Steve Hsu - In No.10 with Boris and Cummings

Steve Hsu completed his PhD in theoretical physics at age 24. He’s since made numerous contributions as a scientist, tech startup founder, professor, and institutional advisor. Little is known about the role he played during Boris Johnson’s ill-fated tenure as Prime Minister of the UK. In Steve’s first interview on the subject, he shares that hidden history, and what lessons can be drawn from it. Later, we discuss the rapidly changing power dynamics of the Western Pacific, as China continues to flex its growing might.

This episode is also available on Spotify and Apple Podcasts.

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

This is the fourth episode from my ‘Emergent Ventures’ visit to San Francisco. I’ve two more talks in the series, which I’m very excited to release over the coming weeks. 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
(02:51) Meeting Dominic Cummings
(10:02) Vote Leave's Strategy
(15:23) Boris and Dom
(26:34) The COVID Response
(35:58) Halcyon Days
(1:01:40) Broken Governance
(1:14:48) China Rising
(1:34:02) Emerging Geopolitical Realities
(1:49:25) Operation Pompous Golem

Transcript

Adam: Today I'm speaking with the polymath Steve Hsu. Steve completed his PhD in theoretical physics at age 24, and over 30 a career in academia made numerous scientific contributions across Particle Physics, Cosmology, and Genomics. He's co-founded a range of technology startups covering areas like privacy, malware protection, genetic illness protection, and enterprise AI. He's acted as a scientific advisor to various institutions across the US, Mainland China and was in No.10 during the Boris Johnson ministry.

So on that last note, Steve, while you've been a prominent public figure, you could say, for decades - the internet generally, and even the AIs ‘Deep Research’, are mostly unaware of your involvement with the UK government while Dominic Cummings was the Chief of Staff at No.10. Can you give us a brief overview of the role that you played there?

[‘DeepResearch’ is the most advanced AI query currently available (in this case I used the ChatGPT version, but Google Gemini and others also provide alternatives). It will scour the internet over the course of 10-15min, and provide a comprehensive overview of publicly available information for the requested topic. Highly recommended.]

Steve: Deep Research did not go deep enough. Somehow they missed some… I think had they read everything Dom has written on this subject, or maybe said…

Adam: Well, it's paywalled, a lot of what he’s said.

Steve: Ah, that's why, yeah.

Adam: Yeah, so it knows that you were proposed to be the Chief of the new DARPA-like organisation, which we can talk about if you're interested later, and it's also seen a photo of you in front of No.10. Then there's the odd news article that has a passing mention.

Steve: Eugenicists in No.10, that's the Guardian coverage.

[Steve is a co-founder of ‘Genomic Prediction’, a company which allows people undergoing IVF to screen their embryos for genetic diseases. Screening for Downs is already widespread and accepted across the West, and efforts to extend this capability to other genetic diseases as well are as close to an unalloyed good as we see these days. I consider rhetoric equating them philosophically to mass-murder and Nazism to be the height of intellectual laziness, or, more commonly, the bad-faith smearing of political enemies.]

Meeting Dominic Cummings

Adam: I’m interested in talking about the media later, but as I said the popular conception, and what's available on the internet, is that you visited [No.10], and you're friends with Dominic Cummings.

Steve: I'm happy to fill in the details. So I'll go all the way to the beginning since this is a long form discussion and so people might be interested in how these things go.

Dom and I have known each other on the internet since, I would say, circa 2005. What happened was that in the lead up to the financial crisis of 2008, you know, the housing bubble burst and collateralised mortgage securities almost destroying the whole world economy.. I had, on my blog, become very interested in this topic because many of my friends, former physicists mostly and mathematicians, work on Wall Street and were the quants building these mortgage-backed securities. So I had a front row seat to this, and many of us were worried that this was going to blow up. So I blogged about it. If you look at my old blog ‘Information Processing’ from that era - many posts on how collateralised debt obligations work, how derivatives work, what's the notional amount at risk, what's going to happen if the housing bubble starts to pop in the United States. So I had written at length about this, and Dom had read it. At the time he was the Spad, which is the British lingo for ‘special advisor’ [i.e. a political appointee to a government minister, not a permanent civil servant], to Michael Gove, who was at the time the Minister for Education. Dom was involved in a lot of actually very interesting reforms around education. Another area that both he and I are interested in is gifted education. He's actually done amazing positive things there, including helping to create a Kolmogorov-like school in London for educating very deeply gifted kids, which is still running.

Dom and I got in contact on the internet, and then we were both invited to a meeting at Google. There used to be an annual meeting there called SciFoo. It was mostly scientists or people in policy related to science. World-class scientists invited [i.e. the event is invite-only] from all over the world. It was a crazy event. It was a little bit like the event you and I were just at, called Manifest, but this ranged all over the Googleplex. You can imagine..

Adam: Just enormous…?

Steve: Probably similar number of people to Manifest, maybe a little bit…

Adam: So 700 odd?

Steve: Yes. Actually, maybe sometimes SciFoo was smaller than that, but anyway but all leading scientists and startup founders, and also like Sergey and Larry [Google’s co-founders] could show up. So that's when I first met Dom in person, and I remember very long conversations with him that went like long into the night. So after that we were kind of IRL friends. First time I visited him in London…

Adam: Just pausing there for a moment. As I understand, that conference, I seem to recall there being some quite interesting innovations happening at that time for governance and more generally. I think this is where the idea of ‘Seeing Rooms’ possibly came about? This idea you wanted your decision makers in a room where there's a technician available, and there's lots of screens, lots of access to information and so on that can be pulled, and we can have an informed discussion. Kind of like what you see with the Joe Rogan Podcast, but better.

[Frankly, it’s a sign of how fundamentally unserious we are about good governance that, every day, Western heads of government are making enormous decisions in circumstances less conducive to information processing than Joe Rogan]

Steve: Yeah, so at every SciFoo meeting there would be a central area, and people who are familiar with the Googleplex would know exactly where this is. It's where they they had the closest thing to an all-hands kind of meeting that they could have at Google. It's not actually big enough, even at the beginning when the Googleplex was first built, to hold the whole company, but it's a big region. They had a demo thing where a lot of the tech founders, and scientists, researchers from Google itself would set things up so that was the first time I ever used VR goggles. I think Dom saw a lot of cool shit there. When he later, well, he was already aware of what No.10 looked like on the inside, realized how big a gap there was between what was possible and what actually was available to the Prime Minister. I think that stuck in his head.

Adam: Sure so I'm keen to get back to the story, but just for context there, we're talking about 21st century tech versus really like 19th century… the rooms that William Pitt would have used basically being unchanged since. We're talking about wood panels, no screens, no whiteboards.

Steve: Yes, and a rat problem too in No.10. Just so you get the full sense of it.

Adam: (grimaces) Sure, so anyway, back to your story.

Steve: So we got to know each other quite well, and we resonate because we're both kind of like, I don't know what, aspie [Aspergers], weirdo, out of the box thinkers or something right? So a very decisive event, which I think has never been really documented anywhere… there were some people who had a background in quantum physics who were originally from the UK. They had studied… I don't want to completely doxx [i.e. make their identity public] these people, but they had done, for example, postdocs at top U.S. universities and they read my blog. They also had read, because I posted things about Dom. Maybe Dom had started blogging, or had written his first manifesto, or something. So people like this contacted me and said.. it's kind of ironic because I'm not the British guy right. It's a British guy who's a scientist contacting me saying, “Hey I'm returning to the UK now from the US and I would really love to meet Dominic Cummings because I really admire his thinking and writing”. So I introduced these people to Dom and they're going to play a very decisive role in the winning of elections, developing new polling methodology (for Brexit?). Literally Brexit. These people are in the film. What was it called the something war, The Uncivil War?

Adam: With Benedict Cumberbatch?

Steve: Yeah that's a great movie, but some of these people are in the movie as sort of composite characters. They actually met Dom through me. I introduced them to Dom, which is very ironic because they were British people but the connection was through me. Very weird.

Adam: That's how the network graph works, right?

Steve: It's how the internet, yeah, social network graph works, but we're all very close friends now after all this.

Vote Leave’s Strategy

So I go over the first time [to visit Dom], and we had a very intimate dinner—he and I and Michael Gove—we talked about the future of British politics where things were going etc. and this is pre-Brexit pre-everything, yeah but everything was about to happen.

Adam: Maybe 2012?

Steve: Something like that. Maybe a little bit later, I could look it all up, but anyway. There's so much to talk about, but for example the polling, all the Cambridge Analytica stuff. Are you familiar with Cambridge Analytica?

Adam: Sure.

Steve: Okay, so all the things that went into the Brexit campaign, Vote Leave [etc.], that are documented in that movie. Various writers at the Guardian including someone called Carole Cadwalladr - their brains were literally broken by Vote Leave and Brexit because the right-thinking elites in London were sure that the proposal would be defeated, but in fact it wasn't. Dom's team, which consisted of these physics guys who literally started a project… the guys in particular that I’m talking about had left academic physics, and had moved back to London to work in AI, actually [DeepMind was one of the world’s top AI companies, and was headquartered in London until after its acquisition by Google in 2014], but were now in touch with Dom. They formed the analytical backbone for the Vote Leave campaigned. They also did an amazing thing, which I think people would find extremely interesting (and Dom has written a little bit about this). They wanted to know, ‘how do you measure public sentiment, and how do you do polling?’

Of course, you might say: “What are you talking about? What some physicists want to know about measuring public sentiment and polling? Surely there are like many social scientists and professional pollsters, and political groups pay millions of dollars for this. Why don't they just go to an established group?” However, as physicists, the idea is like, ‘we're gonna do everything from first principles and, very likely, a big chunk of what those guys think is true is bullshit’.

Adam: Right, just groupthink.

Steve: Yeah, so these guys went through the whole literature on polling and elections and stuff and pulled out from that literature the things that they thought were high-confidence true, and they built their own polling function, they built their own ad-targeting platform and capability. Although the Cambridge Analytica story is totally false… So there was an organisation called Cambridge Analytica. It was claimed played a big role in what Vote Leave did. I know the real people who did the stuff for Dom and they did not use Cambridge Analytica. They actually were quite sceptical of Cambridge Analytica. Although they did develop new techniques which, in the later election, when Boris was in office, showed that they had exquisite predictive power for what was going to happen. In that campaign, Vote Leave, they only used very basic ad-targeting on Facebook, but they knew how to do it right.

Adam: Sure. Even to the extent that Cambridge Analytica [were involved]… there’s some extent to which, what that company was doing, even if the stuff you guys were doing was novel at the time (Steve: not me personally).. [and hence generated a lot of media attention and conspiracy theories] it’s just sort of accepted practice today.

Steve: You mean the use of Facebook?

Adam: Right, the use of Facebook, (Steve: Obama did it a little bit, yes), being more deliberate about the way you target particular demographics. This is like, 10 years on, maybe not with quite the same analytical depth, but it is essentially just modern campaigning.

Steve: Right. The specific techniques of how to place an ad, or whatever, has now become pretty standard. I think where they still have some advantage is in taking the raw data in polling, and then extracting high-confidence interval ranges of what's likely to happen.

Adam: What the key issues are, that maybe everybody else is sleeping on [usually because of group-think leading to a lack of curiosity].

Steve: Right. Their use of focus groups for example, they’re still way beyond the typical guy [i.e. company] that you hire.

I don't know if this is what you want to talk about, but this is a framing of my relationship with Dom. So another anecdote, maybe you're not that interested in these personal anecdotes, if so I can I don't have to tell them but I think some people would find it fascinating. (Adam - sure). So at one point I was in London for some meetings. There's some prominent people in London who are investors in some of my companies, so I was there for that purpose. I had an Airbnb in Camden where there's a canal. I don't know how familiar you are with London, but Londoners will know what I'm talking about. I'm hanging out in this Airbnb, it's really nice because it overlooks a canal, it's a beautiful Sunday afternoon, and Dom says, “Hey, can I stop by and chat? I'm on my way to meet with Boris.” This is before Boris became the Prime Minister.

Boris and Dom

Adam: Possibly even just before he announced his candidacy for the Conservative Party Leadership?

Steve: Could be, yeah. The details are fuzzy, but I remember that Dom came to my Airbnb with one sheet of paper. On the sheet of paper where the demands, which he jokingly called the terrorist demands, that he was gonna make to Boris, saying that if you want me to be your special adviser, these are the things that we have to do. On that sheet were things like: we are going to create an IARPA-like thing to advance British science, we are going to reform high-skill immigration in the UK, which I should say, before people jump down my throat, it was reformed in a way which is not satisfactory, but that's because Dom and Boris got kicked out. It's not the original plan was probably would have pleased the people who are currently very critical of what eventually happened.

Anyway, there were a few issues that he was very focused on, and I think on the insider political thing he demanded full power to hire and fire all the spads, or something like this. So that was an historic moment.

Adam: Sorry, just for context, there's something like a few dozen Spads in No.10?

Steve: At least.

Adam: Right, but these are these are political appointees, as opposed to the actual, you know, the civil service bureaucracy itself.

Steve: Correct, but they're the ones who do the actual work. As you know, Dom's theory of how governance works, at least in the UK, is that the politicians themselves only care about being on camera and building their brand. They're really pretty disinterested in the issues.

Adam: Sure yeah. So I'm keen to get into that, but let's finish this, the story. So Dom's on his way to meet with Boris, he's meeting with you first.

Steve: Yeah, and we're having some beers and discussing all this stuff. In my mind, it's one of these things, it's a historical moment, that meeting between him and Boris, and what he actually cared about, because he said, ‘I'm going to do this job, I'm going to become possibly the most hated public figure in the UK for a while. Why would I do it?’ It's not just raw power, or self-aggrandisement, or narcissism. He's none of those things. He had specific policy goals, and of course finalising Brexit was one of the top ones as well. He had specific things he wanted to accomplish. I'm defending Dom against critics who think he's just some narcissist or egomaniac. He had very specific things he wanted to accomplish, and those were the important things to him.

Adam: Okay, so just kind of fast forwarding a little bit now. Boris becomes leader of the Conservative Party and Prime Minister. Dom is ensconced at No.10 as his Chief of Staff. There's these various kind of policy goals which, you know, there's going to be a range of listeners that kind of agree and disagree with some of these goals, but the specifics of these individual policies aside, what I'd really love to get into is just specifically the bureaucracy itself, and the nature of what happened very soon after. There's something like a six-month window between, you know, or less than a year between when that team comes into power… There's this quite famous memo in January that comes out about hiring weirdos and misfits. Hire a bunch of Silicon Valley type people, superforecasters, and so forth. And then two months later, the biggest crisis for Britain, possibly since, like, Suez or something like that, hits. It [Covid] is completely unexpected, and you're there.

Steve: First of all, I realised that I left something out, which is that the whole thing about calling the U.S. Housing bubble and forecasting that, were it to start popping, there would be a kind of global financial crisis. It's something Dom took from my writings. I think, maybe, he personally didn't see it anywhere else on the Internet. He took those things to Whitehall, because at that time he was Michael Gove's Spad. So he was right in the centre of government in London. And he, you know, he just couldn't get any response. People would read it and say, like, I don't understand what this is all about. I don't have a background in finance. And then the crisis actually happened. So he had a memory of, you know, if you scour carefully, you can find some smart people who actually have an idea of what the impending disaster is, and are accurate. But, the problem is that Whitehall, as an information processing entity, is just not very capable of dealing with stuff.

Adam: I recall a similar kind of thing in New Zealand in the lead up to our own COVID response where even after… I think the markets were actually quite late to COVID in some respects. It wasn't, I think, until late February that you saw that big crash, after ‘Italy's been locked down’, after ‘there are cases virtually all over the world’, that anybody kind of looking at this can see that something is going to, like, the dam's about to break. But even after the markets crashed in February, I was still laughed at when I said, ‘Hey, we should maybe start preparing for COVID here’. I would say in general, my impression is, and part of what I'm interested in talking more with you about, is that the U.K. Government was significantly less prepared than that (Steve: yeah). New Zealand, you know, maybe was a little later than we could have been, but once we spun up you know, the nation got in behind Ardern's and the government's work. We get to zero COVID, and we reopened essentially long before the rest of the world. We had structural advantages in that regard, but meanwhile, the U.K. records the highest excess mortality in Europe during the first wave. Real GDP that year fell 10 percent, which is like a staggering blow really, and the highest fall in the G7. Yet at the same time, it's also the first nation to approve and deploy a vetted proper vaccine that same year. So it's quite an interesting contrast there, where perhaps things aren't [consistent across the government]… Yeah. So I'm just interested. Can you unpack for us? The system's clearly unprepared at first. Can you walk us through how this how this goes?

Steve: So I know you want to get into the fullness of COVID.

Adam: Sure. More background (laughs).

Steve: I just want to say one more thing, because I think I've never commented on this publicly, but it is interesting. So it's not related to COVID. But if you remember, in order to finalise Brexit, there were all these shenanigans they were fighting. The Boris government was fighting. They were against the media, within the Conservative Party, and Labour, there was a huge battle going on. Dom was really the chief strategist of this battle.

Adam: Sure. At least on Boris's side, because there's Theresa May… but she was floundering, unable to kind of draw these things together.

Steve: So the thing I want to say is that, of course, I was watching it in the media. So I would read what The Guardian was writing or The Times of London, or what BBC was saying. But then at the same time, I knew from Dom what was actually going on. If you remember, they sent Parliament home. They did all this… I don't know if you followed this?

Adam: You mean in the lead-up to the election in 2019? [as opposed to during Mar-Apr 2020 due to COVID]

Steve: Yes. So there was a sort of like, it's very interesting, because of their system. The Prime Minister can basically say, ‘We're having a pitched battle on this day. We're going to call a snap election’, and then it could be decisive how things turn out. So it really is like a military strategy. It's like ‘We're going to do this, but before we do this, we're going to poke the other side into a frenzy, so they won't function properly, and then we're going to smash them in this election’.

And I've never you know what, if you have inside connection… I've never been inside The White House, you know, like part of the political team in the White House or something like this. In the US, things are usually not… you can't call a snap election, so it's not quite the same thing. But what was interesting is the divergence between what experts in the media, or even academic experts were saying about what is happening in the UK in the lead-up to this snap election was 100 percent wrong. I had it from the horse's mouth because Dom would say, ‘We're doing this, and we're doing this, and also my geeks, my team, my nerds are forecasting this’. Of course, all the other polls in the opposite direction.

Then they won this resounding victory, and they had actually forecast to very high accuracy how many seats they were going to win. (Adam - it was quite a large majority). Yeah, I mean, it was a historic event, and that I think the full story of that has never been told as far as I know, because Dom was doing stuff where like the public interpretation, again by experts, you know, political experts, BBC commentators..

Adam: They were commenting on essentially unprecedented actions that they had no context to understand the implications of.

Steve: Exactly. He had the balls and the out-of-the-box thinking to do stuff, which, if you look in the historical record of how those actions were interpreted at the time, the other side just got it completely wrong. And Dom was just like a chess master, destroying them. And of course, the proof is in the pudding in the election. His team, this tiny team, basically got the polling right. They basically had a very accurate idea of what was going to happen. And the others, the official polling and all the other people were just totally off. So anyway, it was a resounding victory for Dom. Okay, so now COVID comes. I was not in No. 10 [at the time], although I was communicating with both Dom and his team about this.

The COVID Response

There was a period of time, probably maybe you were one of these people, like I was, where I actually was pretty confident that it was coming to the US. I had studied SARS quite carefully, because I and another theoretical physicist wrote a paper on modelling viral transmission globally in the wake of SARS. So we had worked through all the math of all this stuff so I was pretty familiar with it. So when the first initial data about COVID was coming out of China, I was absolutely sure it was going to spread around the planet.

Anyway, in that lead up, there are a bunch of people looking at every bit of news. What's happening in these Northern Italian towns? What's happening in, you know, anywhere there's data? We're all trying to like, what does it say about r0? [i.e. how contagious it is] What does it say? So there's a bunch of people studying it like this, including some of the guys on Dom's team, and we're communicating. So now I can't remember the name of the there's actually a committee, a sort of science committee.

Adam: COBR [aka COBRA].

Steve: Cobra, which met regularly. Okay. There was a pre-existing pandemic plan for the UK, but that plan was not appropriate for the r0 and the lethality.

Adam: It's just a generic, ‘what happens if there's a pandemic, okay off we go’.

Steve: Yeah, but it had herd immunity as the central component of the plan. The momentum was just such that within COBRA, had Dom not interceded in a very strong way, that's where they would have gone. They would have gone to herd immunity. Now, I think you and I talked about this, so let me get into this. At the time, looking at the numbers, looking at the emergency room capacity, intensive care capacity, all of these things in the UK. Dom's team, and also me as well, we were looking at this and we said, ‘if they go, if they use a herd immunity strategy, a huge number of people are going to die in the UK, especially old people’. We did all these calculations and briefed Dom.

The decisive moment, which has been reported, I think at least in some places, is there was a moment where Dom brought a whiteboard, wheeled…. He and his, the geeks who, these are the actual geeks that I introduced Dom to many years earlier. They roll a whiteboard into Boris's office because there's no whiteboard. There are photographs, actually, of this whiteboard which you can find. And Dom is explaining like, you know, he's got a distribution, he's got a tail of the distribution, ‘These are the people who are going to die.’ And he's, you know, using a calculator to show Boris.

Adam: Because there's no tech.

Steve: There's no tech. You know, if you're watching a James Bond movie, like MI6 has some holographic thing that pops up and they're showing Bond. Nothing like this exists okay! There's literally a whiteboard, and not enough markers. You know what I mean?

Adam: Right. Right. So, well, so just coming back to the existing plan.. So obviously, Dom's team is reasoning under uncertainty here. You know, there's limited information. Nobody's quite sure how big this is going to get. But I'm just interested, just from what you understand: What was the view of the Ministry of Health at the time? Beyond simply ‘we’ve got this plan’. Were they modeling similar? Maybe they didn't have a bunch of physicists, but?

Steve: So, if you remember, there was a very prominent epidemiologist in the UK, I can't remember his name now, but he was on TV every night, on British TV. So there was a, there was a lively public discussion of all this stuff. I hope I'm not misremembering, but my recollection is that guy didn't know what he was talking about.

[Not being in the UK at the time, I’m unsure exactly who Steve is referring to here. GPT returns a number of plausible suggestions. Certainly it’s true that while many experts who became media darlings globally during the crisis performed admirably (especially given the uncertainty at the time), others were much more mercurial in parroting the ever-changing prevailing wisdom of their social class, or veered sharply away from their core area of expertise to become policy advocates]

So, one of the things that (this is getting into the nerd weeds a little bit) is if you're a physics guy, you probably have pretty good intuition for math, but also you've seen it applied to real-world situations in your laboratory. You know, you do some half-assed calculations, and tell your grad student like ‘Oh if you build it this way, this is going to happen’. Then he does what you said, and like ‘Oh no it doesn't work at all, boss’ and then it's like, ‘Oh shit’. Then you realise, ‘well okay, these equations are totally off because we made this assumption’. So you get to the point where you can see a calculation that ‘okay maybe it will work as described in that model, but there's a very good chance that model is way off’, and that was my impression of what the public experts were saying, like sort of publicly advising…

Adam: Just way too certain about their calls.

Steve: Yes, their calibrations were.. the level of conviction they had in what they were doing was way too high. For us the most simple robust thing is ‘if it spreads at this rate, and this is what happens to old people that get it, and if it gets concentrated into old people's homes, or some retirement home, it's gonna… So our calculations were very simple actually, we didn't use supercomputers, but we were very sure that something bad was going to happen. That emergency rooms or intensive care units were going to be overwhelmed and stuff like this. So that gave Dom the courage and the urgency to have this whiteboard meeting with Boris.

So he did all this stuff and Boris [responded]. Boris actually prior to that had been, remember, poo-pooing COVID. Remember he did all this stuff where he's like ‘ehhh, I don't…’, but that convinced Boris. Then the UK pivoted to this lockdown strategy. Now, here's where we talk about decision-making uncertainty. I will be the first one to say, after everything that's happened with Covid, I'm not sure we did the right thing. I'm not sure that this decision was the right thing. There's a counterfactual where, okay, maybe a bunch of old people died right away. Okay, definitely a problem, but maybe they did reach, in that alternative world, maybe they did reach herd immunity. They got the economy back to normal, they spared a lot of people these terrible lockdowns, and maybe it was a little bit more like the way Sweden handled it. So I don't know. I don't know the answer. At the time, we were very excited that we were able to get Boris to quote ‘do the right thing’, because in our minds this was the right thing. But I have enough epistemic humility, and I'm confused enough by all the crazy data that was generated globally by the Covid event, that I'm not sure. I'm actually not sure. I don't know how to think about it, but at the time there were some headlines among in the nerd world about, you know, like ‘Dom saves the UK’. Meaning that Dom corrected them from a disastrous strategy. But I'm not sure now.

Adam: Particularly, I suppose, given the state wasn't geared up for it. I've heard these crazy stories about getting some military plane sending it over to China, filling it up with like masks and stuff because… Once that decision was made, regardless of whether or not it was the correct call and compared to some counterfactual universe, Once it was made, the standard processes of the state essentially were incapable of following through on it. The procurement process for masks was going to take like what 12 weeks or something, there were big shortages across like all manner of different components, and there was need to move rapidly, utilize capabilities that were ostensibly for some other purpose, like military [logistics].

Steve: Yeah, for example, the guy he put in charge of a lot of this stuff was an S.A.S. guy. He was a super competent guy who could get stuff done.

Adam: Can you give me a picture of, beyond Covid? I mean, given your expertise in this specific matter, your post-SARS model you developed and so on, you're contributing to the covid advice, albeit remotely… you were in the US at the time?

Steve: I was in the US. I think I was in London that Fall or Summer, right before Covid happened, but I wasn't around during Covid. It wasn't that easy to travel.

Adam: Yeah. So obviously a wild time. Look forward someday to maybe interviewing someone who was on the ground during those meetings, but your involvement in the Boris Johnson Ministry, or with Dom, was broader than that right?

Halcyon Days

Steve: Yeah so I guess I was in London like just before… So when they first got into power those were halcyon days! Crazy days, because everyone in the government knew Dom had the power to appoint and fire all the spads, and as I was kind of starting to say earlier, the cabinet ministers are all fake in the UK. So what do I mean by that? This is an assertion that Dom makes. Of course, it's not in the interest of any politician in the UK to admit that this is true. Normie commentators who purport to know how government and politics works in the UK, they are not really gonna like take the bull by the horns and discuss this Dom theory, but I will discuss it. So he says, and you can read this on his Substack, he says the politicians in the UK are playing a certain game. The game is reach the highest levels of power, hang out at that high level, and then decamp to some place where they'll pay you, a lot, for your connections, and sure you know lobbying capability. It's the same game maybe all politicians are playing, you'll say?

Adam: (laughs) I was gonna say it's a very cynical take, but I mean perhaps not so much in retrospect, given what's happened to the Conservative Party since.

Steve: Well this is a very specific, he's written, I don't know, 10 000 words on this.

Adam: Right. There is no doubt that there is some extent to which that is true of every government.

Steve: Right right, but he will go into more specifics. The individual minister, minister of education, MOD, whatever portfolio they have… his claim is that that person has very little understanding, and possibly very little interest in the functioning of what comprises that portfolio other than how it affects their public figure, the public perception.

Adam: Sure, and actually further than that. In the UK they actually relatively little power right as well, unlike let's say in Singapore or the US where the Secretary of the Interior is the Chief Executive. Those roles are wrapped in one. In (for example) New Zealand, Australia, and the UK, the role of the executive—the ability to hire and fire to make day-to-day decisions and operations—is devolved to a permanent bureaucrat, a civil servant. The minister's formal role, and formal powers are essentially to exercise oversight. But there's an argument to be made here that modern ministries are so complex, especially if people are kind of coming and going, and they don't have a background in, let's say, health or defense or what have you — that the amount of information that you would need to properly exercise that oversight, especially when you only have maybe one or two people that you can trust that are actually appointable, and how do you know what which people [and which specialised skills] you actually need anyway to help kind of wade through the mass of potentially hundreds of thousands of pages of different documents and so on. I think this is covered quite well in the classic show, Yes, Minister.

Steve: Yes, so one compressed version of this is: ‘Yes Minister is actually a documentary’. It's just as you said. You have a political figurehead, you have an actual permanent civil servant, who's really the guy who does stuff, and then you have a little connection between them who are the Spads. The spads actually are younger people, and if they're competent they can spend late nights studying, ‘okay what is the civil servant saying to us? How are they really doing it?’ I can kind of go in there and talk to people, I can do more than the figurehead guy…

Adam: Who's just spending most of his or her time trying to get media publicity.

Steve: Yeah, so that's Dom's picture of how things work. That is how things work from my experience in No.10. So, during those halcyon days. Just imagine this: Boris comes in; he's appointed this evil mastermind guy (Dom). There's tons of coverage of that day when Boris first enters No.10. Dom is standing there with an OpenAI t-shirt. It's actually stained. On this one I'm a little embarrassed for Dom, because there's like a big stain on his OpenAI t-shirt. This was when OpenAI was not known at all.

Adam: I think maybe they were still yet to beat OG at DotA [actually just after, but you take my point]. They were this is hyper-niche online…

Steve: Only nerds knew about OpenAI. I'm not going to comment on how Dom got that t-shirt but it's an interesting story because there's certain connections between Dom and [OpenAI]. The picture I'm trying to paint of those days, those halcyon days, when Dom was effectively running the UK. I mean literally running the UK.

Adam: Sure, because Boris is sort of like…

Steve: He doesn't care. I'm telling it like it is. I think we can talk about like, ‘would Boris deny this?’, ‘would the people who want you to believe that the system is more sane than it actually is — would they deny these statements? Sure. You can find many people who will deny these statements. I will tell you what I actually experienced. I'm telling you what I actually experienced and saw in No.10 because I was there. I mean that's the whole point of this podcast right? If I tell you like ‘yeah, Hitler was high on cocaine and stimulants, and that's part of the problem why they couldn't win on the eastern front you know’. You might be like ‘oh well this August historian at Yale says that that isn't true’. I was fucking there. I saw him snort the cocaine. That's what you're gonna get from me.

Adam: That’s why I had you on.

Steve: If you want the conventional interpretation just go to BBC and listen to some commentator, who can't say anything outside the Overton window or this little thing comes [around his neck], yanks him off the stage, and you never see him again.

Adam: Okay, so Boris is essentially like the ministers we about talked before, primarily concerned with his public image, particularly since he's just come off this big sweeping win.

Steve: The system lets him do that because there's a permanent, there's a cabinet secretary. There are these people, who are the permanent, and generally competent, people you know who are there to run it. So, as Dom would say, when they photograph the Prime Minister and his Cabinet walking down the street behind the gate down the street, coming in for an important meeting, rushing out of your expensive Rolls Royce, to like dart into [No.10]. Those guys are not actually running the country, but they want you to think they're running the country.

Adam: Do they [the politicians] want to think they’re running the country?

Steve: No. This is the interesting thing. Some of them may actually be cynical enough to know, ‘I just need to make sure nothing bad is going to happen to me PR-wise, optics-wise’.

Adam: Right, but I don't actually have any interest?

Steve: (caricaturising a Toff) ‘It’s horribly complicated. This horribly competent young woman who they've sent us from the Ministry, I can leave things to her’.

Okay, but the connector between that figurehead person who just wants to be on TV, and the career civil servant, is the Spad, and Dom had all the Spads.

Adam: So just one more point that might be important for the audience here is the picture of ‘the ministries essentially being run by the permanent secretary’ is not true to quite the same extent at No.10. The Prime Minister has all kinds of powers, around appointments, and to direct to them, that his or her Ministers don't have. In some respects I think this creates an enormous bottleneck for power, because if something needs an approval (like outside the standard operating procedure) the minister can't approve it, so it has to go to No.10 or it languishes.

Steve: Dom has also written at length about this, and we've discussed this at length. So it all bottlenecks up to the Prime Minister because he does have extraordinary powers. But, this is the power of social convention to control the ape brain: the guy who is Prime Minister is constantly being told by these wise people around him that ‘well, that just isn't done’. You get a bunch of appointments from this Ministry and ‘you just sign here. You sign.’ Dom, because he's a nerd, and he can think outside the window, and closely read stuff. He realised the PM can just say, ‘No, I will just strike this, put Joe Smith’, and you know ‘I can do whatever I want; who are you cabinet secretary, whoever just came in with all these papers?’

Adam: The power does actually, in those cases, lie with the Prime Minister.

Steve: Legally. But you have to have the mental strength to use it. Dom realised this. Most of the politicians who get into that position don't really realise it, or maybe they don't care. He weaponized that, so he’s like ‘No we're not going to do that’. So he literally ran the UK for a while. The enormity of this I think is lost on a lot of people. His enemies—if you look at the puppet show, Spitting Image—to attack him, they like to pretend that he ran the UK and they want to make him look like an evil genius. No he actually did in a sense run the UK because he had Boris' confidence at this time, they had won this enormous victory.

Adam: One could say that he had a mandate, or Boris had a mandate which he delegated to Dom.

Steve: Exactly, and a manifestation of that is in those days, if I came to visit and I went to No.10. I would sometimes go dress like this [in shorts and a t-shirt]. Now just imagine this. I don't know how many people in the history of the United Kingdom have sat in a chair in a meeting in No.10 dressed like this, but I have done it repeatedly, and the whole vibe at the time was this crazy dude, Dom who's a genius, has seized control. He's bringing all these smart but weirdo misfit dudes.

Adam: Silicon Valley types.

Steve: Yeah Silicon Valley. ‘Wow this is an American. His accent. Look, he looks Chinese, but oh he's got this weird American accent. They were just kind of used to it, and all the spads were like ‘Oh you're Steve Hsu?, let's talk about this’. Even the ministers. So the point is, I'm just illustrating this both because it's just an amusing memory for me, but because it illustrates that he actually had total power.

Adam: Can you can you walk us through.. You talk [about] these halcyon days, before the relationship breaks down between Dom and Boris, all kinds of things are getting done. Without getting into the policy analysis side of whether or not a particular action was good or bad, but rather can you unpack that a little bit? What did that look like in the day-to-day?

Steve: Let me describe two things that we worked on, and I worked on personally. One had to do with genomics and NHS, and one had to do with high skill immigration. I had multiple meetings on those topics with the actual ministers, like Minister of Health, Home Office. Not just in No.10, I actually had meetings in other parts of Whitehall, where like enormous like conference room tables with bureaucrats, civil servants, sitting around this table and here comes this weird American professor dressed like this. A lot of stuff like that happened and we had a very, just to take an example, we had a very detailed plan for making it easy for companies in the UK to, if the individual is above a certain talent level, very easy for that person to immigrate to the UK. Now ultimately this was not realised because the time it takes to push stuff through is measured in years, and if you get kicked out then some weirdly altered version of what you're trying to do gets [implemented]. So when I mention this stuff, people on X [Twitter] attack me. Angry British conservatives attack me because the actual immigration reforms that were done under ultimately under Boris and such, I think arguably are really bad, and they weren't what we proposed.

Adam: Which was really around streamlining the process?

Steve: It was around streamlining, and in the US system there's something called an O-1 Alien of Extraordinary Ability, and it's a pretty rigorous process. If you can demonstrate your qualification of that, then you're eligible to come and be a permanent resident of the United States. So we tried to develop

Adam: This is one of the key planks of how the US succeeds in a way, right? To siphon the world's talent.

Steve: Right, but we need to expand the O-1 capacity in the US.

Adam: But the UK example is kind of modeled on this? ‘we want the best people in the world coming to the UK’.

Steve: Exactly, so what Dom and I did was, we literally built a rules based system. So it's like, ‘if you have a degree in this subject, from one of these hundred global universities, like a PhD in machine learning, neuroscience, whatever, you're in. There’s like this very simple process.

Adam: We just need to make sure you're not like a foreign government spy, or these sorts of things.

Steve: Exactly, and DeepMind doesn't have a problem because they have a lot of resources, but like AI startup in you know King's Cross or Cambridge could just hire this guy, and there's a very simple process to get him in. It would only apply to you know, at most, 10,000 people coming into the country a year, or maybe less, but those are key people, and we wanted to make it easy sure for that to happen. Ultimately the system I'm describing did not get implemented, but we were working on it.

Adam: Can you just like walk through, I mean you're talking like halcyon days, your friends in power. You're obviously like well aligned.

Steve: and I'm not, you know, I'm not being compensated for this. I'm just doing it because I think well it would be cool if we could accomplish this. Dom's my friend and I'm like, you know, helping out. That's all I'm doing.

Adam: Right. So if I get what you're saying, or I try to summarise the idea of these ‘halcyon days’? In a way, there's obvious alignment between the Prime Minister and his Chief of Staff. They’ve been elected on an enormous mandate. They're still in what’s often referred to as the grace [or honeymoon] period following an election where there's kind of high political capital, it’s being deployed, and a degree of willingness to sort of get with the program. [As time passed] did you see that ebbing away, the idea of almost like running the government a little… or portions of the government like a startup?

Steve: Moving fast. Breaking things.

Adam: I mean, yeah, that's an interesting way of putting it. There are some parallels here potentially to DOGE.

Steve: If I could add some. Yeah, I was just about to say that. So there are very strong parallels between that finite period, we didn’t know why it would be a finite period, we just knew that eventually something was going to end it, it’s very similar to Trump 2 [i.e. his second term]. Trump 1 was not like this. What we’re experiencing now, they have been acting the same way. The opposition is on its heels. We’ve got a mandate. We’re going to bring people in from the outside, and we’re going to crank. That’s what they’ve been trying to do as well.

Adam: There’s an interesting dilemma here in some respects, right? Because the government, you know, thinking of Yes Minister, there's this great period where the two main characters, Sir Humphrey and Hacker, are talking about essentially what the role of government should be, and one view that the civil service character (Humphrey) provides is this idea of government's not about good or evil, or any of that. It's ultimately about chaos and order, and providing stability, that the private sector can then build upon, and individuals in their own lives and in their families and in businesses can build on. Business requires stability in order to be able to make investments. Just on a philosophical point of view almost, like, okay, maybe the startup government can be done correctly, you know, if you've got the right people and it's the right time, but it's very difficult beforehand, for even intelligent people, highly motivated people to know exactly how that should be done. I think DOGE, I think it's fair to say at this point in June is, you know, something of a disaster. Elon Musk's been, you know, out. There is a range of questions about the legality; what's been done. Some clear waste has been found and cut, but in other cases it's like, you know, this idea of moving fast and breaking things is damaging, you know, there's been defunding, unintentional defunding of like long-running longitudinal scientific studies that are like difficult now to bring back. So, there's been all kinds of damage done. Like, how do we think about this? What do we make of all this?

[Statecraft wrote a nice summary debrief on DOGE for those interested]

Steve: I just want to clarify. So, this attitude that there should be stability and predictability in laws and regulations and basically the operating environment that businesses are going to have to deal with. That is a desirable quality of government. You don't want a government that like every other day someone has an idea and then they just totally shift the landscape, and you're vertical, right? You don't want that.

Adam: Or inconsistent rulings because then you don't have the rule of law, you have the rule of like some random Tsar or something.

Steve: Or Trump, right. Okay. So, when people said in the Dom context, move fast and break things, they didn't mean like we're always going to be doing this, oscillating around in government. He had specific things [he’d identified] that were broken in the UK. You have to move fast to fix them while the opposition is still on its heels. That's the attitude. Now, with Trump, I, I'm not… look, I agree with you, a lot of damage has been done. Actions have been taken which ultimately are not going to be seen to be good.

Adam: So, don't move fast and break government in general, but rather in policy areas where, you know, there's a clear mandate of need for change, and high conviction. You actually do need to be quite agile in those spaces.

Steve: Like, oh, 10 million-ish people came across the southern border under Biden and no one knows where they are. No one cared. They just, oh, I have high conviction we're going to actually seal that fucking border off and we have to move fast and do it before the political opposition says, ‘oh, this is inhuman. How can you possibly have borders? I mean, Singapore doesn't control their borders, do they? Oh, I guess they do.’

You know, I mean [if] there's a definite thing you want to accomplish, but because things are so fucked up, you actually have, even though you have the mandate, you have to move fast to get it done because otherwise they're going to block you. An activist judge is going to block you. CNN is going to run a huge documentary on how and why this is going to happen, why this is bad. You know what I mean?

Adam: I think that's, that's almost true of any government, any policy.

[Change is always hard. People are normally quite comfortable with the status-quo, even if it’s globally suboptimal. Drawbacks are often felt immediately, while benefits can take time to fully materialise. A good strategy needs to compensate for this, and speed is often a key factor determining success or failure.]

Steve: But the thing you're trying to do should be thoughtfully arrived at, not, ‘oh, my 17-year-old, my 19-year-old Doge intern thinks it's good to kill this’. It's like, okay, let's flip a coin. Oh, yeah, like kill it.

Adam: There's a lot of policy by [ChatGpt] o3. AI has said do ‘blah’, but it lacks context. [i.e. garbage in, garbage out].

Steve: That's not what happened under Dom.

Adam: Okay. So thinking, I'm really keen to dive in specifically to, because a lot of what Dom's written about online is not so much about any particular set of policies about ‘we want to achieve this, deliver this particular kind of novel thing, or change this law in some way’, but rather the functioning of the UK government itself, like the systems that were set up, under, let's say, William Pitt in the early 19th century or Northcote Trevelyan later in that century, and that have remained relatively static since then. This kind of scheme, this ‘Yes, Minister’ scheme that we were talking about earlier. There's undoubtedly some advantages to that.

On the other hand, you should expect that as the environment shifts, like things have changed since the 19th century, [and all else being the same, we should assume static systems have probably become outdated]. We talked a little bit about the lack of technology in decision-making. We were talking in the context of COVID, perhaps the lack of specific kinds of technical expertise. Within the bureaucracy, like Ministry of Health, for example, which I'm sure was full of like really talented doctors, policy-type people, and lawyers, and all of the day-to-day management of the bureaucracy, I'm sure is maybe generally well covered. But there are kinds of expertise that you have, for example, that you're talking about of like mathematical ability, ability to reason under uncertainty that are lacking. I think also maybe most summed up in the character of Dom himself is a kind of a lack of the ability of sort of internal reformation type skills. Would that be fair to say? My impression from afar, anyway, is that these institutions have sort of been locked in place, maybe since the end of the Second World War, and are kind of creaking at the edges. You see this in the NHS, and, you know, in defence, and in procurement, and so on.

I think there's some degree to which the popular perception of this is that ‘it's all on parliament, or the government in power, not giving them enough money or what have you, maybe some particular rules’, as opposed to a function of the bureaucracy not being efficient and moving with the times. And just to conclude this point, I think one of the really interesting points that maybe is somewhat missed is the idea that, you know, we're talking about this evil genius, Dom, he's come in, you know, a lot of people hate him for various reasons. Some of them perhaps justified from their perspective. But, that Labour, like the recently elected Labour government under Keir Starmer in 2024, has come out publicly (admittedly unnamed sources close to the Prime Minister), but saying ‘Dom was right’. That's a direct quote. ‘Dom was right’.

Steve: Oh, I didn't know that. Good.

Broken Governance

Adam: About this particular kind of idea that I've just outlined that ‘something is deeply broken, but we're not quite sure what it is’. We have certain values and things that we care about in terms of stability and so on. I think, you know, we're in this time where, whether it's in the US, whether it's in the UK, certainly I write a lot about the subject, the view that there's some kind of institutional reform required. I think that the electorate even senses this, you know, there's been some quite dramatic changes in the UK over the last, well, I mean, like virtually every year. Reform is now polling at near where Labour was when it won the election [Labour won an outright majority of seats in Parliament with 33.7% of the vote]. So with its kind of historic victory last year. So uh, how, what are we to make of all this?

You've got quite a unique inside into all of this, the ways in which a new approach was almost bubbling to the surface during these ‘halcyon days’ you talk about. But then, you know, that rapidly fell apart. The system reestablished itself.

Steve: We’re back to this fundamental question of what's wrong with governance in the UK. This notion that you have these political figureheads and they generally don't they just sort of sign off and what the civil service brings them is not necessarily bad if the civil service is doing a good job. In a lot of things Dom would say actually what the civil servant was proposing was actually the rational thing. The problem is a second factor, which I think you didn't mention, which Dom goes on and on about. I saw it firsthand. There are many talented, energetic, young people in the UK. I actually am just amazed when I go over there. They still have a pretty functional meritocracy at Oxbridge. To get into Oxbridge, you can't just be a legacy. They interview the candidates, and it's pretty much based on scores. So they have a pretty functional system sucking in the graduates from this nation of 60 million, and sorting them a little bit, and educating them pretty well, and then a lot of them end up in the city, in London trying to do something. A lot of them are idealistic. The problem, Dom would say, is the way these bureaucracies function. These talented people come in, they might give it a few years, five years, but they all bleed out. By the time you get to the top, you have a totally different character of person who is willing to make a lifetime career in civil service. I don't know enough to really fully judge, but I do know that the young people that were attracted by Dom and who worked in No.10 were killing themselves, super idealistic…

Adam: By killing themselves, you mean in terms of working long hours? [Killing it, killing themselves, etc. has positive connotations in SF where this was filmed]

Steve: Working long hours, very dedicated, really idealistic people, but were ultimately not a good fit for the bureaucracy as it stands. By now all of them have bled out, have basically left. I don't know if that's the main problem, but it is a problem that Dom points to for that system. I can verify that, from what I saw, this is true.

Adam: There's an irony here that, just as we're in this period of history where things are rapidly developing, you know, we were talking about Cambridge Analytica earlier and how you know 10 years on that's du jour, but you know now we've got AI, who knows what next year's advancements are going to be, but the leadership, through no fault of their own, in some respects, grew up in a world that's pre-social media, maybe pre-modern computing in some respects. So I think a misalignment of sorts is almost inevitable in that kind of circumstance. So a lot of these people come in, as you say, they're [then] burned out, but they're…

Steve: I wouldn't even say it's necessarily burnout that gets them out of there. What it is, is that their ideas are not listened to, they're not acted on. So after a while they're just like ‘I'm wasting my time here, I better go do something in the private sector or something’, but if the system did allow them to have input, and reacted the right way to their good ideas, I think a lot of these people would happily stay, and make the bureaucracy much more capable.

Adam: I've had this idea of like you can we need to be creating almost recreating the midshipman in a way. You give people actual authority. Actual delegated responsibility and authority to achieve something, which so often we fail to delegate those two things together - responsibility for something and the authority to actually make decisions around it. But you do it in a really like limited window, because at the moment there's you don't get any until suddenly you get an enormous amount. Instead we [could] give people very narrow tasks like ‘I want you to deal with this particular thing’ and grow people in that way.

Steve: This observation you made is standard operating procedure at all startups. So the thing that's great about working in a startup is that ‘I'm going to give you a responsibility, but I'm also going to give you autonomy to make decisions about how to fulfill that responsibility’. The two together. If you don't give the second thing it's just a frustrating exercise for the worker, but if you give both, then it's a growth trajectory.

Adam: So you mentioned being super impressed with the degree of talent and enthusiasm, and patriotism? (yes) of young people in the UK at the moment.

[Famed startup incubator founder Paul Graham talks about this a lot, such as on Conversations with Tyler]

Steve: Well I can't say this in general, but this is true of the people that Dom attracted into No.10, and there are plenty of good people.

Adam: Well let's suppose that what you were seeing before is still true, certainly my impression is that there's still an enormous amount of talent in that island, and as you say nothing's changed dramatically in terms of the pipeline of Oxbridge filtering and and training.

Steve: I think it's still functional.

Adam: There are these various green shoots that I can see of [young people organising]. I'm sure criticisms could be levelled at them that they're too idealistic in some respects and they need to have… I’m sure that's probably true in some respects, I don't know. Where to for Britain from here? Reform is polling high. They don't seem to have necessarily much of a team behind them. Possibly destined to just remain at most a plurality of the electorate? Dom's still active behind the scenes in some respects I'm told. What do you see for the next five years for Britain?

Steve: I'm not expert enough to really have a useful opinion on this, but I will say this. I was quite intrigued when Dom recently, I think in a substack, signaled his willingness to work with Reform. I think he made that observation you just made that, the way they're polling, they have a shot at being the leading party. Yet they probably don't have the strategic depth.

Adam: It's still largely a party based around like Farage himself, yeah.

Steve: I have to say this very carefully. In the past Dom has not, even to his close friends, not ever signalled to work with Farage. In the beginning it was a really tactical thing. Vote Leave did not want to be put in the same bucket with Farage, because Farage was viewed as maybe too far right, maybe racist, whatever. So there was a tactical decision to separate what they were doing from what Farage's… I forgot what Farage's thing was called (Adam: UKIP). UKIP, right. I think it was even more than that. I think Dom didn't feel like he could, or should, work with you know Farage. But now he's signaling that Reform needs a, I don't want to say evil genius, a genius political strategist…

Adam: A Machiavellian perhaps? A consequentialist? Maybe is there something there of like, ‘if I don't get involved.. I still don't like Farage..’

Steve: No comment that he doesn’t like Farage.

Adam: …but he had a strategic view before that it wasn't right. Now possibly he's coming to a view that ‘if I'm not involved, or my team, outcomes for Britain will be worse?’

Steve: So if I'm Farage, or the team around Farage, I would say have some meetings with Dom. Doesn't cost you anything to have some meetings with this guy. See if it's possible for you to work together, as the individual definitely has valuable political and strategic insights that your team may not have.

China Rising

Adam: Okay so I am keen to pivot, quite dramatically, to the other side of the world.

Steve: Enough about this little island thing. Although we love it. I mean, you know.

Adam: (laughs) I'm from a little island Steve.

Steve: We love that too. I mean when you're in power in New Zealand, those halcyon days, I'll come visit you.

Adam: Okay, all right. So I think one thing that I would be interested in your views on, perhaps then, but certainly now, is on China. You have some involvement with a University in Beijing, I believe.

Steve: I have no official involvement with anything in China right now, but I do visit there to give, like, lectures in physics or things like this. But I no longer have any… for some reason people keep like… my Wikipedia entry still lists me as an advisor to BGI, but I haven't worked with BGI in about 10 years. Just to set the record straight. I mean it's Wikipedia. I think my official web page at Michigan State, which was set up in like 2012 when I moved there, hasn't changed. It still says I’m a scientific advisor to BGI.

Adam: Okay. That's a whole subject in and of itself, like information theory, the nature of the internet, how much do we trust things and so on, legacy information still being referenced by AI. It's a fascinating subject, but okay. So you don't have any specific involvement in institutions in China, but you do visit there, it’s an area you have a lot of interest in.

Steve: I have a deep interest in understanding what is happening. So imagine it's 1930, or 1935, and you're sitting at home in Peoria, Illinois, and you're reading the newspaper. Say you're a professor or something at the university of Illinois, and you're reading the newspaper, and they're writing all about this guy Hitler, and they're re-industrializing Germany, and they've tamed inflation, and Weimar is over, and ‘wow this could be a problem for the United States, they're they seem to have really good planes and tanks, and you know the Krupp steel is really doing gangbusters stuff’, and like, well, okay you could just read about it in the New York Times and trust that the New York Times knows what they're talking about. Or, you could go to Germany and walk around. If you go to Germany are you just going to eat some strudel, try the beer, go to Oktoberfest, and come back? Or, if you go there are you going to say, ‘well maybe I'd like to meet my counterparts the professors at the University in Munich, maybe I want to be introduced by them to some of the officials in the governing party, maybe I want to meet high-level people business leaders, Krupp steel officials, armaments manufacturers, why has the Messerschmitt got a better engine than the shit that we make in the United States? I might want to know…

Adam: By analogy, this is what you've been doing in China?

Steve: Yeah, you might want to know some things right? I don't trust any experts. You know, to me the world looks like a bunch of retards talking to each other. Same thing with reasoning about covid beforehand, same thing reasoning about financial markets, and the housing bubble before that.

Adam: Go to first principles, as much as you can.

Steve: Yes, gather your own data, or find reliable data sources. Do your own analysis. Of course, it's a luxury to do this right. If you don't have the resources to go to 1935 Germany and look around, you’ll just have to rely on what the fucking New York Times writes for you, or like the supposed expert on German politics who's a professor at your university.

Adam: Good for getting news most of the time, but not necessarily analysis.

Steve: Yeah, I'm not saying this. It’s not normative. It's not like you should do this, or you're an idiot. I just like doing this. So in the modern world, it's remarkably easy if you try and you have some background, people can look up who you are and say, okay, this is a serious guy. Okay, he's dressed like a goofball, but he's probably a serious guy. Yeah, I'll take a meeting with this guy. So you might even like get to talk to some intellectual leaders of that country. Anyway, I think it's possible to get a deeper feel for what is happening in that country, in this case, say China, than what you would get from just ordinary analysis that you can just read in the newspaper, yeah.

Adam: So, I mean, it seems to me, and particularly after this conference and, you know, some of the presentations I've seen, and really I think more generally in the information feeds that I'm seeing that people on the whole, like the general public, maybe even decision-makers, have quite an outdated view of China. That there's still, like, it's still a nation of cheap labour costs, sweatshop manufacturing, it's plagued by internal issues, like, you know, some housing bubble, local government debt. I think it's probably fair to say that there are still some elements of truth to this. Yes. But that on the whole, they're missing 10 years of 5% per year annual growth. And I've seen, I mean, maybe it's still somewhat of a meme these days, but large factories that are dark because there's nobody in them. So you’re talking about really a completely different world to the one of certainly 20 years ago. Maybe still there are elements of 10 years ago still around, but we're entering into a new era for China, and consequently relations.

Steve: So let me just comment that. So you and I were, prior to this taping, at a festival here at Lighthaven, called Manifest.

Adam: A prediction conference.

Steve: Yeah, but you know, I think it's broader than that. If you think of the range of talks people gave, only a limited fraction really were about prediction markets. The sponsoring company Manifold Markets, no connection to my podcast, I let them use the name, I'm not suing them, but they sponsor this festival, and it brings together quite an interesting fun mix of people and talks. It's a self-organised meeting, so the talks are on all kinds of things. It's extremely interesting, but among the people that attend Manifest are superforecasters, so people who are seriously outliers in their ability to use these prediction markets, there were top AI researchers from the big labs here, including the people who actually do the training of the models, AI safety researchers, rationalists who focus on AI risk, people who are at the cutting edge of genomic technology..

Adam: Semiconductor manufacturing.

Steve: Yes, we had a talk by Dylan Patel, who's one of the experts following the U.S- China competition in semiconductors. So it was quite an excellent group of people. The talks were excellent. My talk, which was followed by a discussion with Noah Smith, the econ economics commentator, was about China. Maybe you can put a link to my slides, because my slides are publicly available and are related to the things that you just mentioned. People can go and read in more detail. The slides are full of graphs and stuff like this. Sorry for that long lead-up, but interestingly when I gave my talk at Manifest, it was in the big outdoor arena so maybe a third or half the people who were here went to my talk. A lot of people. Then for days after that people would come up to me and say, ‘Wow great talk. Really shocking. Some of those numbers you showed, and some of the inferences from what you said were shocking to me.’

Adam: People aren't aware that Chinese electricity production is higher than the US.

Steve: It’s 2x and one is going [exponential] like this, and the other one's flat. So I'm just trying to say that in this highly informed population of people who are at Manifest. Even there, people were shocked by a lot of the stuff that I covered in my talk. So yes, there's a big gap I think between even sort of 90th percentile or 80th percentile well-informed person in the US about China, and the reality. There's quite a gap.

Adam: Right, and that reality, in broad terms, is one of obviously continued hyper-centralisation amongst the central party, so you do get these Shanghai lockdowns being continued way past they should be. That's certainly a reality…

Steve: Slight correction there. First of all, the government, there in the way that things are actually done, is not that centralised. There’s directional guidance...

Adam: Sorry, I'm thinking just in terms of the CCP and the concentration of power in President Xi but absolutely a very devolved system when it comes to the province level.

Steve: When it comes to how is an individual city handling covid, how is an individual city handling the mandate that directionally [they are] supposed to support, electric vehicles, and alternative energy, the individual policies on a province-by-province, or city-by-city basis, is very diverse. Those mayors and governors have autonomy. They're judged meritocratically on whether they succeed or not [as to whether they] rise further.

Adam: There’s like an enormous evolutionary algorithm running in China.

Steve: So that Shanghai lockdown, which was probably a big mistake, that was the local government.

Adam: Oh okay. I thought it was imposed from Beijing.

Steve: No, I don't think so. Maybe. It's sort of revisionist history. Maybe the central government was responsible, now they're saying it wasn't — it was the local guy who made the decision, but in any case I do think one big misconception about China is that it's some kind of North Korea totalitarian. It's not at all. It's consciously a system where directional guidance is given from Beijing, but the individual actions are distributed, and locally made, decentralised. So individual companies are competing in market competition, individual local governments are competing to try to do the right thing for economic growth (Adam - within a province even). Exactly, and each province some province would be the size of Germany or something.

Adam: Yeah, and there's an extraordinary amount of economic dynamism in particular as a result of that. So there’s some interesting implications for this idea of like, on the one hand, nevertheless, you do have a high degree of, at least within the central party, concentration of power within Xi.

Steve: Ultimate power is concentrated, but the way that individual [decisions]… you know this is the usual problem, the paradox, about how to run a sophisticated economy. You couldn't possibly centralise it, because imagine every decision passed Xi’s desk. Should the factory in Wuhan expand by 100 hectares?

Adam: The Soviet style.

Steve: Yeah it's never going to happen that way.

Adam: They call it Communism with Chinese Characteristics. Sometimes referred to as ‘Capitalism with Chinese Characteristics’.

Steve: Yeah exactly. So by some metrics, depending on which economist you talk to, there's a claim that actually the governance in China is actually more decentralised than in sort of modern Western countries.

Adam: Yeah and even the US. I think one of the strengths of the US relative to let's say the rest of the anglosphere. I'm thinking here specifically even of New Zealand where our decisions are all concentrated in Wellington, all major decisions around like special economic zones and whether you provide tax incentives, and these kinds of decisions. Instead, I think one of the reasons the US has been so successful economically is just that different states are trying different things, and then they adopt what works from each other.

Steve: Exactly. The difference here though is what states do in the United States is pretty limited, actually — what the governor actually controls and can do is relatively [limited]. In China it's more so. You have many governors, many mayors of huge cities, that are all kind of making potentially independent decisions on what to do. Now of course there is a herding effect because they all talk to each other and they're like ‘well, what was the guy in Shanghai doing?’ but in principle they have more control, more impact on their local region, than a US governor, and they're somewhat independently competing with each other.

Adam: I’m trying to paint a general picture here so we can go on to talk about the implications for the Pacific and so on. But essentially the summary of your slides that people can click the link and read online, and pause the video here. It's something like: ‘you're talking about just extraordinary manufacturing capability across all manner of critical industries, increasingly competitive even in semiconductors where the US has attempted to kneecap them, potentially to their own detriment in a way because instead of controlling CUDA and the chips that AIs are trained on in China, instead China is investing enormous in resources into catching up. Huawei's new 910c chips are let's say maybe a generation behind.

[CUDA is Nvidia foundational software that acts as the intermediary between AI models and the chips they run on. Biden’s export controls, modified but essentially continued under Trump, are forcing Chinese companies to invest in alternatives. Thus, rather than locking in an effective monopoly as generations of ML researchers develop on CUDA, and controlling the foundations of our time’s most important technology, America foreign policy has likely thrown that away, at the expense of its own companies]

99%, I think, of the commercial drone market is Chinese [DJI alone has a 90% market share]. In general you could say that maybe their capacity to project power is not the same as the United States, but near to their shores at the very least they are competitive. I had Nuño Sempere on here, the legendary forecaster of Samotsvety, saying that one of the views of Manifest was, and amongst his kind of super forecasting fraternity, was that the likelihood this decade of boots on the ground in Taiwan was something between 15 and 60 percent, and that's an operationalisation that doesn't include other forms of warfare like a blockade, or what have you. So in one way or another, highly likely that something is going to be going down in that space. It may well be that the next round of elections in Taiwan, you've talked about a bit of the split between the opposition parties. splitting the vote in the recent presidential election. Possibly that's now changed. But one way or another, whether it's peaceful reunification or [compromise and] an easing of tensions – things are going to change quite dramatically in that space over this coming decade. That's going to have quite profound implications for Japan, some of whose islands are claimed by China, certainly Vietnam and Philippines, who could say the same. India, there's this long-running almost cold war, phony war, at the Himalayan border. So can you just sort of – can you walk us through what that looks like? Then, how does a nation like New Zealand or Australia respond to that?

Emerging Geopolitical Realities

Steve: Well, I think the general trend is one of overall growth. Growth in the aggregate comprehensive power of the People's Republic of China, for all the reasons you mentioned. Manufacturing capability, the human capital pool is deepening a lot. They're catching up, or have surpassed, the West in a lot of technological areas, and they have a will to power. They have a feeling that there's a certain place that they should occupy in the world stage.

Adam: The Middle Kingdom.

Steve: Yeah, but which is different from saying ‘we are going to invade and take over other countries’, but still there's a certain level that they think they need to get to, and they're committed to it. So the classic problem is if you're a smaller country next to a bigger country, in your case not next to, but for ASEAN countries, you're a small country next to a big country. And the big country feels like, nah, I have a sphere of influence. I can kind of throw my weight around a little bit. And the little country is uncomfortable with that. And then typically what happens is there's an offshore balancer. In Europe, the offshore balancer for a long time was the UK. So the UK strategy was ‘I will never let one entity dominate the continent because then we could be in trouble'. I'll always align with the next most powerful entity and we'll keep a balance of power in Europe’.

Adam: Napoleonic Wars, World War I.

Steve: The problem is that the ability of the US to be a meaningful offshore balancer is decreasing because the US economy is not as much larger, maybe is now even smaller than the Chinese economy [in PPP terms since 2016]. The projection of power that far across the Pacific is very expensive. With the advance of things like drones and smart and cheap missiles. I want to emphasise it. Right now the focus is on these little shitty quadcopter plastic things, but those things are getting smarter and smarter and, okay, they can threaten you know armoured personnel carriers and tanks and soldiers, but the real tech is a missile with a brain as smart as the drone's brain. That is what you really should be thinking about. Something that can hit your ship from 2 000 kilometers away.

Adam: Without knowing where exactly it is to begin with? This has been like the saving grace in some respects — the ocean is so enormous you can like fire a dumb missile [and it’ll miss or be vulnerable to various countermeasures].

Steve: Now if you do your technical due diligence and you ask, ‘what is the level of satellite coverage of the Pacific by the Chinese?’ It's extremely high, and that is even ignoring their ability to put up either rockets or drones at high altitude. I don't think there's any credible way that the US can hide the position of its aircraft carriers from the Chinese, and the range at which they can be attacked is on the order of 2 000 kilometers (en masse). So launch 100 missiles, and those missiles only need to be given the general coordinates, because the ship is moving of course, but the missile if conditional on the missile being smart enough, that in the final 30 seconds or whatever, the ship moved maybe 10 kilometers or something during this launch phase, but the missile reacquires it because the missile has enough compute, and autonomy reacquires a ship and then hits a ship.

Now we saw in India-Pakistan, missiles launched from beyond visual range. Now a fighter jet, I hope even the retards who don't understand anything about military technology, I hope they will acknowledge that a fighter jet which is maneuvering is a harder target than an aircraft carrier which is moving at most 30 knots or something, and also that an air-to-air missile is much smaller than an anti-ship ballistic missile. The radar system in the cone, and the power and compute of an anti-ship missile is 10x more sophisticated than an air-to-air missile right? Missiles are small, [but the size of the cone in the larger missiles provides much more room for more advanced capabilities]. So the idea that you have some way to hide your carrier from that final reacquisition is very implausible actually. We saw in India-Pakistan that many Indian jets were shot down, and they didn't even detect the missile. They were hit by missiles that they couldn't even detect.

So my point is: maybe there'll be some balance where the missile defence systems of the ship are somewhat effective, but a ship can only carry so many defensive missiles, and if I spam enough at the ship, I will be able to sink them. In the war games that the Pentagon and actual think tanks play — generally in those war games, many carriers are sunk. So that's the situation. Imagine you're the US president, and China's pressuring Taiwan. Maybe they're even invading Taiwan. What's your reaction? Even the Pentagon planning, you know, when they advise the president, they have to have a very clear understanding of the enemy's capabilities. So they say like, ‘oh no we'll be able to kick them out because XYZ’, and if that turns out to be wrong, it's a disaster for the US. So, we're entering to a very very delicate, unstable phase right now. It's unclear how it's going to evolve but I think the general trend is that the Chinese are very serious about the Western Pacific, and because of their superior manufacturing capability, I think their military shipbuilding capability is 200x the US capability, it's hard to imagine the balance of power isn't going to shift more in their favor in the western pacific in the next 10-20 years.

Adam: Notwithstanding what might occur in 50 years. People point the One Child Policy, really [the true impact of] all of that is quite far into the future.

Steve: We could discuss further out, but with some degree of clarity 10 to 20 is really more the period where we can say something. Even that is clouded by the rate of AGI development. So something could happen in 10 to 20 years that [upends everything].

Adam: So given all this, I think fairly clearly there’s an unstable equilibrium over Taiwan. One way or another, you know that there's going to need to be a deal made, or there might be conflict, hopefully not, but some kind of overture, you know change in the geopolitical status of that area. Yeah, like what does what does Japan, what should Vietnam, what should Australia be thinking? I know those examples are a little different given the proximity.

Steve: Right, I think that New Zealand and Australia, being part of the Anglosphere, have a very different cultural take on, or intuitive understanding of, this change in power dynamics in Asia. There are a few intellectuals, for example in Australia, defence thinkers, who I think are like well-calibrated but many of them are not. I think those countries are just much more likely to align with the United States. Among Japan and Korea, which are two countries that really matter a lot in this discussion, I think they're much more realistic. They're closer to China. They have a better understanding of the capabilities.

Adam: They're not speaking English, or too plugged into the American Internet.

Steve: Yeah. They know the Americans lie about what is going on. They know the Americans pressure them to do things that are in their interest. They have a much more realist view. So the CEO of Toyota, or the CEO of Samsung can just have a conversation with the president of South Korea, or the Prime Minister of Japan and just say privately, ‘you know what? Their cars are better than ours now’, or, ‘you know what, they’ve fully caught up in high bandwidth memory. Their phones are better than Samsung phones’. So they can, and they have to face these realities because they are, for example, this is not widely publicised, but Toyota worked for 20+ years on the Prius hybrid technology platform. They've now abandoned that in favour of the BYD hybrid technology because it's vastly superior. The BYD cars can have thousand-mile ranges hybrids with thousand-mile ranges. Toyota-branded cars, hybrids that are being sold in China now—China is their biggest market, I think it's probably the biggest market—are using a platform license from BYD.

Now you wouldn't know that, as an American, because ‘America's great’, right? But but, oh, and I guess Japan is great too, or something? I don't know. Toyota's great, but Chinese can't innovate? You know, whatever, some kind of bullshit worldview. In Japan, they're not going to be able to hide this. Imagine you're the Prime Minister of Japan and you're like ‘Well, uh, should we just antagonise the Chinese because the US president wants us to antagonise the Chinese? What do you think about that?’ The Toyota guy will just quietly explain to him ‘Well, we make most of our profits in China, and not in the United States. Their auto tech is actually better than ours now. We're totally dependent on them for batteries. We're totally dependent on them for rare earths. We're totally dependent on them for the software systems in our cars now. So the conversation is much more realistic than what you're going to hear here.

Adam: So Steve, I'm curious in the context of this coming realignment of East Asia that we've been discussing, New Zealand and Australia are outliers. In both cases I believe, certainly in New Zealand's case, China is our biggest trading partner [this is even more true of Australia] but we're politically and via language, and in Australia's case via defense treaty, aligned with the US. It's a going to be quite a difficult… and a US that is, especially under Trump, but in general perhaps. wanting to you know draw lines of allies and push its weight around more. How do we navigate this?

Steve: Yeah I think very carefully. We hear a lot about the Thucydides trap, that's more about the rising power, and the incumbent hegemon possibly fighting or coming to some kind of agreement for coexistence. The other classical dilemma, either from game theory, or any kind of geostrategy, is the smaller country that is in the sphere of influence of a bigger country, and what can they do about it? If you talk to leaders of ASEAN countries they're very explicit about this. They talk about this openly. They have to manage really carefully the relationships between [themselves and the US], but they usually they say things like, ‘well we want to be friendly to the United States, but China is our neighbor and they're not going away.’ America could go away actually. It wouldn't take much for an isolationist president to come into power.

Adam: Not unprecedented historically right?

Steve: Exactly. The other thing could happen as well is that the geostrategic military-economic analysis from the US side reaches a point where they say, ‘Yeah, I guess we can't really defend Taiwan against China. Yeah, we can't really defend Japan.’ Not that the Chinese would invade Japan, but they could blockade Japan, for example.

Adam: Japan might be forced to give up, let's say, even Okinawa or something, or some of the smaller islands in that in that area.

Steve: I think that stuff could be settled out and it's not existential for either party. It's not existential for Japan. It's not existential for China, the state of those little islands…

Adam: But there's some degree of wanting to avoid salami tactics, though, right?

Steve: Well, this is this is part of how actively engaged does the offshore balancer have to be with what is happening in this distant region. Very distant in this case because it's the Pacific we're talking about. It's not across the English Channel or something. At some point American strategists are going to realise, ‘they can slice salami over there, what are we going to do?’ Our debate over whether to ‘allow them’ to slice salami is totally immaterial, because we can't actually do anything. We can go zero-to-100 and start a global thermonuclear war, which is not good for anybody. Short of that we can't do anything. If you look at these FONOPs… Do you know what FONOPs is? Freedom Of Navigation Operations.

Adam: Sailing through the Taiwan Strait, and so on.

Steve: Or these little islands in the South China Sea. We still engage in these fake FONOPs, even though the ships doing this would be destroyed, I think based on the technology, would be destroyed in 10-minutes if Beijing wanted to destroy these ships, but for symbolic reasons we sail them around.

You could say, ‘Steve you misunderstand, it’s all symbolic. We understand the ships are totally vulnerable, we understand we could not actually operate in the South China Sea for even one day’… because they can see everything, right? Unless you pretend missiles don't work, but we're pretty sure they work. Yeah, the U.S. Presence in the South China Sea probably could be eliminated very quickly. Now, you can say it's just symbolic right now, okay? We're just trying to reassure our allies that we're there for them, even though it's not clear how we could possibly be there for them, right? if push came to shove. Fine.

[While I largely buy Steve’s argument here, I do think there’s something to be said for the normative elements of these FONOPs. Not that the ship itself isn’t highly vulnerable, but, given the strategic ambiguity of how the US as a nation might respond if a military ship was attacked, FONOPs limit opportunities to place de facto limitations on freedom of navigation.]

Operation Pompous Golem

Forget about the South China Sea for a second. Let's go to the Red Sea, okay? We had an official FONOPs operation running in the Red Sea for a year and a half, okay? Now, who was the enemy there? Was it, like, some very powerful country like Iran or something that was not letting ships transit the Red Sea to the Suez Canal? I can't remember the details. I think it was some guys in sandals in an impoverished country with some second-hand Iranian missiles. Now, what happened to that FONOPs operation? So, again, I'm a fact-based guy, okay? We don't have to debate this. We can just look it up. At the beginning of the FONOPs operation, which was begun by the Biden administration, the name of the operation was ‘Operation Prosperity Guardian’. The goal was to restore freedom of shipping through the Red Sea and the Suez Canal. We don't have to debate it. We don't have to do some crazy, really difficult math. We can just ask, ‘Has shipping been restored in the Red Sea a year and a half later?’. It was not restored at all under Biden. The Trump people, the hardcore, you know, like, red-blooded American war hawk guys were all saying, this is just because Biden's a wimp. When the ‘real president’ gets in, we're going to bomb the fuck out of these Houthis and FONOPs will succeed. Freedom of navigation will be restored.

Okay, so what happened? Again, I'm not making any of this up. Go look it up. If you don't look it up and check the facts, then do not fucking talk to me. You're just a retard. We bombed the hell out of them, but we also almost lost a bunch of ships. You can find online interviews with the commander, I think, of the Eisenhower after his tour of duty. Those guys were scared. It was touch and go. Lots of stuff happened. They lost two F-18s. They lost a bunch of drones.

Adam: [regarding the F-18s] because the ships were having to maneuver so violently that they fell off the deck.

Steve: We expended maybe a billion dollars in anti-missile missiles, SMs, to shoot down, you know, incoming like $300 or $3,000 drones and Iranian missiles. So ultimately, Trump, cut a deal just recently with the Houthis. Now, what was the deal? Did Trump say, ‘Guys, I've been bombing you so much, and I'm getting tired of all this winning and bombing of you because I'm a nice guy. Let the ships go back through the Red Sea and I'll stop bombing you’? That's what all the idiot, war hawk, ‘Go U.S. Military!’ people thought was going to happen. That's not the deal that was cut. The deal that was cut was, ‘stop shooting at my military ships, and I will stop bombing innocent women and children in Yemen’. Now, that was the deal that was cut, and the deal doesn't say anything about restoring shipping in the Red Sea, which has not been restored.

[The May ceasefire is often misreported as having been wider in scope than Steve suggests here, such as being an agreement that the Houthis would only target Israeli shipping in future. In fact, Steve is correct. The agreement only applies to US ships, and two Greek commercial vessels were sunk earlier this month.]

So the ‘FONOPs’ part of the FONOPs… The phony Operation Prosperity Guardian…

Adam: The pun would work well as text, I think. [FauxNOPs]

Steve: Yeah, phony-FONOPs. No. They thought they were going to succeed because it's the U.S. Military. This is the most powerful Navy in the world, of course we can succeed right? They thought they were going to succeed, but they didn't succeed. What was called ‘Operation Prosperity Guardian’ should really be called ‘Operation Pompous Golem’. If you know what a golem is, then you should understand this was really ‘Operation Pompous Golem’. The day after Trump did the deal… which is like, you stop firing at my ships, but we actually haven't restored freedom of navigation in the Red Sea, and to this day, it hasn't been restored. The next day after that deal was cut between Trump and the Houthis, they fired ballistic missiles at Israel.

So how impotent is the U.S. Military when it comes to naval operations against anyone who even has some missiles? Some missiles, and some will to power, is enough to defeat the U.S. Navy, okay? I submit to you that what they will face in the South China Sea, or anywhere near Taiwan, or anywhere near Japan, I mean, if you figure out what a 2,000 kilometer range is and you draw a little circle, it covers all of these, maybe it doesn't cover New Zealand, but I guess they can hit, they can hit Darwin with some of their longer range anti-ship missiles. Yeah, but certainly these countries like South Korea and Japan.

Adam: We just recently had a flotilla. One of the myths that's been long-standing is this idea that, ‘Yeah, China builds a lot of ships, but they're not very well trained. It's not really a blue water Navy.’

Steve: That's all like 10, 20 years old.

Adam: Yeah, exactly. It's another such case, right? Because they were like fire… they went all the way to, you know, off between New Zealand and Australia. Yeah, did training exercises. Did a bunch of training, live fire.

Steve: If you follow military technology, like I do, you realize their type 55 destroyers are probably the most advanced missile ships. They are building a few carriers, but they build carriers not because they think they're going to project power against the U.S., but it's just so ‘if there's some incident in the Middle East and we need to protect our shipping or something - we can project power that way’. But for a real peer-conflict, it's really these missile ships. They're just basically just giant trucks that can fire [missiles].

Adam: A bit like the case in Russia recently, right? In a certain sense, the future of warfare, in both cases, seems to be these kind of motherships with just an enormous amount of munitions that can be at least partially autonomously guided and then you just overwhelm.

Steve: Right. So, my whole point of diverting our attention to the Red Sea is that against a very pitiful opponent, armed with some, not first tier, not second tier, third tier missiles and drones, the U.S., after a year and a half, two administrations, not just some wimpy Democrats, but, you know, hardcore, ‘I will bomb civilians and kill lots of Yemenis’. We, FONOPs failed. If the Chinese wanted to blockade Taiwan or cause real economic problems for the South Koreans and Japanese who import 90% of their energy. They could easily do it, and we would not be able to restore shipping there.

Adam: So I think if there's one theme that maybe ties this whole discussion together is that the world has changed dramatically, and we need to be looking at it with clear and a new, fresh, set of eyes in order to be able to understand how we should operate.

Steve: Yes, accurate empirical data, and first principles reasoning. Otherwise, you might make some mistakes. That's maybe the take-home message from me.

Adam: Steve Hsu, thanks for talking with me.

Steve: It's been a pleasure, thanks.