth of the cost or something!

[Madrid’s system, predominately constructed 1995-2007, had an inflation-adjusted cost per mile between 1/5th and 1/20th of similar modern Anglosphere systems. Something is uniquely broken about the way we deliver infrastructure. An excellent Works in Progress article covers this story in depth, for those interested]

Samo: It's a developed country. A lot of the challenges that Spain has, like expensive energy, an aging labor force, well, they're similar to America's challenges. Honestly even New York is better off position in terms of energy cost than Spain. Actually, Spain is relatively cheap energy for Europe, but that's a certainly better position than say Germany is.

Adam: It's a broadly comparable nation in many respects, and yet…

Samo: Exactly. So there's some small and clear… there's significant differences. There's significant possibilities for savings there. I think that again just going with completely new options rather than contracting the same old companies, I think that's at least a band-aid, right? Because what do informal..? Informal arrangements can be good because you can get things done that are not explicitly called for, but then also over time the more experience you have working with the government, the more you know how to basically press all the right buttons to get all the money to fall out. So there's some value in having an inexperienced contractor. If it's a new company that came together, or if it's a company that previously worked mostly for the private sector, because if it mostly worked for the private sector rather than the government, you might not know anyone that works there. That's a negative. However, they don't know how to press the bureaucratic buttons to make the money fall out, and that's an advantage from the government side. So I think there is something to be said for that.

Wasteful Processes

Adam: But we've made the process to get a government contract is just, in many cases, so phenomenally complex that even for another wing of the government at times, the amount of process required to get some money to to support a road project for example is uh… the process itself requires a consultant experienced in the process to interact with another wing of government, and this is completely normal.

Samo: I think governments spend too much time fighting themselves and this is a little bit a consequence of the fact that the responsibilities have been ever more finely divided to provide political careers for ever more people, right? Because the more you devolve responsibilities… Let me propose: There's something to be said for checks and balances between branches of government. I'm not sure a city actually needs that many checks and balances. I think having a pretty powerful mayor to be empowered to do infrastructure projects, and not even needing to answer too much to other levels of authority. If they do something that's like truly out there, I'm sure a court will like find them guilty, whatever corruption, or they'll just be voted out, right? Nothing bad really happens with a powerful mayor in the United States. It would not happen in Germany, and it would not happen in New Zealand. I think we've taken this logic of fragmenting power at all costs much too far, because the point of a process... If a process leads you to a mile of subway costing half a billion dollars, of what use was the process? It should just be scrapped and delegitimised.

Adam: (sarcastically) But we’ve made sure that value has been created!

Samo: Well, astronomic amounts of money are wasted on the appearance of responsibility, and that's a big problem.

Adam: That's a good way of putting it. There's there's some ideas here I have around the relatively short tenure of most bureaucrats at this stage [time]. The labour market has become very fluid. People don't have large families. It's easy to move from one place to another. In general, people come and go all the time. I think we still live in some respects in a world in which the people are, hiring practices and so forth, are reliant on people's reputations preceding them. In a world where somebody works at an organisation for, let's say, 15 years, then you can judge them on their fruits. Whereas if you're only there for 2 years, then the incentives sort of change a little bit, particularly for longer running processes like infrastructure, where [as a senior executive] what you're trying to avoid is any enormous fuck-up becoming seen to be your fault, right?

Samo: Yes.

Adam: So something happens, inevitably, there's some enormous issue because of the sorts of dynamics that we’ve been talking about. The natural response, I think the totally intuitive response, is to impose some new process to make sure that that failure can't happen again in the same way, because otherwise that would be the end of your career, regardless of whether or not globally—or within the system—that is optimal or it destroys, on net, enormous amount of value the new process. So I'm not sure how to break out of this particular issue. It just seems to be a natural consequence of incentives. But it doesn't happen any everywhere though, so…?

Samo: It doesn't happen everywhere at the same pace at least. You know, doing something behind schedule is considered a massive fuck up in many countries. Right?But not in others. I think some of that's down to political culture. Some of that is down to incentives. If you look at something like Lee Kwan Yew’s Singapore. If someone did something behind schedule in order to minimise risks, being behind schedule would itself be something that would be held to account, and that would not you know not be considered good performance, either in the civil service, or if we're talking about the political wing within the People's Action Party that governs Singapore. But we don't need to go to Singapore. A sense of urgency, speed, and effectiveness pervaded large aspects of the civil service in the United States in the 1950s and the 1960s on a whole variety of tasks and missions such as at NASA.

Adam: Or the interstate highway system.

Samo: Sure. I don't want to overfocus on the cliches, it's just that in the case of NASA the difference between the NASA of 1969 and 2025 is so obvious… but yes also the interstate highway system right? I think there's something rather sad or ironic about us attempting to revitalise the United States which is… since I'm based in the United States, and I often see this discourse where there's an attempt to use the competition of China as a way to revitalise us. Thinking that, if the interstate highway system was partially built on national security grounds because every 5 miles of highway have to have a mile of highway that's straight enough so that the plane can land in wartime or… (no actually that was the official regulation) or that it's possible and easier to move supplies across the United States, right? If it's possible to build energy infrastructure and if it's possible to develop space technology and all of that in a post-Sputnik spur of the moment… I think that we want to believe that the default response of the western political system when encountering a geopolitical challenge, or an economic challenge even. Let's forget geopolitics, let's just talk about what an economic challenge how much a $10,000 electric car is. You know, in fair and open competition, no Western company, including Tesla, let alone General Motors or BMW, can beat Chinese companies in terms of quality and cost of an electric vehicle. They just make the best. They make them the cheapest. In the United States, you have to slap 100% tariffs on them just to barely, barely make the Western car companies work. Okay, that's an economic challenge, right?

Economic Challenges

Adam: and this is not just cars to be clear (laughs).

Samo: No, it's not just cars. It's cars, it's phones, it's literally everything you can think of. Anyone who thinks that Apple just designs better phones, even if they're built in China, than what Chinese firms can design, I think they're sleeping on the quality of Chinese electronics and have been maybe for 5 years, maybe even for 10 years.

Adam: Not to mention Apple itself has become something of a dead player, would you say?

Samo: Of course, yes. But the broader point I'm making is even if we are blessed with a peaceful 21st-Century, economically we must reform things, or first-world countries, the developed world, will become known as the formerly-developed world. Just as we talk about the former communist countries of Eastern Europe, we might come in a few decades to talk about formerly developed countries. Because if you don't manufacture anything, if your financial sector is actually a sham, and it's not related to anything truly investable…

Adam: Or it’s uncompetitive, because it has its own problems.

Samo: Let's say maybe tourism is good, but you know historically it worked out better for Britain that was exporting textiles, than for Portugal that was exporting wine. For 300 years Portugal exported the best wine in the world. Britain first exported textiles and then it exported steel and then it exported airplanes and cars. Today it's sort of like, okay, Spain and Portugal are still exporting wine. I'm not quite sure what Britain is exporting, maybe nothing.

Adam: Financial services, money laundering.

Samo: I don't think that really leads to building spaceships or AI systems anymore. Not really. I think China's bet is in a way better [position]. Still though, the very straightforward point here is we've really neglected… we are in a severe economic bind, without our own industrial policy, which is a developmentalist policy. Which requires us to not have this perspective that, ‘oh we're already rich we just need to divide everything equitably’. Our perspective would be in the very near future we might find ourselves to be the global poor and very old poor people at that.

Adam: Is this in part you're projecting a sort of a maybe even like a preference cascade of sorts as consumers switch very rapidly—suddenly the Chinese phone globally is the cool one.

Samo: The dominant cool phone, yes.

Adam: iPhone which has sort of dominated the segment…

Samo: Have you used Siri recently. Do you have an iPhone?

Adam: No.

Samo: Okay. Well, I want everyone to go talk to Siri right now if you have an Apple phone, and compare Siri to ChatGPT.

Adam: Right. Just night and day.

Samo: Night and day. Okay. Let's say there's a Chinese company that clones the iPhone, or makes something nicer. Maybe makes a little nice phone that opens like a book, talks to you like your college sweetheart, or your grandma, or Patrick Stewart or whatever voice you like. And you can just like, you know, you take up your phone, you literally press a button, you're like, "Hey Xiaomi, book me a restaurant. Book me a nice romantic restaurant tonight." It has your chat history. It knows what your wife or your husband loves right? What your boyfriend or girlfriend loves. It knows where you are, and it just uses a normal chatbot [i.e. an LLM-based agent] to book that restaurant.

Adam: Right, and suddenly like a $3 trillion company is like in deep trouble, and you can see this potentially [happening], like there’s not some [enormous moat].

Samo: Well this happened to Nokia. Let's remember Nokia was completely wiped out by the smartphone, by the iPhone right? Under Steve Jobs Apple was a Live player, I think currently it's a very successful, very rich, Dead player. Who uses the vision pro?No one anymore.

Adam: I don't know what it is.

Samo: Oh see, well for a little, it was a [popular] new product launch. The VR goggles.

Adam: Oh, those things.

Samo: Yeah, those things.

Adam: I've never seen them once in my life.

Samo: Exactly. I used to see them occasionally at San Francisco parties. People would show them off, and they were convinced they were the next great thing. I was like, well, I wish it all the best, but it wasn't.

Adam: So really, you work your way down through the list of the top US companies, or the top companies across the West in general, and you think okay well there's Nvidia, currently

Samo: Currently the undisputed front runner.

Adam: Right. But [even there] Huawei is coming up close behind.

Samo: And Nvidia doesn't make the chips. That's TSMC, and they are a Taiwanese company. But did you know that about half of TSMC's facilities are on the Chinese mainland? Unsurprisingly, Taiwan's main business partner is China. And if when you look at SMIC, which is the Chinese equivalent of TSMC, do you know who the senior people are that work at that company?

Adam: Ex-TSMC employees.

Samo: Exactly. Ex & retired TSMC employees. So any thought or idea that the Chinese will not catch up on chips. Wrong. What we need to do in the Western world is figure out an economic system. Now at this point, I think trade is great, but at this point it might just be necessary to make something like some sort of Western economic block. You know European countries responded to the challenge of the United States by attempting to produce a common market. I don't know, maybe it's time to merge the market of, let's call it NATO plus, right? Maybe that's not a bad defensive strategy. We all put some tariffs on China and then maybe our own companies build something good. That would be hell to negotiate. That's of course not going to happen.

Adam: It would also be an existential risk for us in New Zealand and Australia. Our largest plane trading partner is China. We're completely dependent on free trade, not just with China, but with the world.

Samo: Well, then maybe New Zealand doesn't have to join it, but…

Adam: We’d find ourselves in a difficult position.

Samo: Yeah, but I do think that globalization as we've known it, is going to some extent breakdown, right? Because the US will inevitably… I mean it doesn't need to be done in the Donald Trump way. It was already happening in the Biden administration. It would have been bound to happen in a Harris administration, too. There will be economically protectionist actions taken by the United States. It's just inevitable because the pain of job loss will be too severe. Now, at first, you might narrate being outcompeted by China as AI or automation driven. But I really think that it's actually, let's put this way: in China, they have no unemployment problems. In fact, every single year the cost of labor increases in China. It's just supply and demand, right? Being the world's factory, even if you're an automated factory, still requires a lot of workers, right? I think the problem though is that the purely reactive political pressure to save a failing company… We've seen this play out before in the 1970s, right? It did not save Britain's industry in the long run. It did not save the rust belt companies, in the long run, in the United States. And you know what?

Adam: You get spiraling competitive decline. New Zealand is the same story actually, we can talk more about this [later].

Samo: Oh, I'd be interested in learning more about it.

Adam: Yeah. This kind of this… in retrospect it's been called Fortress New Zealand. It was before my time, but an attempt to almost develop an autarchy of sorts, you know, with rampant problems with debt and so on. You need you need to be exposed, the companies need to be exposed, to the global market in order to continue to be a competitive and have export discipline and all that kind of stuff.

[I should write a piece about Fortress NZ someday, the details of this mad period are far less widely known than they should be]

Samo: Exactly. I mean in a way it's much better to try to subsidise your company so they export in a certain way, because then they're exposed to the discipline of global markets. Their products just have to reach a certain level of quality. That's in a way better than to try to subsidise them to continue producing subpar goods and flooding your domestic market with them.

Adam: So you’ve certainly convinced me already of this idea of like, ‘we needing to reinvigorate a certain degree of ambition’ ultimately.

Demographic Challenges

Samo: Let me put it this way, it would take extraordinary ambition and action for us to even continue with business as usual. The default prognosis is not business as usual. No, the default prognosis is that more and more western countries will actually default on their debts until maybe the big final domino falls like 20 or 30 years from now.

Adam: Right. Big issues with not just debts, but liabilities more generally.

Samo: All kinds of liabilities, retirements and growing aging populations. Again, the simple reality is that no matter how you know even if you took the maximalist mass immigration position, which is politically untenable, as the governing class of Canada and the United Kingdom are learning, if you look at the voter polls and how the voters feel about it. But also economically (or Northern Europe, same story), when they did studies in, say, Denmark it turned out that unless you are selective about your immigration, on average it’s actually net negative for budget¹. If you are generous in your social services, you actually have to have quite productive workers before you breakeven per marginal additional worker. That can still mean high immigration. Arguably Canada had the best of both worlds from about 2000 to 2010. I think there just something happened in the immediate post-Covid moment. Maybe it was an attempt at a kind of economic stimulus. Maybe it was some other reason, it just became less selective, right? The result then is worse economic performance.

[Canada has had negative GDP/capita growth since 2022, at a time when the world economy has been growing at ~3%pa]

Adam: Ultimately even leaving aside those dynamics, there's nowhere in the world that's producing enough young people anyway, right?

Samo: Well they might be for the next 10 or 20 years, but in 50 years, no one will be. In 50 years, if your country has young workers, it's not because it accepted them from Nairobi or from Kerala. It's because it makes them at home. Maybe you make them with unskilled labors. Maybe it's humans making them. Maybe you're growing them in CCP-sponsored baby vats, but you're making them yourself. Actually, countries might become protectionist in a different sense. Today, when you look around the world, the only country that taxes you when you renounce your citizenship for 10 years after you leave is the United States. Can you imagine if a country like India adopted such a law? How much revenue would it have? That's a lot of revenue for India, right? Imagine in India when the average age is 40 or 50 and they struggle to fund themselves, right? Because they became old before they became rich.

Adam: Right, and you could even see the moral argument would be quite straightforward, right? It would be: ‘the state has paid for everything for you for your whole life (or rather between the state and your parents). But yeah, educated, healthcare, all the infrastructure used, everything was paid for, and now you know ‘all we're doing is asking that you pay for some of that back’.

Samo: This political argument would just fly everywhere in Europe. Imagine you say ‘Until you were 25, you got free college education, and you benefited from the NHS. Now you want to go start a startup in California? Well, you can go start your startup but we're going to still tax you. We’re going to tax your billion dollar software company in San Francisco.’ That would be very appealing, but there'll be an interesting situation where maybe countries will be struggling to hold on to people. And you know what there are countries that already are trying, sort of, to keep their people and are sort of unhappy. I mean South Africa's been dealing with brain drain, but even much more developed countries like the United Kingdom have been dealing with brain drain for a long time.

Adam: New Zealand has an open borders agreement essentially with Australia which benefits from a whole range of economies of scale, and has cities of five and a half million people with developed tech sectors and so on, and an enormous number of particularly the most highly educated Kiwis end up there. So a degree of ambition in the face of all this. I don't get the sense that you're a particularly pessimistic man, and I'm not either. But I think we're in agreement that like a high degree of ambition here is important in terms of, let's say, institution building.

Ways Forward

So let's imagine a live player is in power, and they've been listening to Samo and myself and are fully seized of the necessity. How do you go about building? What does a live player do to build a well functioning institution? What are the factors that determine success? Do different scales require fundamentally different approaches? I know that like every individual institution is ultimately, whether it's a company or a public institution, is going to have its own unique set of challenges, but this is obviously an area that you've studied quite a lot. What what are the high-level themes here in terms of how we should approach this in order to be successful?

Samo: Most Western countries have a variety of different challenges and limitations. There are some common issues and common problems. I do think that our conception of climate resilience and policy just has to change on an international level. There is a real element to, let's call it first-world group-think, where all the bureaucrats in all the developed countries are copying each other's homework, and they take cues from whichever country that's in this category of special nice countries that we all like, that are democratic and have human rights, or so we say. The first one to regulate something often is sets the tone for it.

Now you can overplay this hand, which is why people around the world make fun of the European Union's claim to being a regulatory superpower. That's an interesting strategy. I feel it's a diminishing strategy. Some of it is basically a unitary challenge, right? Western countries are independent. They still have a lot of state capacity compared to the developing world. They just don't have that much state capacity compared to a place like China let's say, or honestly even a place like South Korea, which is also a democratic and developed country. It is the case though that some of these challenges are almost at a pan-Western level. These countries are interdependent, but intellectually clearly they're very connected. Socially, they're connected. Economically, they're connected.

Adam: The English speaking world, certainly.

Samo: The English speaking world for sure, but then, we can almost count the European Union as part of the English speaking world. What's the language of Brussels? It's not French anymore. It's not French. It's this Euro-English type thing. Still though I guess what I'm saying is that if you pick a particular country, I think one of the first principles should be that essentially, at the very least…

Here's an interesting anecdote. In Estonia after the collapse of the Soviet Union, they, for completely understandable reasons upon independence, basically blocked ex-Soviet officials from being civil servants. It happens to be that Estonia basically had the Soviet MIT of computer science. So they hired a bunch of people into government who are highly educated people in this tiny country of one and a half million people. They happen to be computer scientists, and they see an information problem. How do we staff our civil service bureaucracy? They solve it with computers, right? Estonia is well known for—you as a citizen can do anything with your government except I think get married or be confirmed dead over your computer. You have a digital identity. There's a single platform. It's beautifully scalable. The servers to do this for Estonia for one and a half million people probably basically cost the same as running it for 150 million people…

Adam: These problems are essentially solved, of scale and so on, as long as its done correctly.

Samo: It’s [only] solved as a computer science problem. So computerization and use of the information technology and some ambition in building up a completely new country that just came into being in the 1990s. Yes it has a longer history, but I'm talking about the state machinery was rebuilt so that it was not Soviet, so that they would retain and build their independence from Russia. Otherwise, if Soviet officials ran it, it might be a country that looks more like Belarus, right? So it was actually a very reasonable policy. Well, the result of this was that Estonia per capita has very few civil servants, right? For a while this was promoted this suite of software solutions was promoted across Europe, and in fact it does improve convenience for citizens. However, when Belgium adopted basically the same computerized systems…

Adam: Do Belgium the country, or Belgium…?

Samo: Not Brussels, Belgium. Belgium the real country has some things to do with the EU but it's not a synonym for the EU. Belgium has about twice as many civil servants per capita as Estonia. So it turns out that if you fire the civil servants and then build software, you get a better result than if you have the civil servants, then buy the software.

Adam: Sorry, can you just unpack that a little bit?

Samo: Of course. What I'm saying, to be quite direct, is that because of its political circumstances, Estonia ended up with having half of the number of people per capita providing its citizens basically the same services as Belgium does, merely by adopting information technology. By the way, we're not even talking about AI. We're just talking about ‘needing to give the government information over the internet once, and the government when it knows something not bugging you with filling out more paperwork’, or being able to do things like e- residency in your country, which Estonia has. They have something like 250,000 e-residents which don't need to physically reside in the country to register a company there, etc. If you just take these [measures] and adopt them, you will in fact fail to replace any unnecessary jobs. Work will expand to meet the available jobs positions you have.

So I think what should be done is an increase in efficiency where you identify which systems were made in the ‘paper-pushing’ era rather than the era of email or ideally… we're honestly at a point where we shouldn't even be talking about the email era, but it shows how obsolete and slow government processes are right? You identify which systems could be automated by software, or AI, first and then you pick your political battle, you actually you develop the system totally independently from the relevant bureaucracy, and then you fire people until it breaks. When it breaks, you stop firing people, you use the software to replace them. Bam. You have probably achieved Estonia-like results.

[While Estonian public service headcount may be lower, public spending as a percentage of GDP is roughly equivalent to the UK. However, the GDP comparison is perhaps unreasonable given that Estonia, for understandable reasons, spends a whopping 5.3% of GDP on defence—more than double most of Western Europe. Finally, it might be noted that Estonian service levels are not consistently high. Within healthcare for example, despite impressive digital results here too, other elements of the system are struggling, with significantly longer surgery wait times than most peer nations, though largely for unrelated reasons like skill shortages].

By the way it's not firing people for its own sake. Those people will end up doing economically useful and productive things. There will be no reduction in services for citizens, because what I'm positing here is, again, that it's quite reasonable to expect that computerization should make bureaucrats more productive, but they should not increase the need for a bureaucracy. So really we should be seeing the most efficient governments to have ever existed right now. In the private sector, if a company did not adopt computers to be more efficient, it would go out of business. In politics, the only way to achieve that is to expend political capital. I understand that it's a costly business. But, if you said something like ‘we're going to we're going to keep all the good parts of the NHS, but we're going to empower nurses to do more jobs that doctors previously could do’, or to say, ‘Oh, we're going to use AI to help diagnose more citizens so more of them get their life-saving medicine sooner.’ That would be a quite popular message, at least if it’s done well.

Adam: Change can be structured in such a way that, the people that matter—the public—actually see the benefits of it prior to an election and can judge and thus either vote for it or against it.

The Failure of DOGE

Just coming back to a phrase you mentioned earlier of ‘firing people until it breaks’. So I'd like to talk about DOGE, because this was a prominent sentiment of that crowd. I'm not sure how aware people would be of this, but it was like coming through Twitter feeds and so on from from people involved.

Samo: They certainly shared the sentiment.

Adam: Right, so what went wrong there? How does that not fit into your model?

Samo: They don't have the authority to fire that many people. There's like the plum book, which is this like book of positions in the US government that the administration can actually fire. It's not that many. Congress controls a lot of these budgets and so on, not the president. So you can't really do that without certain amount of constitutional overreach. I think in the US it's… we talk about, in the British world, parliamentary supremacy. You know, no one quite calls it that, but in practice, in the US, you have Congressional supremacy. Congress just is on paper in charge of so many things. It's incredible. It's just that the incentives to do anything as a congressman are very close to nothing.

Adam: Everybody loves their own congressman because they're looking out for their own patch, constituency, but approval ratings for Congress as a whole…

Samo: That's how you get the so-called big beautiful bill, which has been named less flattering things, but it's basically this giant pork barrel. Look I think in most of the Anglosphere the Prime Minister and their party, if they have a coalition let alone an outright majority, they could actually enact a lot of these reforms. [Though] they don't choose to. In the US though, it's basically a whole separate election-cycle. How you get the president, and how you get Congress, this is the midterms. That's very different than say how we get the Prime Minister of the United Kingdom.

Adam: Right so just coming back to DOGE?

Samo: So I guess what I'm saying is: if I was President, be it a Democrat or a Republican in the United States, I think my first task would be to beat up my own party and remake it in my image. Recreate a proper governing political party that could coordinate the President and Congress. It’s not even clear that the president would be the senior partner in that, right? An ambitious speaker could actually do something like this. Then once you had that coordination, then you could do the kind of comprehensive reforms that Eisenhower was capable of doing in the 1950s, or that FDR could do in the '40s. Even in more recent history you had things like Reagan or Bill Clinton. Look if you had the amount of coordination between the president and their own party in Congress as you had under either Clinton, or the first Bush, right? George H. W. Bush, that would be almost unprecedented functionality in the modern American system. That is beyond what was achieved by Obama. That is beyond what was achieved by Trump.

Adam: Right. So in this model, DOGE is just running [an impossible task]. Say this is the context. Nevertheless, there high promises are made or expectations of this institution are held within the cabinet, by Trump and etc.

Samo: I could restate this. I could restate what I said. I could say that for Elon to be able to deliver on the cost cutting promises, for him to be able to deliver: congressmen, Republican congressmen would have to take massive political hits, in their own constituency. What is their incentive to do that? Well, if Donald Trump personally singles them out, and campaigns basically against them, which is a pretty big move, but what else could possibly be offered to those congressmen?

Adam: Right. So instead…

Samo: One person's political benefit, another person's political cost, this doesn't make sense as a political trade does it?

Adam: So instead you end up in a situation where the institution DOGE rapidly runs out of political capital, basically?

Samo: Or at least it does what it could do. They did succeed at some cuts. You know, here I'm saying this is a neutral observation. Neither for it nor against it. It also did take some political scalps. It took out some political enemies which, let's be realistic, in the Western World while we want to act for the spirit of the public, and public benefit, the truth of it is that also there are such things as political and ideological fights. I'm not even convinced… some of the stuff that was cut in USAID was just culture war right? They cut programs they didn't like internationally because they promoted a different ideological message than was their priority. You could call that a saving, but you could also just call it a change in a cultural policy. And we could stop pretending that the US does not have a global cultural policy. It obviously does.

So I think it achieved some efficiencies. It achieved some policy changes that were useful to this administration for their purposes. But yeah, it rapidly ran out of things to do would be my claim. It didn't even run out of legitimacy. The PR was going strong. It's just, you know…

Adam: So USAID, because it was international and the money was going internationally, you didn't have the same dynamic with Congress that you're talking about.

Samo: It was much easier, yeah. People have joked about the US president being the dictator of the world and the mayor of America, right? Like way more power internationally because who is taking the political cost for something like a tariff Well, it's that president over there…

Adam: Or a drone strike or something.

Samo: Exactly. And if your country is relatively small and weak, you can't really impose externalities on other countries. If you're strong like Russia, or the United States, you can. By the way, Russia's a very weak country globally today compared to where it was. But it fits the bare minimum of how strong you have to be to be able to basically throw your weight around and to export your problems abroad. I can't think of a weaker country that can succeed at that. I don't think France can impose its problems onto the EU quite as easily. It's pretty much held in check by at least Germany, probably ever more so Poland. Okay, maybe Germany is strong enough to impose its economic mismanagement on the rest of the EU.

Adam: Turkey perhaps, wouldn’t that be…

Samo: Turkey is pretty strong actually. It's become stronger.

Adam: Not as strong as Russia, though surely?

Samo: Nowhere near as strong as Russia, but again, they do make cheap drones, right? They have cheap electricity, and they've intervened militarily in Syria.

Adam: It has relatively weaker neighbors.

Samo: Of course, it's always a relative perspective where you are in the world, right? But globally speaking… look in the United States changes to international policy are, in a way, easier than changes to domestic policy. The two are interrelated in interesting ways. Say if I wanted to discipline American companies that are manufacturing civilian nuclear reactors, I would just go out of my way to facilitate South Korean companies building nuclear reactors in the United States. Because South Korea is…

Industrial Policy

Adam: When you say discipline, do you mean because the CEOs pissed you off pissed or something something of this nature?

Samo: Just fiscal discipline? Just like, ‘oh, you think you're the only one who can build my nuclear reactors because you're building my nuclear arsenal?’ No, actually, we have this other allied nation that we want to support anyway, and that we want to have in our orbit anyway. Their president will be so happy with us if we take 15 Korean contracts to have them build nuclear reactors here. What I can promise America's voters, what I can promise America's citizens, is that they will have a cheap electricity and I have saved so much money.

Adam: Right. This wouldn't even need to be done in any necessarily targeted fashion. You could just open up the contract internationally…

Samo: Or at least say that you will as a negotiating position, and you know what? Maybe Bechtel will suddenly discover they can do the same job for 20% less. Maybe they'll come back to you and say ‘please, please don't take this contract from us, we can do it for 20% less’.

Adam: It's not to say that their profit margin would currently [cover that lower margin], but it's just like they're under no pressure at all. Employees aren't under any pressure. Management is not under any pressure. Their subcontractor costs [aren’t either].

Samo: As politically powerful as Bechtel or Boeing is, let's be real, they're less politically powerful than actual civil servants. That's why I think first putting pressure on the government contractors is the right way. Now, I said that we can incubate almost a startup-like approach to make more SpaceXs. I think the Western World has room for something like seven or eight giant companies of the size of SpaceX covering different key aspects of, let's call it, ‘big infrastructure’ or ‘big government need’. Aerospace, navy, energy, etc., all of that. We can we can nurture those, but it takes 10 to 20 years to get there, to nurture such a company into being. That might exceed your political mandate, but shopping around among your country's closest allies to find an ally that is incentivised to do a good job, like South Korea, that's not a bad idea actually. That can immediately help you somewhat discipline the companies. Ideally, of course, if you're the US, you want the US to build the world's best nuclear reactors, not South Korea, right? But South Korea sure is better than China.

Adam: But I mean it's just you there's there's arguments there. Clearly if they can do it that much better that they can move to a foreign country and build something..

Samo: Well, they built the UAE reactor. They’ve built many reactors around the world.

Adam: Well there's an element of needing to to do the Singaporean playbook almost, right, and like, ‘okay these guys… our expertise in this area is clearly atrophied. We have to learn from them. If we want to build it back up, let's get a head start by [bringing in the best].’

Samo: Of course. Maybe the right move is that you want to build a chip fab. You ask Samsung or TSMC. You don't have them partner with Intel. You just have a nice big contract among new startups in the semiconductor space in the United States. You pick one of those that is the best one, or you pick a few of them, and partner on three, four or five projects. You run the contract. Samsung or TSMC get their money for the chip fab. And instead of you propping up Intel, you've created a new semiconductor giant in the US. A small one, but a real one, because that company that was partnering with them will have seen technology transfer. Then you simply allow them to compete in the next round five or six years from now on the next round of government contracts to build its data centers or whatever, chip fabs and so on.

Adam: All right. We've talked around the strong do what they will and the weak suffer what they must to some extent here. Obviously I'm this is… I'm from New Zealand. What's your advice, given everything that we've talked about here, changing geopolitics… what should we be really focused on?

Samo: I think New Zealand is. First off it's you know a beautiful beautiful country. You guys can probably export a few more movies about hobbits. However, that's much more similar to Portugal exporting wine than it is to Britain exporting textiles. I think that New Zealand is, what I know of the country is that your costs are relatively high for all sorts of goods, partially because you're pretty isolated. Probably like the most out of way developed nation in the world.

Adam: Yeah, fair unquestionably. Yes.

Samo: I mean much more so than say Iceland, for example. Okay. I think in that case it perhaps pays to invest into things where that physical distance doesn't really matter. Would it be so beyond the scope of New Zealand to invest into cheap, abundant, green energy, and then have the Prime Minister of New Zealand go around talking to big data companies said, "Hey, do you want a country that's in the Five Eyes that can host your data center, and provide you with cheap electricity?”

Adam: Yeah, and we'll derisk it all the way through for you, planning and so on.

[Samo arrived at this conclusion independently, but at the time of filming I’d already started on writing this piece at the insistence of Gwern]:

Samo: From the perspective of a big training run, it really doesn't matter if a server is in New Zealand, or in Canada, or in Argentina, or in Siberia. Information moves at the speed of light. There's a tiny bit of a lag issue for…

Adam: Serving video, or something.

Samo: Yeah, but you know what, there's enough people in Indonesia, and Indonesia is close enough, etc. Also, the training runs really do matter, and the training runs can be distributed anywhere. So there’s massive demand for data centers, and data centers could just go anywhere there's cheap electricity. There's already talk of geothermal being a possibility for this kind of development in Iceland. If I remember correctly Iceland does some aluminum smelting which is also an energy sensitive industry.

Adam: Sounds very familiar. With New Zealand: enormous amount of geothermal potential and I think something like 15% of our generation goes to aluminium smelting.

Samo: So maybe a new New Zealand ambition would be something like ‘producing five times the total electricity that you produce right now using groundbreaking new geothermal’. Why could not the global leader in geothermal energy be a New Zealand company, for example?

Adam: I think it is already. We’ve got the expertise, it’s just [capital and demand].

Samo: So maybe scaling geothermal. That fits the signature of being a green, pleasant, environmentally friendly country because it's zero carbon energy. That's so important. Not only is it zero-carbon energy, it's not intermittent like solar is. Solar is, of course, a great success, but energy storage remains a great challenge. Because of that, it's just not yet appropriate for sustaining industry. You get various problems like Spain's energy grid experienced a few weeks ago, right? The whole country had a blackout uh for a week. It started for a variety of reasons, but one of the risk factors was certainly the high reliance on intermittent energy sources. So anyway, geothermal in principle makes a lot of sense. I would propose that you guys should not avoid nuclear, but I know that's probably extremely unpopular. I think that, you know, nuclear energy is a good investment. It's something to buy.

Adam: Times are changing, right? Like in Europe, for example.

Samo: Oh, in Europe for sure. In Eastern Europe especially, they're eager to build nuclear power, but also nuclear power is a good investment. It's a relatively expensive source of energy. I think it's a good idea to buy nuclear power plants when you're rich, assuming that you might not be able to build them in the future, right?

Then, further things that could be done. This is going to anger maybe environmentalists, but have you considered looking at undersea mining for example? You guys are on an interesting geological part of the world, basically on a sunken continent if I remember my geology right.

Adam: Essentially.

Samo: Yeah. There's like some relatively shallow seas. I don't know. I know that there's rare earth metals in the Sea of Japan that Japan is planning on exploiting. Maybe there's stuff New Zealand could do.

Adam: I think whatever it is that we end up doing, it certainly needs to be looking closely at what our comparative advantage is and doing something that's really quite niche. That's always been where we've been successful. We make something like 90% of the world's goalie hockey goalie equipment is made in New Zealand. There's these industries, very niche industries that we do very well with.

[There’s an excellent lecture by the late Sir Paul Callaghan on this subject that I cannot recommend highly enough for patriotic Kiwis. New Zealand is obviously well known for it’s farming, but one of our biggest export areas (even bigger than dairy) is hyper niche manufacturing like sleep apnea and airport baggage-handing devices, and crystal ossillators for GPS chips. ‘If you ask ‘what the hell is that’, it may stand a chance here’]

Samo: Maybe you guys should try marketing some particular kind of liquor after all, right? I don't know. Luxury goods, I mean look, Dubai suddenly decided they can make chocolates out of nowhere, and now I keep hearing about them. These marketing ideas…

Adam: Hobbit-branded chocolates, yeah. Some food for thought there anyway. Look, again, thanks very much for coming on Samo Burja.

Samo: Pleasure to be here.

1. See here for a non-paywalled summary of the Danish immigration statistics.