Transcript

Rob’s intro [00:00:00]

Rob Wiblin: Hi listeners, this is The 80,000 Hours Podcast, where we have unusually in-depth conversations about the world’s most pressing problems, what you can do to solve them, and my favourite documentary Yes Minister. I’m Rob Wiblin, Head of Research at 80,000 Hours.

Nuclear security is a fascinating topic that has been much more salient during 2022 than any of us might wish. But it can be hard to find super informed people who are willing to speak candidly about it.

Fortunately though there’s Jeffrey Lewis. Jeffrey has worked on nuclear weapons and proliferation for decades while remaining a fun and freewheeling academic who hosts a very banter-packed podcast of his own called Arms Control Wonk.

This is very much not a dry technical episode. If you stick around I can say with high probability that you will be entertained by some very spicy personal takes.

The backstory is that a few months ago Jeffrey said on his own podcast that he thought me and other people like me were well-meaning but getting some basic stuff about nuclear security wrong.

Unfortunately he didn’t say what I was wrong about!

So I couldn’t pass that up, and reached out to get him on the show to tell me why I’m wrong to my face.

Because I’m super confused about this issue and if anyone can clear things up… they’d be doing me a massive favour.

I also solicited questions on this topic from regular listeners on Twitter and got some really outstanding ones, so Jeffrey was able to field many of the things you’re most likely to be wondering about nuclear war.

Among other things Jeffrey tells us:

• Why mutually assured destruction is a myth of sorts
• That the US officially says it will not just win but dominate a nuclear war
• Why US presidents secretly think they won’t use nuclear weapons but why Jeffrey doesn’t believe it
• The silliest mistakes the US is currently making with nuclear weapons
• Why a huge constraint on US nuclear weapons reform is rivalry between the the US Army, Navy, and Air Force.
• Why the US and China have somehow never agreed to any treaty about nuclear weapons despite it being in both their interests.
• What bothers him most about the culture of his field — actually multiple things that bother him about the culture of his field
• What the effective altruism community is mostly likely to get wrong about nukes, as well as the most useful things it can bring to the table

One notice before that is that on our other show, 80k After Hours, we’ve released a new interview with Marcus Davis, a founder of Rethink Priorities — a global prioritisation group that has over 40 staff now and who produced some of the work Jeffrey is very broadly responding to in this conversation. Among other things, Marcus talks about careers in generalist research where you study really neglected topics. If you’re interested in global priorities research or cost-benefit analysis in new areas, go check that out on the 80k After Hours feed.

All right, I recommend strapping on your seatbelt and returning your chair to an upright position, because without further ado, I bring you Jeffrey Lewis.

The interview begins [00:02:49]

Rob Wiblin: Today I’m speaking with Jeffrey Lewis. Jeffrey is the director of the East Asia Nonproliferation Program at the Middlebury Institute of International Studies. He’s also the founding author of the popular Arms Control Wonk blog, which he started all the way back in 2004, as well as now being a co-host of the Arms Control Wonk Podcast.

Years ago, he did his dissertation on China’s approach to nuclear deterrence, with international relations legends John Steinbruner and Thomas Schelling as advisors, among others. And today he’s a particular expert on the nuclear strategies and arsenals of China and North Korea. Thanks for coming on the podcast, Jeffrey.

Jeffrey Lewis: I am so happy to be here. Even if it is a bit early in California.

Rob Wiblin: Yeah, one delightful thing about having moved from California to London is I very rarely have to do early morning interviews now. As a night owl, I’m very happy to be doing them more often at 8:00PM instead.

Jeffrey Lewis: And happily, I’m a morning person.

Rob Wiblin: Oh, excellent. Cool. Well, I hope we’re going to get to talk about ways of reducing the risk of nuclear weapons being used in war. But first, as always, what are you working on at the moment and why do you think it’s important?

Jeffrey Lewis: A lot of my work these days is creating open source information about nuclear programmes. Just tracking who’s building what, with the idea that maybe if we know what countries are building, we’ll have some insight into why. So I spend a lot of time looking at the changes in China’s nuclear arsenal, getting ready for the other shoe to drop in North Korea, and just generally keeping a wary eye on the Russians.

Rob Wiblin: I see. So the idea is to use open source intelligence to get thicker details on what exactly is going on in those countries, and then reverse engineering their overall strategy from that?

Jeffrey Lewis: Yeah. A really good example is we spent a lot of time looking at machine tools in North Korean military factories, which sounds really boring. In fact, there are horizontal boring machines, which are boring in both ways. But one fascinating outcome of doing that kind of research is that you will learn things about the status of their nuclear and missile programme that they would never tell you — but they will show you, because they’ll show you the machine where they build the component. And you can figure out when the machine was built, when it was likely imported, and you know the building it’s located in, you know when the building was built. So you can start to reconstruct a history of when important decisions were made about their nuclear and missile programme.

Just as an example of that, we can see when North Korea imported the machines that led to the big breakthroughs they made in 2017 in terms of their liquid propellant rocket engines. So people always are like, “What happened in 2017? How did it go so fast?” And we can say, “Well, 10 years earlier they built a new building and they filled it with machines, and this is now just the inevitable outcome of that thing that you missed at the time.”

Rob Wiblin: Right, right. Yeah, that’s fascinating work. We might get a chance to come back to some of it later in the interview.

Jeffrey Lewis: It’s a weird gig.

Misconceptions in the effective altruism community [00:05:42]

Rob Wiblin: Yeah. The number one reason I wanted to talk to you today, and I guess we shouldn’t waste any time getting into it, is some pretty exciting and spicy comments you made on your own podcast. Saying that you think the effective altruism community — which includes me and many people listening to this interview — potentially have some pretty wrong ideas about nuclear security issues, basically because the issue is so complicated and hard for domain experts to explain well. I guess you excuse us somewhat — saying that you think that a lot of people think about these issues wrong, so it’s not just us — but nonetheless, you’d love it if we could get some better ideas. When I heard you say that, I was super excited, because I feel incredibly uncertain personally and see myself at risk of making big mistakes in this area just due to a complete lack of specialist knowledge.

One complication with this conversation is I didn’t fully realise that there were any clear or consistent positions that the effective altruism community took when it came to nuclear weapons. And in fact, I might have thought the biggest problem we have is that we’re at a bit of a loss to know what to do in this space, to the point that it’s pretty hard to make progress. But maybe that’s just my perception because I’m in the really uncertain camp personally.

Anyway, you said this on your show: “If you look at the work that the effective altruism community does on nuclear risk, it’s as misguided as the Strategic Air Command’s original approach to nuclear weapons.” What ideas or activities did you think of yourself as reacting to here?

Jeffrey Lewis: Well, I think the first thing I should say is it’s probably right that I should never reduce an entire community, but I had an impressionistic reaction to a few different pieces I’ve read. And when you come to a field new, you do tend to observe cultural differences, differences in trends.

So an example of the kind of thing that popped out at me, that really struck me, is: in a lot of different things I’ve seen in the effective altruism community’s writings, there’s a huge emphasis on counterforce versus countervalue targeting — as though that is a meaningful distinction that can be analytically used to understand nuclear policy and targeting and pathways to war. And to me that’s just crazy. Because while these are real terms that have meaning, those terms are fossils from an era that doesn’t really exist anymore. There’s an entire history behind those terms, and if you know that history, you can use those terms — and I use them and other people use them — but that history is so complex and weird and specific that they’re really limited in what they mean.

And so just to explain that a little bit, or unpack that: if you look at what the United States says about its nuclear weapons today, we are explicit that we target things that the enemy values, and we are also explicit that we follow certain interpretations of the law of armed conflict. And it is absolutely clear in those legal writings that the United States does not target civilians intentionally, but that in conducting what you might call “counterforce,” there is a list of permissible targets. And they include not just nuclear forces. I think often in the EA community, people assume counterforce means nuclear forces, because it’s got the word “force,” right? But it’s not true. So traditionally, the US targets nuclear forces and all of the supporting infrastructure — including command and control, it targets leadership, it targets other military forces, and it targets what used to be called “war-supporting industries,” but now are called “war-sustaining industries.”

Rob Wiblin: So that would include all kinds of industry, potentially?

Jeffrey Lewis: Factories, yeah. And so there of course is this constant effort to scrub targets, and move things in and move things out, and reinterpretation. But there just is not this clear distinction. And my favourite example of this, which is now archaic — I mean, here I am talking about fossils — but when the Soviet Union deployed its nuclear-armed anti-ballistic missile system around Moscow, there was a particularly difficult-to-destroy radar that had several dozen warheads assigned to it, and it was in the suburbs of Moscow. So yes, we don’t deliberately target civilians, but on the other hand…

Rob Wiblin: But when bombs start landing in Moscow, it might not be a distinction that the Soviets were super interested in.

Jeffrey Lewis: And it’s really important to understand the reason that this distinction came about was because — and honestly the reason it popped into my mind, and I perhaps uncharitably compared the EA community (which contains a bunch of really nice people) to General Power, who nobody thought was very nice, so it’s a little unfair — but the whole reason that the idea of countervalue existed is because very early on, Strategic Air Command had a fairly simple idea, which is: in a war, we’re going to use everything we have and we’re not going to hold anything back.

And that horrified a lot of civilian strategists, including Bill Kaufmann — who’s sort of the hero of the book Wizards of Armageddon and was my advisor’s, John Steinbruner’s, advisor. And he argued that that’s suicide. If we start hitting Russian cities, they’ll start hitting US cities. So he was one of many analysts who started to argue that we should try to at least refrain initially from hitting cities in the hopes that the Soviets would do the same.

And it’s important to note, we didn’t know if they would do the same. We didn’t know if the same thought had occurred to them; it was pure speculation. And initially Strategic Air Command resisted it. And then they realised that if they embraced counterforce, their list of targets would grow very, very long and they could have all the bombs they wanted. And the McNamara Pentagon — which initially embraces new cities and counterforce, and it’s still with us today in that sense — ends up being not very happy about what they thought was a rather clever trick to rein in the generals.

Rob Wiblin: Right, right. OK, so that’s one area where we’re potentially confused. I guess my goal today is just to get as much information as I can out of your brain — of lots of instances where the audience, or the effective altruism community, or I might just be confused in how we think about nuclear strategy, how we think about nuclear arsenals, and how we analyse the problem when we’re trying to figure out what (if anything) can plausibly be done to improve the situation.

This is obviously a super wide-ranging topic, but one way we can at least slightly constrain it is that on this show we’re probably more worried about nuclear conflict between major powers — so between China–US, Russia–US, rather than a handful of nukes from North Korea — just because the severity of the climatic effects, the severity of the direct effects would be so much larger. All that said, what’s another idea that stands out to you that you’d like people to think about differently instead?

Jeffrey Lewis: I have a list of other ideas, but I wanted to lodge a second complaint, which is more conceptual. I observe that there are a lot of secondary and tertiary sources cited in the literature that I come into contact with. And there are a lot of inferences drawn on the basis of the accounts in these secondary and tertiary sources, which are often really, really wrong. I can think of a specific example where a person asserted that building reprocessing facilities to separate plutonium — which you need for some bombs — was difficult. And the warrant for the claim was that only a small number of countries had built such facilities, and I think Wikipedia or something was cited.

The problems with that claim were: first, more countries had built these facilities than were acknowledged, and that was an issue of knowing what the right place to go look up something like that would be. But then secondarily, and this is a deeper conceptual issue, the US has had a policy over a long period of time of discouraging countries from building such things. And the countries that do build them have been fairly — although not completely — restrained in exporting such capabilities. So it’s not so much that it’s difficult technically for a determined proliferator; it’s that it’s not the kind of business you would get into casually, because of the political constraints.

Rob Wiblin: I guess that’s just a general issue with generalist researchers, or people who are smart amateurs in an area. It seems like a reasonable inference perhaps to draw from not many countries having done something: that it might be difficult. But then firstly, the claim that not many countries have done it might be wrong — if you are not informed enough to be able to tell that that’s mistaken — and then also the interpretation can be quite wrong, even if it’s a plausible one.

Jeffrey Lewis: But it’s difficult, because where would you go to look something up like that? I know as an expert I would say, “Well, you need to go look at the facilities at the IAEA Safeguards, and then you need to know all of the nuclear programmes that are outside of Safeguards” — and I would just make my own bespoke list.

That’s much harder to do if you are a generalist researcher or you’re new to the field. And that ultimately is why I did feel that maybe I should be a little bit kinder in my comments. Because at the end of the day, if my community isn’t producing easy-to-access factual data that explains things clearly, it’s not helpful to throw stones at people who get confused by our crappy work.

Rob Wiblin: Yeah, I think this is just a super general phenomenon. The issue is if you’re an amateur coming into an area trying to make sense of it, you don’t even know the things that you’re plausibly getting wrong, because you’re lacking so much context. So you don’t even know potentially what questions to ask or what things to check more deeply. I’ve found almost the only way to deal with this is to make your best hash of something, and then you have to take it to someone who has spent decades in the field, who can then bring the red pen to it and point out all of the stuff that’s seriously misguided.

Jeffrey Lewis: Although in our field, you have to be careful because people will lie to you.

Rob Wiblin: Oh, really?

Jeffrey Lewis: There’s an enormous amount of gatekeeping and linguistic gamesmanship that serve to make it very hard to get at the truth of a matter.

Rob Wiblin: Tell me more about that.

Jeffrey Lewis: I can give you just a trivial example of it. The US does not have a launch on warning posture. We have a launch under attack posture. And if you look up the definition of launch under attack in the DOD glossary, it used to say — I don’t know if it still says it — “launch on tactical warning.”

But if you say “launch on warning,” people act like you’re some dangerous hippie who has no idea what you’re talking about. Yet the distinction ends up being incredibly complicated, which is: we have the ability to launch on warning, but we say that we do not rely on launch on warning. And that the policy, which is to say a thing that could change is that we would wait for confirmation of a nuclear detonation. But that doesn’t mean that we don’t have the ability to launch on warning. And it doesn’t mean that the system isn’t structured to create incentives to do that.

Rob Wiblin: Or that we plausibly could in the heat of the moment.

Jeffrey Lewis: And so when you say, “We have a launch on warning policy,” and someone says, “Oh no we don’t, we have a launch under attack policy,” what happens is you have this debate that ensues that typically prevents the person from understanding the situation and leads to lack of clarity. Rather than somebody saying, “Here’s the well-thought-out explanation of what we’re really doing.” And that’s why I say it’s gamesmanship, and to a large extent gatekeeping.

Rob Wiblin: OK, we’ll maybe come back to that issue in the next section.

Nuclear deterrence [00:17:36]

Rob Wiblin: But I got the sense that you think perhaps that we, and many people, think about nuclear deterrence in quite a mistaken way these days. Is it possible to elaborate on that?

Jeffrey Lewis: Well, this is really where Bill Kaufmann gives you John Steinbruner gives you me, presumably. I’m like the really bad photocopy of Bill Kaufmann at this point, blurry and smudged.

I think one of the mistakes that exists generally is that there is a sense that states are the important unit of analysis, and that states are relatively rational, and that you can sort of understand what a state does based on what it values.

I tend to look at this through the lens of decision-making. I see governments as being just systems that aggregate the preferences of individuals. So you have a public, and you have bureaucrats, and you have politicians, and you have military officers, and they are cooperating and competing. They have interests that are about the thing they’re fighting about. They have other interests that are not the thing that they’re fighting about, but which impinge on their decision-making. Maybe you don’t care about nuclear weapons, but you do care about getting reelected. Or you don’t care about nuclear weapons, but you care about not looking weak because you care about some other issue.

My work has traditionally focused on how those structures and norms and practices aggregate preferences to produce outcomes. Which is what happens if you have John Steinbruner and Tom Schelling on your committee. Tom Schelling spends a tonne of time doing these wonderful thought experiments to show how rational individuals can do crazy things.

My favourite example — which is nobody else’s favourite example, but I like it — is: he was giving a lecture once, and when he walks in, there’s no one in the first two rows and the rest of the lecture hall is crowded. Now a normal person says, “Oh, they must have reserved the first two rows.” But Tom speculates, what if everyone wanted to sit as close as possible, but no one wanted to sit in the front row. So the first person comes in and sits in the second row and everybody else fills the second row, and then they fill the rows behind, and then only then they begin filling the front rows. So what he creates is a rational individual choice, which is a lot like his work on housing discrimination, that in aggregate produces an irrational outcome — because everybody would be happier if they were sitting two rows closer.

John had a very different take on that, but it was the same kind of idea. John’s sensibility was not that people were having these sort of micro-motives that produced macro-behaviors, but rather that the world is too complex for us to understand. So he enjoyed the example of a tennis player. If you were to try to mathematically model the act of striking a tennis ball with a racquet, that is an impressive piece of mathematics. Tennis players are dumb. They don’t do the math.

Rob Wiblin: Didn’t go to grad school.

Jeffrey Lewis: Right. This is a field that used to be called cybernetics. The idea that you could have what today we describe as algorithms that govern our behaviour in a complex situation. And you might not optimise all the time, but over time you would do pretty well. So I think each of them were concerned with how individuals operating in big structures produce these outcomes, whether that involves housing discrimination or nuclear war.

Rob Wiblin: As I understand it, a lot of people, including me, think about nuclear issues at a pretty high level of abstraction — where we’re going for the sort of pure game theory that I might have learned in second-year microeconomics. That’s a very natural way for me to think about this issue: if X does this, then Y will respond that way. And it’s quite a clean analysis.

A different way of looking at it would be to say that if you have a much thicker understanding of the powers that different individuals in the system have, and the different committees that something would go through, and how the information flows from the radar to the president, then you can instead think about it at the institutional level: How is a decision actually being made here? And given that, what decision is likely to be spat out by the system? And that’s at a much lower level of abstraction than some of the purest game theory.

Jeffrey Lewis: I think that that’s, again, the joy of dealing with both Tom and John at the same time. Because Tom really liked those game theoretical situations, but he was also aware of the ways in which they break down in reality.

I think the simplest example of this is: how do you even know you’re under attack? It’s not a video game. It’s not that you’re told that there’s an attack, and you’re told this accurately, and you’re given the correct number, and you’re given it in a prompt and predictable way. You know, we have radars and we have space-based sensors that look for infrared plumes. So that kind of data that comes in, there are often false alarms. There are things that are missed, there are ambiguities. It’s handled at an extremely low level, and it’s reported up a chain of command. And by the time the president’s told it, the president’s not looking at the radar image. The president has a person conveying words that are meant to describe this complex phenomenon that’s out there happening in the world.

And so when you realise how little control each individual really has in that process, on the one hand, it’s kind of a bummer, because it’s like, “Oh my God, we’re all going to die.” But on the other hand, it becomes a really rich and complex story.

And John F. Kennedy got this really well when he had ordered a stand down to all U-2 flights during the Cuban Missile Crisis, because he knew this was exactly the kind of thing that could get out of control. And he did not have total control, but a U-2 flight still went off and somebody got shot down. And Kennedy’s response, which is like peak Kennedy, was, “There’s always some son of a bitch who doesn’t get the message.” And I think that’s a really good way of looking at dynamic complex systems.

Rob Wiblin: Right, yeah. So I suppose in theory, it seems like you could try to incorporate all of this stuff into a game theory matrix, but by the time you’ve done that, it’s so enormously confusing that common sense and thinking through the process without a table might serve you just as well.

Jeffrey Lewis: I think you have to do both. I do not in any way, shape, or form reject game theoretic models to try to understand how decision-makers might act. I just think it is important to be familiar with the particular animals in the zoo. Particularly because I see a lot of predictable decision-making pathologies.

Rob Wiblin: Yeah, what’s an example?

Jeffrey Lewis: I will tell you. One of Tom Schelling’s great books was written with this guy, Mort Halperin. There’s a triquetra of important figures in my life, and Mort is one of those three.

Mort tells a story about the time he’s in the Defense Department that illustrates the pathology. It needs a name; I don’t have a good name for it. But he gets this speech that the Secretary of Defense, or maybe it’s the undersecretary, but a speech a senior official is going to give. And unlike a typical Defense Department speech, it has pages of detail about Chinese internal politics, and this figure named Lin Biao, who was Mao’s heir apparent, who will later die in very strange circumstances. Mort cuts almost all of that out and brings the speech back, and the feedback he gets is “too much Lin Biao.” So he takes out a bunch more and he brings it back, and he’s told still too much Lin Biao. And Mort, in frustration, says, “I have taken out 95% of it.” And the answer, and this is the decision-making pathology, is: “No one knows what you took out. They’re only going to know what’s still in there.”

I see all the time decision-makers talk themselves into terrible decisions, because the terrible decision they’re making is not as stupid as the four other things that were put on the table. I guess it’s the decision-making equivalent of the Overton window, where you make one proposal that’s absurd, and suddenly a really bad idea looks reasonable by comparison. And you imagine your opponent is going to see that.

I wrote a novel about a North Korean nuclear war, and in it, the South Koreans talk themselves into what they see as a limited and de-escalatory missile strike on the North Koreans because it’s much smaller than the plan they have on the books. But the North Koreans don’t know the plan on the books — they only know the missiles that hit them.

Rob Wiblin: Does that sort of thing happen just because bureaucracies are messing up? I think there’s some famous Yes Minister scene where they talk about how they get the decision that they want by ensuring that the least stupid decision that they put on the table is the one that the bureaucrat prefers. Is that also going on?

Jeffrey Lewis: Yes. Yes Minister is the finest documentary on decision-making that I have ever seen. I cannot say enough about that television program. It, to me, fundamentally captures the dynamics that we see play out all the time.

And one thing I would say is that that gets cleaned up in histories. Often when we read about accounts after the fact — and I’m now thinking of a particular account in my field, and I just don’t feel like starting this fight — I know of many cases where individuals take one side in an internal argument and they lose. And then as a public official, they’re required to defend it. Then in a memoir, they make the decision sound like everybody agreed and it was all very straightforward and it made perfect sense. And it completely obscures the nature of the fight, which is often quite petty and has very little to do with the issue that’s on the table and much more to do with issues that surround it.

So the thing I loved about Yes Minister is that I know it’s a comedy, but I have seen that behaviour in Washington time and time again. And honestly, the work of Armando Iannucci, I don’t know if you’ve seen The Thick of It? That, to me, is much more true than any serious drama on those subjects I have ever seen.

Rob Wiblin: Yeah. We’ll stick up some links to some top clips on Yes Minister, and maybe from Iannucci as well, for anyone in the audience who has had the misfortune of not watching them already.

Dishonest rituals [00:28:17]

Rob Wiblin: Earlier you mentioned that people lie to outsiders from this community. Do they perceive themselves as lying because they have an agenda that they’re seeking?

Jeffrey Lewis: No.

Rob Wiblin: OK, can you elaborate on the lying?

Jeffrey Lewis: Well, let me say, some of them may. There is always a little bit of, “I’m smarter than you and I can get away with this.” But normally what it is is that there’s a ritual that must be performed in how one says things, and that ritual is typically enforced.

So one trivial example is if you say, “The United States is going to use nuclear weapons,” you will quickly be corrected. You will be told, “The United States uses nuclear weapons every day for deterrence. The word you’re looking for is ’employ’.” And indeed it’s called the “nuclear weapons employment guidance.”

So it’s not, I suppose, that people know they’re lying — although there are individuals I think who are dishonest — but there is much more than that. There is a kind of culture of, “This is how we talk about these things. This is how we phrase things. This is how we describe things” — and it makes it very hard to dissent and to disagree.

Rob Wiblin: I see.

Jeffrey Lewis: I recall I had an argument with someone once, where I was arguing for a new policy as an outsider, and the person kept telling me that that’s not how they do it. And finally I got frustrated and said, “I know that’s not how you do it. The way you do it is stupid. Do it a different way.” And that person looked at me like this was the most bizarre suggestion. So there really is a kind of orthodox practice that one is expected to adhere to.

And so when one comes in from the outside, I always get a chuckle. Not a negative chuckle. But I see in the EA language a lot of “de-proliferation” or “anti-proliferation.” And that’s the kind of stuff that would immediately get you treated to a very tedious lecture about why the phrase is “nonproliferation” and how nonproliferation really only refers to this very narrow area. And you would just be listening to this with your eyes rolling in your head, like, “This is good faith engagement. OK.”

Rob Wiblin: Right. I think listeners might be able to pick up that you are someone who is happy to provide people with the benefit of their forthright opinion. Is that in part a reaction to this? It sounds almost like there’s a culture of stonewalling people who are not insiders in this community, and you’re just like, “No, I’m going to crash through that, and I’m not going to be intimidated.”

Jeffrey Lewis: No, it’s a function of a couple of things. I grew up in rural Illinois, and that is a place of extreme politeness, because otherwise you’ll get stabbed. But it’s definitely not a place where once the politeness falls, you don’t keep playing these weird linguistic games. So just culturally, it was very odd to me.

And early on I enjoyed it, because I had these fantastic mentors and I’m fairly OK with words, so I could learn those games and I could play them really effectively. But I think just over time, when you learn something like that, it loses its interest. It’s very predictable. I almost never go to talks in my own field anymore, because I know what people are going to say. It’s like the old joke about the Comedians’ Convention where somebody says ‘Number 47!’ and everyone laughs — that’s what it’s like. So I found it tedious, and it’s antithetical to thinking clearly and creatively. At some point I just thought, “Do I really want to do that with my life? Do I want to go to the Pentagon every day and…”

Rob Wiblin: “Do I want to be that person?”

Jeffrey Lewis: And it’s like, “No. No, I don’t. I can live in California and not wear a tie and say exactly what I think.” And that’s a good life, for me at least.

Downsides of generalist research [00:32:13]

Rob Wiblin: Excellent. OK, coming back to a kind of list of misunderstandings or bad ideas, are there policy suggestions or ideas for making the world better that you’ve seen come from the effective altruism community or adjacent groups, which you think are seriously wrongheaded in some way?

Jeffrey Lewis: I have not seen a lot of EA-specific recommendations. I have seen a fair amount of EA writing that tries to put numbers on things, which is an effort I generally support, that tries to understand what things are likely to happen. But what I haven’t seen, and I have had a couple of conversations with friends of mine in the field about, are what bets we might plausibly make that would be influenced by this particular type of thinking about the world. But I don’t know that those exist yet. Maybe I’m wrong. Maybe there is the nuclear equivalent of mosquito nets, but I don’t think there is.

Rob Wiblin: Yeah, I think sadly there’s not. The reason that we have generalist analysts is that the EA community as a whole — and I guess global priorities research as a discipline — is still very significantly at the stage of trying to prioritise between different problems in the world.

So you need to have generalists, because we don’t want someone who only knows about nuclear issues. We want someone who knows a bit about that, but can compare it with risks from pandemics or bioweapons or new weapons that might be developed — and compare that with global health potentially, and compare that with climate change, and so on. So someone who’s doing that is pretty liable to make errors in each of their individual analyses, especially if they don’t check it with a domain expert in each case. But I think I’ve also seen a case where domain experts try their hand at this higher-level global priorities research stuff, and they make terrible mistakes, because being a generalist is a specialism of its own in a way.

Anyway, when you’re doing the global priorities research stuff, you’re very interested to know what is the annual risk of nuclear war, because that really features very prominently in the equation. So a lot of it has been trying to do this risk assessment, figuring out how bad is the present situation. Then of course, that’s almost an easier question than figuring out what stuff you want to change in order to make it better.

Jeffrey Lewis: John Steinbruner had an analogous interest, but instead of imagining these different priorities that are competing — which is a perfectly reasonable thing to do when you’re trying to allocate funding — he was really interested in three, or sometimes four, sets of problems. The three big ones being climate change, nuclear war, and pandemics. His thinking was that they were essentially the same problem, in that they were challenges that were too big for us to solve at a state level. That a set of solutions pursued by sovereign states on individual bases would not fix these three problems.

So his intuition, which I think is right, is that the big things we need to do — not like the little policies, but the big conceptual stuff about how we think about our place in the world, how we organise our diplomacy — are pretty much the same across those three areas. Because they’re all cases where we have to learn to pool our sovereignty with other countries, work with countries that we might not like in pursuit of some shared interest.

And honestly, it’s the reason that I find nuclear risk to be such an interesting field: because it is, in some sense, the first real problem that human beings face where it’s big enough that we might not make it. It’s not the only problem; I’m just saying it’s the first. We had violence for a long time, but it’s the first time that the scale of violence is such where you’re like, “You know what? I don’t know if we’ll survive this in a meaningful way.” And so you realise that no matter how much you hate the Soviets, that you have a shared danger that compels you to cooperate. And to me, that’s a wild idea. That’s a big thought.

Rob Wiblin: It’s challenging for people to get their heads around it. And it did take decades, I think, for people to fully embrace the idea that you had to cooperate with a group that you thought was evil, and they thought that you were evil. But nonetheless, you were better off together than apart, in an odd way.

Jeffrey Lewis: I don’t know if you ever saw the original movie, The Day the Earth Stood Still.

Rob Wiblin: No, I haven’t.

Jeffrey Lewis: So Americans, we’re bad at big ideas. So when we confront big ideas, we either throw it in a Western or throw it in a science fiction movie so that we can get our very pragmatic little heads around it. The Day the Earth Stood Still is a parable about nuclear weapons, where an alien shows up and says, “You settle your disputes with violence, and that’s your business. But now that you do space travel and nuclear weapons, your violence threatens to spill off your planet. And so, you have some short amount of time” — like 72 hours — “to sort your lives out, or we’re going to destroy you.”

And the alien is nuclear weapons. It’s this idea that there is this external threat that compels you to cooperate even with people you don’t like. And it’s a very powerful idea. It’s the reason Ronald Reagan used to always say how our differences with the Soviet Union would fade if we were ever confronted with aliens.

Now, I think that’s a very optimistic reading of human history. As someone who’s also a member of a Native American tribe, no, some of us will cooperate with the aliens. Sadly, that’s what we’re like.

Rob Wiblin: And some of us will claim that the aliens don’t exist probably, as well.

Jeffrey Lewis: Yes. But that tale, I think, does work for people to try to think about how you could be facing a threat that is too big for you to solve on your own, and it forces you to work with people you don’t like. Which is tough.

“Mutual assured destruction” [00:38:18]

Rob Wiblin: Zooming out a little bit, are there any bad ideas that you’re constantly seeing among the general public or in mass media? That you just bang your head against a wall, like, “I can’t believe I’m having to read this nonsense again.”

Jeffrey Lewis: Every time someone says “mutual assured destruction,” I want to take a hammer and crush my own skull.

Rob Wiblin: OK. I occasionally use that expression.

Jeffrey Lewis: There is no such thing. No such thing.

Rob Wiblin: OK. Tell me more.

Jeffrey Lewis: The important thing to understand is that the context that gave us counterforce and countervalue, that early 1960s effort to impose some civilian control over Strategic Air Command also gives us this phrase, “mutual assured destruction.”

What happened was, once Strategic Air Command embraced counterforce, and said, “This is great, we can still hit cities, but we’re going to have 20 times as many nuclear weapons, because I’m hitting the factory or the Kremlin,” the Kennedy administration was like… Can I swear on this podcast?

Rob Wiblin: This is a podcast for grownups, we’re talking about nuclear war here. So yes.

Jeffrey Lewis: Proverbially, they were like, “Shit. We have now created this situation where we’ve justified much larger increases in nuclear forces than we really wanted to.” So largely a fellow named Alain Enthoven comes up with this idea of a thought experiment. So this is a thought experiment. This is not a real thing; this is just a way of thinking: How many nuclear weapons do we really need? Is there some theoretical limit after which more is worthless?

So he imagines that we have a force of one-megaton bombs. Now, we don’t have a force of just one-megaton bombs, but it’s a thought experiment.

Rob Wiblin: So one megaton, just for reference, is not that big, is that right?

Jeffrey Lewis: No, no. It’s huge. Hiroshima was 10, 15 kilotons.

Rob Wiblin: I see, OK. 100 times that.

Jeffrey Lewis: Exactly. EA people can do math. That’s good. It’s nice. Numeracy, it’s the new literacy.

So he imagines you have this force of these very large bombs, and you just drop them sequentially on the Soviet Union to maximise damage. Literally not our policy, but this is a thought experiment.

And then he draws two curves: one showing the damage each successive bomb does to Soviet industry and to Soviet population. And as you might imagine, the first bomb you drop on central Moscow does a lot of damage. The second one on Leningrad at the time does not quite as much, but a lot. But eventually the curve levels off, and it levels off at about 400 bombs. And the thought experiment is complete. The view is: look, the Soviets don’t know what we target. They don’t know what we’re going to do. But if they know that we can put 400 megatons on the Soviet Union, then after that we’re just making the rubble bounce and we have enough.

And so the Kennedy people try to cap the arsenal, calling this “assured destruction.” We just need to have the ability to assure the destruction of the Soviet Union. And the people who want more nuclear weapons hate it — they say it doesn’t have a theory of victory. And so to mock it, it’s a calumny: Donald Brennan and Herman Khan and others start calling it “mutual assured destruction” — which is to mean suicide — or they call it “assured vulnerability.” These are like attacks.

And “assured destruction” is just a Kennedy thing. It lives on into Johnson, but it’s not anything any future president really accepts. Carter has something called the “countervailing strategy,” which is about us winning a nuclear war. Reagan, his guidance talks about us “prevailing” in a nuclear war.

The reason I think mutual assured destruction is so sticky as a concept is, even if it’s not a real thing, it is also the case that even at the time that Ronald Reagan has a paper that he signs that says, “the policy of the United States is to prevail in a protracted nuclear war of the Soviet Union,” he also announces that a nuclear war can never be won and must not be fought. So his policy is to win a nuclear war. And he says, “I don’t believe that.”

I do think that presidents often look at this whole business and say it’s insane. And then they silently dissent or defect. But they don’t change the policy; they become disconnected. And so I think the reason it sticks is because we often want presidents to admit publicly that they have defected from this, in the hopes that if they say things like, “I believe in mutual assured destruction, even if it’s not really a thing,” it would lead to policy change.

Rob Wiblin: OK, I don’t quite get how it’s not a thing. Maybe on paper it’s not in the policy anymore. Maybe it never was called that in the official policies exactly. But isn’t it still the case that when Russia and the US are fighting a proxy war in Ukraine, the fact that they could both destroy one another — and that if things got to some point, they would both bomb one another and both of their societies will be ruined — doesn’t that, and their risk aversion, kind of constrain their actions and affect what they do?

Jeffrey Lewis: That’s just deterrence. The whole idea of mutual assured destruction is you accept the possibility of reciprocal deterrence. And our policy is not to do that. Our policy is to be able to prevail.

Rob Wiblin: I see. I thought that the US policy to some degree was to accept mutually assured destruction, and maybe build into the treaties with Russia that there’ll be a sufficient arsenal that both countries would be wrecked. But actually that’s not what’s going on. The US, does it still have an idea that it could win a nuclear war with Russia?

Jeffrey Lewis: Yes. Yes. Go look at the declassified OPLAN.

Rob Wiblin: How would that play out?

Jeffrey Lewis: So the idea is you deter, and if you failed to deter, I believe the next phase is “Shape.” Actually, I have a declassified OPLAN in front of me. I can look this up. Sorry, this is so goofy. I really hate the way they phrase these things.

There’s phase 0, which is “Shape.” Phase 1 is “Deter” — that’s where you’re attempting to deter a threat.

Rob Wiblin: So “Shape” is kind of shape the situation through your general foreign policy all the time. And then at the point where there’s a conflict on the table, then you’re engaged in deterrence.

Jeffrey Lewis: Exactly. Phase 2 is “Seize the Initiative,” which is, I don’t know, Deterrence+. And phase 3 is “Dominate.”

Rob Wiblin: And how exactly would the US dominate Russia in a nuclear war?

Jeffrey Lewis: So that is redacted. [laughs] So all we have is, “During this phase, USSTRATCOM, in collaboration with the applicable GCCs, conducts operations, integrated across all mission areas, in order to defeat adversaries and return to phase 0 on terms acceptable to the US. The USSTRATCOM, as directed by the President — redacted.”

Rob Wiblin: I see. So you can see why people think that actually the policy is mutually assured destruction.

Jeffrey Lewis: Yeah. Because it’s freaking madness.

Rob Wiblin: But isn’t that actually what’s happening? That the reason the US and Russia tread carefully around one another, the reason that they actually are deterred from getting too close to shooting one another, is because mutually assured destruction kind of is the result?

Jeffrey Lewis: I agree with you to a point, but this is where my work maybe causes us to diverge. As I said, I believe that presidents silently defect from this. I think they get the briefing and they say —

Rob Wiblin: “What the hell is this?”

Jeffrey Lewis: — “That doesn’t seem very sensible to me.” But because it’s a political loser, because it’s abstract, because it’s complicated, they allow the plans to be written, they spend the money to acquire the forces, they select senior officials who believe in the mission. And what I think they have concluded is that because they are silently defecting, they know they’ll never do it.

Rob Wiblin: I see.

Jeffrey Lewis: And I think that is a terrible mistake. Because all the time, we see presidents get boxed in because they overestimate their power and they underestimate the structural constraints that will exist in terms of time, their own stress level, the advice they’re getting.

There’s a great example of this in Fred Kaplan’s other great nuclear book, where there’s a war game that’s being fought. And the war game occurs at two levels: the so-called “deputies” level and then the “principals” level. The deputies are the deputy secretary of state, the principal is like the secretary of state. So you do it once with the number twos and then once with the number ones — you play the game twice.

And in the game, Russia uses one nuclear weapon — which I don’t think is very realistic, but we’ll just roll with it. The deputies decide one nuclear weapon is not going to make a difference to the war. Terrible risk of escalation. We’re just going to go ahead and win the war. And the punishment for the Russians is they lose the war and everybody thinks they’re awful.

They play it again at the principals level. And the principals decide they must respond with a nuclear weapon. They feel that their credibility is somehow on the line, in a way the deputies don’t feel. The deputies can be more dispassionate in some sense. But they’re terrified to hit the Russians back with a nuclear weapon because the Russians using a nuclear weapon against Poland is one thing. Us actually hitting Russia with one of our nuclear weapons, potentially very escalatory. And so they flounder around for a while before deciding that they’re going to nuke Belarus, which is not even a party of the conflict in the game.

And to me, what that illustrates is when you are in a situation that you have not fully anticipated, and you’re kind of making it up as you go along, and if you’re feeling some stress, and if you’re getting some advice that’s maybe not great, it’s maybe hard to silently defect anymore when you have invested all of this time and energy in creating this system.

So that’s why I say, even though I don’t think mutual assured destruction is a thing, it’s not a policy, it’s something the US rejects. I think presidents, when you have them in a calm moment, they know that’s madness. I am not worried about what they’re going to do in a calm moment. I worry what they’re going to do in a not calm moment when that happens.

Rob Wiblin: What do you mean by a not [calm] moment? So I suppose a typical person, maybe me, worries about the situation where there’s a false alarm. That they think the US is being attacked by hundreds and hundreds of nuclear weapons, but actually they’re not. And the call comes at 3:00 AM, and then maybe they’re stressed out and they decide to retaliate mistakenly. Is that what you mean by a not [calm] moment?

Jeffrey Lewis: That is definitely a moment. I have a different set of scenarios in my mind, but what I would say is that a 3:00 AM phone call is probably not going to result in a nuclear war unless it happens against the backdrop of a bunch of other crises — which by the way, we have seen. 1983 was a terrifying year. There were false alarms, but nothing lined up. But what’s terrifying about 1983 is how the background kept getting amped up, and that a false alarm at the wrong time might really have been misinterpreted.

Rob Wiblin: I see.

Jeffrey Lewis: So I do worry. But more than that, I worry about escalation. The US is terribly worried about a particular scenario, which I think is quite unlikely, but it’s the Russians use a single nuclear weapon. That’s the scenario in the war game I mentioned, and they do it to signal. And we are really committed to the idea of signalling back. To the point that the Trump administration created, and Biden has kept, a low-yield nuclear weapon — which is still like a Hiroshima-sized nuclear weapon — on our submarines, with the express purpose that if they were to use a nuclear weapon in Europe, we would hit a target in Russia with one of our own. And we would do this off a submarine, which by the way also carries all the strategic warheads that we would use for a big strike.

The Russians have been absolutely clear that if they see submarine-launched ballistic missiles coming at Russia, even if it’s only one, they’re going to assume that there are more that they can’t see, and they are not going to wait. They’re not going to wait for the thing to explode and then measure its yield and then decide what to do. They see something come up out of the water, it’s go time. So we have this fundamental unresolved issue, where we have a debate on our side about whether or not they really mean that, and they have a debate about whether we’d really do that.

So when I say a “not calm moment,” I mean a situation where there is a crisis — where either the Russians have used a single nuclear weapon, or maybe we just think they have — and the president is now asked to execute the strategy which he publicly supports. And it’s on the piece of paper he signed, and he’s put the weapons there, and he’s being told it will be fine. And maybe it will be fine. Maybe it’ll be 100% fine. I 100% worry it’s not. And that’s what I mean by not a calm moment.

Rob Wiblin: OK. You are saying “calm moment,” C-A-L-M.

Jeffrey Lewis: Yeah.

Rob Wiblin: I was hearing C-A-L-L. But that was a fantastic answer. I was hearing “call” and “not call.” But that was a very instructive answer nonetheless.

Jeffrey Lewis: Two countries separated by a common language. But see, people hear your accent and they think you’re smart. People hear my accent and they’re like…

Rob Wiblin: Oh, I’m not British. I’m Australian, I’m afraid.

Jeffrey Lewis: It doesn’t matter.

Rob Wiblin: I don’t know. They know I’m not from where they are, but they often can’t place it. But yeah, we’re jumping around all over the place.

Budgetary considerations for competing parts of the US military [00:51:53]

Rob Wiblin: I really want to ask you about something that I’ve heard consistently, from listening to some of your show, talking to other people about nuclear weapons and the history during the Cold War: this impression that nuclear policy was not set by strategic considerations as much as it was set by budgetary considerations for different competing parts of the US military.

Jeffrey Lewis: No, it’s just a total coincidence that the budgets for the three services are exactly the same, and that the Navy and the Air Force are both in the business, and that we have missiles on land and missiles at sea, but they’re not the same missile. Just a total coincidence.

Rob Wiblin: Yeah, can you elaborate on what effect that has had? And maybe also, if you’re a general — one of the people leading one of these forces, the Air Force and the Navy — don’t you take your responsibility to the United States, to its security, seriously enough to —

Jeffrey Lewis: No.

Rob Wiblin: OK. [laughs] Go on, go on.

Jeffrey Lewis: This is like the old joke about the two US nuclear labs. In the US, we have two design labs, and they used to say it about each other. It would go something like, the folks at Los Alamos would say, “The Soviets are the adversary, but Livermore is the enemy.”

We have seen consistently in the US very high levels of inter-service rivalry, and we think that’s generally a good thing, right? That’s our competition, free-market idea. The last real effort to unify those services was the Truman administration, which really wanted to weaken the Navy and abolish the Marine Corps, and that led to the very famous Revolt of the Admirals, which was a pretty severe crisis in civil-military relations. As it happens, I think Truman was wrong about the Navy, but right about civilian control of the military.

So no, each of our military services has their own unique camouflage patterns. What I would say, charitably, to those folks is that because they have distinct service cultures, they sincerely believe that their way of war and their contribution is the best one. And you know, you pick, right? So if you go into the Navy, it’s because you believe in the need.

Rob Wiblin: So you go into the one that you back? Yeah.

Jeffrey Lewis: And you’re inculcated with a worldview. So I don’t think people are purely craven, but I do think that it’s very telling that the thing that I say that makes people the angriest is when I suggest the US is about to buy a new ground-based ICBM. We have a perfectly serviceable missile on our submarines that you could put in those silos. It would cost a fraction of replacing the ICBM with a new one. And the Air Force hates it. And I’ve had this argument over about 10 years now, and every two, three years, the Air Force has a new reason they can’t do it. It’s never the same reason.

I think what it ultimately comes down to is that it’s just this service pride and prerogative and autonomy that we see play out. I mean, honestly, we see it in counterforce. The reason that the Strategic Air Command embraced counterforce when it was presented to them was precisely because the Navy had started to get in the business of deterrence, and the Navy systems were inaccurate and they could really only hit cities. And so this was a moment in which Strategic Air Command could say, “See, we have something better.”

So you have all these Navy officers who talked about how terrible nuclear weapons were, you shouldn’t target cities. Until they get inaccurate missiles — and then they talk about how minimum deterrence is the only way to go, and we don’t need all these Air Force things. Until those missiles become accurate enough, and then they go right into counterforce. It’s hard not to be cynical.

Rob Wiblin: Right, right. I mean, I think that this is an area where someone who’s kind of idealistic like me is at risk of producing just completely useless analysis. Because my assumption coming in would be that we need to figure out a way in which you can preserve the United States’s security while minimising the risk of accidental nuclear war, or minimising the risk of escalation. Whereas, actually, the question you need to ask is how can we minimise the risk of escalation, the risk of nuclear war, without damaging the Navy’s budget? That’s the actual issue at play.

Jeffrey Lewis: Yes.

Rob Wiblin: And that might call for very different options.

Jeffrey Lewis: Yes, I think that’s exactly right. If you look at one of the systems that the Obama administration was able to eliminate, it was the sea-launched, nuclear-armed cruise missile, which had been in storage since George H. W. Bush announced a series of initiatives in 1991. I think in ’92 it went into storage, so it sat there for two decades. The way that the Obama administration was able to retire it is that the Navy hated it.

Rob Wiblin: I see. Why?

Jeffrey Lewis: Because in general, the military is not the problem with nuclear weapons. There are elements in the military that are very attached to their nuclear weapons, but Navy officers who drive attack submarines — which is where this thing would be deployed — don’t want to spend the time training to handle nuclear weapons for what they consider to be a fairly unlikely scenario. They want to go out and practice sinking ships and other submarines. That’s what they’re interested in doing.

Rob Wiblin: Mmm, that’s cool.

Jeffrey Lewis: Yes. So the idea that they would have to select a few boats every year to go do this training, to upload this archaic system that they thought they were never going to use, made them really hate it. It was actually the civilians that Obama appointed that wanted to replace the system.

Rob Wiblin: OK, right. What are some distortions in the US’s military arsenal today that are generated by these budgetary competition concerns?

Jeffrey Lewis: The existence of land-based ICBMs strikes me as one of the major ones, because we often hear that the United States cannot ride out a nuclear attack. We have to be able to launch under attack, because our missiles and silos are vulnerable.

Tom Schelling wrote a lovely article in the ’80s where he just pointed out that if you didn’t have missiles and silos, you would not have this problem. He actually called land-based missiles an embarrassment. You could put them all on submarines, which is what the French do. I mean, the French have a small aircraft-delivered capability. But it’s what the British do. If we had two or three times the number of submarines at sea, attacking the US would be hopeless. And you could even leave some of them in port, if you wanted to make sure that the Russians had to strike a US city. Right?

Rob Wiblin: Yeah. So which branch does the land-based silos? Is that the Army?

Jeffrey Lewis: No, it’s the Air Force. The Air Force won this in a battle with the Army.

Rob Wiblin: OK, well, if the Air Force can do land-based silos, why can’t we just get the Air Force to do submarines instead? Maybe they can get the budget to build their own set of nuclear subs.

Jeffrey Lewis: You know, turnabout is fair play, because the Air Force was very against the Navy having its own air component.

Rob Wiblin: Right, right.

Jeffrey Lewis: That was a huge fight. Yeah, the Army got air defence missiles. So missile defence systems are Army — not that we have nuclear-armed ones anymore, but that was their mission — and the Air Force got the strike missiles. That was a huge fight, by the way.

Rob Wiblin: Did the Marines get their own nukes to play with? Not the Coast Guard? No, they got screwed? OK.

Jeffrey Lewis: No. Although the Marines are now, for the first time, getting longer-range conventional missiles, so maybe someday. But yeah, that stuff matters. That service prerogative absolutely matters. I’ve seen an Air Force general fight with an Army general about how useless missile defence is, and you will not be surprised that the Army general was in favour of it, and the Air Force general was like, “You go with offence. You don’t shoot the missile down. You deter and you defeat, or dominate.”

Rob Wiblin: Yeah. It makes me feel like I’m approaching decision-making and bureaucracy within 80,000 Hours wrong. I should just be trying to maximise the podcast budget, finding excuses to hire dozens of people to expand our empire.

Jeffrey Lewis: I mean, you joke about that, but if you look at think tanks in DC, they all exhibit steady growth, right? There is an ideology of growth. No one sits down and says, “You know, I think eight people is the right number for this think tank.”

Rob Wiblin: It’s an oddity, I suppose. Yeah. Or you come in and say, “This is 10% too big.”

Jeffrey Lewis: Yeah, no one ever does that, right? People build empires. Maybe Washington is an exception this way, but I don’t know. I see this stuff on display in other places.

Rob Wiblin: For sure. Another cynical explanation I’ve heard for distortions in nuclear policy is that there are particular companies that make a whole lot of money manufacturing all of these weapons and systems, and of course, they have advocates in DC defending their interests, I suppose, and explaining the value that’s being provided, and they might have larger budgets than the nonproliferation folks do. Do you think that is an important issue as well, anywhere near the same level as these inter-service rivalries?

Jeffrey Lewis: I don’t think it’s as bad. I mean, it’s definitely the case that defence contractors want to make money and that they lobby for things, but my sense of them is always that as long as the money shows up, they’ll do anything. So if we had a sensible agenda for reducing nuclear risk, what we would need to do is simply account for the fact that they’re not going to argue themselves out of business, and so we’d have to imagine what their business lines looked like.

And honestly, it’s what the Clinton administration did with the nuclear test ban, right? The US stops nuclear testing in the ’90s. Clinton signs a treaty that prohibits all nuclear testing, and it corresponds with a massive investment in what’s called the Stockpile Stewardship Program, in which the labs — which are privately run; at the time, the two design labs were both run by the University of California, but now they’re run by a consortia — are promised generous funding for these new activities. And the labs got behind it.

Where the effective altruism community can potentially add the most value [01:02:15]

Rob Wiblin: I see. OK, pushing on a little bit, where do you think the effective altruism community can potentially add the most value to this space, given our particular strengths and weaknesses?

Jeffrey Lewis: Well, definitely the fact that there is funding gives you power. The thing about my community is it is in a constant state of funding contraction. So just giving money doesn’t fix the problems, because it just replicates the problems that we’ve had in the past, but I would be not honest if I didn’t say people will pay attention — because you are potentially the difference between them being employed or not.

What I think substantively you bring — what I’d like to see that influence used for — is the analytic rigour and method I see on display. The insistence that we try to estimate the likelihood of nuclear war. The insistence that we try to understand what the pathways are with some effort at quantifying them, that we be explicit about models of change. All those things that you all do naturally — or maybe ideologically, because it’s how you look at the world — those are the things that are missing in my community. Which, by the way, is why people get worked up about it, because as I said in the post, it’s why dinosaurs don’t like asteroids.

Rob Wiblin: I suppose my natural way of thinking would be to say: First we analyse what is the background rate of nuclear war now. And then we might consider various policy changes and try to figure out: Has that reduced it by 10%? Has it made it more volatile? I mean, obviously these things are all estimates. We’re not robots who just go through following these things without any critical analysis. But the desire to roughly put numbers on things, to sense check them, and also to aid communication between people — to make sure that they’re at least broadly understanding what one another is trying to say.

Is that threatening to people who are part of this existing way of thinking about things? I suppose it would potentially damage some of their analysis, or they just don’t feel comfortable doing it personally?

Jeffrey Lewis: I think they just don’t feel comfortable doing it personally. Sports analogies are so overused and I try not to use them, but I really do think a lot about the 1970s, when the first efforts were made to look at baseball analytically. And baseball fans hated it. I mean, just despised it. And it was clear we were not collecting the right kind of data and weren’t understanding what resulted in teams winning games. I always find the resistance to it fascinating, because don’t you just want to win the game? And the answer is, “No. I want to win the game playing in a certain style, because I have this emotional attachment to it.” I think it’s just hard for people to let go of those things.

Rob Wiblin: Right, right. Yeah, just spitballing, another thing that I think the effective altruism community might be able to bring to the table is just a very strong open-mindedness about what might be good. We really don’t have any dog in this fight at this point. I certainly don’t, and as far as I can tell, other people don’t. There’s a lot of open-mindedness to different approaches to funding different ways of moving things, and maybe that’s a benefit. I suppose things might not stay that way in decades’ time, but at this point, I think nothing’s really been crystallised.

Jeffrey Lewis: I think that’s right. There are two aspects of that that I find valuable. One is being non-ideological. These debates have gotten pretty stale. And as somebody who enjoys generating new policy ideas, let me tell you, the thing that I do that gets the least positive reaction is suggesting that the way we’ve been going about something is wrong and we should try something different. There’s even this phrase which comes out of the military: “the good idea fairy.” The idea that you’re coming along with good ideas —

Rob Wiblin: Wait, what? I’ve never heard this concept. That’s so great.

Jeffrey Lewis: The good idea fairy. Like: We have this way we’ve been doing things, we’ve been fighting the good fight, we know the right answer — and now you’re coming along and you’re suggesting we change everything. People really don’t like that.

Rob Wiblin: So the good idea fairy is a term of derision for someone who thinks they’ve had a new idea.

Jeffrey Lewis: Yes.

Rob Wiblin: Oh, wow. Huh, OK. Just as an aside for listeners, and you might be interested too: there’s this term that’s slightly come into vogue in policy analysis in the UK called “cheems mindset.” And cheems mindset is a mindset that people argue that lots of folks in policy analysis have, where if you ever suggest doing something much better or very different than what’s already been going, then you’re immediately accused of being naive and just not understanding the complexity, and not understanding that in reality, nothing could ever work — things couldn’t really be better.

This is kind of an instinctive pose that a lot of people take, and I don’t entirely understand why. But one explanation that’s given is that it allows someone who’s an insider to maybe seem wiser than people who are coming in from outside suggesting that the world could be better, because they of course know all of the unintended side effects, and they understand that, in fact, things could never be changed in any significant way.

Jeffrey Lewis: I think that is both the dominant narrative for insiders who are in favour of the current approach to nuclear weapons, but it is also surprisingly dominant among those of us who think that the current approach is not optimal. In part because there’s a certain guild quality. I think very fondly of studying under John and Tom, and I like to think that they’re right, and I like that kind of continuity. And they were pretty open-minded — that’s the second piece you get out of this, which is you could argue with them and they would change their mind. But for a lot of people, changing your mind is like giving up on your team.

Rob Wiblin: Huh. Is there an issue that, as I understand it, a lot of people in this field are now approaching the end of their careers? So maybe the dominant age is to be in your 50s and 60s. And perhaps it’s people in their 20s who tend to bring the energy for new ideas, because they want to have a new position, they want to have something to contribute, and they haven’t already absorbed all of the existing knowledge.

Jeffrey Lewis: Right. And the reason for that is because of the funding models that we have. You have a small number of large charitable organisations which are fairly risk averse. So I mean, the MacArthur Foundation’s now leaving the field, but you still have the Carnegie Corporation of New York. And I mean, I have zero complaints about how I have been treated in my career. MacArthur paid for my graduate school. These groups owe me nothing. I am so grateful.

But I got all of that through my advisor. Whenever people say, “Let’s fund young people,” that never means to give a cheque to a young person. That means give a cheque to an old person, who will then tell the young person what to do. They’re just very reluctant.

So simply funding different people, funding younger people, taking chances on things that are unusual. I like to joke that a lot of projects are too good to fund. That open-mindedness, that willingness to embrace risk, that willingness to place bets that maybe don’t have a high likelihood of payoff, but if they do would be enormous — all those represent diversity in grantmaking. Whereas right now, we have I think a very homogenous grantmaking style that privileges a small number of people, of whom I am luckily one, but that’s not a healthy view.

Rob Wiblin: Yeah, there should be more going on. So the situation where the EA community — and philanthropists who are inspired by an EA-style mindset — might be able to have significant influence over the nuclear security space through having access to funding that otherwise might not be available, that’s a relatively new situation. And as that’s become evident, I think that community is professionalising, and it’s beginning to hire in subject matter experts — exactly to fill the gaps that you’re troubled by, that could allow someone like me to make really stupid suggestions.

But it sounds like there’s a tricky balance here, because you’re saying the existing discipline and the ideas are somewhat sterile and risk averse and stale. And there’s a possibility that you’d hire in exactly the expertise that you need to correct your errors, but that those folks will convince you that nothing can ever be done, nothing could be better — you just have to fund the same stuff as always.

Jeffrey Lewis: This very famously happened, and a story I probably shouldn’t tell on the podcast, but let us just say that there are lots of instances where people come into this field wanting to make a big difference and a big change, and they surround themselves with starry names. And those people immediately begin the… is it “cheems mindset”?

Rob Wiblin: Cheems mindset. I’ll stick up a link to the blog post and I’ll share it with you, so you can see the post.

Jeffrey Lewis: Where they say, “Ah, the big stuff you want to do, we need to set our sights lower. You know, it’s complicated. Things are this way for a reason.” And obviously, some of that is valuable.

Rob Wiblin: It’s always partly true.

Jeffrey Lewis: Right? It is always partly true, but you can be captured. And I suppose, like all things in life, there isn’t a simple rulebook. It is an issue of balance and course correction, and there’s a certain art in knowing when you’ve gone too far in one direction or the other. But yeah, I think it’s a fundamental tension.

Gatekeeping [01:12:04]

Rob Wiblin: Yeah. OK, I just want to come back to this issue, where you wrote this in a comment online. You said: “My community is dominated by the obscure and arcane, and kept that way with an absurd amount of gatekeeping.”

What’s the goal of that? Do people perceive themselves as engaged in gatekeeping? Or is it maybe they notice that someone is talking about things in an imprecise way, and they’re like, “This person is an amateur. They don’t have all the established knowledge that I do, and I need to make sure that they understand and put them in their place?”

Jeffrey Lewis: It’s the ideology of hazing. I mean, how many people have you met who went through some hazing experience, and they are absolutely convinced that because they went through the hazing, everybody else needs to go through the hazing too. You know, you spend years, whether you’re working professionally or you’re going to school to learn the language and the games.

And you’re not doing it overtly, right? You’re not going to graduate school saying, “I am going to learn how to play these language games.” You go to graduate school and you are educated, and you stop thinking of it as hazing. You start thinking of it as your inculcation into the field, and I think it’s a very human thing to then replicate that. It also has the side effect of selecting for a particular group of people with a particular set of ideas, and it keeps the riffraff out. So you are able to filter dissent and silence dissent, which you want to do because you’ve been selected into the group, and you agree with the ideas.

Rob Wiblin: Right. You’re a winner of the current system, so why would you want someone to shake it up?

Jeffrey Lewis: Yeah. And it’s not that people aren’t nice and generous in that community. I mean, I learned to play these games from a lot of people who hate my views. It’s not like they are terrible human beings. So I’m not trying to single out individuals as being awful. I’m just saying that when I see students — or people from the outside, because you don’t necessarily have to be young or a student — try to engage with my community, I observe the pathology on display. And again, I don’t think it’s driven by any kind of hostility or hatred, but it stinks.

Rob Wiblin: Yeah. So this phenomenon exists in most places to some degree. I’m sure I’ve done something like this, probably many times, where someone says something not quite right, and I’ve heard so much of the debate back and forth that issues that are actually not that substantive feel very important to me.

Jeffrey Lewis: Like I say “mosquito nets,” and you’re like, “Actually, it’s mosquito mitigation, because nets are not the only strategy.” And it’s like, “Oh, c’mon.”

Rob Wiblin: “It’s very important that you get that exactly right.” Do you think that this issue is worse in the nuclear security space? And if so, is there an underlying reason for that?

Jeffrey Lewis: I think it’s worse because of secrecy. A lot of things you can only acquire by having a clearance, or by being told things you’re not supposed to know — but in that case, you don’t really know if what you’ve been told is true — and also by the arcane and slightly technical nature of the subject matter. So the OPLAN is classified. How are you going to know what our nuclear strategy is, because you’re not allowed to see it? Well, I shouldn’t say strategy, right? That’s me misspeaking. What is the employment plan?

So just that combination of secrecy and the technical nature of it I think naturally lends itself to that kind of long period of apprenticeship where you really don’t know what you’re talking about until you’ve been doing it a long time. And at that point, who wants to throw that all away in dissent? Unless you are like me and you want to live in California, or you’re like Dan Ellsberg, who went crazy, in a good way. Mort Halperin, one of my mentors, he and Dan were very close — Nixon accused them of being drug buddies at one point — and Mort had a very similar experience, where having gone through this whole process and mastered it and being an assistant secretary at like 29 years old, was just like, “What am I doing?” But almost nobody does that.

Strengths of the nuclear security community [01:16:14]

Rob Wiblin: Yeah. So I thought coming in that you might shit-talk me and my crew more than you and your crew, but it seems like it’s kind of been the other way around. So I want to hear some good things about your community for balance now.

Jeffrey Lewis: OK. There are lots of good things about my community. That’s a very broad question — so good things like, why can it be a nice place to work? I think the main good thing is that for the most part, the people in my community are very much like the EA community, in that their primary motivation is making the world a better place, and that’s a pretty unusual thing for a person.

Some people do pretty well by it, but a lot of people I know make very little money, and they make very little money the entire course of their career, and they retire having spent a career where they’ve lost most of the fights they were in. They haven’t succeeded in making the world a better place, and yet they have dedicated the 80,000 hours that they have solely to this thankless, seemingly impossible task. That can sometimes be a little aggravating, but I think you have to really like that about people.

Rob Wiblin: Yeah, absolutely. It does seem like there’s a lot of more moral seriousness in the question. It tends to attract people who take the issue extremely seriously, and I guess maybe even the gatekeeping is partly a reflection of how seriously people take the issue.

I suppose a benefit of gatekeeping to some degree is that, while it might make it hard to have new ideas or to have some revolution in thinking, it does keep out bad ideas. You might think this is an area where we don’t really want to be playing around too much. Maybe some risk aversion is appropriate, and we haven’t had a nuclear war yet, so why change what’s already working?

Jeffrey Lewis: Well, it’s a funny thing. I’m not someone who actually is a fan of the disarmament frame for reducing nuclear risk. I think our goal really ought to be reducing the risk of nuclear war, but — and this is a really important but — the idea that we are going to base our security on these weapons forever means that deterrence will eventually fail if you do it long enough, and it will fail catastrophically. I am one of these people who thinks the goal is to reduce the risk of nuclear war. In the near term, that probably doesn’t mean disarmament, but over the 100- or 1,000-year timeframe, it probably has to mean disarmament, right?

So you can say I’m using nuclear weapons to establish a system of deterrence to buy myself time to transition to something else. I need deterrence to function. I need not to have an accidental nuclear war so we live long enough. But even though I’m not a huge fan of disarmament as an organising principle in what we do, I genuinely believe that we’ve got to get there at some point. I don’t want to do this for 1,000 years.

Rob Wiblin: Yeah. So the idea there is that to function, deterrence relies on the possibility that people will use nuclear weapons. If you knew that you weren’t going to use them, then deterrence would cease to function as a mechanism to scare people, and so there almost has to be some minimum irreducible level of risk each year of things going wrong. Sure, you might be able to use deterrence for 100 years until you figure out something better, but it’s a system that is built to fail sooner or later.

Jeffrey Lewis: I think that is the most important thing that people fail to grasp. We have these very rational models where we imagine it would never make sense to do this. And that’s fine, except if that were true, then deterrence would cease to function — because you could just go right up to the brink every time and no one would want to push the button.

Tom Schelling called it “The threat that leaves something to chance.” Unless there is something left to chance, it can’t function — and once something is left to chance, then over very long timescales, you are in some trouble.

Rob Wiblin: I guess a listener might have the idea of: Let’s assume that there’s only two actors. They’re completely rational, and hypothetically they have total control over the missiles. And they have, at any point in time, two options: one to launch the missiles; one don’t launch the missiles. It’s always in their interest not to launch the missiles, because they’ll see missiles flying back. So you could indefinitely have this deterrence functioning, preventing anyone from pressing the button. Why is that the wrong way to think about it?

Jeffrey Lewis: It’s the wrong way to think about it for a few reasons. Well, first, there’s not two — now there’s three, right?

Rob Wiblin:
Right, right.

Jeffrey Lewis: We have two serious arms races playing out. First, political leaders are not as good at decision-making as that model suggests. I spend a lot of time talking to not just people in my field, but I really enjoy talking to politicians and other people who are much more senior. And it’s not that they’re dumb, but they are selected for other talents.

If you are the president of the United States, you are expected to be as good on the issue of nuclear war and peace as you are on financial reform, as you are on policing reform. And you have to go seamlessly from one meeting to another, managing all these issues, so you are working at a level of abstraction. I find that the best politicians have a few simple ideas about the world that seem to work across a whole bunch of issue areas. That’s not optimising, right? That’s satisficing, and I think that entails some risk. So the decision-makers I think are not so perfect.

And which, by the way, we see right now. I mean, Vladimir Putin absolutely believed that Ukraine would collapse, right?

Rob Wiblin: Yeah.

Jeffrey Lewis: If Ukraine had had nuclear weapons, I believe Vladimir Putin would have invaded anyway, because his assumption was Zelenskyy would flee and he would take over the country with the nuclear weapons. And he would have been wrong about that. Just would have been wrong.

Rob Wiblin: And we don’t know how that would have played out, yeah.

Jeffrey Lewis: Yeah. That’s one thing. The second is: I love game theory, I love models, but they rely so much on the quality of the information you have. And the scenario is not that Joe Biden wakes up and thinks, “I want to start a nuclear war with Russia today.” It’s that he wakes up and he is told that he does not have a choice, because the Russian attack is coming, and it is better for us to go first than to go second. And it’s regrettable, but it makes a big difference in the outcome if we go first. So he has to decide if that’s true, and maybe he decides to run the risk, maybe he doesn’t. You just don’t know.

Rob Wiblin: Yeah. I suppose, in the simplest model, it does make sense on paper, but if you actually look at the world as it’s playing out in real life, it’s extremely clear how this doesn’t apply. Both what we’ve seen in Ukraine this year makes it patently obvious, and then also what we’re seeing in Taiwan now. Is it clear that the risk of things accidentally escalating and the US getting involved in a war with China or Russia is not zero? It’s not zero, that’s for sure.

Jeffrey Lewis: Yeah. As Kennedy said, “There’s always some son of a bitch who doesn’t get the message.” The thing is, it works most of the time. I spend a lot of time thinking about what the risk of a nuclear war between the US and Russia or the US and China is. My answer is, it’s 1 in 100, 1 in 1,000. It’s a small number. It’s a really small number on a day-to-day basis. It really only is when you compound the risk over time and account for the severity of the event that you realise that —

Rob Wiblin: Then it becomes evident what the problem is, yeah. Who’s doing the best work in your field? Is there anything that you can really recommend to someone who wants to get a bit more on top of these issues? What should they read or listen to?

Jeffrey Lewis: Well, the first thing I would say is the old stuff is really good. We’ve been at this for 70 years, and there is a tendency to privilege the work we do now, but one of the reasons the work we do now is not all that interesting is because it’s already been said once and it’s been said pretty well, and it’s just that people didn’t like the answer.

The stuff Tom Schelling wrote is fantastic; the stuff John Steinbruner wrote is fantastic; the stuff Mort Halperin wrote. I particularly love Carol Cohn, who wrote a wonderful mid-’80s feminist set of observations about this community, because at the time it was entirely male. It’s not just the feminist piece of it, which is fantastic. It’s also, because she was an outsider, it was her ability to dissent from things people said and did, and to just say, people do these things. She noticed that men — which by the way, when she said this, I was like, “No, this doesn’t happen,” and then I realised it’s true — we always touch the bombs. We pat the bombs. She’s like, “Why are you doing this?”

Rob Wiblin: Really?

Jeffrey Lewis: Yeah. Oh, you take me up to a missile on display, I’ll go up and pat it. And I can show you pictures of presidents and world leaders patting missiles. And she’s like, “You know that’s a guy thing, right?” I was like, “What do you mean?” “You touched it.” Yeah, so that kind of stuff is all really great.

Today, the people I think doing the best work is not so much deterrence theory, as it is really just getting factual data about the world. Individuals doing open source work on what are countries building. We do that. Federation of American Scientists does that.

And then I think some people doing historical work to really try to understand how we got here, which was classified. People like Alex Wellerstein or the team at the National Security Archive. I think that kind of foundational work is really, really valuable — both to give us data about the world we’re in at the moment, but then also help us start to build theories of change. What would make the world safer and what is actually achievable? I don’t know, that’s kind of where my head is at the moment.

Rob Wiblin: In prepping for this episode, I asked for questions from the audience for you, and we got a lot of very good ones. To be honest, I felt a little bit shown up by the audience on this episode. I’m not sure when I can remember last thinking, “Wow, these questions are better than the ones that I’m coming up with.” So yeah, do you mind if we just do a grab-bag of things that people sent in?

Jeffrey Lewis: That would be awesome.

Disarmament [01:26:58]

Rob Wiblin: All right. One question was: “Are there any efforts to decrease nuclear risk that are likely to have the opposite effect?” I suppose that they might be thinking of things that plausibly the EA community could end up funding by accident.

Jeffrey Lewis: Yeah. I think there probably are. It’s difficult to just randomly throw stones, because I don’t know what people are likely to embrace. One thing I would say is that a bet my community made that I think was misguided was to emphasise disarmament as a frame. It’s interesting. When that bet was made, it was made very consciously initially by George Schultz, Bill Perry, Sam Nunn, and Henry Kissinger.

Rob Wiblin: Are those some of the folks behind the Nuclear Threat Initiative?

Jeffrey Lewis: Yes. And I love Senator Nunn, and Bill Perry is one of the finest people I’ve ever met. So I disagree, but I don’t begrudge them this. They were very clear that the decision to embrace rhetoric about disarmament — which none of them had been particularly vocal supporters of previously — they felt like it was a way to mobilise resources. That we’d have a compelling vision, that we’d get people to take the steps that we need to. Senator Nunn would say, “It’s like a mountain. You maybe can’t see the top of the mountain, but you know if you’re going up or down.”

So it was a pragmatic bet that they made, and what I think happened was, it was the wrong bet. It didn’t really inspire people. It raised expectations that we were unable to meet, which I think in part explains why the Treaty on the Prohibition of Nuclear Weapons exists — the treaty prohibiting nuclear weapons — because it raised expectations about disarmament, which then didn’t come through. So you get this kind of adversarial treaty that tries to hold the nuclear weapons states’ feet to the fire, which I get. But I think it also results in funders feeling like they followed that elimination as rhetoric with a lot of money, and then didn’t get a lot of return.

I believe in disarmament, because I don’t want to do this for 1,000 years. It’s just the narrow bet of: Is this the frame we should have used? I don’t think it was. I was actually involved — I was working at the time with the Open Society Institute. This was before Barack Obama gave the speech in Prague where he said he was going to seek the peace and security of the world without nuclear weapons. And we focus-grouped and polled people about that particular frame: “How do you feel about nuclear disarmament?”

It was a really interesting thing we learned. If you ask people, “Do you support nuclear disarmament?,” it’s like motherhood and apple pie: everybody’s for it. Republicans, Democrats, it’s very popular. The minute Barack Obama, a living partisan figure, says he’s for it, the numbers among Democrats go up, and the numbers among Republicans go down. The minute Mitch McConnell says he’s against it, the numbers for Democrats go up, and the numbers for Republicans go down. And independents don’t really exist, right? They just go home to their partisan camps that they refuse to admit that they’re part of.

So what we learned from that was that people’s feelings about nuclear weapons are less important than their partisan affiliations, and once you make something a partisan issue that dominates. So we did this study, we gave the study to the Obama administration. We said, “Please don’t give the speech.” They patted us on our heads and thanked us for our interest in national security and sent us on our way. Gave the speech.

Then a few years later, Richard Lugar, the senator from Indiana who helped get the New START treaty through, a Republican, was giving a talk at our board meeting. And without knowing anything about the study, proceeded to make all the same arguments, saying that the hardest part for him in getting the New START treaty through the Senate was that Obama had embraced the idea of disarmament. By making it a partisan issue, identified with the President, Republicans that he thought he might be able to get to support the treaty were unwilling to support Obama’s treaty.

Rob Wiblin: Right, right.

Jeffrey Lewis: When you say bad ideas, it’s not that I think disarmament is a bad idea. But that’s clearly a bet that we made as a community that didn’t pan out the way we wanted to, and I think it left us in a much more dangerous place, because we have failed to do all the important things that we should have done. Now we’re in a situation where we’re looking at running simultaneous arms races with the Russians and the Chinese.

Rob Wiblin: Yeah. There’s quite a lot of groups that are on board to a greater or lesser degree with the idea that we should be aiming for a nuclear-weapon-free world. I guess even if it was accomplished, it could potentially backfire, because then one country could arm themselves very quickly and engage in a first strike. I don’t know whether that’s very likely, but it could be more likely than the risk of an accidental nuclear war is now. So that’s one kind of perverse impact.

Also the fact that it does just seem like a heavy lift to get to zero makes me more inclined to think that our aspiration should be… Currently New START allows, was it 750 warheads to be on alert at any point in time for the US and Russia?

Jeffrey Lewis: 1,550.

Rob Wiblin: Oh, 1,550. Where did I get 750 from? Maybe that’s for each country.

Jeffrey Lewis: No, no, it’s 1,550 for each country.

Rob Wiblin: Oh, OK. Cool. I’m just wrong.

Jeffrey Lewis: Which in real life is more like 1,700, because each bomber counts as one, no matter how many nuclear weapons it carries.

Rob Wiblin: Oh, OK.

Jeffrey Lewis: Welcome to the gatekeeping.

Rob Wiblin: OK, yeah. Very helpful. So why don’t we try to reduce that number? At 300, it would still be a pretty severe threat of retaliation, probably sufficient to stop anyone from engaging in a first strike. So why don’t we get it down to 300, and then of course we can think about getting it below that as well? As I understand it, we don’t have any treaty with China with regards to nuclear weapons control, some agreement not to have too many — so maybe we could get any agreement with China. That would be a great step forward. That’s my amateur take on this. Am I thinking in the right direction?

Jeffrey Lewis: That’s basically where I am. I look at: our goal is to prevent a nuclear war. Disarmament is a way to do that over a very long timescale. I’m actually more positive about disarmament in a case where one country builds a nuclear weapon. Because I actually think when you get to disarmament, what happens is you realise that these are just weapons, and a country being able to destroy one city doesn’t necessarily give them the ability to win a war, right?

But that’s all very long-term stuff. That’s about our attitudinal change towards nuclear weapons. That’s like a couple-hundred-year project. In the near term, the goal is for deterrence to function stably so we don’t kill each other. And part and parcel of that is reducing the number of the weapons, so that if deterrence does fail —

Rob Wiblin: Why does that help?

Jeffrey Lewis: Well, one thing is, it would be great if we had fewer nuclear weapons than were necessary to trigger a nuclear winter. That would be one yardstick. But also, I’m not one of these people who believes that once a nuclear weapon gets used, we’ll learn our lesson, because we’ve done that twice now and we didn’t learn the lesson, right? I’m not a big believer in teleology. What lessons we learn are the lessons we choose to learn, so I worry that even a small nuclear use sets us up to a scenario where it’s very dynamic, and then our risk profile changes in a really negative way.

Rob Wiblin: Right, right.

Jeffrey Lewis: I’m all about reducing the risk of use for the US–Russia and the US–China relationships. Arms control is typically the way to start, for two reasons. One is, reducing numbers has this benefit of just making things less destructive. But if we go back to my fundamental interest in decision-making, what we’re really trying to do is build a constituency in the United States — and in China, and in Russia — for that big idea we talked about: that you have to cooperate, even with your adversaries. The way you do that is by negotiating and creating agreements in which you pursue mutual interests.

Reducing numbers is not just about reducing the aggregate number; it’s about reducing the most destabilising system. But more deeply, it’s about building a system that enshrines that shared interest. Honestly, it’s what I think happened with the Iran nuclear deal, and it’s one of the reasons I’m so alarmed at its collapse: we created this community in Iran that actually supported, in a very tentative way, a different relationship with the United States — and then we destroyed that. If you’re an Iranian policymaker and you stuck your neck out for the Iran nuclear deal, would you do that twice? I wouldn’t.

That’s very true in China, by the way. In my dissertation, I interviewed a lot of Chinese who were involved in arms control negotiations in the ’90s, and they feel like they had the rug pulled out from under them.

Rob Wiblin: What happened there?

Jeffrey Lewis: So in the 1990s, the United States and other countries negotiated a ban on nuclear testing, and people thought the Chinese would never agree to this. The Chinese reorganised internally, bureaucratically, had huge bureaucratic internal fights, and the people who supported the ban on nuclear testing won. China signed the treaty, the US signed the treaty. The US refused to ratify the treaty, which then destroyed that coalition in China.

And to me, that’s a very common story. By the way, there’s an analogy in the US. In the 1990s, the Clinton administration did a nuclear posture review, and the guy who ran it was named Ash Carter. Ash Carter wanted to get rid of land-based ICBMs, and he got destroyed — just bureaucratically annihilated. Ash later becomes the US Secretary of Defense, but when he is up to be Secretary of Defense, he had learned his lesson. He was a full-throated supporter of a triad, including replacing the ICBMs.

Now, he will tell you his views evolved. I think he learned his lesson, in the same way that the Chinese learned their lesson, which is: you have to be careful when you stick your neck out for diplomacy. Because if the other side screws you, then inside the government you look weak and naive, and you don’t get that promotion you want. That’s why I support arms control. It’s building those communities organised around an idea of shared interests.

Rob Wiblin: Someone in the audience asked why the US and Russia invested so much in making lots of land-based nuclear missiles, because couldn’t we just put them in submarines and wouldn’t that be much better? I feel like you’ve probably already said what you have to say about that one, unless the answer is different in Russia, maybe?

Jeffrey Lewis: Oh, the answer’s totally different in Russia, because in Russia they had a much weaker navy, and so they are much more heavily reliant on land-based missiles, which is why they have invested in mobile missiles. The amount of money they give the navy is less, so the submarines that they built lagged behind ours. They were not as good technically. They’re pretty good now, because they’ve been at it a long time, and now they have a force. But they have always placed much more emphasis on land-based forces, precisely because their navy is much weaker than ours.

Rob Wiblin: And I guess that makes them more skittish, because the land-based missiles are more vulnerable?

Jeffrey Lewis: That’s right, and you also develop operational concepts that justify the forces you have. We look at putting multiple warheads on land-based missiles as destabilising. You have to go first, right? Because when it’s in the silo, it’s one target and very vulnerable. Once it’s out, it’s 10 warheads, but that’s because we have this balanced system. The Russians have to justify their force and they say, “Oh, no, no, no. Having multiple warheads is stabilising, because if even one gets through, then that’s 10 warheads.”

You create an entire intellectual life around the force you have, which has its own notions of stability and signalling and escalation control. From an arms control perspective, the thing that we have to be worried about is we have to make sure that their ideas about stability and our ideas about stability add up to stability. It’s like waves, right? Do the waves make a really big wave together, or do they cancel out?

Nuclear winter [01:38:53]

Rob Wiblin: Yeah. OK, another question: “In the context of massive nuclear attacks, why isn’t the danger of nuclear winter widely seen as making nuclear retaliation redundant, as Ellsberg suggested on his episode on The 80,000 Hours Podcast?”

Jeffrey Lewis: Because people don’t believe in it. I was looking at one article that was critical of nuclear winter, and the first thing I noticed is the first person it cites is like Fred Singer, who’s also a climate denialist. I definitely think that in any issue like this, there’s a certain amount of ideological motivation, which I think naturally happens. I mean, if you’re a scientist working on an end-of-the-world problem and you think somebody’s about to end the world…

Rob Wiblin: You don’t equivocate.

Jeffrey Lewis: You suddenly have a partisan view: “I don’t want to die.” But then there’s a natural reaction, right? You also then get people who say, “No, the climate’s not changing,” or “No, it’s not people,” and there are a lot of people who say, “No, there’s no chance of nuclear winter.” I do think there are real debates to be had about the models themselves — because the inputs are really unclear, the mechanisms are difficult to understand — but those debates are different than the ideologically driven “no, it’s not happening, you’re just a hippie” objections that I see thrown at nuclear winter. I think it’s pretty plausible.

Rob Wiblin: You think it’s pretty plausible? Not knowing for sure how bad it will be, but credible?

Jeffrey Lewis: I mean, you’re going to get fires.

Rob Wiblin: Is the 1,700 warheads on each side sufficient?

Jeffrey Lewis: Probably.

Rob Wiblin: Probably, yeah. I see.

Jeffrey Lewis: I mean, the issue is how much burns? If we hit silos in the interior of Russia, will you get big forest fires out of it? Are you hitting leadership targets in cities? Are the cities going to burn? The thing I like to point out is that Hiroshima burned, Nagasaki didn’t. What determines whether there’s going to be a firestorm is really hard to model. And for the longest time, the US didn’t model fire damage at all — they just did blast damage, because they could calculate it. Fire damage was just way too complex. That’s why I say you could have a real good faith intellectual discussion about how much certain things —

Rob Wiblin: How likely, how bad…

Jeffrey Lewis: But I think that’s not the debate that’s occurring. Just like you could have a serious debate about a climate model, but that’s not what’s happening. That’s not bringing a snowball onto the Senate floor.

Rob Wiblin: Right. I suppose having that debate would require some technical knowledge about the climate models, and some of the issues and the uncertainties. Whereas just saying, “Yep, you’re a stooge for the nuclear industry and you’re a hippie” is something anyone could do. I could participate in that.

Jeffrey Lewis: That’s true. Although, a surprising number of scientists still engage in such behaviour, right?

Attacks against US allies [01:41:46]

Rob Wiblin: Huh, OK. Here’s another question: “The US threatens to retaliate with nukes against anyone who nukes certain US allies” — I guess they’re imagining here maybe Taiwan, maybe the Baltic States — “How credible is this threat, and why?”

Jeffrey Lewis: It’s the oldest problem in the book. I don’t know. Charles De Gaulle put his finger on it when he asked if the US was willing to trade New York for Hamburg. His answer was, “It’s not credible at all. France is building its own nuclear weapons.” The German decision was, “It is credible. We don’t need our own nuclear weapons. Also, maybe everybody hates us after the war, so we won’t build them.”

Rob Wiblin: “Let’s just be quiet.” Yeah.

Jeffrey Lewis: Yeah. But one thing that’s really interesting to me is that our language about this is actually a little bit misleading, because we talk about a nuclear umbrella all the time: not really a thing that exists. We have security commitments that we make to countries — an armed attack on NATO is an armed attack on the US if it’s in the Euro-Atlantic area, an attack on Japan — but we don’t actually have a commitment to use nuclear weapons on their behalf. We have these two things: we have a security commitment to them, and we have nuclear weapons. And the Russians can put two and two together, but a lot of what we do is try to make that commitment credible.

In Europe, the thing they settled on was dual-capable aircraft. We have German pilots, Belgian pilots, Italian pilots, Dutch pilots at air bases with US nuclear weapons, and we claim that in a war, we are going to put a US nuclear weapon on a Belgian plane and they’re going to go fly it off and drop it on the Russians.

Rob Wiblin: OK. Is that right?

Jeffrey Lewis: Maybe that makes it credible. It doesn’t seem very credible to me.

Rob Wiblin: Why is that?

Jeffrey Lewis: Because that particular weapon system is not one that the US has any interest in using. I have had senior US officials say we would never drop a B61 off the wing of a plane. It exists only as a symbol of our commitment.

We don’t do that in South Korea or Japan. In South Korea and Japan, we just say, “We consult with you. We have this dialogue that we do to show you how committed we are.” What I end up saying is, our commitment to their security can’t be separated from the nuclear piece of it. You either believe we’re going to defend them or you don’t — and that’s based on a whole bunch of broad factors, like the political relationship, our interests at stake. Then, if you think we’re going to come to their defence, you kind of have to assume we might use nuclear weapons because we have them. But the problem is, if you think we might not, then I don’t think all the DCA in the world will change your mind.

I don’t say that to be derisive about the decisions people have made. It’s a hard problem. It is.

Rob Wiblin: What’s DCA?

Jeffrey Lewis: Oh, dual-capable aircraft: aircraft that can carry nuclear weapons and conventional weapons. Exactly, sorry. See it right there, I did the gatekeeping thing myself.

Rob Wiblin: Is the DCA thing set up such that the Belgian pilots can just grab the nukes and launch them themselves without the US?

Jeffrey Lewis: No, no, we have to put them on. We have to put them on, and arm them.

Rob Wiblin: Why? Why can’t they pick them up? Is there some security system on them?

Jeffrey Lewis: Yeah, the fear is that in the Cold War, they would’ve grabbed them and gone and done something bad. And by the way, the US has a bunch of these things sitting in Turkey right now, Incirlik Air Base, which was central in the failed coup against Erdoğan. That’s a really dicey situation.

Rob Wiblin: Right, right, right.

Jeffrey Lewis: So yeah, the US wants to retain what’s called “positive control.”

Most likely weapons to get used [01:45:11]

Rob Wiblin: Someone in the audience asked: “It would be great if you could ask Jeffrey what types of nuclear weapons are most likely to be used.”

Jeffrey Lewis: Traditionally, I would have said the big strategic weapons — that the idea of tactical nuclear weapons was to kind of talk ourselves into the fact that we might use them, but if we were going to use them, we’d go big. I kind of no longer feel that way.

Rob Wiblin: Why’s that?

Jeffrey Lewis: Because the Russians and the Chinese are investing really heavily in medium- and intermediate-range missiles that can do both nuclear and conventional missions. That introduces a certain amount of ambiguity into a conflict that worries me. The Russians are hitting Ukraine with these Iskander missiles, which are nuclear capable.

You could imagine a situation in which there were more nuclear threats, or more fears of nuclear use, in which the US got involved and was targeting those missiles. And now, you’re targeting nuclear forces. And maybe they move the nuclear warheads around — they’re doing it because they’re afraid that you’re going to target them, but you think they’re doing it because they’re going to launch them.

It just complicates things in a way that I find really alarming. I’ve kind of reversed that view, and started to think that all the little stuff is just starting to get integrated with our conventional operations in ways that I don’t think we’ve really thought through very carefully.

The role of moral arguments [01:46:40]

Rob Wiblin: What role do moral arguments play when strategic planners make policy? Do they squarely face the implications of things like countervalue targeting — that is, killing many millions of civilians — or do they tend to avoid thinking about the topic?

Jeffrey Lewis: They tend to avoid thinking about the topic. There is a really strong masculine urge not to appear weak and sentimental. I always find it really hard. I used to go back to Hiroshima every year, because I’m on the governor of Hiroshima’s round table on nuclear disarmament. And going from Hiroshima to DC can be very difficult, because it’s hard for me to go back into the kind of abstract language and denial of the consequences that’s socially required in those conversations. You see examples of that, like the civilians who would be killed in a counterforce strike used to be called “bonus.”

Rob Wiblin: OK. I guess this is a predecessor to “collateral damage,” in terms of euphemistic language?

Jeffrey Lewis: Yeah. I mean, we still use “collateral damage,” but “bonus” is now frowned upon for reasons that are probably obvious.

Rob Wiblin: Pretty obvious, yeah.

Jeffrey Lewis: So no, I don’t think that policymakers really squarely face the moral implications. To the extent they do, they come through as legal.

Rob Wiblin: They’ll say, “This will be a war crime, so it wouldn’t be permitted to happen.” I mean, I guess they might say we have to be willing to set this stuff aside in order to make it credible that we would retaliate. So if we seem sufficiently softhearted that it would never be believable that we’d be willing to use the weapons, then in fact that makes war more likely.

Jeffrey Lewis: Yes. I think that is the canonical argument. The rejoinder to it is it’s very difficult to base a credible threat on an uncredible action. If you’re artificially excluding some considerations, is that still something that is likely to be credible? Then honestly, are you still making a good decision?

But yeah, there’s an interesting perspective shift that happens here. And I just did it myself, so I should be explicit about it. When we are talking about what will deter our enemy, we of course do not know, and the models that have been created are not terribly impressive. We guess at what Putin values; we don’t really know.

And so what officials usually do, in my experience, is they substitute themselves. We are actually talking not about what it will take to deter Putin; we are talking about what will give us the courage to stand up to him. It’s all about our own courage more than it is an actual calculation of what the adversary will do, because it’s an interior discussion.

I think a lot of the language that we use about nuclear weapons is the equivalent of having a stiff drink. It’s about making yourself courageous enough to say you would do that thing and to try to express that credibly.

Rob Wiblin: I suppose you like thinking about decision-making processes more concretely. Imagining a situation where the president is being pressured to some extent — through circumstance or because of who’s advising them to use nuclear weapons — to what degree do you think that they might hold back, because they just find the moral consequences so unpalatable, so disgusting, that in the moment they would almost consider that they would rather just die and not retaliate?

Jeffrey Lewis: I think that some presidents would feel that way. I think some presidents have a very strong moral streak and would, in the moment, do something that they deny they would do.

Rob Wiblin: And we want them to deny it. I think.

Jeffrey Lewis: Right. On the other hand, Donald Trump, I don’t know. And I don’t say that simply implying that I know he would — I really don’t know what he would do. I think that there’s a certain amount of him that is bluster and that he might hold back, but then I think there’s a certain amount of him that is callous and selfish, and he might not. So I just think it’s important to remember that the individual people probably matter. That people make different decisions.

Rob Wiblin: Yeah. To what extent do you think it will be affected by what side of the bed they got up on in the morning? It’s an alarming thing to imagine that just the mood that they happen to be in in the moment could be definitive.

Jeffrey Lewis: Yeah, there are plenty of examples where staff at a lower level have to decide: Do you wake the president up? Do you want the person to wake up and make a decision now, when they’re cranky, or do you want them to get a good night’s sleep because they got a long day ahead? North Korea shot down a US reconnaissance aircraft in 1969, killed everyone on board, and they decided not to wake Nixon up. They thought, “He needs to sleep, and it’s going to be a long day tomorrow, and it can wait.”

Rob Wiblin: It can wait a few hours.

Jeffrey Lewis: They were sincere. Nixon was someone who famously would get drunk and have tantrums, and they often organised their decision-making structure around his human frailty. So I think it matters.

Salami tactics [01:52:01]

Rob Wiblin: Yeah, yeah. For those who don’t know, salami tactics is this idea that your adversary will go one kilometre past your border and stop, and then you’ll think, “Well, do I really want to destroy the world now? Do I want to press the button now?” And you think, “No,” and then they go another kilometre, and then they stop. And you’re like, “Do I want to destroy the world now?” And the question is, at what point does that end? To what degree does the ability to use salami tactics make nuclear weapons far less useful for deterrence than you might think on paper?

Jeffrey Lewis: First of all, you should play that clip from… That actually may be the Yes Prime Minister era, right?

Rob Wiblin: It’s so good.

Jeffrey Lewis: It is fantastic. I love the salami tactics argument, because I do think it is very hard to imagine a rational scenario in which one uses nuclear weapons, and I think deterrence is probably less effective than people believe. The examples I always give are Argentina invaded the Falklands. They concluded Britain wouldn’t use —

Rob Wiblin: I was going to ask about that. Aren’t they declaring war on NATO?

Jeffrey Lewis: It’s not in the North Atlantic. That’s why it’s the North Atlantic Treaty Organization.

Rob Wiblin: Got it. OK.

Jeffrey Lewis: But also, Egypt invaded the Suez when Israel had nuclear weapons. China and Russia fought a border skirmish that the Chinese initiated. And people say, “Well, those don’t count as deterrence failures.” And I’m like, “Yeah, that’s the point. They don’t count. The aggressor was able to successfully figure out that there are some things that count and other things that don’t,” and when you start to play that game, it’s very unclear where that line is.

But then this brings us back to, if it was a purely rational endeavour, you would salami tactic your way to no nuclear war, and then nuclear deterrence just simply wouldn’t work. I think it is that irreducible risk that animates the system. We’re talking a lot about Tom Schelling today. One great thing he said was, “We think of the nuclear brink as a precipice, and that is wrong. Because you know where the edge of the precipice is. It is a curved slope and you are inching your way out. The further you inch out, the less sure you are about when or where you’re going to slip.”

That’s my answer to salami tactics. I think, at a purely rational level, they’re real, and it shows why deterrence is not all it’s cracked up to be. But the answer, which is good and bad, is there is still an irreducible risk that there must be some point at which — because the president got up on the wrong side of the bed, because the information is wrong — things get out of control.

Rob Wiblin: First, to me, the main thing that saves us from all of this is that you can imagine Russia considering invading one kilometre past the border in the Baltic states, and then just stopping and seeing: Are you going to respond? Obviously, the rational answer is no, you’re not going to destroy the world now. But there’s a 1-in-1,000 chance that you might; there’s some tiny chance. And even at a 1-in-1,000 chance, the value of crossing the border one kilometre is not worth it. It’s not worth a 1-in-1,000 chance of your society being destroyed.

And it’s that asymmetry between the magnitude of the gains that one can get by violating these rules, relative to the enormous downside — where the downside is just many, many, many, many multiples larger — that means that it’s very rarely in any country’s interests to take that risk.

Jeffrey Lewis: I think that is probably one of those really crucial things to say about deterrence. The way that I express that idea is that nuclear weapons simplify a lot of calculations. There are a lot of complex games you could play where you’re imagining, but because there is that irreducible risk of everything going wrong with this enormous cost, it simplifies things for policymakers in a lot of cases.

The other issue that I think we should be aware of is that it’s easy to say you drive one kilometre in, but we have to keep in mind that the person getting driven in doesn’t know you’ve only driven in one kilometre, right? They don’t know that you’ve stopped. They know that you’ve crossed the border, and then they get a report that says they’ve crossed the border, and maybe four hours later they get a report that says they haven’t got to the next town. They’re like, “Well, maybe send out a reconnaissance unit and figure out where are they?” Or you fly a plane, and maybe the plane sees you, but maybe the plane doesn’t see you. Maybe they see you and they’re like, “Oh, we can bomb them, because they’ve broken down.” You just don’t really know, and so again, because you don’t have that perfect information, you do still have that 1-in-1,000 chance.

Rob Wiblin: Yeah, totally.

Jeffrey Lewis: It’s the Lin Biao pathology. They don’t know what you didn’t do. They only see what you did.

Jeffrey’s disagreements with Thomas Schelling [01:57:00]

Rob Wiblin: Yeah. What were some of your main points of disagreement with Thomas Schelling?

Jeffrey Lewis: Oh, there’s one that I particularly cherish. When Tom thought about the risk of nuclear war, the thing he always compared it to — at least in our conversations, and he did it once in a published paper — is when he’s driving, he’s confident a pedestrian isn’t going to step out in front of him, and that’s deterrence. He’s very comfortable with deterrence.

I was like, “Tom, people get hit by cars all the time. What are you talking about?” And I think that goes to that fundamental difference. He was confident that he was never going to hit a pedestrian, and that a significant portion of that was the deterrence of driving, and a small part of that was him thinking he was a pretty good driver.

I’ll never forget, I threw a dinner and Mort Halperin was there, and Tom and Mort had had a couple of glasses of wine, maybe more than a couple. And Tom, who at that point is in his late 80s, says to Mort, “I’ll drive you home.” And I’m like, “Oh my God. No. This is not how I want to win the argument about pedestrian fatalities.”

Rob Wiblin: Interesting.

Jeffrey Lewis: He was just, at some level, very confident that people were deeply sensible and rational, and I don’t share that sense. I actually looked up statistically, what are the percentage of serious accidents involving pedestrians and cars? I wanted to find it for Massachusetts in the early ’80s when he wrote this, just to get a sense of what was his reference risk. It turns out, and you have to look at licenced drivers, from a driver’s perspective, about one in 1,000 drivers every year would hit a pedestrian.

Rob Wiblin: Yeah, so not that low a number.

Jeffrey Lewis: A 1-in-1,000 risk of a nuclear war. “Tom, that seems high to me.”

Rob Wiblin: Yeah. It seems to me like the game theory on something as serious as this needs to build in… You say people are rational, people choose the sensible option. But then one in every 1,000 times they just happen to choose the other option for unfathomable reasons that are outside the model, and you want to have a strategy that is robust to that.

I guess, in the Prisoner’s Dilemma case, you deal with that by potentially having forgiveness. I won’t go into this a lot, but you can have this kind of grim trigger, where if someone defects on you, then you continue defecting forever. But that’s a bad idea, because you can get stuck in a loop of defecting indefinitely. You need to have some forgiveness process whereby people can redeem themselves and get back to cooperation.

Basically, it seems like we need strategies that are robust to people doing something randomly stupid.

Jeffrey Lewis: I think that is a really important point, and the way that I think Tom dealt with that, on a personal level, was by being willing to change his mind, by being willing to adjust. I think, for him, it was almost — I don’t know what you’d call it — a personal choice, just a way he looked at the world.

Rob Wiblin: Stylistic choice.

Jeffrey Lewis: Yeah. You could persuade him that sometimes his stylistic choices were really particular. He has a famous example about if you have to meet a friend in New York, but you don’t know when or where, where would you meet? I was always unimpressed by this example, because he’s like, “Of course it’s Grand Central Station at noon.” I’m like, “Tom, that’s like a man who used to take the train from Harvard or New Haven down to New York City once a month or something. I’m from Illinois. I’d probably be…”

Rob Wiblin: “I hadn’t heard of Grand Central Station.” Yeah.

Jeffrey Lewis: Yeah. Empire State Building, right?

Rob Wiblin: Times Square.

Jeffrey Lewis: But he admitted in his book, “This is probably pretty culturally specific to myself and my group of friends.” I’m like, the answer is yes. The problem is, I am trying to deter someone with whom I have an adversarial relationship. Largely probably because I don’t share a lot of the same cultural touchpoints.

Why did it take so long to get nuclear arms agreements? [02:01:11]

Rob Wiblin: Yeah. Why did it take decades after the first nuclear weapons for countries to sign nuclear nonproliferation and arms control agreements?

Jeffrey Lewis: That’s easy, because nuclear weapons were good. I mean, we conceived of them as weapons, right? It was a very efficient way to conduct strategic bombing. We thought they were good when we had them. They were good when our allies had them. It was only bad when the bad guys had them.

I think the intellectual change that caused us to stop thinking of them as just another weapon that happens to be really big was when China built them. We could sort of adjust ourselves to the fact that the Russians might have joined us, now we were in this deterrent situation. But they kind of look like us, and they’re rational. The Chinese, it was just a lot harder for us, I think, emotionally, to wrap our heads around.

Rob Wiblin: Because the US didn’t understand China?

Jeffrey Lewis: Yeah. Mao seems crazy. I mean, look, I’m going to be honest, Americans are pretty racist, especially 1964. China’s poor. It’s strange to us. We don’t have any relationship with it. Mao seems nuts. Mao kind of is nuts. It’s a tumultuous place.

So you really see that there’s a kind of shift. There’s something called the Gilpatric Committee. Which, by the way, Roswell Gilpatric had his love letters with Jackie Kennedy stolen out of his safe — DC, small place. After the Chinese nuclear detonation, he’s put on this commission, because this is what presidents always do. Something bad happens, you have a commission, which means you make the bad story go away, because you say, “Well, I’m waiting for the commission.”

That’s really where we get a lot of our modern arms control and nonproliferation agenda. It comes from the Gilpatric Committee. At least US support for it. I would argue that the political impetus that makes that seem like a palatable idea is Chinese acquisition of nuclear weapons — because once they can do it, anybody can do it, and that’s really scary.

Detecting secret nuclear facilities [02:03:18]

Rob Wiblin: How good do you think current systems are at detecting secret nuclear facilities?

Jeffrey Lewis: That is a really tough question to answer. It is the thing I think about the most. In the past, we were pretty good at it because nuclear facilities were big and they had a lot of signatures — like a gaseous diffusion plant was hot, a reactor has to be cooled. Centrifuge plants for enriching uranium of the kind that the Iranians built in secret are much more difficult to detect.

On the other hand, the US caught the Iranians three different times building secret centrifuge plants, so three for three is a really good number. The big question is, did we have a 90% chance of detecting those? Will our luck hold for numbers four and five? In which case, I think you need the Iran nuclear deal, which is why I support it. Or are we just so good? But they’re small and they don’t have a lot of signatures, and so I think they should be pretty hard to detect.

Rob Wiblin: Huh. I would’ve thought, just in the modern age with so much stuff being logged on the internet or being logged by computers, that if you were really smart about it, you could find some indication that a country was trying to build one of these centrifuge facilities.

Jeffrey Lewis: What typically happens is you pick up procurements.

Rob Wiblin: Right. How many people are buying that kind of centrifuge?

Jeffrey Lewis: So you have to buy the components, right? Because no one will sell you the centrifuge. This is where the famous aluminium tubes came from in Iraq. There’s always some ambiguity. The tubes, and the ball bearings, and the vacuum systems — are they for centrifuges or are they for other things?

But the example I would give is the US detected, in the early 2000s, North Korean procurements that they correctly assumed were for a large centrifuge plant. The North Koreans began constructing that plant about the same time. The US did not identify it as a site of interest for four or five years, and did not positively ID it as a centrifuge plant — which by the way, we’re still not sure, because we’d never been in — until about 2010. And the outside community, we’d heard rumours of this place. We didn’t find it for another 10 years after that.

Rob Wiblin: OK, yeah. I mean, North Korea is pretty extreme.

Jeffrey Lewis: Yeah, that’s true. It is. But the problem is when you look at the facility itself, it doesn’t give itself away. You might see stuff going in and out, and you’re like, “I think you’re building a centrifuge plant.” But can you prove it? Where is it? That gets really, really hard.

Where Jeffrey would give $10M in grants [02:05:46]

Rob Wiblin: If you had $10 million to give in grants related to nuclear security, who might you disperse the money to? Asking for a friend.

Jeffrey Lewis: I would mostly spend it on data, and a little bit on decision-making and theories of change.

Rob Wiblin: What sort of data?

Jeffrey Lewis: So the open source work, like we do — although there are other people who do it; I’m not trying to take the $10 million. But open source work that tracks nuclear programmes. The example I like to use is: We’re having a big debate about the nature of our relationship with China and the possibility of an arms race. We are the first people who publicly discovered the existence of several hundred missile silos being constructed out in the Chinese desert. And then two other groups subsequently found other fields.

Being wrong about how many nuclear weapons and what kind the Chinese are building has huge consequences for understanding where they are in their thinking about an arms race. You can’t plausibly imagine an arms control regime unless you have some idea of what they have and some guess about why they’re building it. So that research, I think, is very valuable.

And I think I mentioned this earlier: archival research. So the National Security Archive, which gets stuff declassified and helps us reconstruct how we got from the past to the present, I find that stuff really, really valuable. Because we’re still arguing about what is the US target? And the way you’re going to get that data is by getting things declassified and by collecting oral histories and going through that data in a scholarly, careful way.

Rob Wiblin: OK, let’s take those one by one. So on the data collection side, one argument that this helps is that inasmuch as there’s massive uncertainty about, in this case, what China and Russia are doing with their nuclear arsenals — inasmuch as it’s extremely opaque to the decision-makers in the US — they’re inclined to think the worst. To think China is arming up, they’re probably doing a whole bunch of stuff we can’t see, and we’re worried about that.

And so even if that’s not happening, hypothetically, then the US will tend to start building up in anticipation of that possibility. Which means then China kind of has to do the same thing as well. So if neither side can really tell what the other side is doing, then that is very conducive towards an accidental arms race that neither side might like. Whereas if there were very strong transparency, then both sides could regularly confirm that the other side is not starting an arms race, and things could remain in that state for a significant period of time.

Jeffrey Lewis: And I would say there’s a civil society piece, which is we can’t just leave it to the intelligence community. Because as we saw in the invasion of Iraq, sometimes the intelligence community gets things wrong. And I’m a believer in civil society. I think one of the reasons that the Iraq War was able to happen is that civil society wasn’t able to participate in a meaningful way in the question of whether Saddam Hussein might actually build a nuclear weapon. So I think it’s a conversation we have to have publicly as well as privately.

Rob Wiblin: Which means you need declassified information.

Jeffrey Lewis: Right. Or unclassified information.

Rob Wiblin: Unclassified. Sorry, I meant to say unclassified. Yeah.

Jeffrey Lewis: Oh, sorry. Did I gatekeep again? Jeez.

Rob Wiblin: No, no, no. That was just me being an idiot. Even I know that’s wrong.

Jeffrey Lewis: But the point you raise is incredibly important, because we might have false beliefs about what the Chinese are doing and why they’re doing it. And that might cause an arms race in peacetime, and in a crisis it might cause us to misinterpret what they’re doing. So understanding their nuclear programme is really, really important. One example of this is China has this no first use policy, right? They say they would not be the first to use nuclear weapons. Nobody in the United States believes this.

Rob Wiblin: Oh really?

Jeffrey Lewis: So there’s a fundamental kind of disconnect. And the Bush and Obama administrations really believed that the way to prevent China from building up was to maintain a nuclear force that was four or five times bigger than the Chinese. They said that would dissuade — that was the term of art they used — dissuade the Chinese from undertaking a buildup.

Rob Wiblin: Why would that dissuade them?

Jeffrey Lewis: As somebody who spent a lot of time talking to Chinese nuclear weapons guys, I was like, “That is not dissuasion. That is incitement.”

Rob Wiblin: Right. It seems like it’s at least extremely ambiguous what effect that would have.

Jeffrey Lewis: Yes. And in particular, the Chinese were saying, “The fact that you have a much bigger arsenal and you’re building missile defences, we’re going to have to increase the size of our force.”

Rob Wiblin: Right.

Jeffrey Lewis: So there was a huge debate, which my side lost. So, we did it. We have big missile defences, no limitations, and we kept the size of our arsenal four or five times bigger than the Chinese, and now we’ve got silos in the desert. So regardless of me just saying that I think I’m right, we had a consequential debate about what was motivating the Chinese. We placed a bet and that bet did not pay off. We did not dissuade the Chinese from undertaking a buildup. They are now in the midst of a buildup.

Rob Wiblin: I’m guessing there hasn’t been a lot of accountability for that mistake.

Jeffrey Lewis: I mean, did anybody suffer any consequences for Iraq? Vietnam? Consequences are not a thing we do.

The importance of archival research [02:11:03]

Rob Wiblin: Right. OK, turning to the other thing you were saying. I didn’t expect this one, archival research. So we’ll send in people to read all these recently declassified documents in order to better understand the history. I guess some people might think that that, while fascinating, isn’t really at the pointy end of changing policy, of affecting the risk of nuclear war anytime soon. Do you want to make the case that actually that work really does matter?

Jeffrey Lewis: Yeah, because the people who say that are the people who then turn around and repeat the mistake that was made 10 years ago that they don’t even know happened. And I don’t think I really appreciated this till I got old, but I meet all these bright young things who roll up and they are like, “I have this brilliant idea.” And I’m like, “I have heard this brilliant idea before. It is in fact a brilliant idea. Let me tell you why this brilliant idea didn’t work.” And I’m not at all saying we shouldn’t try again, but we can learn, right? We can maybe figure out why it didn’t work then. Maybe it would work now. Maybe the conditions are different.

Particularly as someone who is focused on decision-making pathologies, I’ve got to tell you, when I read about the Truman administration, it does not feel black and white to me. It’s not an old movie. It’s not a period piece. They do the same things we do. They are the same human beings. So we have this big dataset, but we just decide, “But those are black and white movies, so they’re not part of our dataset.” So we throw the data away, and we don’t necessarily understand how we got where we are, and we don’t really have a tonne of insight into how things are playing out.

And by the way, I think that’s true of the protagonists. Maybe somebody like Joe Biden, because he’s really old, remembers the good old days. But presidents make the same mistakes over and over and over again — and it’s because they’re like, “Well, this is now and I’m new and I am different,” and then they proceed to do exactly the same thing over and over and over again. So I don’t think people learn all that effectively, but I think we can if we make a choice.

Rob Wiblin: Yeah. I suppose understanding all of the history and learning the right lessons is very challenging and can involve a lot of work. Especially given that there’s many decision-makers in this process who are not specialists, who are not going to actually go into a PhD to learn all of that.

Jeffrey Lewis: Right.

Rob Wiblin: But I imagine it’s all made significantly worse by the fact that so much of it is secret and classified, and the decision is extremely opaque and people are maybe trying to hide what happened because it’s embarrassing. It makes it very easy for the same thing to happen again, and for people to not realise at the time that they’re doing what happened 10 years ago.

Jeffrey Lewis: And then that becomes the benefit of patterns, because people tend to tell the same kinds of stories. So often I will hear someone’s version of events, and I will recognise the narrative and that will cause me to be a little sceptical — because it’s a shopworn narrative I’ve heard before, but it also then guides me to look for the little inconsistencies.

So I’m not someone who really believes there’s a past. I don’t think the past is a fundamentally different place. I think that it makes sense to think of the past as being this parallel and related place that does largely overlap with the world we live in today.

Rob Wiblin: A really famous episode in US nuclear policy — sufficiently famous that even I know about it — is that in the late ’40s and ’50s, as I understand it, the Air Force massively overestimated the number of nuclear weapons that the Soviet Union had. Conveniently for them, I suppose, because it then justified a massive spend on rapid construction of lots of bombers and development of ICBMs and so on. It sounds like possibly we’ve made a similar mistake again, in a way, of allowing ourselves to be lulled into thinking that an arms race is taking place when maybe it’s not. Actually, no. I suppose the situation with China is different. I should scrub that bit.

Jeffrey Lewis: Well, we don’t know. No, we actually don’t know. One of the reasons I think civil society’s finding of those silo fields is so important is the US government had found them, and they believed that every single silo is going to have a missile, which is a huge buildup. And if they’re right, that is a really significant development, but it’s a specific kind of significant development.

When we looked at the way the silos were arranged, they struck us as being much more like a scheme the US had in the late ’70s called the shell game — where, because silos are very vulnerable because you can’t hide them, you build lots of them, because they’re very cheap to build and you put a smaller number of expensive missiles in them. And statistically, that is the best way to ensure their survivability. It sounds crazy, but you can build so many more silos for the cost of one more missile that you get more survivability from one missile and a bunch of silos than from two missiles and two silos. Just how the math works.

And so we said, “You don’t actually know that those are full, and that’s a very different kind of posture.” And my understanding is that had not been given a serious hearing in the US government at all, even though it strikes us as an important and consequential possibility. And when we really pushed DOD on it, to their credit, they were like, “You are right. We don’t know. We’re going to assume they’re all going to be filled because that’s our job. But you’re right, we can’t prove it.”

Well, it matters, right? I get that we can’t prove it right now and that it’s a thing that requires study, but whether they plan to build several hundred or 50 hidden in several hundred has an enormous impact for how we think about stability and balance and deterrence.

Rob Wiblin: And what their intentions are. Yeah, you mentioned earlier that no one believed, or maybe no one believes, China’s no first use policy. Do you believe it?

Jeffrey Lewis: I do, stipulating that I think I understand what it is. Which is to say it’s not a promise they made to us. A lot of people are like, “They say they’re going to be nice to us, but they could change their mind.” What it actually is is an ideological position. Because when China conducted its first nuclear test, there was a movement to ban atmospheric testing. And most of the countries that supported that in the world were China’s friends, so the Chinese were in a weird spot: they needed to atmospherically test to develop the weapons, and it also made their friends angry.

They needed a propaganda response. And the propaganda response was to say, “Hey, how about this? How about we all agree not to use nuclear weapons first against one another?” And there are bureaucratic reasons they supported that. The advocates for nuclear weapons in the Chinese programme were scientists who wanted to do tests, but didn’t necessarily want to build a lot of nuclear weapons because they wanted to spend the money on R&D.

So it acquired this quality of being a thing that has been said publicly by people like Mao. And so if you’re a colonel, you can’t really argue against it. But it’s not really a promise to us. It was a statement about how they thought about nuclear weapons, which is that their theory is the US uses nuclear weapons to push them around and it makes sense for them to say, “By the way, you can’t use them first. That’s not a credible threat because we have them too.” If China got into the business of nuclear coercion at that time, then they’re kind of subject to nuclear coercion, right?

So I feel like, yes, it is an ideological statement that Mao made that no leader really wants to revisit, and that does impose some constraints on them. At the same time, it was left to the military, which over time has come up with different answers to work out, like what does that mean in practice. And so I believe it, but what I believe is a little bit more complicated than just they wouldn’t.

Rob Wiblin: Something more subtle.

Jeffrey Lewis: Yeah. You can never ask an academic about his dissertation. That is death. Four hours later you’re like, “I want to go home.”

Rob Wiblin: No, I think it’s super interesting. I thought that they might have developed an ideological attachment to the idea of a no first use policy from the idea that they’re taking the high road — that they’re the people who all along have been against aggressive use of nuclear weapons, and they look down on other countries that are more frivolous about them or use them for coercion. And maybe they would get attached to their own positive self-image.

Jeffrey Lewis: In a way they do. Although what they say is that the countries that threaten others with nuclear weapons do so because they’re the ones who are truly afraid of them, right? So it’s this phrase, “A frightened dog barks the loudest.” They think if you are out threatening countries with nuclear weapons, the reason you’re doing that is because, what are you really afraid of? Nuclear weapons.

And so their argument for the reason that they feel that, even though the US has so many more nuclear weapons, they’re not coerced is because they’re not frightened. And because they’re not frightened, they don’t threaten people with nuclear weapons.

Rob Wiblin: I see.

Jeffrey Lewis: Because everybody knows deterrence will hold. And it has a certain smug quality to it, which I think is not totally false.

Rob Wiblin: It’s like, “Take a chill pill, guys. We can all be calm together.”

Jeffrey Lewis: They’re like, “I also have a button.”

Rob Wiblin: “La-di-da.”

Jeffrey Lewis: It’s like, “I’m going to push the button. I’m going to push the button.” Like, “I also have a button.”

Jeffrey’s policy ideas [02:20:03]

Rob Wiblin: Yeah. OK, for our final section, let’s push on and talk about what you think are good ideas or good marginal changes that we could make, all things considered, to US strategy, or I guess possibly NATO strategy, or the arsenal. And for this section, I want you to imagine that you have a reasonably free hand changing things — that we’re not just going to think about what’s possible given all of the institutional constraints and all of the different interest groups and so on.

Yeah, if you were one of the top advisors to the president and they were willing to go along with your ideas, what is a change that you would make to US nuclear policy?

Jeffrey Lewis: There are three things I really want to do, and they go from little to big. Little is: I really think the president should say, “We would never use a nuclear weapon if a conventional weapon would work as well.”

And the reason I say this is, earlier I talked about how often there’s this kind of disconnect between what the president really thinks and what the plans are. And a lot of what we do is try to surface those tensions to resolve them in positive ways. I really think we’ve gotten to a point where nuclear weapons have an almost metaphysical aspect to them. They’re lousy weapons. The military tends to not want to use them in conflict, but we imagine that we’ll invoke them because they have this magical power to make the other side give up, which I think is crazy. I think that’s total non-falsifiable metaphysical mumbo jumbo. Would it? Maybe? I don’t know. There’s a bunch of unanswered questions.

What I really think the value of the nuclear weapons is is the destruction they do, and they should be treated like any other weapon. I don’t think Kim Jong-un cares if he gets killed with a nuclear weapon or a regular weapon. So by saying we’re going to compare these to other weapons, they have real serious political downsides, so we would only use them if they provided some weapon-like quality that was irreplaceable. My guess is the number of targets in the world would drop to 10.

Rob Wiblin: I guess you mean really hardened silos or something like that that couldn’t be attacked otherwise?

Jeffrey Lewis: Yeah. Yeah, like hardened command structures. And that would force us to have a really straightforward debate about what are these things for, and what do we think they do? So that’s one thing.

Rob Wiblin: OK, what’s number two?

Jeffrey Lewis: Number two is ratify the Comprehensive Nuclear-Test-Ban Treaty. The US, China, and Russia are all modernising our nuclear test sites. If you look at the open source information, it is busy, busy, busy at all three. All of them are preparing to resume nuclear testing in case the other ones do. I think we’re going to stumble into a renewed era of nuclear testing, which will accelerate the arms race. The best way to do that is to ratify the damn treaty that says you don’t, and to try to put in place some additional verification measures and confidence-building measures to stop that outcome from happening.

Rob Wiblin: Who has ratified that treaty and who hasn’t?

Jeffrey Lewis: It’s a multinational treaty, so there are many signatories. The big issue is there are a small number of countries that must ratify it for it to enter into force.

Rob Wiblin: I see. And the US is one of them?

Jeffrey Lewis: The US and China are big holdouts. There are other holdouts. So simply getting the US and China in — the Russians are already in — wouldn’t solve all our problems, but it would put us in a position to do maybe some stuff on the side that could stop the three of them from resuming testing. Which I think we’re not going to want to go back to a world of testing, but we might.

Rob Wiblin: Wow. OK. And what’s number three?

Jeffrey Lewis: Number three is a little bit conceptual. We’re back to Tom Schelling. Tom Schelling thought the greatest accomplishment that academics made, and the only good thing we ever did in arms control, was convince people that missile defences or ABMs were destabilising.

People think defences are good, and how can my defences be bad? But the reality is, if I have a shield and a sword and you have a sword, I’m going to hit you with my sword and then parry or block your thrust, right?

Rob Wiblin: Yeah.

Jeffrey Lewis: So that view became dominant in the 1960s. It’s why we got the ABM Treaty, and it was the foundation for all the arms control that came after it.

The US left that treaty in 2003. The Chinese and Russians have made it clear that they cannot possibly imagine limiting their offensive systems unless there’s a corresponding limit on the defensive systems that they have to beat. And that makes sense, right? It’s two sides of the same equation. But in America that is political poison.

So if you give me a free hand, a magic wand, that intellectual change of accepting that a world in which we have defences and offences is a world in which they both have the same thing — and that’s a world of an unconstrained arms race. That’s bad. And if you want arms control to work, if you want to limit offensive numbers, you have to have some kind of corresponding limit on defences.

Rob Wiblin: Yeah. Is there any way of selling that to the public or the armed forces? I suppose maybe you could treat it as a budgetary measure, that the stuff is too expensive maybe?

Jeffrey Lewis: It’s fantastically expensive, but Republicans love it.

Rob Wiblin: Do they not believe that it’s counterproductive?

Jeffrey Lewis: They believe defences are stabilising because defences are good. Which, by the way, is what the Soviets believed initially too, right? Defences are good. It’s a very simplistic argument.

Rob Wiblin: But the argument is very obvious. I mean, it doesn’t take —

Jeffrey Lewis: But Ronald Reagan didn’t get it. Ronald Reagan was like, “Defence is good. We would make nuclear weapons impotent and obsolete because I will build a leak-proof defence and then we will abolish all the nuclear weapons.”

Rob Wiblin: Right.

Jeffrey Lewis: I mean, it sounds goofy, but remember, a majority of Republicans don’t believe in climate change, evolution… And I’m sure you can pick things Democrats believe that are goofy. We all have our ideological commitments. And for Republicans, it’s that defences are good.

Rob Wiblin: Well, those are three very interesting. They sound like great changes. Is there any way that we could sell the Nuclear-Test-Ban Treaty to the public or to politicians?

Jeffrey Lewis: Yeah, I think so. Obviously the atmospheric test ban was easier because there were fission products showing up in children’s milk, which was bad.

Rob Wiblin: I see.

Jeffrey Lewis: So this really closed the last bit of banning all nuclear tests, which really included underground nuclear tests. But yeah, I made a tongue-in-cheek suggestion, which is the only time I got an angry call from a member of Congress’s office. Senator Kyl from Arizona was talking about how the US needed to resume nuclear testing, so I suggested that we put the nuclear test site in Arizona and name it after him, and his office called me, very angry.

Yeah, I have often suggested that if you want to get the CTBT ratified in the United States, we ought to do a survey where people go up and knock on doors and just ask people’s opinion about how they feel about the possibility of greenfielding a nuclear test site down the road. Tell them, “Very limited shaking — no more than two, three times a year. You feel that rumble? That’s the feeling of freedom.”

Rob Wiblin: We need to wheel out a force more powerful than the armed services, more powerful than nuclear weapons.

Jeffrey Lewis: That’s how I feel.

Rob Wiblin: Well, those three are great. So I’m tempted to ask, is there a fourth or a fifth or a sixth?

Jeffrey Lewis: I mean, I’ve got more, but boy, if we get one of those three things in my lifetime, I’m going to be real happy.

What should the US do regarding China? [02:27:10]

Rob Wiblin: Very chuffed. What approach should the US adopt towards China regarding nuclear weapons? All things considered.

Jeffrey Lewis: That’s a big question. So, here’s the nub of the problem. When I got into this business, China’s arsenal was small. It was focused solely on deterring US nuclear coercion, and it was pretty relaxed. It was kept off alert. And over the years, I have argued that we are taking that for granted. And that if we don’t engage with the Chinese seriously, if we don’t accept some limitations on missile defences, if we don’t find a way to improve our relationship, that cannot last forever.

Now we’re in the period in which it’s no longer lasting, and we’re seeing really rapid change. What I don’t know is how much of this is just purely service driven. The Second Artillery is now the Rocket Force, and they’re much more powerful and they get their way. And with getting their way is money and budgets and cool toys. I don’t know how much of it is Xi Jinping and his outlook, and I just don’t know how much of it is just the general collapse in our relationship with China.

So what I like to say is that we have to at least begin a dialogue with them about some concept of shared stability. I don’t think we have the same ideas about stability. And that basically is us agreeing that we don’t seek to coerce them with nuclear weapons into accepting something like Taiwan being independent, and them admitting that they’re not trying to catch us in terms of parity.

But over the long run, I think what we have to try to figure out is, what is life going to look like after Xi Jinping? One thing people in Washington are really bad about — and I’m guilty of it too — is we act like Putin and Xi Jinping are forever. These guys are both within five years, I think, of the life expectancy for men in their country. And they’re rich and powerful, so they’ll live a little longer than that, but they’re not going to be with us in 20 years, right? So we need to look past the immediate situation, and try to imagine what it looks like if you focus over a 10- or 20-year time frame. What does it look like to build a relationship where even though we have real disagreements, we still are focused on that shared interest in not having a nuclear war? People don’t like having that conversation, but it’s crucial.

Rob Wiblin: So I, as an idiot, do not understand why the US and China don’t already have some agreement regarding nuclear weapons. Because it just seems like even if both sides kind of hate one another, it’s in both sides’ interests to have fewer nuclear weapons on high alert, to not have to do massive amounts of research into new weapons, to count on what the other side is doing, to be in an arms race, to be testing stuff all the time. It seems like there should be a large space for a negotiated agreement that is in both sides’ interests — from a security point of view and from a money-saving point of view. And yet there isn’t.

Jeffrey Lewis: Well, that was sort of the Chinese position. Because when this started, they had 100 nuclear weapons and they were off alert and they had a no first use pledge, and they didn’t have missile defences. And the US had all these things, had a big alert force.

The Chinese view was like, “You need to accept some limits.” And one of the things the Chinese wanted was for the US to pledge not to use nuclear weapons first against them. Another thing is that they wanted the US to say that it accepted what has come to be called “mutual vulnerability.” Which is funny, right? MAD, the calumny: mutual assured destruction and assured vulnerability, the two parts of the calumny. The mutual vulnerability survived and the assured destruction part is gone.

They wanted the US to accept that. They wanted the US to ratify the CTBT. They wanted to engage. And the US position was like, “There’s nothing in that for us. We would be the ones giving all these things up and you’re only promising not to do something, and we’re going to just dissuade you by keeping a giant force.” That’s the bet we made.

Rob Wiblin: But what you’re gaining is that they’re not going to go do this thing that they’re now doing.

Jeffrey Lewis: I tried to make this point many times over the course of a decade.

Rob Wiblin: Lost the argument.

Jeffrey Lewis: But it’s very hard to argue that you should give something up in the hopes that this other country won’t do something.

Rob Wiblin: Yeah. Well, now that they’re doing it, now can we have an agreement where they agree to stop doing it?

Jeffrey Lewis: I’m hoping we’re in that sweet spot between it’s just scary enough that we’re motivated, but it’s not so scary that we freak out and build 10,000 more nuclear weapons. And that’s a delicate little window to try to manage.

Rob Wiblin: Window of opportunity. Well, fortunately the relationship is otherwise blossoming, so this should be straightforward.

What should the US do regarding Russia? [02:31:42]

Rob Wiblin: What could we potentially do regarding Russia? As I understand it, the New START treaty expires 2026, is it? 2025? And it can’t just be renewed, so we’ve got to have some new agreement. But they’re not really heading to the same dinner parties at the moment. Not a lot of chitchat about this. What are the least bad options here?

Jeffrey Lewis: Yeah, it is very difficult. Arms control, threat reduction, all that stuff is always hostage to the general geopolitical relationship.

Rob Wiblin: That actually is a mystery to me, or it is very strange.

Jeffrey Lewis: Right. Because you’re just as dead in a nuclear war when you’re not getting along as you are when you are getting along. Yes, you’re right.

Rob Wiblin: So maybe the background relationship is terrible, but that actually doesn’t change the calculation around the nuclear treaties almost at all.

Jeffrey Lewis: It should make it more salient.

Rob Wiblin: Right.

Jeffrey Lewis: The fact that a nuclear war is now more likely should make you more —

Rob Wiblin: You need it more.

Jeffrey Lewis: Yes.

Rob Wiblin: It’s OK. It is not that way.

Jeffrey Lewis: In America, we have this idea that you only sign treaties with good guys and your friends. So it’s just harder to do. After the Soviet Union invaded Afghanistan, Carter pulled the SALT II treaty at a time when we needed it more.

I think the fundamental issue with Russia, which is also the issue with China, is the Russians are modernising their force to defeat the missile defences the US has and that they fear the US will build. And until the US is willing to cap those defences, the Russians are going to keep building to overwhelm them. So unless we can find a way forward on limiting — I hate even calling them defences, because they’re not defences — but limiting ABMs again, you’re going to have the arms race. It’s just going to happen. So that’s the thing that’s got to change.

Rob Wiblin: Reading between the lines here — so setting aside anything that we think about China and Russia as to how they act in general, how they act towards their population, how good their systems are — it seems like from a purely nuclear escalation point of view, the US has been the bad guy lately. It’s the group that is not willing to come to the table to make these agreements on reductions in arms numbers or in preventing arms races, preventing testing races.

Which I suppose from an impact point of view is good, because the people listening to the show have more influence over the US than they would over China, so there’s more possibility of fixing US policy if that is the key broken piece here.

Jeffrey Lewis: Yeah. I mean, good guy/bad guy language gets pretty loaded. What I would say is we were the strongest party, and that gave us the most ability to shape the environment. We chose to really go all in on missile defences and not really limit our nuclear arsenal significantly, and so that shaped the environment. So that was a choice we made.

We also have a broken political system where we don’t do treaties. We struggle to make diplomatic agreements because the other party criticises you for being weak. And so we really find it hard to make and keep agreements — especially agreements that are sustainable, because a sustainable agreement is where you give something up in exchange for something, right? Where both parties get something. But in the American political system, whatever the other party gets, which makes the agreement sustainable, is a concession — and that’s a bad thing; that’s a thing you’ve given away. So it’s very typical to see US politicians, particularly Republicans, use words like “appeasement” and compare deals to Munich. I mean, that is our political system.

And so that combination of being the strongest and being the most broken…

Rob Wiblin: I just don’t understand the attitude that we need to come up with an agreement with these other countries that provides no benefit to them. And yet it’s strange that that isn’t happening.

Jeffrey Lewis: As Dick Cheney said, “We don’t negotiate with evil. We defeat it.”

What should the US do regarding Taiwan? [02:35:27]

Rob Wiblin: Yeah. Do you have any comments regarding what we should do re: Taiwan specifically?

Jeffrey Lewis: Taiwan’s a tough call, because we don’t want to throw a democratic country to the wolves. There are millions of people who live in Taiwan, they have an elected government. It’s very hard to say, “I’m sorry. Our relationship with China is more important. Goodbye.”

On the other hand, it is a neuralgic issue for the Chinese government. And while I think their position is unreasonable, the empiricist in me says that it’s not possible for Xi Jinping to be chill about it. So, I think this is one of those situations where we really have to muddle through.

And if you look at the fundamental bargain in the Taiwan Communiqué — which I find people always get confused about, or it’s expressed poorly. To me, it seems very simple. The bargain was, we accepted that the Chinese were never going to let that island go. And our requirement is that however it’s settled, it has to be settled peacefully. It cannot be settled by force. Our red line is you don’t get to blitz the island, right? So, Taiwan doesn’t get to be independent. We get that this is a make-or-break issue for you, but whatever you do, you cannot use force. And no one really likes that, but that strikes me as being sustainable.

Rob Wiblin: As good as we could do. Yes.

Jeffrey Lewis: I mean, there are a lot of situations that are miserable, but sustainable. And I think we often flirt with defecting from it because it’s so unsatisfactory. But it’s really hard to justify having a nuclear war over Taiwan.

Rob Wiblin: It is, yeah.

Jeffrey Lewis: But as a friend of mine used to say, “There are principalities involved.”

Rob Wiblin: Yeah. Muddling through, I think, is horrifically underrated. Muddling through is what we do most of the time, and most of the time it’s all right.

Advice for people interested in working on nuclear security [02:37:23]

Rob Wiblin: OK, you’ve got to run, but one final question, or one final theme. People in the audience, some of them are interested in going into this discipline, trying to make humanity safe from nuclear weapons. What would you like to say to those folks?

Jeffrey Lewis: Do it. It is, in many ways, a wonderful field, because for all of the annoying bits about the gatekeeping and the technical piece and the arcane bit, you never stop learning. I learn something new every day. Every day I find some new story, I understand something a little bit better. It is a giant, complicated field that will intellectually reward you every day.

And it’s one of the most important problems that we face. I accept that there are probably several of these problems that we would put on the table. But if you really want to dedicate your life to trying to work on a problem that matters, I don’t think anyone says this one doesn’t matter. This is a big one.

I have found it to be an enormously rewarding career with the proviso that you probably need to have a slightly dark sense of humour and maybe a tad cynical view of the humans.

Rob Wiblin: Yeah. Well, I came into this conversation hoping to learn some stuff, and I feel like I’ve learned an awful lot. I’m not sure exactly why, I’m going to have to think about this, but I’m also coming away somehow more optimistic. It’s possible to make a difference. Maybe you’ve just given me a bit of clarity where previously things felt very hard to understand.

Jeffrey Lewis: I hope so. And it’ll probably all turn out to be a mirage.

Rob Wiblin: My guest today has been Jeffrey Lewis. Thanks so much for coming on The 80,000 Hours Podcast, Jeffrey.

Jeffrey Lewis: It was a pleasure.

Rob’s outro [02:39:13]

Rob Wiblin: OK if you enjoyed that I think the episode you’d most like next is:

• #93 – Andy Weber on rendering bioweapons obsolete and ending the new nuclear arms race

Then perhaps check out:

• #118 – Jaime Yassif on safeguarding bioscience to prevent catastrophic lab accidents and bioweapons development

And finally:

• #125 – Joan Rohlfing on how to avoid catastrophic nuclear blunders

Just a reminder that on our other show, 80k After Hours, we’ve released a new interview with Marcus Davis — a founder of the research consulting group Rethink Priorities, which helped produce some of the work Jeffrey is very broadly reacting to in this conversation.

Among other things he talks about careers in generalist research where you study really neglected topics. If you’re interested in global priorities research or cost-benefit analysis in new areas, or the effective altruism community more broadly, do go check that out on the 80k After Hours feed.

All right, The 80,000 Hours Podcast is produced and edited by Keiran Harris.

Audio mastering and technical editing by Ben Cordell.

Full transcripts and an extensive collection of links to learn more are available on our site and put together by Katy Moore.

Thanks for joining, talk to you again soon.