Jake Sullivan
Biden's NSA on Playing the Long Game
After five long years since his last ChinaTalk appearance, Jake Sullivan returns to the show.
We discuss…
Sullivan’s experience managing crises, implementing grand strategy, and cultivating leadership skills during the Biden administration,
The art of crafting aggressive industrial policy, from chips to rare earths to infrastructure,
The risk of miscalculation in the Taiwan Strait, and whether Pelosi’s Taipei visit was a mistake,
Russia’s nuclear brinkmanship and the development of Biden’s posture on Ukraine,
Whether Trump can succeed at ratcheting down tensions with China.
Listen now on your favorite podcast app or on YouTube.
A reminder: this is the conversation I wanted to have with Jake, not the one you want me to have. For other recent interviews that get more into the Biden administration around the withdrawal of Afghanistan, the pace of arming Ukraine, and America’s handling of Israel’s invasion of Gaza, see all these other shows he’s done this year.
Playing to Win
Jordan Schneider: Jake Sullivan, Biden’s former National Security Advisor, currently a professor at the Harvard Kennedy School, and my near peer in podcasting. Jake, welcome to ChinaTalk.
Jake Sullivan: Thank you for having me. Now I have a whole different vantage point on being a guest on a podcast, so I’ll spend my time silently judging you.
Jordan Schneider: Great, we’ll have an impromptu masterclass.
You’re calling your new show The Long Game. What are your reflections on how crises interact with the goal of maximizing national power, or however you want to define the long game?
Jake Sullivan: Part of the reason we’re calling it The Long Game is that it’s incredibly important for us to lift our heads up and out of the smoke of immediate crises and ask, how do we put the US on the strongest strategic footing going forward? That really is the ultimate essence of the long game — how do we marshal and husband the sources of American power in service of our national security, our prosperity, and our values? That’s the ethos behind The Long Game.
Now, to your question about the interaction between crises and the long game — it takes an enormous amount of discipline, especially in my time in the seat when we were dealing with a lot of crises and a lot of different types of incoming, to say we’re going to set aside the time, the effort, the resources, and the top-level attention to actually focus on the long game. I’m actually quite proud that through Ukraine, the Middle East, Afghanistan, Chinese balloons, and lots of other stuff, we set aside the time to really invest in our alliances, invest in our industrial base, and invest in a set of technology policies that advanced America’s capabilities and helped protect those capabilities from being used against us.
That requires discipline — a huge amount of discipline — and I tried to bring that discipline with me to work every day. I was also lucky to have a lot of people who were assigned to the long-game things and not just the crises, who had no trouble banging on my door and saying, “Jake, don’t forget we have to be working on this issue,” whether it was chip controls or infrastructure projects in Africa or what have you. Having that kind of team around you — people who keep you honest and say we’re not going to let you just get swallowed by the inbox — that was very important as well.
Jordan Schneider: Is there a design fix around this? Should there be two NSAs — the firefighter and the long-term person? Why does this need to be one job?
Jake Sullivan: That’s a great question. Maybe it’s right that if you think about the close advisors to the President, there shouldn’t be any reason you couldn’t have two people essentially dividing the job. My overall reaction is that it’s probably a process fix better done with other senior people who are devoted to specific crises, not just someone who’s like, “I come in and handle all crises.”
A National Security Advisor should actually have both the dispensation and the instinct to say, “I’m just not going to spend a lot of time on this particular set of issues. I’m going to have the Principal Deputy do that, or an envoy do that, or someone else in government do that, because I need to be allocating more of my time to the longer-term things.” That’s probably a better way to do it than formally dividing the role into two.
One thing I reflect on from my time in government is, should I have just more consciously and systematically said, “This is something I’m putting on someone else’s plate, and I will check in on it every now and again, but it’s not going to be something I’m responsible for”?
Jordan Schneider: What were the things that you decided you needed to own and drive personally very aggressively, and why did you make those calls?
Jake Sullivan: First, I felt I was unusual as a National Security Advisor in that I thought I had to be at the table as an advocate, a designer, and to a certain extent an implementer of domestic industrial policy. I thought it was important that we have the national security perspective on that and that we’d be pushing that forward on chips, on clean energy, on infrastructure. That was one thing that I put a lot of effort into, particularly in the first two years.
Second, I thought the intersection of technology and national security was going to be defining. We created from whole cloth a new directorate on Technology and National Security, and I saw it as my responsibility to stay on top of that set of issues — yes, semiconductor export controls, but also other areas, such as biotech, quantum, and as I mentioned before, the clean energy transition.
Third, I believed from the beginning that a defining feature of modern geopolitics is the competition between the US and China. I thought it was my responsibility to play a central role in the design and execution of our China strategy and in the management of the US-China relationship.
Those were areas that I felt, regardless of what else was going on anywhere else in the world, I had to be focused on. Tied to all of those — industrial policy, technology policy, US-China — were allies. That meant husbanding and stewarding our relationships with core allies: the G7 plus India, Korea, and Australia. I put a lot of thought, energy, and effort into trying to elevate those relationships so that they were in the strongest possible state.
Jordan Schneider: That is a lot of stuff.
Jake Sullivan: Yes. I didn’t even mention the war in Ukraine. Or the multiple overlapping crises in the Middle East, or the drawdown from Afghanistan. There’s a lot going on. But the things I just named — they’re also interconnected. They’re not just different piles of work. At our best moments, they were a coherent strategy that we were trying to drive.
Jordan Schneider: You were one person both pushing the long-term stuff and being the crisis manager — what are the upsides to having that all be on your shoulders?
Jake Sullivan: First of all, if you just take it from the top, the President of the United States has to do everything. There is no division of that job into multiple different units. At some point as you go up any institution, any organization, you get to a CEO or a cabinet secretary or a President or a National Security Advisor. The question is, can you have a multi-headed monster at the top of the National Security Council enterprise? That’s difficult because you’re trying to execute and run a process that allocates time, energy, resources, priority, and then unity of vision and coherence of execution across everything. At the end of the day, you need a single point of accountability for that. That’s why it has to be basically in the job of the National Security Advisor.
But I also believe that theories of delegation and relative levels of personal engagement on different issues have to be an important part of how someone in that role thinks about what they’re doing day to day.
On Age, Experience, and Getting Ground Down
Jordan Schneider: What do you think the pluses and minuses were of being the second-youngest National Security Advisor when you started the job?
Jake Sullivan: It’s interesting — I’ve thought a fair amount about this, and it may strike some of your listeners as surprising. I was 44 years old when I took the job. That’s only a couple years younger than Kissinger was when he took the job.
Jordan Schneider: He was 45 and a half. ChatGPT made us a chart. Not sponsored, by the way. McGeorge Bundy was 41, which is insane. Condi was 46. You’re not an outlier on the young side. The mean is 53.
Jake Sullivan: There are a number of National Security Advisors around my age. When I took the job, I felt very young in a way because you tend to think of the black-and-white photographs of people with gray hair. But that’s a solid list of folks in their mid-to-late 40s entering this job — in some cases not with a lot of government experience at all coming before them, in other cases, like Condi, with quite a substantial amount of government experience.
Pluses of being on the younger side include energy, stamina, the capacity to really dig in and do the job full bore, full scale, 24/7. Being willing to push the envelope and be creative and dynamic and say we’re going to do things a different way, we’re going to have a theory of the case and try to execute against it — that youthful vision, energy, and creativity matters. That’s not just the single person acting as National Security Advisor — it’s also the team they build and the dynamic they build.
In my case, I felt I was able to build a very flat structure where everyone could come in, challenge, question, raise ideas, and feel like they had a voice. There wasn’t some oracle up top with hierarchy. That made for a much better — well, first of all, it was a better working environment, just more fun to work in that way. But I also think it allowed us to develop more interesting, more creative, more dynamic strategies in critical areas.
As for downsides… having done this job for four years. I understand deeply, to my bones, the value of experience — both positive experience and hard-won experience. There’s no substitute for it. There’s actually no substitute for experience as National Security Advisor. It is a truly unique type of seat to sit in that in some ways nothing can really actually prepare you for.
Jordan Schneider: More on that then. What was year-four Jake Sullivan able to do that year-one Jake Sullivan couldn’t?
Jake Sullivan: Actually, the most interesting part about experience is that to essentially metabolize experience, you need distance. Ask me in a year and I’ll give you an answer to the question. I’m being a bit glib, but what I mean is it’s less that in year two I was suddenly able to do things I couldn’t do in year one. It’s more the accumulation of that experience and then stepping away and being able to say, “Okay, I now have lessons learned — things that worked well, things that didn’t work well.” As this conversation goes on, I’ll point out some of those, I assume.
Sometimes I think that actually the right way for someone to do a job like this is to come in for two years, leave for six months, and then come back for 18 months — or in for 18, out for a year, back. I don’t know, whatever it adds up to. Leaving even for a little while just gives you an angle on what you were doing in the trenches that you just don’t have when you’re sitting in those trenches. That’s an interesting model for how to think about this going forward: you serve for a while, then you step away for a little while, then you come back and say, “Okay, now I’ve actually had the chance to metabolize that experience.” But anyway, that’s an idle thought for others to consider down the road.
Jordan Schneider: On what dimensions do you feel like experience gets accrued?
Jake Sullivan: The first dimension is living through a crisis, and actually having to stare square in the face the hard trade-offs and the imperfect choices, the lack of clear information, the need to form assumptions in the shadow of uncertainty. You can read all about that, but until you actually have to live through it, you’re not going to fully understand what both the opportunities and limitations are. That’s one.
Second is converting vision into action. How do you actually turn the idea of an industrial strategy into results? You’ve got to go through the thick and thin of that and make progress, but also come up short, to really be able to expose and understand what the obstacles are, what the modes of operating are, and how we could have done things differently — faster, more ambitiously, more creatively. Those are a couple of examples of where having to actually do something is what teaches you what works, what doesn’t, and how to be most effective.
Jordan Schneider: What about on the management side?
Jake Sullivan: On the management side, it comes down to a combination of how do you get the best out of people. Experience can come both ways, to be honest with you. One thing that I observed over the course of my time in government is that you get ground down. Your kindness, your patience, your sense of joie de vivre just get ground down. You become more impatient. Government is hard, so you become harder. I had to constantly fight against that. That’s an asterisk, a proviso — set that aside. That’s how too much experience, too much time in the trenches can actually degrade you rather than enhance you.
But how you actually get the most out of a team, how you run a process, what works and doesn’t work with respect to trying to surface and crystallize options for the President, how to make the government as a whole all pull in the same direction — these are things that require trial and error. It’s a very human exercise. Any particular group of people is going to have its own psychology, its own operating capacity. You can take experience from one administration and it won’t map neatly onto another. But there are some broader lessons that you can learn from it.
Jordan Schneider: In 20th-century American history, we have a bit of a pendulum. On one side, we have FDR and Trump — presidents just winging it. On the other side, we have the reactions to that — Truman being like, “Man, this FDR guy was crazy, we need to create the NSC,” or the Biden administration saying, “This Trump guy was crazy, we need to bring in a Yale Law School person to organize things.” After doing this job, how do you see the trade-offs of both models?
Jake Sullivan: This is going to maybe be true to brand, but I actually think that the right answer is to try to land in the middle of that pendulum. What I mean by that is — rigorous, fair, honest process is really important to the long-term strength of the United States and to the discipline of strategy on the big picture. But I also believe that a President and a National Security Advisor need a theory of what’s really important, and they need to get after it.
I had two people that I really looked at to try to approximate — not emulate, because I couldn’t live up to these two guys — but approximate. They were Brent Scowcroft and Zbigniew Brzezinski. Scowcroft for process and for being an honest broker and for not coming in and just calling the shots, but rather teeing up the debates of principles to the President. Brzezinski, because he had a theory of geopolitics and a worldview that helped shape and drive decision-making, even as he dealt with a bunch of crises on the edges and margins of that. A blend of those two strategies is the best way to pursue statecraft.
Breakneck Industrial Policy
Jordan Schneider: Are there dimensions on which you wish you’d had more freedom of maneuver? What felt the most constraining?
Jake Sullivan: You’ve had Dan Wang on your podcast — I thought his book Breakneck was just incredibly thought-provoking and interesting. His overall thesis is that America is run by lawyers — of which I am one — while China is run by engineers. There’s a truth to that. Democratic administrations are even more run by lawyers.
It would have been great to have more freedom of maneuver to actually just build at speed and scale than we were able to accomplish, because we had every conceivable obstacle to being able to do that — from the defense industrial base to infrastructure to building a semiconductor fab in Arizona. That would be one area where more freedom of maneuver, more capacity to move fast, would have helped.
Two: The US is funny. We’re the richest country on the planet, with deep and liquid capital markets and a massive federal budget, and yet our ability to mobilize capital in an intentional way to serve national security ends, both domestically and globally, is wanting.
I would have loved more freedom of maneuver to be able to offer a value proposition to the countries of the Global South for building infrastructure and competing with the Belt and Road Initiative. It was hard to even get nickels and dimes to be able to do that.
That was maddening because for pennies on the dollar, we could have competed more effectively and can to this day compete more effectively. We’re not doing it because we can’t get that money.
Jordan Schneider: In that vein, what homework would you give researchers, future administrations, and anyone listening?
Jake Sullivan: We really need a deep and rigorous study of what the objectives of industrial policy are, what the limits of industrial policy should be, what tools work and don’t work, and then once it’s being applied, what are the obstacles to actually executing in a way that delivers results on a reasonable timeframe. To the extent those obstacles need to be adjusted, how do we adjust them?
That entire chain of questions — people are looking at some of it. Some of my former colleagues, like
and , who were central to the CHIPS Act effort, are really doing a deep study of some of this today [now on substack at ]. But there is not a body of work on this, in part because industrial policy was basically considered unacceptable to work on. We’ve got to bring it front and center, and not just with the core economics profession — that’s got to be a dialogue between national security professionals and economists coming up with a range of answers to those questions that are rooted in empirics and evidence and rigorous study.I would ask anyone out there who’s thinking about contributing to the national security literature of the future — this, to me, is a set of questions for which we need better answers than we have. Those answers will be in large measure defining of whether we’re able to pursue an effective strategy.
What If’s on Rare Earths and US-China Escalation
Jordan Schneider: The CHIPS Act is a long-game play, but there is a world in which you implemented the October 7th controls and China decided to play the rare earths card early. If you had woken up in that world, what would you have done?
Jake Sullivan: First of all, we wouldn’t wake up the next day and think about what to do, because we thought about that before we did the controls. We thought about retaliation risk. We thought about how to structure a strategy to protect our most advanced technology without getting on an escalation ladder that could end in harm to us or a downward spiral, or as we’ve just seen recently, the need for us to basically blink and back down.
We executed the controls in a way consistent with this theory of “small yard, high fence,” precisely to avoid the massive counterreaction that we have seen once President Trump decided to slap on 145% tariffs. That was a credible and sensible strategy.
Now, it’s totally fair to ask — and this gets to the point of hard-won experience — we knew that rare earths were a supply chain vulnerability for the United States. We knew that China was weaponizing it. We were explicit about that, open about that. We took steps on it. I could send you a note with a whole list of things that would all look like perfectly credible steps one would take to try to reduce that supply chain vulnerability. And yet it obviously didn’t end that vulnerability or even really, truly dramatically reduce it. Why?
I’ve reflected on this question, and there are a few reasons. One reason is that it is a new set of muscles for the government to intervene in markets where the markets have failed or where a competitor like China has taken advantage of those markets to dominate supply. We were trying to build that muscle.
Second, it’s a dynamic game. We made investments in US firms, we also made investments in allies, and China was counterreacting by cratering the price or driving a particular processing plant out of business. We hadn’t yet fully gotten up to speed in terms of that reaction-counterreaction.
The thing that concerns me right now is the Trump administration is taking good steps on this and they’re obviously motivated, but it’s still linear from where we were. We need to go really nonlinear. We need a much more aggressive strategy, in my view.
If I’d woken up the next day after the imposition of the semiconductor export controls and they’d played the massive rare earth card, I would have said drop everything — we need to solve this within the shortest possible span of time and essentially have Operation Warp Speed to get it done.
That is what the Trump administration should do. They’re moving in that direction, but now is the moment for more alacrity. I say that as someone recognizing — not enough alacrity in our administration, in the end, not enough alacrity. I’m worried that that remains the case today.
Jordan Schneider: Zooming back, we have two countries that are competing, which are deeply economically intertwined. Say you solve rare earths — it seems to me that insofar as the two countries have an enormous amount of economic interaction with each other, there will still be ways for either country to squeeze and cause the other one pain, even if it’s not commercial. We saw Volt Typhoon — “What the hell are you doing in our water treatment plants?” I’m sure you told them to knock it off and they didn’t.
It’s a game worth playing, but what does the incremental resiliency-building help the US when it comes to these negotiations? To use a wrestling metaphor — if you’re winning points in the round, but your opponent could still pin you if they really felt like they needed to.
Jake Sullivan: As far as homework — let’s fully map the vulnerabilities as far as we can see them. Let’s ask where they are. Let’s test this hypothesis you put forward that there is simply no end to interdependence and that there will be some fundamental vulnerability that can be exploited by China or, for that matter, by the United States. That is worth a deep dive, not just all of us talking back and forth about it.
I have three answers to this question.
There are different forms of vulnerability. Some act very fast, and the pain can be applied asymmetrically, powerfully, and swiftly. Some act much more slowly and give you time to adjust, or they act in a way that harms the country imposing that action, so there’s some degree of deterrence. That’s one answer — let’s not just have a list of all the interdependencies, but we need to isolate the examples that are like rare earths. How many of those are there actually?
I would prefer to have fewer vulnerabilities rather than more. If I can take some bullets out of your gun, could I still end up dead? Maybe. But if you gave me the choice between you having a full magazine and less than a full magazine, I would say let’s make investments to do that.
This is really important — the more a country shows it has the muscle to be able to adapt and adjust if another country tries to weaponize interdependence, the more of a broader-based deterrent effect it ends up having. Having the wherewithal to relatively rapidly identify and then close a gap, even if you know there are other gaps out there, has a knock-on effect on those other gaps too. For all those reasons, we should still have a resilience strategy, though I do acknowledge that this is a very legitimate question.
Part of the answer to it has to come down to a cost-benefit analysis of trying to deal with this resilience. In the rare earths case, the cost-benefit analysis seems to be pretty straightforward. This is not a massive market. It doesn’t require us to move heaven and earth to resolve the vulnerability. It requires some determined, concerted action, and we should take it.
On Yards and Fences 固若金汤
Jordan Schneider: The closest thing we got to a Sullivan Doctrine was given in September 2022 — “Given the foundational nature of certain technologies, we must maintain as large of a lead as possible.” There’s the maximalist version of that, where everything gets cut off, you’re Stuxnet-ing private companies. Then there’s the “small yard, high fence” version of Biden vintage, where every day you put out new export controls, and the next day we’re recording an emergency podcast talking about all the loopholes. I’m curious about your reflections on what the calculus was and what was constraining it. Was it domestic stuff? Was it political economy? Was it worries about retaliation? How did the level end up getting set?
Jake Sullivan: One element is retaliation risk. We talked that through, right? If you just say to China, “We’re essentially cutting you off altogether,” then you’re going to induce a really dramatic reaction. That’s what happened when we fired a bazooka of 145% tariffs — they fired a bazooka back.
A second element is trying to actually have some discipline about a threshold above which you have national security concerns and below which you’re just talking about broad-based commercial or consumer applications. We wanted to have the American idea that we’re not in favor of a total technology blockade. We are just in favor of focusing on those elements of technology that have genuine national security and strategic implications for us.
A third aspect is what you said about political economy — you’re trying to bring allies along, you’re trying to bring your private sector along, you’re trying to organize a government that has very different views on this.
If you and I sat down for an hour and went through ‘22, ‘23, and ‘24 — do I think we got all the calibrations right, all the levels? Of course not. First of all, in ‘22, they knocked out the interconnect and had the H100, and we had to update the controls in part because that interconnect speed criteria didn’t make sense. We were learning as we went — the stockpiling of some of the manufacturing equipment, those timelines we could walk through. But we were executing a novel strategy, trying to make sure that we were nurturing and sustaining our advantages, protecting our most advanced technologies, while at the same time dealing with this other set of considerations.
I had breakfast with a guy I respect a huge amount in the technology field who was basically like, “My only objection to ‘small yard, high fence’ is that it should be ‘big yard, high fence.’” He had an argument — we should control legacy chips, control it all. I actually think the experience of this year and how things have played out compared with the experience of the last three years is a good argument for “small yard, high fence” — for a particular style and strategy of a pretty aggressive policy that is conducted with a degree of care and precision. That was a more sustainable policy for the United States over the longer term than just letting it rip. Now, it’s hard to say for sure that that’s right. There’s no algorithm for this, but that was the judgment that we reached, and that’s why we proceeded the way that we did.
The last thing I would raise on this is, having watched President Trump do what he has done with our allies, was the right answer just FDPR — the Foreign Direct Product Rule — from the start? Basically, don’t negotiate with our allies over coordinated semiconductor manufacturing equipment controls. Just tell them they’ve got American content, you’re not selling, done. I think about that sometimes. That comes down to a question of what are the longer-term costs of treating your friends that way? I believe that there are genuine longer-term costs for that that are real and strategic and meaningful. But I can’t prove that. That’s something we’re going to have to watch over time because now we’re running a real lab experiment.
Jordan Schneider: All right, let’s talk about China. There was kind of a big yard with FDPR. Should we have been more aggressive with FDPR, for example, to really put maximal pressure on Huawei? Do you think China would have freaked out?
Jake Sullivan: The yard could have been smaller than it is, I guess. Here’s what I would say about some of these arguments — like, if only you do this extra thing on Huawei, then that happens. It reminds me of the “real socialism has never been tried” argument. Look, it’s possible you’re right — just one more crushing sanction or export control.
By the way, I’m being glib because you may be right. But the question that one has to grapple with is some of the assumptions that underpinned the Huawei controls that were put in place in the Trump administration and the statements about what would happen as a result of those controls — not the next, better version later down the line — didn’t quite bear out. We have to grapple with that, too. You may be right that just taking more steps in the Biden administration could have made all the difference.
Jordan Schneider: I identified the biggest “what if” is FDPR on equipment earlier, because the big chart is the one where SME exports double and double again and double again. It’s going to keep doubling. That seems to me to be the big fork in the road.
.jpg){kind=link}
Jake Sullivan: It’s a big question, and I’m not sure what the right answer is. We had a thesis with respect to our allies — you work with them rather than coerce them. That obviously meant China got access to more manufacturing equipment than if we had just coerced them. But in a net assessment of the overall health, vitality, and strength of those relationships and how they would play out in — to coin a phrase — the long game, that’s a big question.
President Trump is testing that question because he’s basically saying, “I don’t buy it. I can go punish all these people as much as I like, and they’re just going to have to keep being friends because they have no other choice.” He seems to think that’ll just be how it is. I happen to think what’s more likely is you get a group of countries that were coming along with us to de-risk from China, who now are sitting there thinking, “I’ve got to do what I’ve got to do today, but over time, I need to de-risk from the United States.” In the end, that is not going to be a winning strategy.
These are the kinds of calls that you have to make in the shadow of uncertainty. I will say we didn’t have knockdown, drag-out fights about FDPR. There was a general sense that we should try to negotiate this with our allies rather than just club them over the head. There are also other ways to get at this challenge than FDPR — like timelines on implementation and how companies went about allowing the stockpiling despite government agreements and so forth. That’s for another time, maybe over a drink.
Taiwan, Pelosi and the Risk of Accidental War
Jordan Schneider: You were present at the beginning of the Xi era. What do you believe about him that’s not the consensus?
Jake Sullivan: He is more improvisational than the theory that he’s just set out a coherent strategy that they are just going to execute. He’s basically like any other leader of any other big, unruly country, and he’s got to make a lot of stuff up as he goes along.
Jordan Schneider: Accidental conflict is something Americans worry about. I’m personally skeptical that if two countries really don’t want to fight, they can trip into a war neither side really wants. Am I wrong here? Is this something I should be more scared of?
Jake Sullivan: It’s such a fair question. It’s almost like a shibboleth — the risk of mistake, miscalculation, escalation. Ships bumping into each other in the South China Sea, and all of a sudden, you’re in World War III. I share a degree of your skepticism. There are restraining factors that can allow disengagement and de-escalation.
But let me give you a hypothetical scenario, and you tell me if it worries you. As you know, the PRC is pressing closer and closer to the island of Taiwan in the air and on the sea, bumping up against fewer and fewer nautical miles offshore, right? They’re doing that with manned aircraft, they’re doing it with ships, they’re doing it with drones. At some point, let’s just say Taiwan says, “We can’t tolerate this anymore. We’ve got to fire a warning shot, or we’ve got to do a fake dogfight with one of these planes to show them that we’re not tolerating the continued encroachment.” Then one thing leads to another and those two planes splash down. You think the next day it’s cool? Does that bother you? Does that worry you?
Because that scenario does worry me. Do I think automatically we’re off to the races? No. But that kind of scenario, in an already unstable operational environment — I don’t know that the risk of a tactical mistake leading to a change in the strategic situation — I’m not at one end of the spectrum on this that you’re pushing against, but I’m not quite where you are either.
Jordan Schneider: It wouldn’t be a nice thing. You’d lose some sleep over it.
Jake Sullivan: That’s the thought experiment to me. Why would you lose sleep over it? Because you’d be like, well, there is a possibility — maybe not that immediately the invasion force comes flowing over the horizon, but rather that it leads to a change in the national conversation on the mainland. It leads to arguments that we just can’t tolerate this, there has to be punishment and so on. Can that contribute to a shift in a negative direction that raises the risk of outright conflict? It can. I don’t want to overstate the case because I take your point, but that kind of scenario worries me more in a way than the US and China bumping up against each other.
India-Pakistan is another example where I think a mistake can lead to very rapid escalation. To me, it’s a little bit more condition-specific than just in the abstract.
Jordan Schneider: But isn’t the lesson of the most recent India-Pakistan crisis almost the other thing? It’s like, okay, we have a game now and we just play it every five years.
Actually, it probably doesn’t feel like that if you’re getting woken up at 4 o’clock in the morning.
Jake Sullivan: You know what? I’ll take that point. That’s fair because I basically agree with you that at the end of the day, the two sides don’t want to go to all-out war, so there are reasons for them to end up not doing so. I withdraw the India-Pakistan example.
Jordan Schneider: Was it a mistake for Pelosi to go to Taiwan?
Jake Sullivan: Look, I want to be fair to the Speaker. I’m going to answer your question, but I want to do it in a fair way. I spoke with her about going to Taipei, and she basically said to me, “All you White House Democrats and Republicans — you’re all too restrained. I should be able to do what I want to do, and nobody should tell us whether we can go to a city.” She was pretty clear and direct in her view.
I believe that the cost to Taiwan of that visit far exceeded the benefit to Taiwan of that visit. For me, it’s pretty simple calculus.
Jordan Schneider: How so?
Jake Sullivan: Well, it led to not just an immediate reaction by China that put a huge amount of pressure on Taiwan, but it led to a change in the operational environment around Taiwan that has not gone back to the way it was before — substantive, negative changes in Taiwan’s immediate environment. On the positive side, some symbolism, I guess.
Managing the Stress of Putin’s Nuke Threats
Jordan Schneider: On the accidental risk stuff, speaking of you getting stressed out — the Putin nuclear scenario is probably the scariest thing you had a 5-10% chance of seeing, I assume. You’ve already given some reflections on this, but maybe from — we had all these NSC management questions — this is the big one. What sticks with you?
Jake Sullivan: This is fall of ‘22. The Ukrainians are on this counteroffensive in Kherson and Kharkiv, and the intelligence community at the most senior levels comes to the President and says, “If there’s a catastrophic collapse of Russian lines and Putin feels that he’s in danger of potentially losing the war, there’s a 50% chance — a coin toss — that he will use tactical nuclear weapons to avert that defeat and shore up his lines.”
By the way, this is not just people guessing out of thin air. These are people who have studied this issue, are watching everything going on, who have amassed a fairly broad-based view across the intelligence community of this judgment. You’re the President of the United States, and you’re like, “All right, we’ve got to deal with that. Can’t be paralyzed by it. You’ve got to keep supporting Ukraine, but we’ve got to deal with that.”
We gathered in the Situation Room. We ran tabletop exercises. What would happen? What would we do in response? What would they do in response to our response? Not very pleasant scenarios, all told. We communicated directly with the Russians about the consequences of taking such an action. Of course, we reached out to, among others, China to get them to weigh in as well.
But this is the kind of thing where the difference between a commentator saying, “I don’t think it’s a very serious risk,” and actually being in the seat, having the responsibility to the American people of taking very seriously what sober, senior intelligence professionals are telling you while also continuing to support Ukraine — that was very real and very challenging.
I do not believe that this was all just BS. This was a risk. If it had happened — the first nuclear use since Hiroshima and Nagasaki — the United States would have had to take meaningful action in response. That action could easily have led to a totally different form of escalation between us and Russia. It’s good it did not happen. A lot of people look at the fact that it didn’t happen and say this was all overblown. I think we had some influence over it, and events on the battlefield had some influence over it as well.
Jordan Schneider: We’ve talked about a few different escalatory ladders — the econ-tech fights, planes hitting each other in the sky. But this is different than what they were theorizing in the ’50s and ’60s. Thoughts? What’s my question here?
Jake Sullivan: I think your question is, how the hell do you deal with that?
It raises a question about risk tolerance, right? You’re walking on a narrow mountain path and there’s a steep cliff off to one side — one side’s the mountain, the other side’s a steep cliff. The path is, call it five feet wide. Do you walk right on the edge, saying, “I don’t think I’ll slip”? Or do you walk up against the mountain?
A lot of the debates over the nuclear escalation thing is, why weren’t you closer to the edge? Why were you closer to the mountain? The right answer on this is you have to keep moving forward. You’ve got to go from point A to point B. You can’t stop providing weapons to Ukraine, intelligence to Ukraine, capacity to Ukraine. But you also have a responsibility to the American people not to fall off the cliff.
Jordan Schneider: Can you talk about psychologically, yourself and your team? It’s one thing to be in the CIA for 30 years and be thinking about Russian nuclear posture. It’s another thing when it’s the week when it’s more risky than any time since the ’80s or even the ’60s.
Jake Sullivan: One thing that I’m playing around with in my head — I don’t want to say it wrong because it’s going to sound somehow like I didn’t take my job seriously enough, and I took my job deadly seriously — is that stress and stakes and consequences normalize. What I mean by that is the human capacity to just adjust to your circumstances.
I was an associate at a law firm in Minneapolis working on commercial litigation. We had these cases that kept me up at night. I stressed out over whether we got something right or wrong. I second-guessed and recriminated, the whole thing. Then I end up in this nuclear scenario. These are night and day. But somehow human beings don’t just have an infinite level of calibrating stress. They just have their bands, and then they’re presented with things. Some of them are life-or-death situations, and some of them are life-or-death decision-making situations. You get up in the morning and you go to work and you do your job and then you go home at night and you execute your responsibility to the best of your ability to do so.
I don’t have a better answer than that, but I don’t feel like, “Oh, I was some special person.” I was just a guy who had a job. This was the problem presented in that job, and I had to deal with it. A lot of it is just about being able to inhabit that mindset and say, “I’ve got to go to work. This is my job today.” But that didn’t mean I didn’t sweat through a lot of shirts.
Jordan Schneider: It’s a good point. The Marines fought on Peleliu and Iwo Jima. You’re going to an office.
Jake Sullivan: Exactly. This is such a good way of putting it. Somebody is walking into the teeth of gunfire at Omaha Beach who, a year before, was a schoolteacher in my hometown of Minneapolis.
The ability to normalize just your situation — this is my job today. That’s so much more real than anything I had to deal with. Yet we all have really stressful and complicated situations in our lives. Even when something is objectively not that high stakes, I don’t begrudge people that they feel super stressed in that situation because they’re just operating within the range we all operate of stress and stakes based on their lived experience.
Jordan Schneider: But there have been very few humans in human history who have stared down a non-zero risk of nuclear war and managed through it. Even then, it still is a unique experience.
Jake Sullivan: It’s a lot. It is a lot. It’s heavy. The other important thing is we’re all just human beings leading three-dimensional lives while dealing with all of this — dealing with family stuff or health stuff or whatever the case may be. I talked before about how you can get ground down in these jobs and lose a certain sense of who you are. There is a way in which the stress and the stakes harden you. You don’t even quite realize it at the time.
Being vigilant for that, to try to remember at the end of the day that it’s your job to be a good and decent person, is a really important thing. That requires more discipline than often in a given day you can bring to bear.
Jordan Schneider: You said on a podcast you still don’t sleep well. Did you sleep well before?
Jake Sullivan: In Trump one, there was a period where I had a really hard time sleeping post-2016. But yeah, I would say I slept pretty well. I don’t now, because in many ways, a lot of the things that we dealt with had no perfect outcomes achievable. The outcome was going to be painful in one way or another. There were some painful outcomes, and then they happen and you’re like, “Well, what if I’d done this? Or what if I’d done that? Or what if we had done this? What could we learn from that?” You turn it over in your head, and then you do all of that with perfect hindsight, which is a super problematic thing to do because you could only make those decisions in the moment.
But I don’t know — that’s just going to be what it is for me for a while. I don’t think that’s a bad thing. In a way, the completion of discharging your responsibility in a job like I had is not to walk out the day you’re done and just go, “All right, someone else’s problem.” It’s to continue to wrestle with it indefinitely. That’s part of the service.
Jordan Schneider: You mentioned earlier that you think the thinking around AI and national policy is really poor. What are the hottest questions you wish there was better thought on?
Jake Sullivan: Let me be absolutely clear. If I said that, I don’t mean to say it’s really poor. I mean to say that we’re in the early innings. There are brilliant people thinking about this, and they’re thinking thoughts way beyond in complexity and sophistication that I could think. My concern is that we are designing strategy — and who’s the “we”? Because it’s a combination of this government plus these big private companies that are driving the frontier, without really fully unearthing the assumptions underlying those strategies.
A world in which we are rapidly approaching AGI and ASI versus a world in which it’s a boundary and jagged pathway towards greater capability — what you would do across a range of different inputs is different in those two worlds. Yet I don’t think we sit and work out what our relative confidence is in one world as opposed to the other. What are the can’t-fail or the must-dos, the no-regrets moves? Then what are the things we have to be able to adjust as we gain more information?
That’s just one example of many where getting technologists and strategists together to really go through the underlying assumptions of where all this is going is important. Most of this conversation floats above at a plane of abstraction. As a result, we pursue a policy with a lot of hidden assumptions that we haven’t fully validated or unearthed. That’s my main concern and something I’m turning over in my head and thinking about how to articulate better than I just did.
Jordan Schneider: Maybe more broadly, the process of gathering information, even though it’s under uncertainty and you’ll never get to much less 80% confidence with some of these things — how did you think about it and the sources, and how did that evolve over time for you?
Jake Sullivan: First of all, when you’re in government, access to information is amazing. My friend Kurt Campbell likes to say, “When you’re in government, the shit comes to you. When you’re out of government, you have to go find the shit.” There’s a simple, crude elegance to that. We could have senior people from the AI labs come in and actually present to us where they were and their capabilities and where they were going and what concerned them — literally in the flesh, just do it. We could also get all of the industry data and analysis synthesized by a team of people at any agency in the US government and supply it to you as well.
The biggest weakness on the US government side in terms of the consumption of information is that we tend to over-prioritize classified information over unclassified information because we somehow think it’s more special. Particularly in technology, it’s the unclassified information where you really find out what the hell is going on. But basically, all you can do is try to bring in as much as you can.
Then I believe in the debate method — basically having people on opposite sides or at different ends of the spectrum on a given question of what’s likely to happen actually debate it out, unearth it, and figure out where they agree and where they disagree and isolate the points of disagreement. That method gives you the best confidence that you’ve actually kicked the tires on all the potential perspectives. Then you just have to decide where you land, where you fall.
Jordan Schneider: Trump two does not seem to see China as the threat that you guys and Trump one seemed to characterize it as. Two questions fall out of that — to what extent do you think, if he really tries, he can change the tenor of this relationship? How structural is the competition, basically? Second, given that this is what they’re seeing, if you had foresight into this, would the bias have been to push harder or push less hard?
Jake Sullivan: The second’s a good question. Increasingly, it’s going to be a salient question because we are going to see more swings in US foreign policy in the years ahead. Future National Security Advisors will have to contemplate dramatic departures one way or the other. We’re seeing that play out right now.
On the first question, here’s what I think — It’s structural. The United States and China are two big, ambitious, dynamic countries with proud people who want to succeed and frankly don’t want someone else calling the shots or having undue influence or having greater capacity than they have. We have two different political systems, two different value systems. Therefore, competition is a feature of the relationship.
What I don’t think is structural is that it has to be conflict. Where we have the ability to influence things is to manage that competition effectively. But the idea that you can just wish the competition away and go with win-win, peaceful coexistence and all these kinds of phrases — I don’t really buy that. Frankly, I don’t think China buys it deep down either.
President Trump can talk about the G2, and I completely agree he looks at this through an economic lens, a mercantilist lens, and not a lens of strategic competition. But at the end of the day, these structural factors will shine through. I have to think about whether it would have made us do something different. That’s a bit of a mind-bender, projecting back and then forward, but it’s a totally reasonable question.
Jordan Schneider: You talked earlier about the personal weight of all this, and you said somewhere that you almost envy the people who can have more confidence in their calls. The irony, of course, is this new book about McNamara at war. He’s the one who projected the most confidence out of any of these folks, and he had panic attacks. It was actually all a cover for all his insecurity.
Jake Sullivan: I was being a little wry when I said that. That’s understated Minnesota wryness. What I meant was I just find it so interesting that people can assert things like, “Yes, it’s this.” You’re like, “Congratulations, I’m so glad that you just know that.”
Going back to the point about hard-won experience — most of these issues do not have easy calls. They have trade-offs and they have puts and takes. I really admire the self-confidence because I could sleep at night super easily then. It’s just a hell of a lot less pain and suffering, but I don’t think it’s right.
Jordan Schneider: I have three book recommendations for you.
Feeding Ghosts — this one is best suited for midnight reading. It’s actually a graphic memoir. It just won the Pulitzer. It’s a personal history with the US-China arc in it that I can’t wait for you to read.
The Social History of the Machine Gun — it is probably the most stylishly written military history I’ve ever come across. It goes through the technology, the companies, the acquisition side, and then the human and strategic implications. Fantastic. Someone’s going to have to write the social history of the drone. We didn’t get to drones today. We’ll have you back maybe. But that book’s a real treat.
To Run the World: The Kremlin’s Cold War Bid for Global Power. [Ed. Check out ChinaTalk’s interviews with the author here and here].
Final thing — what feedback do you have for ChinaTalk? What do you want to put on my plate or our team’s plate?
Jake Sullivan: Actually, bringing on some more people on the future of manufacturing, supply chains, industrial strategy — all the things that we were talking about a little bit ago that deserve so much more depth. Getting folks in to really unpack a lot of the kinds of questions Dan was posing that Chris Miller has been thinking a lot about. I know you do this, but doing a dedicated series on this subject with a specific emphasis on the particular challenge China poses and what the United States should be using government tools for to deal with it, and what we should not be, so that we don’t try to out-China China. That’d be a piece of advice I would give.
In studio!