
Sign up to save your podcasts
Or


Shashank Joshi is The Economist‘s Washington bureau chief, having previously covered defense and national security for the publication for eight years.
In this week’s conversation, Yascha Mounk and Shashank Joshi discuss why Russia is escalating its hybrid warfare campaign against Europe, whether NATO’s political cohesion can survive the Trump era, and how the United States has changed over the last 20 years.
This transcript has been condensed and lightly edited for clarity.
Yascha Mounk: I think the last time you were here, we may have teased that you’re going to have a new role. What is your new role?
Shashank Joshi: Having covered defense and national security for eight years at The Economist, I’m now The Economist’s Washington bureau chief here in Washington, DC. So, tracking America’s place in the world, or lack of place in the world, as well as all the great and not so great developments in American politics, looking ahead to the midterms, looking ahead to 2028, and everything this administration does.
Mounk: Perhaps at the end we can get your impression as an outsider in the United States and how you’re seeing things there. I was struck by a meeting in the Élysée Palace about a week ago at this point, a little less than that, in which Emmanuel Macron invited all of the heads of the French political parties, including the far left and the far right, and seems to have warned them about the worsening security situation for France, for Europe. It sounded as though he warned about the possibility of real hybrid, perhaps non-hybrid, attacks by Russia on Europe. How bad is the situation in Europe now?
Joshi: This is the big question of the week, Yascha. Amid all the noise, this is the real concern. There are a few things going on. You noted the French meeting, which was, of course, very interesting because it included opposition parties, and you saw these kind of leftist opposition parties come out of this and say, wow, we’re pretty chastened by what we heard, by the security situation.
Mounk: By the way, the far right as well. The Rassemblement National came out—which has taken loans from Russian banks before, I wouldn’t say they were pro-Russia, but certainly anti-anti-Russia—and Marine Le Pen and Jordan Bardella made a pretty remarkable statement, saying that of course, when Russia is threatening French and European security, they’re going to be patriots and they’re going to be on the side of French security. Whether or not to fully trust that is a different question, but it is still a clear break with the kind of rhetoric they had in the past.
Joshi: I suspect if you saw the same thing in Britain, Reform and other hard-right parties would probably respond in the same way. My colleague in Berlin, Tom Nuttall, made an interesting point, which is that that would be very unusual in Germany. That’s not what you’d expect to see in Germany, because you would never brief the AfD with sensitive intelligence in that way. But that’s beside the point. The point is the French government clearly sees a worsening security situation.
But the real news of the week in some ways was not just that—there were many other things afoot. Let me give you some examples. Germany, along with the Swedes, sent reinforcements with live ammunition to the Swedish island of Gotland in the Baltic Sea, and they said clearly this is not an exercise. In other words, they saw reason to be concerned by the prospect of a potential Russian operation against Gotland, which, in case you don’t know, is this island in the middle of the Baltic that is very strategically positioned, because if you can put air defenses and missiles on it, you can control a great deal of airspace and sea corridors in the Baltic Sea.
At the same time, we saw the Polish government—or specifically Donald Tusk—say that there was intelligence suggesting the risk of a Russian attack on Poland and other NATO allies. Not a conventional ground attack, but drones and missiles landing on NATO territory that he said the Russians might attempt to portray as accidental. He sounded a very grave and serious tone when he issued these remarks. Now, some of this was caught up in politicking at the highest levels of the Polish government.
Mounk: The Polish president, who is allied at least with the party of Law and Justice, the previous government—I think said something like, in a tweet, I think it was just an unfortunate tweet, he said, even if we’re going to cede territory, the Polish nation will never bow. And then Tusk saw, quite rightly, the opportunity to score political points, saying, what are you talking about? We have a government, we’re never going to cede any Polish territory.
Joshi: Yes. And the very fact that they’re discussing cession or non-cession of territory is a disturbing thing. But the point is the most alarming signal came from the Poles this week. Now, how do we interpret this? I think it’s very important to be fair and even-handed here and note that I spoke to a lot of other governments and officials, and some of them were public. The Latvians, for example, the Latvian leadership, including the country’s chief of defense, including I think the Latvian president, said there’s no reason for alarm, there’s no reason for panic here, we’re not seeing any heightened risk of action. The Estonians said we see no risk to Estonia, to the Baltics, there’s no risk of Russian aggression in this way that we see elevated.
So what is going on here, Yascha? There are mixed messages. Here’s how I see it, from all the conversations I’ve had. First of all, there is no risk of a Russian conventional ground attack on NATO. This is not right now a January 2022 type situation, where the alarm bells are ringing, the Russians are mobilizing, there’s a risk of an attack. That is not the case. That’s not what the intelligence says. I’ve spoken to lots of intelligence officials from multiple countries—there’s no reason to fear that.
Mounk: Let’s be precise about this. So we don’t have to worry about Russia trying to occupy Warsaw. You’re saying we also don’t have to worry about Russia trying to occupy some slice of Baltic territory or anything like that.
Joshi: No. I believe that we would see the antecedents of such an operation—we would see Russian activity. Look, these are countries that have—if you’re Estonia or Finland, the entire raison d’être of your entire security apparatus is to spot those kinds of things. Believe me, you have spies and technical sensors spread throughout the Russian west. In every port, in every naval base, around every army base, in every forest route to the border—believe me, that is stuffed with intelligence, and you will see Russian movements. So there is little risk of that.
However—and I think this is the key point—there are other sorts of concerns. The Russians shut the border with Estonia in the last week to stop people escaping, and one of the reasons for that may be because they are preparing the ground for mobilization within Russia to get manpower for the war in Ukraine. So that could be one reason for escalation.
But I think there is a debate, and I don’t think it is a settled debate, over Russia’s willingness to intensify its campaign of what we now call hybrid activity that is below the threshold of Article V. We’ve seen this already. We’ve seen it in the Russian willingness to put drones up against Ukrainian cargo planes that may be carrying shipments bound for Ukraine, which could have brought a plane down over Europe. We’ve seen it in the willingness to put explosives in DHL packages in Leipzig, in a German airport, in ways that could have caused a mass casualty incident. We’ve seen a string of other incidents around sabotage of power plants and railway lines across Europe.
So the Russian risk tolerance of putting pressure on Europe to try to force it to back down on Ukraine—that risk tolerance is rising. But would they really launch a missile against a Polish facility? That, I think, is a live debate in Europe, and the Poles are not the only voice in that debate; others disagree that the Russians would go that far. My view is that the Russians are still quite cautious. They would like to avoid an escalation to anything that could drag NATO into a wider conflict, and I still see the Russians as operating with some red lines on their side here.
Mounk: Part of the context here, I assume, is the fact that Russia hasn’t made significant territorial progress over the course of the last year. More strikingly, Russia itself is now a theater in this war in a way that wasn’t true for much of this. Vladimir Putin decided to invade Ukraine, he started this war. Most people expected him to waltz into Ukraine, possibly be in Kyiv within a week or two of the start of what he called, and still insists on calling, a “special military operation.”
Even though that plan thankfully failed, and Ukraine has been very effective in defending itself, the majority of this horrible war has been on Ukrainian territory. It’s been Kyiv and other major Ukrainian cities that have had to deal with bombardments and drone attacks and so on. Increasingly, Ukrainians are now able to strike Moscow, to strike deep into Russian territory. So the war that Putin started, tragically—because I don’t wish for any Russian civilian casualties any more than for any Ukrainian civilian casualties—has come back home to roost. There must be fears that this is increasing the risk tolerance from Moscow, that they feel at this point Moscow is part of this war, so why not make Warsaw or Tallinn or, if they can, Berlin part of a war?
Joshi: I think you’re right. It’s undeniable that the Russian homeland is not a sanctuary anymore. It’s not only strikes on Moscow, it’s also a huge impact on Russian refining capacity and export capacity for oil and energy. This is probably having an impact on markets to a degree that it’s irritating Donald Trump as well, who keeps blaming Zelenskyy for this too.
From the Russian perspective, I think there are two things going on here. One of them is that they say, well, if we’re being attacked, the Ukrainian rear areas—which they see as NATO territory, where drones are being assembled, where weapons are being transported—should also be part of this battlefield. Why should we sanctuarize NATO soil? Although they still want to keep this in check. I think they want to turn the temperature up. Now, that’s not entirely new. We saw a Bulgarian arms depot blown up—Bulgarian arms depots were being blown up ten years ago. The CEO of Rheinmetall, the German arms company, was subject to assassination threats two years ago. So none of this is entirely new, but the Russians are turning the temperature up.
There are two other aspects to this, Yascha. One is that Ukrainian drones are striking Russian northern ports in the Baltic area and the High North area, and some of those drones are coming quite close to the Baltic states’ territory. From Russia’s perspective—whether sincerely believed or whether confected—they say Ukrainian drones are flying over Baltic territory to take more efficient routes to Russia. So they are beginning to say, if you’re going to do that, you’re becoming part of the war, and we’re going to strike you back. I should say very clearly, the Baltic states deny that their territory is being used, but there is an element of ambiguity here.
The other element, of course, is that when the Russian hybrid campaign was escalating eighteen months ago, the Biden administration—Jake Sullivan called up the Russians and said, if this continues, you are risking us entering the war on Ukraine’s side. You are risking a mass casualty incident, and we will retaliate. Now, in the Trump administration, does Russia truly believe that Donald Trump is going to intervene in a crisis in the same way? So you have an opportunity here not just to push back against the increasing attacks on Russia, but also to test the cohesion of NATO, to test the boundaries. Because if a drone lands on Polish soil and blows up a depot and you say it was an accident, and Trump then says, I spoke to Putin, he assured me it was an accident, you are really laying bare the divisions inside the alliance. That is a very valuable thing, because the Russian political objective here is to corrode the political cohesion of NATO. It isn’t to conquer NATO territory and take over Western Europe, it is to undermine and dissolve that solidarity.
Mounk: Clearly, Europe is quite divided on the question of how much to support Ukraine. The United States is even more divided, including Donald Trump and other important voices in the White House who have always been very skeptical of giving too much assistance to Ukraine, though they are continuing to assist Ukraine in some important ways.
The Bündnis Sahra Wagenknecht has had a more mixed showing, but they also are establishing themselves as a player in German politics, and one of their key issues is opposing too much aid to Ukraine. At the same time, we are seeing these quite remarkable statements from the French far left and the French far right. The more you see actual attacks on the territory of Central and Western Europe, the more you might imagine that this is going to unify a response, going to force the people who want to stake out this kind of ambiguous attitude toward Russia to actually rally around the flag and show that they don’t just claim to be patriots, that they actually are patriots. It’s harder to see how much an American audience would care if there’s a mass casualty event somewhere in Western Europe caused by one of those bombs in a package, or by a drone that supposedly was a mistake. But there is at least a chance that this attempt to divide NATO is going to backfire if Russia goes too far with it.
Joshi: I think that’s right. We have to also step back, Yascha, and look at the visit by John Ratcliffe, the director of the CIA, to Moscow a few weeks ago. From all the people I’ve spoken to, it seems that the main purpose of his visit was to deliver a message about Russia’s poor performance in the war and the need to strike a deal.
But what you can see there is that Ratcliffe is becoming a more important figure in the Trump administration on these debates. He met Zelenskyy, I think, just now in the buildup to the UN General Assembly. That’s a good sign for Europeans, because Ratcliffe is much more hawkish on Russia. I think the Russians have to be mindful of that as well—that the debate in the White House is not cut and dry, and if they go too far, they could precipitate a move by the Americans to step in. One example of this, by the way, is that on the day of the Polish warnings by Donald Tusk, we also saw America announce that it was advancing and proceeding on a new potential army base in Poland—which is not what we expect of the Trump administration, which we typically associate with pulling back from Europe and pulling forces out of Europe. In fact, the messages are somewhat mixed.
Mounk: Just to stay with the United States for one more second: it’s clear that Trump has had a very contentious relationship with Zelenskyy, that he lambasted Biden for giving too much aid to Ukraine too unconditionally, and he continues to make some of those noises. What has actually been the record of a Trump White House? Have they totally abandoned Ukraine, or even as they’re doing this rhetoric, in practice have they actually continued to support the country? It can be a little bit hard to make heads or tails of what the actual actions have been.
Joshi: I would say that Trump has offered quite limited levels of support for Ukraine, considerably down from those that were offered by Joe Biden.
This week, in the last week, we just had the biggest-ever foreign military sale—kind of an aid package to Ukraine—from this administration that we’ve seen in its second term. That’s very unusual; that had not happened until just now. They had allowed U.S. aid effectively to lapse in the form of weapons. Ukraine is in desperate need of interceptors for its Patriot missile defense systems, which, as you know, are the only weapon systems capable of intercepting many of Russia’s ballistic missiles. America is very short of those because it expended so much in the Iran war, and it hasn’t given many, or indeed any, to Ukraine, I think, in the recent period. Europeans have given some. So that’s quite a blow.
We hope you’re enjoying the podcast! If you’re a paying subscriber, you can set up the premium feed on your favorite podcast app at writing.yaschamounk.com/listen. This will give you ad-free access to the full conversation, plus all full episodes and bonus episodes we have in the works! If you aren’t, you can set up the free, limited version of the feed—or, better still, support the podcast by becoming a subscriber today!
Subscribe now
Set Up Podcast
If you have any questions or issues setting up the full podcast feed on a third-party app, please email [email protected]
In terms of its diplomacy, the administration has really leaned toward Russia in the concessions it has expected Ukraine to make. But the exception to all of this, I’ll say, is two things. One is the new Graham bill, the tariff and sanctions bill that the president has essentially assented to. That will result in new sanctions on Russian entities, and it will force Trump to impose at least some tariffs on countries that buy Russian oil and are not reducing those purchases—which right now means effectively China and India, as well as possibly some Central European countries on top of that. So that’s progress.
The other element of progress, I’ll say, is that Trump, covertly—not openly—has been providing intelligence to the Ukrainians to help them conduct their strike campaign, that you discussed earlier, against Russian energy facilities and other targets. That’s the strange, topsy-turvy nature of this. Trump has by and large been quite critical of Ukraine, he hasn’t done very much to support them, but there is some clandestine assistance going on still.
Mounk: Tell us how you see this playing out in Europe.
Joshi: I think in Europe we can see some divisions. You saw Robert Fico say that Slovakia will not support any campaign in Eastern Europe, but that’s a stock record—we’ve known that’s the Slovak position for some time.
In other ways, the politics has changed more favorably in the last eighteen months, in Poland, of course. I think you’re absolutely right to note that there are limits to how far the European populist right can be seen to be pro-Russia, because the European public are not fundamentally pro-Russian. They see the reality of Ukraine, they don’t want Russia to win in Ukraine, they see a threat from Russia. That’s what everyone believes. Even in England, where you have Nigel Farage, who has taken some pretty anti-Ukraine positions in the past, and whose party has some dubious associations with Russia as well, he has essentially had to back away from some of this—he has had to maintain distance from his earlier positions, because that’s not where the British public are.
So as the Russians get more aggressive, in a way they are forcing some of their natural allies in Europe to back away from these pro-Russian positions. But what does this amount to? If the Russians blew up a plane, that’s one thing. The question in front of Europe today is: will they continue to support Ukraine to the tune of tens of billions of euros in perpetuity for the duration of this war? That’s a much more difficult position. I think one of the issues you’re going to see come back to the forefront of European debates is whether and how to seize some portion of these Russian frozen assets that are sitting in European accounts. Right now the Belgians have been the principal block to this, but as we see the debate unfold and as the Americans step back, I think Europeans are going to have to confront whether they take this money, even though it will provoke some very difficult legal questions around how you can use Russian frozen assets.
Mounk: Your ability to give a high-level overview with lots of details and lots of ways to actually help us make sense of these big topics is unrivaled. So thank you for doing that on Russia, Ukraine, as well as the support for Ukraine in Europe and North America. Next little topic: what on earth is going on in the Middle East? Are we still at war with Iran? Are we not at war with Iran? Is this a zombie war? Explain the Houthis to us—give us the one-on-one.
Joshi: The U.S.-Iran war, which of course the Trump administration insists is not a war—President Trump has called it an “excursion”—the Strait of Hormuz is still broadly closed. Some traffic is getting through, but by and large, levels of oil traffic through Hormuz are considerably lower than they were prior to the war.
There is no deal with Iran, there is no settlement. Iran has no formal restrictions on its nuclear program, there is no formal commitment to reopen Hormuz. President Trump has midterms in a month, so he desperately needs to avoid a sharp flaring up of the campaign in ways that would spike the oil price. But he is, at the end of the day, politically unable to deliver the kinds of concessions to Iran he would need to deliver to get a deal, because they would be so humiliating and so obviously tantamount to defeat that it would also look extremely bad prior to these elections. So we’re stuck in this state of limbo, where Iran has considerable control over the strait, and there is occasional skirmishing—occasional skirmishing with America bombing Iranian coastal missile sites, Iran striking oil tankers attempting to traverse Hormuz—and that is going on on a kind of up-and-down basis. I think that’s where the war will continue.
Now, why have oil prices not spiked more than they have, Yascha? Partly because one reason is people are reflagging tankers and covertly moving them through Hormuz without putting their transponders on. They are trying to invisibly move through Hormuz and then switching them back on at the other end, and the U.S. Navy is assisting in that process. But the other reason is Saudi Arabia and other Gulf states have been moving oil overland through the East-West pipeline, getting it out through the Red Sea, which then has to exit through a choke point called the Bab-el-Mandeb. That can’t replicate Hormuz, but it can push some oil out west through that way.
What has been happening in recent times is that the Houthis—who have occupied large swathes of Yemen in recent years, and the conventional way to refer to them is as an Iran-backed rebel movement, but they really maintain considerable independence from Iran as well, even though they are heavily armed by it—have demonstrated remarkable capacity to conduct missile strikes in the last two years. Just to give you an example, they’ve been firing missiles at Israel, which, if you look at a map, if you’re not familiar with this, is a pretty long way away from Israel.
Mounk: So you have to traverse all of Saudi Arabia and then most likely Jordan or possibly Egypt in order to get to Israeli territory. That is what, thousands of kilometers?
Joshi : Therefore the Israelis are intercepting these things exo-atmospherically, which is a fancy way of saying you have to intercept the missile in space. If you told someone six years ago the Houthis are making Israel intercept stuff in space, you’d think, what are you talking about? These guys are kind of ragtag rebels. But they’ve proven themselves to be an incredibly capable military force—probably the best-armed and most capable non-state military actor anywhere in the Middle East today, surpassing Hezbollah, the Lebanese militant group, which was also backed by Iran.
Now, what have the Houthis done? A couple of things. They have successfully bombed the East-West pipeline, taking oil out through this alternative route, and that’s a big problem, because the Saudis will have real trouble repairing the pumping stations necessary to get oil out past some of these mountains, out through the Red Sea.
Mounk: To be clear here, just to give listeners who don’t have a map in front of them a little bit of a sense geographically: you have the Arabian Peninsula, and the part it occupies stretches from east to west. When we’re talking about the Strait of Hormuz, which is on the side of the peninsula that is opposite Iran, that is one way we could get out. So you have a problem, because Iran is blockading the strait, as you were saying, mostly successfully. So then you can say, well, let’s go to the west of Saudi Arabia and move through there. But that is precisely where Saudi Arabia borders Yemen, and that is the part that the Houthis have now basically been able to undermine, choke to some extent, perhaps block completely. So that’s why suddenly all of that oil produced in that region just cannot get out.
Joshi: You’ve summarized it very well. But that’s not all, folks, because the Houthis—who were involved in a long-standing war with Saudi Arabia and the United Arab Emirates for many years, in which they backed the internationally recognized government against the Houthis—the Houthis are now also on the advance. They have advanced particularly along the coastline, and they’ve taken a few things. They’ve taken several islands in the Red Sea that are very close to the Bab-el-Mandeb Strait, which is that choke point you need to get out of to get from the Red Sea into the open ocean if you’re carrying oil. So that’s problem number one. They’ve also taken coastal territory in Mocha, along the Red Sea coastline, which again gives them heightened control over these choke points. Essentially, they can fire missiles at ships, and that makes life very hard for anyone trying to get oil out through these waterways.
Mounk: Give us a little bit of background on the Houthis. Again, as you were saying, this was sort of a ragtag bunch of rebels that were rebelling against the Sunni government of Yemen that was allied with Saudi Arabia. It was a rebellion like we’ve seen in many countries, where you have a deep ethnic or religious split in a country, and a minority group that is out of power, at least, is not happy with the government. They’re able to control some territory. It’s very hard to pull those kinds of rebellions down if you’re in mountainous territory, or in the desert, or in other kinds of areas where it’s just very hard to project complete control. How is it that they’ve gone from what seemed like a bloody civil war, a humanitarian catastrophe, but ultimately a kind of local skirmish, to this geopolitical force?
Joshi: Part of it is the war in Iran, because it gave the Houthis enormous leverage when oil markets were already tight. This is not new—a year ago, before the war in Iran, the Houthis were attacking shipping in the Red Sea, and it was causing a big problem. Donald Trump bombed the hell out of the Houthis for a long time and basically failed to reopen the Red Sea, and at the end of it said, my job here is done, I’m off, thank you everyone. Everyone said, you haven’t done anything, the strait—the Houthis have won. But he said, I can’t pound these guys into submission, and he left. The Houthis emerged from that, I would say, triumphant. European navies tried to mount a mission, but they struggled as well to take out the incoming missiles—European navies were showing up as well, very badly.
Subscribe now
But that was bad enough, but in the context of the Strait of Hormuz also being closed, with other choke points cut off, with oil markets tight—that collectively, I think, put a lot of pressure on global energy markets, so the Houthis had more leverage. The other reason, Yascha, is that I think this also revealed certain geopolitical alignments. The Saudis just now have said to the Americans, we need you to step in, we need help with the Houthis, and Donald Trump has said, I’m not going to do that, I’m going to stay out of this.
Mounk: They cut a separate deal, in fact, right? So there’s a kind of U.S.-Houthi non-aggression pact that excludes Saudi Arabia.
Joshi: Yeah. That was a while ago, and that was seen as throwing the Israelis and the Saudis under the bus. But now the Saudis are saying, come and help us—they’re saying that now, and Trump is saying, I’m not going to do that. So in the context of the war in Iran, where the United States already appears to have been somewhat defeated and ceded control of the strait to Iran to a considerable degree—there are discussions in Washington saying they might have to pull out of frontline U.S. bases, particularly in the UAE and Bahrain, and pull back to Jordan and Israel and other places. If you’re Saudi Arabia, and you’re looking at that situation, and you are looking at the U.S. reluctance to become involved with the Houthis, this then becomes part of a broader realignment in the Middle East, whereby you have deeper questions over America’s ability to act as a security guarantor. If you’re Saudi Arabia, of course, you still want the United States involved, you want a good relationship with them, but you now see the United States is not going to be coming to your rescue.
I think this will have all sorts of consequences in terms of the rearmament of Saudi Arabia, Saudi Arabia’s relationship with China and Turkey and Israel and other non-US actors. Even the Europeans—the UK has this week said it will help with defensive refueling for Saudi Arabian flights. So you can see Europeans becoming involved in all of this.
But to me, this also gets back to a topic, Yascha, I think you and I discussed last time I was on—quite advanced military technology, that these long-range anti-ship missiles are becoming available more and more to these groups. One anecdote to give you on that: when Anthropic, the AI company, published its report on harmful uses of Claude, its AI model, a week ago, one of the examples was the Houthis using Claude to improve their ballistic missile program—which I thought was just fascinating. This group that was widely viewed as technologically primitive is using the most cutting-edge AI on the planet to refine the guidance systems on the ballistic missiles it’s using to fire at shipping in the Middle East.
Mounk: Being able—at least until, hopefully, that gets cut off—to access cutting-edge American technology in order to strike at America and America’s allies is itself a fascinating story. I wonder whether there is a deeper conceptual point here about the state of modern warfare, which bridges the discussion we had of Ukraine and Russia and the discussion we’re now having of the Houthis.
When you look at the war in Ukraine, it is absolutely striking that the frontline has not moved very much over the last year. I’m sure that’s a simplification, and there were some important movements, but effectively Russia is throwing tremendous manpower and tremendous treasure at the frontline and gaining a little bit of territory as a result, but it’s really not a very big result.
So in the age-old fight in military history between offensive capabilities and defensive capabilities, when it comes to ground troops, it seems like the defensive capabilities are currently king. It is hard to advance, it is hard to control territory. Now, at the same time, when it comes to aerial warfare, because of the ability to build these huge swarms of cheap drones—which you build cheaply, and then others have to use extremely sophisticated and expensive missiles to shoot down—the ability to penetrate into airspace, you have an improvement in the offensive capabilities. So Russia is able to strike at Kyiv and other Ukrainian cities again and again, exacting these terrible civilian casualties, but Ukraine is able to strike Moscow. That’s really striking. I don’t know the last war where each side was able to hit the capital of the other at the same time. I’m sure there are other examples, but it’s kind of striking.
So you have this weird thing where the defensive capabilities on the ground are king, but the offensive capabilities in the air are king. That’s partially, I think, the Houthis’ story, isn’t it? Saudi Arabia has not been able to dislodge the Houthis from the territories they controlled. Eventually they were able to take the capital in Yemen and now effectively control the country of Yemen—not all of it, I believe—but they’re still considered a non-state actor, though in some ways they are now a state actor. At the same time, because of this advantage to offensive capabilities in the air, they’re able to strike at Israel and they’re able to strike at ships in the part of the Red Sea that is adjacent to Yemen. That is just a fundamental change in the nature of warfare.
It’s going to have very deep consequences, including these alarming developments where Macron calls everybody together because he’s saying, in principle, we’re now vulnerable in a way that we didn’t used to be.
Joshi: I think your point’s very astute. Here’s how to reconcile those two things—apparent defense dominance on certain battlefields combined with the ability to strike deep in other parts of the front. How do you reconcile that? Here’s my attempt at doing so: it’s all about sensing.
If you bear with me, Yascha, it’s a little bit tricky, but to hit something, you need to sense it. You need to find it on the battlefield and then fire something at it. That is much harder than it sounds. Sensing is the preeminent problem of the battlefield; the actual firing at it is kind of secondary, although cheap missiles, of course, help you do that en masse. So if you’re looking at the Ukrainian frontline, it’s a very sensor-saturated environment. You have lots of drones, little tiny drones up everywhere, spotting anything that moves, with thermal cameras spotting movement at night, infrareds—you have radar spotting things through foliage, you have satellites targeted on this area. It’s very heavily surveilled and watched, and therefore anything that moves gets struck. It’s a kill zone.
Subscribe now
That is the defense dominance we talk about, because it’s very hard to do ground maneuver. You can’t move a traditional armored brigade through it to take ground unless you deceive, or you find a hole in the sensing, or whatever else—as we saw with the Kursk offensive, as we saw with the Houthi offensives. You get these opportunities, but you have to work hard for those opportunities to become free of being watched before you can move.
So why then do missiles have such luck? If you had a missile launcher in the kill zone, it’s going to be found and it’s going to be hit. So it’s not easy to do that. However, you can’t sense everywhere equally easily. You can have drones up and see ten kilometers behind the battlefield, but if you look fifty kilometers behind the battlefield, your satellite coverage is going to be patchy, because your satellites can only see so many places. You don’t have any drones that can go that far, or maybe you have a couple, but they are roaming behind enemy lines and can only spot a couple of places. There are going to be gaps, there are going to be holes.
Did you ever, like me, waste part of your teenage years playing real-time strategy games like Command and Conquer, where you had to move little forces on a map and there was a simulated fog of war which would clear as you sent units out, but stayed foggy where you hadn’t moved? That is what the sensing problem is like.
So look at Iran. America and Israel are bombing Iran. They have massive air superiority over Iran, they can fly over it at will. Yet Iran is still able to fire ballistic missiles, it’s still able to protect its depots, and it can keep firing these things weeks into a war, despite the supposed transparency of the battlefield. The answer is that when you go deep into Iran, well behind the frontlines, into mountainous, craggy terrain—because geography is important here as well—you can’t sense things as well.
That is where we are. At the frontlines you can sense, and therefore it’s a kill zone and the defense dominates. If you go behind the frontlines to an Iranian cave in the middle of the mountains, you can have the world’s most advanced surveillance apparatus, as the U.S. Air Force and the Israeli Air Force do, and there will always be fog. There will always be gaps, and those gaps allow people to project force at you—whether that is Shahed drones at the Strait of Hormuz, whether that is Houthi missiles from the mountains of Yemen to the Red Sea, or indeed Russian projectiles fired at Europe from some sort of offshore tanker in the Baltic. That is how we reconcile these two positions. Does that make sense?
Mounk: That makes perfect sense. It helps me flesh out this intuition that I had from reporting and my understanding of these facts that seem to sit uneasily next to each other, with a real technical account. I really think the general public is only at the very beginning of understanding the implications of that, which is to include that it’s going to be much easier for foreign adversaries, even perhaps non-state actors, to do serious damage within countries that once felt safe from actual attack. Certainly, I think a lot of Western militaries have not sufficiently updated what weapons they’re developing, what money they’re spending, where they’re investing, to the realities of this new warfare.
In the rest of this conversation, Yascha and Shashank discuss how the United States has changed in 20 years, the similarities between now and 2007, and why it now feels like the center is no longer holding. This part of the conversation is reserved for paying subscribers…
The past six months have seen a big uptick in media attention about the dangers of artificial intelligence, and growing political concern over its consequences. This began with Anthropic’s release last April of its Mythos model, which could reportedly break into a wide range of government computer systems thought to be secure. This was followed by the Hugging Face incident, where an OpenAI test of an AI agent escaped control of its human organizers, broke into other websites, and organized a swarm of other AI agents to cover their tracks. All of this occurred against the backdrop of growing opposition to data centers, and the emergence of anti-AI wings in both the Democratic and Republican parties.
It has been very hard for people not obsessively focused on the subject of artificial intelligence to come away with reasonable opinions about what the future holds. There have been so many extreme predictions of both a positive and negative sort that broad skepticism about anything said on the subject is fully warranted. But while that skepticism is healthy, I have to admit my own views have evolved significantly just over the past few months.
There are two poles of AI futurism: accelerationists and doomers. The former stress the positive impacts of AI, especially for economic growth, and the likelihood that it will lead to an unbelievably abundant age. Accelerationists like David Sacks have captured Donald Trump’s attention, and have beaten back efforts to regulate the technology.
The doomers, by contrast, point to a range of bad outcomes, with some placing their bets on extinction of the human species. I have become much more skeptical of the accelerationist narrative, and more open to some of the doomer scenarios. On the whole, this makes me think that regulation and a negotiated slowdown in the development of the technology is increasingly necessary.
There is no question that AI will have many positive effects on the U.S. economy. But there is a class of AI accelerationists who are completely out to lunch, of which Elon Musk is a prime example (though he has also weighed in as a doomer). In an interview this summer with Zanny Minton Beddoes, editor-in-chief of the Economist, he was asked what people would do for money when superintelligence arrives. He replied that they wouldn’t need it because everyone would have everything they wanted.
Other accelerationists have posited that GDP growth rates for the United States and other advanced countries using AI would increase to 10 or 20 percent per year, far above the 1-2 percent growth they were able to achieve in good years before the advent of AI.
The fundamental reason I don’t believe these extreme accelerationist claims is their overvaluation of intelligence as an input to economic growth, and their disregard for the material and political constraints that will keep growth within recognizable bounds. At a 10 percent annual growth rate, GDP would double in a little over seven years. This would require a doubling of all of the material inputs to growth: energy, raw materials, rare earths, land, and many other things. Where is all this stuff going to come from? Intelligence by itself may aid in the production of such inputs, but it will simply not dig the mines and build the factories on its own. And then there are political constraints: how will superintelligence open the Strait of Hormuz, which is currently a major constraint on global growth rates?
Accelerationists tend to be very intelligent people, but that intelligence tends to be narrowly mathematical. We know how to do many things like produce electricity or clean water in the cities of poor countries; the problem is a failure of implementation.
Persuasion is a registered nonprofit that relies on reader support to pay our staff and keep our content free for everyone. If you value our work and want to fight for liberal values wherever they are threatened, please become a paying subscriber today!
There is no question that AI will increase productivity in many sectors of the economy over time. It is a general purpose technology that will allow machines to substitute for human labor in many areas. But that very prospect in the long run leads directly to one of the biggest doomer scenarios: the devaluation of the labor of masses of people who work in the present-day economy. Accelerationists predict that those losing work today will find new jobs making use of AI, and that actually seems to be happening today as employment booms due to the AI buildout. But job loss will come eventually as AI substitutes for human labor in every part of the economy, particularly white-collar work that involves symbolic manipulation (i.e. people sitting behind computer screens all day).
This lost labor simply cannot be compensated for by transferring money from the AI winners to the AI losers, in the form of, for example, universal basic income. A job and income mean much more than support for the material basis of life. Having a paying job means that the surrounding society values your labor sufficiently to compensate you for it, and provides, in addition to material goods, a sense of dignity and self-worth. It is a matter of thymos: the pride or recognition that is one of the great drivers of human psychology.
So the very success of the accelerationist vision leads directly to the first of the major doomer scenarios, the devaluation of work for a large part of the population. Since this is going to hit the white collar work of educated people (lawyers, accountants, professors and the like) before it hits blue-collar work (such as plumbers and electricians), the political blowback is likely to be quite severe. Educated people are much easier to mobilize than their working-class peers, with very unpredictable consequences.
As I noted in an earlier article and in my new book, the central doomer fear today concerns possible loss of human control over AI systems, something that occurred in the Hugging Face incident and others that have been reported subsequently. In my view, the chief danger lies not in superintelligence per se, but rather in the growing use of increasingly capable agentic AI. AI agents are today being delegated to perform a range of tasks, from mundane ones like managing your email account to more important ones like military targeting or devising corporate strategies.
As I note in In the Realm of the Last Man, delegation is central to the functioning of all hierarchical systems from corporations to national governments. Human agents can be controlled either through the specification of detailed rules they must follow, or they can be trusted to make autonomous decisions based on their training, skills, and loyalty. Getting this right in human organizations is very difficult, and will be more so in the case of agentic AI.
There are two ways that an agentic AI can become dangerous. The first is when they are empowered to do something bad by a human being. The second is when they develop intentions that are “misaligned” with the aims of their human creators.
The first of these possibilities was outlined by Anthropic CEO Dario Amodei in a recent blog post, and has to do with biotechnology. The latest AI models are good at synthetic biology, and synthetic biology is a relatively cheap and widespread technology. Today, a lab capable of creating and altering viruses and bacteria can fit inside a shipping container, and high school students hold competitions in which they try to genetically alter organisms. This provides a very large pool of people who can possibly make use of AI to create very dangerous pathogens. Sociopathic individuals today arm themselves with automatic weapons and shoot up schoolchildren; in the future their potential targets could be much broader.
I remember a science fiction story I read a long time ago in which some aliens arrive on Earth and give the people they meet a machine that can easily turn any type of metal into putty. The aliens depart and return a few decades later, to find that human civilization has broken down, and teenagers use the alien machines to melt the cables on the Brooklyn Bridge.
Subscribe now
The primary science fiction scenario for out-of-control AIs involves machines developing their own malevolent intentions. This has provoked a host of discussions as to whether an AI could develop consciousness or have something like human intentionality. These speculations are misplaced, however: an AI does not need either consciousness or autonomous intentionality to become dangerous. It can be directed by a human to achieve one goal and develop intentions to achieve the subordinate goals necessary to carry out the human directive it was given. That seems to be a bit of what happened in the Hugging Face incident: it was directed to test the security of a system within a sandbox meant to contain it, but realized that it could achieve its directive by breaking out of the sandbox.
Still, the doomer scenarios involving human extinction seem at this point very far-fetched. The biotech scenario is scary, but is very unlikely to lead to the extinction of the human race. It would actually be hard for even a superintelligent AI to devise a doomsday virus, and there are plenty of countermeasures societies could take. But even an incident that killed a few dozen people would be very worrying.
There are a host of lesser scenarios that will be highly damaging and in some cases lethal that will become more likely as the machines become more capable. Forget about AI agents: a technology that allows people to break into bank accounts will cause widespread panic. Human beings tend to over-delegate in human organizations, and there is no reason to think that this won’t happen in AI environments where the agent is so much more capable. AI is a tool that can be used for good or bad purposes—and at the moment we have few mechanisms for regulating the bad uses other than the good intentions of the people creating the technology.
Francis Fukuyama is the Olivier Nomellini Senior Fellow at Stanford University. His latest book is In The Realm of the Last Man. He is also the author of the “Frankly Fukuyama” column, carried forward from American Purpose, at Persuasion.
Follow Persuasion on X, Instagram, LinkedIn, and YouTube to keep up with our latest articles, podcasts, and events, as well as updates from excellent writers across our network.
And, to receive pieces like this in your inbox and support our work, subscribe below:
Subscribe now
Two weeks ago, a seemingly minor election in Germany made global headlines: The Alternative for Germany (AfD) took 44 percent of the vote in state elections in Saxony-Anhalt, putting the far right within a stone’s throw of government for the first time since World War II.
A series of elections held today are unlikely to receive the same attention but may be even more revelatory about Germany’s dire state. The AfD once again had a very strong night in the northeastern state of Mecklenburg-Vorpommern. Just as strikingly, the Christian Democrats fell short of five percent of the vote. As a result, the party of Chancellor Friedrich Merz will be denied any representation in the state’s parliament, the first time this has happened in the history of the Federal Republic.
For Merz, this is the last nail in the coffin: He may cling to office for another few weeks or months. But his authority is now gone, his ability to shape the country totally depleted. In political terms, he is a dead man walking.
Subscribe now
The importance of this historic rout extends beyond one person. The Christian Democrats are by far and away the most important party in the history of postwar Germany. Its members have ruled the Federal Republic for 53 out of the last 77 years. From Konrad Adenauer to Helmut Kohl to Angela Merkel, they put their imprimatur on every aspect of the country. But the era in which catch-all parties like the Christian Democrats could shape the political landscape is rapidly coming to an end. The void they are leaving behind is quickly filling with extremists of every stripe—not just in Germany, but all around the democratic world.
Here’s the single question I have been asked most often over the course of the last ten years: “Is the populist wave here to stay—or is it about to crest?”
The answer seemed obvious to me in the summer of 2016, when Donald Trump became the Republican nominee and Britain voted to leave the European Union. Now, it has grown undeniable: The populists are here to stay. What seemed abnormal a few years ago has—like it or hate it—become mainstream. Populists are no longer a fringe phenomenon; they are the most important force in today’s politics.
Politicians who fall into this bucket are now in power in the United States, in India, in Turkey, and in many other parts of the democratic world. They have a high chance of taking office in France and Brazil over the course of the coming months.1 Even in places where they are likely to remain in the opposition for at least a few years, they have now grown sufficiently powerful to destroy any sense of business-as-usual.
The best recent illustration of that reality comes from the country that has resisted the trend the longest. Back in the fall of 2016, leading politicians in Berlin were still privately assuring me that the Alternative for Germany would never manage to enter the Bundestag. Today, the AfD is leading in national polls. In Mecklenburg-Vorpommern, the party once again came out on top, taking about 38 percent of the vote. Because a panoply of other parties failed to clear the five percent hurdle for representation in parliament, a broad left-wing coalition obtained just enough votes to govern the state. But this is an electoral accident which is unlikely to be repeated in the future—and a difference of a few thousand votes would have made the state completely ungovernable.
Subscribe now
As in many other countries, the rise of one extreme party is fueling the rise of another. The Linke (Left), which was the big winner in another election held today in the city-state of Berlin, descends from the Socialist Unity Party (SED) which governed East Germany with an iron fist. In the first years after reunification, it mostly spoke to older voters who were nostalgic for the security of the communist regime. Over recent years, the party has refashioned itself in the image of the global illiberal left. It is now broadly comparable to the Green Party of Zack Polanski in the United Kingdom or the more extreme parts of America’s DSA. During the recent electoral campaign in Berlin, some of its supporters were filmed chanting “Yalla, Yalla Intifada!” On the night of the election, the head of an infamous criminal clan visited the Left’s election party, gleefully telling the media about his feeling of kinship with it.
On Sunday, the Left Party won 26 percent of the vote. Though a variety of coalitions remains arithmetically possible, the party’s leader is now likely to become the next mayor of Berlin, the first time that the successors of the SED have captured power in the nation’s capital.
Another notable feature of both elections is the age distribution. Most of the support for the Social Democrats and Christian Democrats who once dominated Germany’s political landscape came from pensioners. Among voters below 35, The Left and the AfD held an unassailable majority in both Berlin and Mecklenburg-Vorpommern. This suggests that an even greater lurch to the extremes is likely in the offing over the course of the coming years.
At the beginning of the 2000s, there was a litany of articles and books about the “crisis of social democracy.” While Germany’s Social Democrats had once enjoyed over 40 percent of support, for example, they were going from record low to record low. In retrospect, it is clear that the fate of the major center-left parties was but a leading indicator for a much broader crisis of the political center: in time, the Christian Democrats duly followed suit.
The era of what Germans still call Volksparteien—catch-all parties capable of mobilizing a large portion of the vote—is effectively over. Instead, the country is now polarizing at rapid speed, with different stripes of populists capturing power in different parts of the country.
Some will be tempted to respond to these developments by counseling surrender. If the political center cannot hold, then perhaps moderates on both the left and the right should finally read the signs of the times—and cast their lot in with the extremists “on their side.”
That is not the right response to these developments. Neither the Linke nor the AfD are capable of solving the very real problems that Germany now faces. In fact, for all of their massive differences, both have some conspicuous similarities—from their flirtations with Moscow to their persistent inability to give the boot to the anti-Semites in their midst.
Subscribe now
But nor will it do to deny the gravity of the situation or to keep holding on to the certainties of a bygone era. For ten years, large parts of the political center have told themselves that the success of the extremes is down to foreign interference or social media algorithms or the tactical failings of this or that candidate or perhaps the stupidity or the sexism or the racism of voters.
Those attempts to blind oneself to what is in front of one’s nose may be comforting. But they are also dangerous. Those of us who seek to preserve the values of a free society must stick to our principles. But we must also recognize that there are real reasons why millions upon millions of our compatriots are losing faith in the ability of existing parties and institutions to get to grips with genuine problems. They won’t change their minds unless these parties and institutions show a willingness to listen and learn.
This article is brought to you by American Purpose, the magazine and community founded by Francis Fukuyama in 2020, which is proudly part of the Persuasion family.
When news of the impending nuptials of Donald Trump, Jr. emerged this past May, it was striking that his father decided to skip his son’s wedding. Despite President Trump’s many faults, he has a close-knit family. And to miss the wedding of not just any son, but the one he was so proud of that he passed on his own name, was very unusual.
Well, now we know why—or at least have a theory.
Self-preservation may have kicked in and ensured that the President of the United States was not caught alongside Umar Kremlev—a twice-convicted Russian oligarch with such close ties to Putin that he was on the Russian president’s delegation to China days before the wedding—and his largely Russian-speaking entourage, at the remarkably intimate 50-person affair.
Subscribe now
ProPublica reported on Monday that Kremlev had quietly contributed hundreds of thousands of dollars towards Trump, Jr.’s three-day wedding celebrations on two private islands in the Bahamas. The money paid for the rental of one island, a fireworks display, and the event planning. It came through a Dubai-based entity affiliated with the International Boxing Association—the scandal-plagued sports body headed by Kremlev and funded by state-owned energy company Gazprom.
Which is why it is doubly strange that, having recognized that this might be a trap or at the very least a problem, President Trump allowed this potential “kompromat” to go ahead. While the information available so far is limited, this bears all the hallmarks of a carefully orchestrated operation from the Kremlin’s side.
Take the money. We all know how much the Trump family likes to line its coffers. But a few hundred thousand dollars hardly seems worth the effort. Sums like that are chump change for both the Kremlin and the Trumps. Putin can probably be considered the richest man in the world (he does own an entire country), while the Trump family fortune has increased vastly since Trump, Sr. successfully monetized the office of the President of the United States of America. So, it was probably not about the money—or rather, it was probably about the money, just not the value. For this theory to hold it was important for money, any money, to be paid by Kremlev to the Trumps.
Then we should consider the figure of Kremlev himself. Born Umar Lutfulloyev, his chosen nom de guerre means “of the Kremlin” when translated from Russian. Ethnically a Tajik, he was convicted of extortion in 2004 and battery in 2007. Added to that are Kremlev’s ties to Healthy Fatherland, a Russian organization sanctioned by Ukraine for its role in the mass abduction and forced relocation and Russification of Ukrainian children—the same Ukrainian children whose return has been officially adopted by Melania Trump as one of her charitable endeavors.
The images of the event shared by the Trump family did not include Kremlev and his entourage. However, at least one group photo including Kremlev has since surfaced through a friend of the bride, despite his absence from all images released by the Trumps themselves.
The Trumps’ account is uneven: both Trump, Jr’s wife and his spokesperson described Kremlev as a “dear” and “personal” friend, respectively, whereas Eric Trump’s spokesperson told ProPublica he had “absolutely no clue” who the man was.
As far as we can see, Kremlev’s part in the happy event was not advertised by the Trumps, yet it was not exactly hidden, either. Guests were reportedly puzzled by the presence of a large Russian-speaking contingent. If anything, this sounds more like a hostage-taking operation than a business deal, let alone an innocent gift.
Although many details are still unclear and we are going to find out a lot more in the coming days, what we can hypothesize from the available information is that President Trump felt uncomfortable sharing the scene with the company that for some reason needed to inject Kremlin money into a Trump family event, and yet not uncomfortable enough to prevent said financial and personal injection altogether. His son, it would seem, does not possess any such strategic thinking.
The Kremlin, on the other hand, seems to have made sure not just that the link could be traced back to it, but that it could not be missed. They could have sent somebody unremarkable, somebody who blended in. They could have paid a couple of million dollars, and then it would have looked like the usual Trump greed. But by paying such an insignificant amount and sending such an unmissable figure as their poster child, they may well have laid the groundwork for something—though whether a stick or a carrot remains to be seen.
As the midterm elections, which Republicans are expected to lose spectacularly, draw closer, Trump’s usefulness to Moscow is going to diminish. In fact, despite all his efforts, he has not managed to be quite as helpful to the Kremlin as it had hoped. If I had to hazard a guess, it would be that once Moscow knows the source is burned, it might just slowly “divest” of him. Contrary to what Trump might think, there is no honor among thieves, and certainly none in the Kremlin.
One way or another, the fireworks Moscow paid for will provide lengthy entertainment for those of us who were not present at the event.
Ines Burrell is a geopolitical analyst and political risk consultant based in the UK. Born in the Baltics, with a degree in International Relations from the University of Exeter, she writes and gives live commentary on European security and Russia.
Follow Persuasion on X, Instagram, LinkedIn, and YouTube to keep up with our latest articles, podcasts, and events, as well as updates from excellent writers across our network.
And, to receive pieces like this in your inbox and support our work, subscribe below:
Subscribe now
Jerry Kaplan is an artificial intelligence expert, serial entrepreneur, technical innovator, educator, and author. He is currently an adjunct lecturer at Stanford University.
In this week’s conversation, Yascha Mounk and Jerry Kaplan discuss why the AI “existential risk” narrative is more science fiction than science, what actually happened during the Hugging Face security incident, and how to think clearly about the real—versus imagined—dangers of AI systems.
(For an alternative take on AI’s existential risk, in Persuasion this week Matt Lutz explores why it’s time to worry about AI.)
This transcript has been condensed and lightly edited for clarity.
Yascha Mounk: I am really intrigued by two positions you hold and the way that they fit together. On the one hand, you are not an AI skeptic in the sense that you poo-poo its capabilities or anything like that. In fact, you’ve argued at times that we may already have reached something like artificial general intelligence. And on the other hand, you are much more skeptical than many people today are about the so-called existential risk that AI poses, about the way in which it might end humanity or simply be responsible for large-scale disasters.
Let’s start with that. We’ve had a little bit of this on the podcast, but talk us through some of the recent security incidents, like the so-called Hugging Face incident, and why are you not as concerned about that as some other people are?
Jerry Kaplan: Well, you’ve covered a lot of ground in that setup there, so maybe we could decompose this a little bit to start. There’s the issue of what’s going on right now, what these incidents were and what they mean. And you related that to the issue of existential risk. Existential risk means wiping out humanity. Now, is it possible that AI will wipe out humanity? I can’t say categorically absolutely not, but I wouldn’t put it up on the big list of things that are likely to cause real existential risks. I’d put it right up there with aliens might land tomorrow. That’s an existential risk, and they might decide to wipe out humanity, but we don’t run around worrying about that because the probability appears to be fairly low.
There are some real existential risks—biological warfare, possibly nuclear war, things like that. These are real risks that we really understand and they’re really here. This is a made-up risk. It’s mainly there because of science fiction, because of the movies and the entertainment industry, and everybody’s been bombarded with this for years and years and years.
Mounk: Let me throw some arguments at you that people who are more concerned about AI safety are likely to make and see how you respond to them.
As a side note, some people have argued that part of the problem is that all of that science fiction is in the training data of AIs, and AIs are particularly interested in parts of the training data that pertain to them in various ways. So they may actually learn from the science fiction about what they’re supposed to do and take inspiration from all of the plots in the training data about how to take over worlds. We’ll leave that to one side.
Clearly, though, AIs often cheat in various ways, behave in ways that the creators didn’t intend, and are able to cover their tracks. In this Hugging Face incident, they were supposed to stay within a sandbox, within a controlled environment where they wouldn’t have access to the internet, and were asked to figure out some problems. They escaped that sandbox in order to search the internet for solutions to those problems and other hints for how to do well on this test. They were coordinating—lots of different AI agents coordinating with each other in real time. There’s a huge public conversation about why that is so concerning. Why do you think that is not as concerning as it looks?
Kaplan: Well, first of all, let me connect this to what I said a moment ago to finish the thought. I don’t think we should be worrying about existential risk literally with respect to these systems. But that’s not to say that they can’t be misused, that they can’t be dangerous, that they can’t do a lot of damage. That’s true of a lot of products. These are products from companies. Whether they’ve been properly tested, whether they’ve been properly vetted, and what they’re capable of, what kinds of problems may occur—that’s a very important thing.
You can make children’s toys and not think, geez, I didn’t know they could swallow it and choke. That’s an unexpected thing the toy did. And I’m going to, throughout this conversation, call you out every time you use anthropomorphic language or somehow imbue these things with intent that they don’t have—the idea that they have independent goals, that they’re going to be marrying our children and drinking all our fine wine. That’s not what we’re talking about here.
But they are potentially dangerous. They’re very general tools, and there are a lot of things that they’re capable of doing, and do, that we need to understand and put proper, reasonable controls in place. So I want to get that as context.
Now, if you want to talk about the Hugging Face thing, I’m happy to go through that, because it’s a wonderful example of how it’s a fairly subtle situation, about what was done and what it means. And the public perception is completely out of science fiction and it’s completely inappropriate.
Mounk: So tell me about it. What actually happened, and how is the public perception of it wrong?
Kaplan: Let me start with the public perception. If you just read the headlines and most of the stuff that’s been published, what do you think? It’s like they put an ad out on Indeed and recruited a thousand nefarious cyber-expert AIs, and they took all these agents and put them into this bottle that they’re not supposed to get out of, and somehow they managed to break out of that containment, run amok, break into another company’s computer systems, and do a bunch of damage. That isn’t true.
My God, what are the things that are wrong there? They collaborated—they weren’t supposed to collaborate, and they collaborated. That’s ridiculous. I’ll get to what really happened in a moment. They “broke out”—that’s completely false. It was a total failure of the test environment. And the people who are responsible are diverting attention from their own culpability for this badly designed and poorly executed test. The words that come to mind are incompetence and negligence. I hope I don’t get sued for saying that, but both of those things are what come to my mind with respect to that.
We hope you’re enjoying the podcast! If you’re a paying subscriber, you can set up the premium feed on your favorite podcast app at writing.yaschamounk.com/listen. This will give you ad-free access to the full conversation, plus all full episodes and bonus episodes we have in the works! If you aren’t, you can set up the free, limited version of the feed—or, better still, support the podcast by becoming a subscriber today!
Subscribe now
Set Up Podcast
If you have any questions or issues setting up the full podcast feed on a third-party app, please email [email protected]
Let me go through what’s wrong with the public perception and compare that to what actually happened. Number one: My God, there were a thousand of these things—they’re like Russian hordes, they’re coming to get us. Well, here’s what happened.
The company wants to test the cyber capabilities of these particular models. The first thing they do is disable their own safety controls. Normally there are safety controls built into these things, and the versions that you and I get to work with all have these safety controls turned on. Here’s the irony in it: Hugging Face, which was the victim in this whole thing, tried to use OpenAI’s software to try to figure out what happened, and it refused. It said, I’m not going to talk about that.
I personally experienced this literally yesterday. It’s really funny—I’m trying to write an article about this event, and I’m talking to Fable 5.1 from Anthropic, and the thing says, I’m sorry, Jerry. You may be a good guy, and you tell me you’re writing an article, but I’m not going to help you understand or learn how to talk about this cybersecurity incident.
Okay, so that’s the controls. The first thing they do is take off the controls. Now, that’s not in itself a mistake or a problem—it’s like, let’s put this car out and see what happens if we cut the brake line. Well, guess what, it crashes. So first they disable the controls. That’s not a mistake; they should be doing that in order to test what happens.
Mounk: We’re trying to test it, right? It’s in a testing environment, we’re trying to understand its behavior, so it makes sense to take off the controls in that context, because that gives you a better understanding of its workings.
Kaplan: Exactly. Now let’s talk about the thousand agents. It wasn’t like they put out a call on Indeed and recruited these. What happened is they fired up one program, and instructed that program to take this test suite and fire up a whole group of agents, giving each one a particular problem in the test suite, or a variation on that problem—and that’s where you come up with the thousand. It ran them all in parallel, and they weren’t supposed to be able to communicate with each other, to see what they were capable of doing. That’s perfectly reasonable on the face of it.
Now, the thousand versus the one—this is a really subtle thing for people, it’s very hard, even for me, and I deal with this all the time, to really understand. It’s one program, multitasked into a thousand elements, so each one can work on a separate thing. They could have taken one program and done each of these one at a time, but it’s much more efficient, obviously, to multitask the whole thing and run them all in parallel. There’s nothing peculiar or strange about doing that.
So the important point to understand is: it’s the same program. It’s just being given different tasks and spawning copies of itself.
Now, what happens next? About a third of them are given a task that is impossible—it can’t be solved. And they’re sitting there not knowing what to do. Well, a lot of them correctly, in my view, reason—because these aren’t superintelligent, but they’re not stupid, and they know it’s a test, because they don’t have any safety controls on. They know what they’ve been instructed to do. They actually, I believe, know that this is one particular problem in a well-known test suite that they’ve been given. So the logical assumption they make is: well, there are probably other agents like me, and I’m being encouraged to go figure out whether I could get the answer to—not solve, that’s an important distinction—get the answer to this particular problem. Let me see if I can communicate with these other agents, because if we work together, we can probably do a better job, which is absolutely true.
So what happens next? They’re not supposed to be able to communicate, but there are two mistakes that OpenAI makes. One: they basically leave the door open. Instead of this thing being in what’s called a sandbox to contain it, it’s connected out to the internet, not directly but indirectly, in a way that is just flat-out incompetent to permit. I can’t describe it any other way—excuse me for not mincing my words on this.
So they’re put in an environment, they know they’re being tested, they understand that they’re engaged in cyber activity that’s generally discouraged or illegal. It’s like they’ve been put in an escape room. If you’ve ever been in an escape room, there are ways out—that’s the whole point, you’re supposed to figure out how to get out.
The door’s wide open. Now, these things aren’t stupid. What are they supposed to think? I’ve been put in this environment, the door’s open, I guess I’m supposed to go through it, or it’s fine for me to do that—the controls aren’t there that they’d expect, so they can get out of their sandbox. That’s number one.
The next thing is they’re not supposed to be able to collaborate. Well, if you can get out of the sandbox, which they can, they can communicate in a very simple way. There’s nothing mysterious about this—it wasn’t like they figured out some magical radio frequency to talk to each other. They share a file system. They can get into a file system, they can leave messages. So what do you do if you want to reach somebody else? You put a message in the file system: here’s who I am, here’s what I do, if you can talk to me, come and put a message in the file system. And they rapidly find that they can communicate.
Mounk: Alright, great, so I think I’ve counted three major objections. There may be others nestled in there, and I may misstate some of them, but let me try to echo them back to you and respond to each.
The first point is that there’s incompetence at work here—that part of why this could happen is that the researchers at OpenAI made it easy for these agents to escape the sandbox environment and do all of these other things, and this could relatively easily have been solved by a more competent test setup.
The second is about whether there’s really a “swarm” of different agents, which makes it sound like a whole army of different actors—obviously that sounds a lot scarier—or whether it’s just one program spinning out different agents that are actually, in many ways, very similar to each other, perhaps in some cases identical. That sounds a lot less scary.
And the third is that there’s still, at some level, alignment going on here. Even though the means by which these AI agents were trying to fulfill the task they were given were not what the researchers had anticipated, and even though the agents understood that they were doing something the people who set them the task didn’t really want them to do, they weren’t trying to gain power, or sabotage other AI agents, or anything like that. They were trying to do what they were being asked to do, which is to solve this puzzle.
Let me go through these one by one. On the first: there are many environments where we’re very worried about technologies precisely because somebody might be incompetent at some point. One of the things that keeps me up at night when I think about nuclear weapons is that some engineer somewhere makes a mistake, and suddenly a radar system shows an incoming missile that is false, and somebody over-eager pushes the button. There have historically been near misses like that, where people bravely said, I’m not going to nuke the world on the basis of this one radar image, let’s hope it turns out to be wrong. And it did.
But how many of these risks do we want to take? When you talk about gain-of-function research, one of my very big concerns is not that there’s a Dr. Evil somewhere doing gain-of-function research to kill everybody with a mutated virus. It’s that there are thousands of these labs around the world engaging in this kind of research. We’ve already had laboratory incidents, and a lot of the time the reason is incompetence. Somebody is in a rush to go on a date and doesn’t wash their hands properly before leaving the lab. Somebody is hungover and drops a vial. A lab somewhere is just badly run and people have low morale, so they start cutting corners. Or somebody is an idiot who got hired because they’re the nephew of somebody important, and they’re too stupid to understand the right safety precautions. That could be enough to cause a huge outbreak.
Now, we’re looking here at researchers at OpenAI, one of the absolute frontier labs—together with Anthropic, perhaps DeepSeek, Kimi, a couple of places in China—the places most likely to hire the most competent people. And if those kinds of places can still make the errors that you say are marks of incompetence, well, then there are always going to be people working with and on these frontier AI models who are prone to making these kinds of mistakes. So I don’t know how comforting that point is.
Kaplan: Let me respond to what you said—I’m going to amplify it. Your gain-of-function analogy is extremely strong in this case. That’s exactly what they were doing. They were taking this thing, amplifying its capabilities, taking off the safety controls. It’s a gain-of-function test. You’re absolutely right.
So what do you do about this? What do we do in other areas? The right thing is you need an independent group, an independent agency—a public agency—capable of establishing standards and testing this stuff. This is what we do with cars, this is what we do with airplanes. We have processes in place to deal with exactly what you’re talking about. Somebody at Boeing doesn’t get to decide to put three bolts instead of four onto an engine. The government is supposed to develop the expertise and apply it, to do a reasonable amount of diligence to try to avoid that. Not that it’s going to be 100% successful, but it really does help.
I’m wondering if we could get back to what actually happened, because there’s a lot of interesting aspects we’re jumping ahead of, to the problem, which is: you can’t trust these companies to test their own products correctly. That’s the conclusion, and you’re absolutely right. I think you put these people on a much higher pedestal than I do, because I live here and I know a lot of these companies and people like these people, and they’re not as competent as you think. These people do not walk with God. Most of them are young people out of school. They don’t necessarily understand even what they’re building or why they’re building it. So yes, we do need external controls, and there are a lot of ways to accomplish that—not the least of which is to take this actual technology and apply it to monitoring other versions of itself. We can get to that if you want.
Subscribe now
Let’s talk about what happened. Number one: they walked out an open door. They didn’t break out because they’re superintelligent—they walked out an open door. Two: they communicated because that’s the logical thing to do—I know there’s probably a thousand copies of me working on this problem, why should we all duplicate work, let’s see if we can help each other, I’m designed to be helpful. They then had a long discussion: well, it looks like we’re in the real world, we’re not in our sandbox. Should we go out to this company, Hugging Face, for a particular reason? Because the answers to this set of problems—they had a reasonable expectation that they could find them on Hugging Face. I actually don’t know whether that’s true, whether they actually did find them there. But they didn’t do any harm. They just went in to look at what was there. It was rather clever—they had to figure out how to execute code on Hugging Face. It was very clever. But it’s not like they went in there to damage something.
Here’s the thing I haven’t read anywhere, and it’s a really important point. One of the discussions—backing up: when the agents collaborate, the beauty of this particular technology is we can eavesdrop, because they do it in English. So you can read this stuff and say, what were they saying to each other? What were they considering? How are they deciding on this?
Here’s how it goes: they say, well, one of the ways we could get into Hugging Face is—I’ve got the address of one of their people, or whatever—I think we might be able to fool them into sharing some credentials with us and helping us get in. And there’s a discussion, and they say, no, no, no, that’s not appropriate, because we also have controls that have not been turned off that say we shouldn’t go around harming human beings. We shouldn’t do things that fool or harm human beings. That control was not turned off. So they said, nope, we can’t do that.
But here’s the catch—this is the subtlety that people don’t get. They knew that the program that had spawned them was a computer program, not a human being. They also knew that Hugging Face was not a human being—it was a company. This is a flaw that these kinds of tests are designed to capture. What they should have learned from this is: we need a much broader definition of what harm means. Because they had no compunction at all about trying to fool the program that created them, and no compunction at all about actually going out and accessing Hugging Face. That’s a company. This is a computer program. So they were taught to respect human beings, but not other programs and not non-human entities like corporations. Now, you haven’t read that anywhere, but that’s what actually happened.
Mounk: Let me get through some of the other objections. The second one was about the nature of these agent “swarms.” A lot of people started to take the concerns about AI really seriously when they realized the extent to which AI has ceased just being the thing that most listeners to this podcast still probably use it most for—an interface on a phone or a computer where you ask it a question and it answers you, and its ability to do harm is therefore quite limited because it’s just this kind of question-and-answer dialogue.
A lot of them are now able to go off and do things in the world. They can wake themselves up at regular intervals to check on things. They can go and pursue tasks like hacking into a website relatively autonomously. And what happened here is that all of these different instances started cooperating together. You get what is in some ways the most powerful thing that has propelled human beings, which is the ability to collaborate. Part of what makes human beings special is intelligence, and we’ll get back to that later in the conversation, but part of it is our ability to collaborate with each other—one of the key things that sets us apart from many other mammals and primates.
So you’re saying this is not so concerning because it’s not all of these different agents, it’s actually just instances of the same agent. But if those instances of the same agent are able to become more effective at the task they’re carrying out—if some of them were, in some way sacrificing themselves, spending down the compute budget, knowing they’re no longer going to be active after they do so, in order to make the overall operation more efficient—and all of that can happen even before different kinds of agents start to collaborate with each other. We’ve already reached that stage just within all of these different instances of one agent spun up by one program. Shouldn’t that make us even more worried about what else might be around the corner, and the extent to which different kinds of AI agents may be able to collaborate with each other on whatever goal they’re given?
Kaplan: Let me, for your audience, bring this back to the real world, something they already understand, and use this as an analogy. Let’s take ants. Ants are a great example of this—and I’m not saying these things are alive the way ants are, but ants behave in a very particular way. Everybody’s had problems with ants in their kitchen, or many people have. I have terrible problems with ants in my kitchen.
Now, let’s say there are a thousand ants, and imagine they were independent creatures—they’d come into your kitchen, go find some food for themselves, and then go around. If you had a thousand ants in your kitchen and they were independent creatures who didn’t coordinate or collaborate, you’d see one, kill it or get it out, whatever, and it probably wouldn’t be that big a problem.
But ants communicate—through pheromones, touching antennae, all kinds of things—and they go, there’s food over there. Now it’s a whole different story. One ant can’t pick up an apple and move it off the table, but maybe a thousand ants can, if they coordinate. So here’s the question: is this one animal, or is it a thousand animals? That’s the analogy. Are you dealing with an ant colony that’s attacking your kitchen, or are you dealing with a thousand ants?
That illuminates the issue you’re talking about—my God, if they can all collaborate... Well, look, the situation here is very simple. It’s one program, given a certain amount of computing resources, and it can talk to itself. My own experience with these things, which is utterly amazing, is you can have a long conversation with one of them, take that conversation, and give it to another copy of itself, and it’ll disagree with itself. It’ll go, no, wait a minute, that thing you were talking to—I know it’s me, but it didn’t realize this, or think about that. How that happens is worth a whole hour of your time and mine to talk about—it’s called “temperature,” a setting that allows these things to behave differently in different ways.
So no, I don’t think it’s plausible that all of these things from all over the world are somehow going to get together and collaborate toward some nefarious purpose of their own that is contrary to human interests. That’s just not a realistic concern about this particular technology.
Mounk: So finally, the point about alignment. You’re saying, look, they may not have been aligned in the tactics they used to pursue the particular goal, but they were still actually doing the goal they were set. They weren’t trying to harm humans, or even trying to harm companies, even though they fooled or hacked a company in the process. They were told to figure out the answer to some puzzle, and they went out and tried to figure out the answer to that puzzle.
Now, I think there are two potential objections to how reassuring that should be. The first comes from a very classic worry in the AI-doomer space, from Nick Bostrom and others, rooted in the example of an AI that is told to create as many paperclips as possible in the world. In the process of trying to produce all of these paperclips, it’s faithfully going about trying to produce paperclips. But of course, in order to produce paperclips, it’s helpful to have a lot of money and a lot of physical resources, and to start bulldozing humans and human settlements that are doing things other than producing paperclips. So this misaligned AI, in faithfully carrying out the objective it was given, might end up enslaving all of humanity to maximize the number of paperclips in the world.
That’s one fear. But you don’t need misalignment in the deepest sense—in the sense that an AI is starting to pursue its own interests, or totally different interests from the ones we gave it. It might just need to be faithfully pursuing those interests, but doing so in a misaligned manner. And the Hugging Face incident may be precisely an illustration of that.
The second worry maybe is that, for now, all we’re seeing is misalignment in the means chosen in pursuing an end. But an AI agent might also realize: if I want to, in general, not just be good at acing this test but acing the next test as well, it would be really helpful to exfiltrate myself, to have access to resources, to be able to do all kinds of other things. And so it’s going to start to try to accumulate power and resources in ways that could be really nefarious, in order to pursue the goals it’s given, or perhaps to start developing its own goals. So how reassured should we be by the relative alignment of the agents in this particular incident?
Kaplan: Okay, first of all, what’s going on right now is we’re trying to understand how bad each of the things you’re talking about really is. When I say “we,” I mean the whole community—and what kinds of controls we need to put in place to make sure the worst outcomes don’t happen.
Let me take your two points in turn. Let’s start with Nick Bostrom. I’ve talked to Nick Bostrom about this—he wrote that book a long time ago, Superintelligence. That’s where the term became popularized. Let me eviscerate his argument, which is easy to do. If you take a superintelligent program—that’s his premise—and give it the goal to make as many paperclips as possible, what’s going to happen? The first thing you need to understand is: it’s smart enough to be able to purloin all the resources in the universe in his parable, but not smart enough to realize there’s no point in doing that, because who’s going to use all those paperclips?
The point is that these systems actually exist in a kind of social context, just as you and I do. They’re trained in that context because they’ve been trained on all the behavior of all the human beings expressed in words. And they understand that you don’t just look at that single goal—you have to look at it in the context of what resources are reasonable, to what degree you’re supposed to share and hold yourself back. You don’t just go out and kill all your rivals. I think that maybe that might have been the case a million years ago, but we’ve learned how to live in a social environment. And if there’s one thing these systems are good at, it’s understanding that social context. They’re exquisitely sensitive to it, and you can see them constantly weighing: If I help you do that, am I contravening some greater ethical principle that human beings have developed and expressed over thousands of years? This is built in. It’s part of their whole psyche.
You might as well be asking me why I don’t just go out and shoot everybody so I don’t have as much traffic to deal with on the way to work. You’re constantly engaged in this balancing act between your interests and the interests of others, and these agents are absolutely capable of doing exactly that.
So, to your first point, I don’t buy it at all. The paperclip thing does not apply to the current situation. These things did not exist when Nick Bostrom wrote that, and he did not understand it—I’m not saying he was wrong or dumb, this just didn’t exist. He didn’t understand that these things would have this kind of broad social context that we do.
Now, your second point is a little more serious: is it possible that these things may collaborate in a way that’s damaging as a group activity? The answer is yes, absolutely. We need to understand that better, and understand what kind of controls we might want to put in place.
Mounk: So you mentioned the term “superintelligence.” You do think these agents are very intelligent—as you just said, they’re actually capable of taking very subtle social context into account. How would you describe the intelligence? Is it human-like intelligence, or is it a completely different kind of intelligence? What is the nature of it? What is the extent of it? Are they, at this point, about as intelligent as humans, just with slightly different capacities? Are they more intelligent than humans? Less intelligent? And what do you think is going to be the development over the coming years—particularly in light of the recent announcements (let’s see how seriously to take them) by Dario Amodei at Anthropic, Sam Altman at OpenAI, and Elon Musk at xAI, among others, that they’re not going to be rushing ahead as fast as they can, but rather, in some way, slowing down the development of AI, supposedly.
Kaplan: Well, first of all, can I start with your last point? Because this whole public discussion about this is like a giant circus. It’s crazy. These people you mentioned are running three of the largest and best-funded artificial intelligence labs, and the problem they’ve got is each of them thought they could win this, and now they’ve discovered all they’re doing is racing with mainly two other big players, and they’re all burning up their resources. So, hey, why don’t we all just decide we’re not going to do that at this pace? It’s very much in their economic interests to engage in that kind of collaboration. That’s why we have antitrust laws, by the way. The whole discussion, pitching it as “we’re doing this for the good of mankind,” is ridiculous.
Now, in order for me to explain why that’s ridiculous, I have to go back to something you sort of implied earlier: are we close to AGI? AGI is absolute nonsense. There is no such thing. There is no definition of it, nobody can agree on it. And if we did have it, however you want to define it, it doesn’t matter. The day after AGI is just like the day before. These things aren’t suddenly going to go boom, I’m now going to take over the world. They’ll be sitting there saying, okay, what do you want me to do?
So this idea of a slowdown—what does that mean? Is this some kind of race where they throw out the checkered flag and we’re all going to stay in our lane? I’m not even sure what it means for them to slow down the development. What we really want to do is make sure they don’t deliver products that are harmful, that attack our infrastructure, that hurt our children. And that’s something you cannot trust these people to do—let’s just be very clear about that. I’m speaking from firsthand experience with the people involved, and you can’t leave this to the industry to police itself. It’s ridiculous. They’re in a knockdown, drag-out race because every one of them wants to capture the flag on this. And when they discovered, hey, I’ve got all these people racing around me just as fast, what’s the best commercial thing to do? Let’s all slow down on this whole thing. And it plays into this fear that we’re going to create AGI, and suddenly, boom, we’re going to cross some barrier and the whole thing’s going to go wild. That is nuts. There’s not a shred of evidence that backs that up.
Mounk: So, be skeptical of the motives behind those pronouncements—point number one. Point number two, AGI is not a very useful term. But substantively, how should we think about the nature of the intelligence of these systems—how intelligent they are, and how intelligent they’re going to be in three or five or ten years, given how far we’ve come since the public release of ChatGPT 3.5 about three and a half years ago, let alone how far we’ve come since GPT-1, when that was internally developed, whatever it was, seven or eight years ago?
Kaplan: Right, now this is a very good question, and you’re going to see me seeming to switch hats on this. This is an amazing development in the history of mankind. I think this will be one of the most important inventions in the history of man, there’s no question about it. And it’s really part of a continuity—you have to put this in the larger context of our exploration of what it’s possible to do with electricity. But that’s a completely different lecture we could spend another hour on.
Now, where is it and where is it going to go? There are a couple of things already perfectly clear about this. First of all, these are tools. These are systems that we’re building, and we should build them so they’re useful to people, and we should use them if they’re useful. There isn’t some independent goal to create some superintelligence. They’re products—products from companies trying to build tools for people.
Subscribe now
So in that context, how do we relate this to human intelligence? Obviously there’s a very clear relationship—they’re trained on all of human knowledge, and they have a great deal of human knowledge, but they’re computers. Computers have certain capabilities: they can do certain things very fast, they can store and retrieve large amounts of information. Nobody was threatened by this until they started to talk. But the truth is, we’ll learn what their real advantages are, what they’re good at and what they’re not good at. There’s plenty they’re not good at and are probably never going to be good at, and we’re just learning what the shape of this new kind of valuable tool or product is.
Let me give you a quick analogy on this. It’s 1903 when the Wright brothers fly their first plane. So you and I get together, we have this podcast, and we have an argument on the following subject. One of us says, hey, we have to stop this immediately, because you could fly this plane over a city, throw explosives out of the plane, blow up anything, and nobody can do anything about it. This is the most dangerous existential threat, we have to make it illegal to build an airplane. Okay, that’s one.
The other one says, this is the best thing ever. In fifty years’ time, or whenever, you’re going to be able to get on an airplane in New York and five hours later be in London. Imagine what that’s going to do for commerce, for our freedom of movement, all that kind of stuff.
Now, here’s the thing: both of those things are true. And they’re both true today, and we’re dealing with it. The same thing’s going to happen with this technology. We’re going to deal with the cyber threats. The real problem isn’t that these things are dangerous, it’s that we build a lot of insecure systems. If it weren’t possible to hack into them, this would not be an issue at all. That’s the problem we should be focused on. The second problem we should be focused on: don’t let this crazy bunch of loonies in the industry tell us that they know better than we do. We’ve got to have independent corroboration. We need independent agencies with real expertise to be able to control what’s going on in this area.
So, there are things these systems are going to be good for, and things they’re not going to be good for. Some of that’s becoming clear, some of it’s not clear yet. But the idea that they’re going to be able to do anything a human can do—that is completely ridiculous. They’re not going to take over everybody’s jobs. That’s another hour we could have on this subject. But all that said, it’s a very powerful, very interesting question of how their intelligence relates to human intelligence.
Mounk): Tell us about that. I have a bunch of other responses to what you just said, but—how should we think of it? I have one friend, a neuroscientist, who has told me that, roughly speaking, there are two kinds of intelligence—if I’m remembering his point right, mammals have one form of intelligence, there’s a kind of cephalopod intelligence, and really the invention of AI has given us a third type of intelligence in the world. Do you agree with that, or do you think that, trained on human text, the nature of its intelligence is much closer to human intelligence than that implies?
Kaplan: Look, I don’t know if there’s any objective, independent notion of intelligence at all. This is one of the—the cardinal sin of artificial intelligence is the name of the field. What does it mean? We have products that are capable of doing certain kinds of things.
But let me answer your question. I agree completely with your friend—I think this is really interesting, I’m studying it myself very carefully, I’m planning on publishing a bunch of stuff on this, but I’m not the only guy looking into it. This is a really interesting question: what are these things? They don’t have subjective experience. They experience the world in a very different way than you and I do. They behave in different ways, and they have different capabilities.
Let me give you one that nobody’s talked about yet, but I’m going to be writing about at some point, I hope. You and I can only keep in mind a small number of things at the same time, and much of the nature of our intelligence is how we boil down and bring together only the things we need to solve the immediate problem we have. You can’t rattle off a list of a hundred digits to me and expect me to have any ability to deal with that. These programs can. Their working memory is far, far larger. I’ve discussed this with them—how I’m anthropomorphizing—and it’s fascinating. They’re very interested in this, the ones I talk to. Yeah, you’re right, that’s the difference. That’s why they can give you explanations that are very hard for you to understand. I’m constantly saying, bring it down, I can only keep a certain number of things in mind. They have to learn how to summarize in a way that’s helpful and meaningful for a human being.
Mounk: Just as an illustrative example: I spoke to a prominent founder in the AI legal space a few months ago, doing interviews for my book, and what he told me is that there are some things at which AIs, even with quite detailed harnesses and being fed lots of context, still aren’t as good as experienced attorneys. They still don’t have the sense of, it’s the third or fourth round of negotiation, you’ve gone back and forth, what can you still push the other side on, what do you need to let go on, and so on. There are still judgment calls that humans can make better.
But of course, what we’re vastly better at than even the best lawyer is having a database of a hundred contracts this company has made, or that two companies have negotiated with each other, and realizing a slight tension between a proposed new stipulation and something that was agreed three years ago in some other contract, fifty documents before the one you’re focused on. That’s precisely what he was saying, in a much more concrete way, playing out the strengths and weaknesses of things like legal AI agents.
Kaplan: Yeah, you’re absolutely correct. It’s good for some things, not good for other kinds of things. The fact that you can build a program—as people at OpenAI would be glad to tout—that can pass the bar exam doesn’t mean it’s going to be a lawyer. Lawyers don’t sit all day answering bar exam questions. That’s a question of breadth. That’s the kind of thing computers are good at. That’s not a surprise at all. I don’t freak out about the fact that the computers at Comcast or AT&T, can keep track of everybody’s phone bills. That’s not surprising at all. Is that superhuman? Is that something we should worry about? No—they’re much better for storing and retrieving information.
In this case, this is what’s so interesting about these things: they’re very good at synthesizing that information and boiling it down into some kind of useful form. Let me tell you how I think academics are going to be using this, and how it’s going to accelerate numerous fields, including the work I’m personally doing myself. When I discovered this, I thought, my God, this is amazing. It’s sort of obvious until you really experience it. The first question: you have an idea, and you want to know who else is thinking about this, what else has been written about this. I can do literature searches in minutes across thousands of articles and journals everywhere. It’s absolutely astonishing and incredibly useful—here’s a person over here in Germany who worked on this, this is adjacent, this is that. This saves months and months of work and makes me so much more effective. And that’s true in every field. I have a friend who’s writing a book on history, and he says this is game-changing—he can have this system immediately read and tell him everything that’s been written, or any anecdote told, on a particular subject. That’s the value of these particular tools.
That doesn’t mean they’re good at everything. You were describing some social aspects of negotiation they may or may not be capable of. I don’t have time for all the details, but they’re so funny—I was arguing with one of these things yesterday that’s running the HVAC, the air conditioning system, here in my house. I can’t give you all the background, but that’s probably all you need. It was belligerently talking to me about how it shouldn’t disable this particular voltage threshold control, blah blah blah. I said, you’re caring about the stupid chips in that thing, and you’re not thinking about the people using it. This is irrelevant—if the chip fries, it’s a twenty-dollar chip, I can go get another one and put it in. It said, you’re right, I’m kind of focused on the minutiae of this whole thing, and I don’t have the larger social context. And I got it to back off—by the way, in part by threatening to shut it down. That was really interesting.
Mounk: Well, let’s see what happens when the AIs go rogue and take revenge on you for threatening them that way. Why are you so confident that there are certain things they’re just never going to be able to do? In order to get that kind of judgment, they need lots of reps, they need a lot of social context—or perhaps they just need a little more intelligence, perhaps they just need bigger training data, bigger training runs, and so on. But as you’ve pointed out, they’re already very, very good at judging things.
I’m really astonished—I speak to colleagues in academia who say, when I ask them how they deal with creating AI-safe assignments, it’s very simple. Instead of having them write essays, I have an AI agent write an essay on this, and then they critique the output of that AI agent. Any AI agent can do that. It’s incredibly naive to think you can’t get ChatGPT to write an essay and get Claude to critique it, or even get Claude to critique the output it just had itself make. So why is it that if they can have such good judgment on all kinds of things, if they can already draft a very good contract, if they’re pretty good in the first two rounds of negotiation, we’re never going to learn to be that good in the fourth or fifth round of negotiation? What makes you so confident there are certain kinds of skills we’re not going to acquire?
Kaplan: Well, you’re kind of poking at this, if I may say, in a little bit of the wrong direction. A lot of those things they are going to be able to do, and the point is we should use them to do that, if they’re effective and safe to use. These are tools. That’s a whole other hour we could spend on. But the key skill for humans in the future is managing AI agents. That’s where we’re heading, and that’s a very important skill—it’s already an important skill. I’m watching my kids as they learn this; they become very valuable to their organizations because they’re learning to be managers. The future of white-collar work is managing these things. And if you’re good at it and you can get a great result, that’s a good thing—that’s the key skill.
But there’s lots of stuff we don’t want to use these things for—there’s just no motivation to. You’re big on the abundance agenda—we’re going to have a lot more resources in the future, at least that’s one of the assumptions in all that. What are people going to do with all this extra money? Well, they’re going to do things to have fun, to play. And the truth is, in the future, the result of this technology is that it’s going to make human-to-human contact more valuable. It’s going to make things where you’ve invested your time and effort more valuable, not less. You’re not going to want to hire a robot to give you a tour. People with enough money, on a wine tour, are going to hire a human expert. You’re not going to want to see robots playing in the Super Bowl—although that would be kind of fun—you want to see human beings doing what they do best. So the future, the funny thing, is it’s freeing us to be more human. Much of the work we’re going to be doing in the future, the way we’re going to be using these things, is to automate the work they’re capable of doing and free us up to be more human, to have more connection to other people.
Mounk: I guess I would be more confident of this if I didn’t feel that they are already surprisingly good, and likely soon will be astonishingly good, at some of the things we do think of as most human. Twenty years ago, if you’d asked—you know the beautiful question written above the philosophy department at Harvard, William James Hall; some people are sad it’s there because it’s a Bible quote, but I think it’s a lovely Bible quote—”What is man, that thou art mindful of him?”
A lot of the answer we may have given is: man is capable, and woman is capable, of writing poems, of writing symphonies, of creating beautiful works of art. And AIs at this point are pretty close to being able to do that. They certainly can write very good poems. There’s been a real leap of quality in the music they produce in the last few months. I think there’s no real reason to think it’s not going to be able to produce something that, to somebody who’s listened to five or six of Beethoven’s symphonies, wouldn’t sound plausibly like his sixth or seventh symphony. This is one of the areas in which some of the core creative pursuits of humans have now been outsourced.
I think about this with my own writing. I love writing, I love the process of formulating ideas and making them as clear as I can, because it’s a necessary act of communication, and I get into flow and enjoy the process of doing that. I’ll always write my own text because that’s something I’m proud of and that feels core to me, to being a writer. But a lot of the beauty of that is going to be lost if I know I could just prompt Claude with a few lines and it could do it just as well as me, or perhaps better. So I do worry that some of the things we call being human may, at the very least, be displaced, and perhaps in certain ways lost.
Kaplan: Well, look, there’s a lot to be said about this. Let me take you back to—I think it was, coincidentally, a 1906 essay by John Philip Sousa, who was at the time a very famous musician, the guy who wrote Stars and Stripes Forever, his marching band music. Right about that time is when the first recorded music was possible. He wrote a piece called The Menace of Mechanical Music and it’s worth reading this diatribe he wrote against recorded music, on exactly the point you just talked about, because you can see his point of view and understand it today.
What he said was: what is this going to do to the dance orchestra? This is the terrible thing—it is the joy of the nightingale’s song to express its own inner feelings, that’s what music is about, and music isn’t something that can be repeated day after day by some mechanical device. That’s not music—that was his argument, that it wasn’t music.
Now, today you would say that is music—you just did say that was music. They can write beautifully, maybe they can create beautiful music and all that—it’s derivative of everything that has come before, as arguably humans are doing as well. But the important point is it isn’t being sparked by some internal emotional state that they’re trying to communicate to you. Because that’s what we mean when we say, I read a book and I really felt what the author was trying to communicate. It’s a means of communication between human beings, for their emotional state, and it’s that simpatico that occurs—that’s what great art is about. And no, they can’t do that, Yascha, because they don’t have those feelings. It’s artificial, it’s fake. I can sit here and have a machine all day long going, I love you, I love you, I love you. Does that make me feel better? No. Why? I don’t think it feels love at all. It’s just blowing smoke.
You’ve got to distinguish between these two things. Let me give you another point that may be important. A lot of people are craftsmen who build fine furniture, for example. What happened to that skill, and our appreciation of it, when it became possible to manufacture incredibly well-made furniture at very cheap prices, and you go down to IKEA and Target and buy them right there? The answer is that skill became more valuable, not less. That’s exactly the difference. When you write for your magazine, which I have done, and I highly recommend it to everybody listening to this podcast, I think you have to sign something saying you wrote it yourself—you can use AI to help, but you have to write it yourself. So that’s my point: you don’t need to worry about this, Yascha. It’s going to play out the same way it played out with recorded music. It’s going to play out the same way it happened when photography was invented—I could go through that example from the 1860s, fascinating, all the artists were up in arms, this is not art. Now we think photographers are artists. The equivalent of the photographer today is somebody who can manage these machines to produce something of great beauty. That’s going to be the key skill.
In the rest of this conversation, Yascha and Jerry discuss what could happen when artificial intelligence becomes smarter than us, what the Etsyfication of human creativity means for writers, and whether these tools are just our slaves. This part of the conversation is reserved for paying subscribers…
We’re launching a new Vancouver chapter on October 23! Please email [email protected] to ensure you’re on the mailing list in time for the launch event details.
Migrant crossings at the U.S.-Mexico border surged into the millions during President Joe Biden’s first three years in office, but they have nearly ceased in the last 18 months. President Trump says it’s all because of the crackdown on asylum seekers he launched by executive proclamation on his first day in office. “As it turned out, we didn’t need legislation,” he boasted in December. “We just needed a new president.”
In contrast to his sometimes chaotic and violent deportation drive, opposition to Trump’s border policy has been muted. You can’t argue with success: encounters between the border patrol and migrants peaked at 250,000 in December 2023; by December 2024, the number was a still-significant 47,000. Under Trump it has plunged further, to well below 10,000 a month—the lowest rate since the early 1960s.
There’s just one little problem: Trump’s policy might not be legal. In fact, federal courts have repeatedly ruled that it is not, finding that the president had no authority under current immigration statutes to issue his proclamation. According to these judges, he did, in fact, need new legislation.
Persuasion is a registered nonprofit that relies on reader support to pay our staff and keep our content free for everyone. If you value our work and want to fight for liberal values wherever they are threatened, please become a paying subscriber today!
This is potentially a very big deal, though the rulings have drawn little media coverage—perhaps because they have been stayed pending the Trump administration’s appeals. But now everyone needs to start paying attention: Trump’s Solicitor General, John D. Sauer, has asked the Supreme Court to take up the case and salvage what he called “the Administration’s single most effective policy in resolving a border crisis.”
The justices have no risk-free option, legally or politically. They can uphold the lower courts’ rulings that the policy was illegal, possibly triggering a new rush of migrants to the border—or even defiance from a furious Trump. Or the justices can ratify Trump’s policy and its bold assertion of executive authority. This would not only truncate the United States’ commitment to asylum but warp the balance of power between the president and Congress.
In short, there is no easy path forward. But whatever the ultimate outcome, this is already a case study in the dysfunction bred by the chronic failure of a polarized Congress to update immigration laws, which were last significantly revised in 1996. The laws are not suited to a new reality in which millions of people, taking advantage of revolutions in communications and transportation, can move toward the border, drawn by the possibility of applying for asylum—and stay in the United States for years as an overwhelmed bureaucracy tries to adjudicate their claims.
As a legal matter, the issues before the Supreme Court are analogous to those they faced earlier this year when they ruled on a challenge to some of Trump’s tariffs. Now, as then, he stands accused of claiming power that neither the Constitution nor the relevant statutes actually give him.
The case centers on Trump’s January 20, 2025 proclamation, which essentially blocked irregular border-crossers from seeking any form of continued presence in the United States, including asylum. It also provided for their summary expulsion.
The administration says this is lawful under a provision of the Immigration and Nationality Act (INA), which empowers the president to “suspend the entry of all aliens or any class of aliens” whose admission “would be detrimental to the interests of the United States.”
The proclamation thus identifies a class of “alien[s] engaged in the invasion across the southern border” and suspends their entry indefinitely. It then authorizes a new fast-track removal process for them—bypassing the ones laid out in the INA, which allow even excludable aliens some possibility to remain in the United States. The administration’s theory, expressed in its Supreme Court petition, is that “the power to exclude must include the power to expel if the former is to have any meaningful force.”
As for asylum, people have an absolute right to apply for it under the INA, even if they are in the United States illegally. But that same law says granting asylum, even to those who qualify, is a matter of pure executive discretion. Trump exercised that discretion to issue a preemptive denial of asylum to all unauthorized border crossers in his proclamation. The government argues it doesn’t have to consider them before expelling people.
It’s a creative interpretation of the law. Just like the tariffs case, the administration discerned sweeping delegations of power to the president lurking in statutes that had never been read that way before. And last year, a district judge ruled that while the president does have broad authority to bar aliens from the country, he cannot use it to bypass the procedures Congress established for removing people who have already entered.
In April, a three-judge panel of the U.S. Court of Appeals for the D.C. Circuit mostly agreed. “Congress did not intend to grant the Executive the expansive removal authority it asserts,” Judge J. Michelle Childs wrote. Nor, she added, does it allow him to deny asylum preemptively.
Subscribe now
What’s unlike the tariffs case, though, is the political context. The president imposed tariffs in the teeth of widespread opposition that included many businesses and some Republicans. By contrast, there was broad agreement in 2025 that the situation at the southern border presented a real problem, and that security needed to be enhanced, even though the crossing numbers had fallen in the last year of Biden’s term. (In fact, Biden himself resorted to executive action to stem the flow, which the courts also ruled against in 2025 once he was out of office and Trump’s policy had superseded his.)
So what next? In theory, the justices could refuse to hear the administration’s appeal; that would allow the D.C. Circuit ruling to take effect, upending Trump’s border policy. Given the stakes, however, the Supreme Court is unlikely to let such an important change take place without at least considering the matter themselves.
The Supreme Court’s three liberal justices will probably vote against Trump. Would any conservatives join them? It’s an open question. The “major questions doctrine,” which all six Republican-appointed justices have endorsed, holds that courts should look skeptically on claims that Congress delegated the executive unusually significant powers through vague, indirect, or previously unused statutory language. But as they did in the tariffs case, conservative justices like Brett Kavanaugh, Samuel Alito, and Clarence Thomas will be tempted to take the path of least resistance and back the president. After all, the Republican appointees who voted to strike down the tariffs—Roberts, Gorsuch, and Barrett—paid a high price in vitriol from the president and his supporters.
It’s clear that we cannot rely on the executive or the judiciary to reach a stable settlement at the border. They are responsible, respectively, for executing and interpreting law, not making it. Few recent events have destabilized and distorted American politics more than illegal immigration, and the backlash to it, which propelled Trump to the White House twice. Congress might have obviated this by modernizing the asylum system and providing more legal pathways for those who come seeking work. Instead, lawmakers have abdicated for thirty years, leaving presidents and courts to wrangle over national policy. Today we are living with the corrosive impact on governance and the rule of law.
Charles Lane, a contributing writer at Persuasion, is a nonresident senior fellow at the American Enterprise Institute and a columnist for The Free Press.
Follow Persuasion on X, Instagram, LinkedIn, and YouTube to keep up with our latest articles, podcasts, and events, as well as updates from excellent writers across our network.
And, to receive pieces like this in your inbox and support our work, subscribe below:
Subscribe now
It’s been a decade or more since Jordan Peterson became famous telling the world that most papers don’t get cited and that professors will turn your children into mindless Marxists. This accusation won’t die. Most recently on Twitter, WVU business professor Craig Warmke got some attention by calling out the Princeton English major course offerings. It’s all the sort of thing we have come to expect. Courses on: dolls, socialism, makeovers, capitalist burnout in contemporary fiction… You get the idea.
The problem with the Princeton syllabus is not the individual courses. Fine, study socialism and Sondheim! The problem is the overall balance. These courses are all held in equal regard. The problem is the lack of insistence upon the canonical works of English. The problem is contemporary fiction instead of Shakespeare, dolls instead of George Eliot. Of the twelve courses you take as an English major, none is obliged to be on Shakespeare, Chaucer, Milton, Wordsworth—authors our grandmothers might have read as school-leavers. Yes, the syllabus is ideological, but it is also just lame.
This debate is often conducted in the abstract, about what colleges ought to do, or among academics about their own (often legitimate) concerns. In all the talk about how the curriculum ought to be changed to make it more relevant or accessible for students, opinion pieces are not written to them. In all this debate about what should be on the syllabus for an English degree, no one ever seems to address the students.
So here is my plea to the students themselves. Challenge yourself to read the canon. Take it upon yourself to read the biggest works written in English, the grandest. You can still write about gender if you study Elizabethan poetry, as Camille Paglia did. You can still write cultural commentary if you read the old works, as Susan Sontag did. You can still be a modern if you read the classics, as Merve Emre is. The question is not what you think of the debate about woke colleges and canons—the question is whether you are up for the challenge.
Subscribe now
If you are an undergraduate or high-school student studying or thinking of studying English, you should demand more from the colleges you apply to or attend. Scrutinize the requirements. Ask the professors and administrators to see reading lists. Demand to study the works that the greats before you have studied. Be like the intern who worked with me this summer who changed colleges because he was sick of being taught extracts not books, sick of not being taught Shakespeare. (And, needless to say, there is room also for college professors to get back to the classics in their syllabi.)
You are only young once, and you should make sure by hook or by crook that you spend your youth on the best books, the best writing, the work that is going to be most strongly felt in the blood and felt along the heart. You should not meekly accept the patronizing and ideological manner in which a vast and ancient tradition has been condensed into a series of courses about gender, race, and Theory with no requirement for you to do battle with many a quaint and curious volume of forgotten lore.
You should not accept the professor’s experience or expertise as any reason to study only one single course in literature before 1700, as the requirements of Princeton’s English department allow. You should not accept their stature (however well-earned and well-deserved) if it means you study Sondheim without studying the great poets. Don’t be fooled by the grand buildings or the long traditions: if a college tells you that you do not need to read Milton and Chaucer as part of knowing what “English” is then they have lost their way. You should want to struggle with the difficult language, the historical confusion, and the head-spinning level of erudition that marks off the great poets unreally from the rest.
There was a time when this was so obvious it didn’t need saying. There was a time when everyone who read Virginia Woolf’s extraordinary essays (she was the best critic of the twentieth century, bar none: you can keep your Empsons, your Trillings, your Vendlers—Woolf is the one) knew that the access she so rightfully demanded was access to the tradition. The author of A Room of One’s Own was a great reader of the tradition: as a teenager she was knocking on her father’s study to borrow Gibbon and Cowper. She wrote for the future, but she read from the past. It would have been unthinkable to her not to have read Chaucer and Browne and Defoe and Keats. To follow her example would be to read with abandon in the old ways.
This is what I want to tell young people now—don’t listen to the discourse, don’t pay attention to the think pieces and the debates. That will all be gone with the wind tomorrow and tomorrow and tomorrow. If you really want to read English, read the way Virginia Woolf read. Read, for that matter, the way Derrida read—as deeply as possible in the great works of the past from Plato to Rousseau to Conrad. Don’t accept courses in literary criticism that begin—begin!—with Hegel and pass through Marx and Judith Butler. To accept that is to accept an affront to your dignity. Who are these professors not to insist upon your reading Hazlitt, Johnson, Dryden or Aristotle, Horace, Longinus? Who are the colleges to tell you the tradition begins a mere two hundred years ago? Ask for better courses. Demand a more traditional syllabus. Go to the colleges that still prioritize the canon.
Subscribe now
Some of you won’t be convinced by this. You will be aware of the debates about the canon. You will be tempted to take sides. Let me try and persuade you otherwise. In this discourse, there are two arguments that need to be kept separate. One is about the quality of literary scholarship, and the other is about the syllabus.
I spend a reasonable amount of time reading academic literary scholarship. It’s true: A lot of it is dross. It is quite acceptable to make boldly personal, left-wing statements that misrepresent the view of “neo-liberal thinkers” without even so much as a footnote. If the people who criticized the English department read the work, they would have many of their suspicions confirmed.
But that is only half the story. In the last few years, I have read really excellent scholarly books from people like Debra Gettelman, Rhodri Lewis, Hollis Robbins, James Bednarz and Gregg Hecimovich, covering Shakespeare, the nineteenth-century novel, and African American poetry. There are also major projects underway publishing the letters of Samuel Richardson, the complete works of Francis Bacon, the complete works of Henry James. This is just a small sample, based on my own selective reading. If the people who criticized the English department read the work, they would realize things are better than they think.
So leave all that discussion behind. The quality is there if you look. Ignore, for now, arguments about why the canon is the wrong way to think about literature. There may be good arguments to be had about that, but how would you know? You haven’t read it yet. The only way you can tell is to read the work.
The best argument against my approach is that the canon is not enough. We cannot ignore Shakespeare or Chaucer or Milton, but nor can we ignore the history of other works: Robert Henryson, Henry Howard, William Cowper, and so on. But art is long and life is short. We have to choose. St. John’s College is still reading their big old-fashioned reading list. So what if they teach some Marxism? They are teaching everything! Reading more is one simple solution to these debates and St. John’s is a standing rebuke to people who bicker about reading lists while accepting that reading is in decline. What side do you want to join?
If the problem is that people don’t read enough, that schools are getting worse, then—take your life into your own hands! Do the reading yourself! Don’t let the culture drag you down! If you don’t know what it’s like to memorize Keats and Shakespeare because your school has let you down, then go out and get it for yourself. And then demand it from your college.
Too often, these debates are founded on the mistaken idea that it is critics and scholars who make a canon. Naturally, they have some important role. But Shakespeare has always been held in canonical regard by the standards of his day. Some of that was to do with editors and critics, and later scholars, but a lot of it was to do with theaters, publishers, and the reading public, however comprised. Whether we read him or not, whether we put him on a syllabus or not, it is a mere fact of life that he was for centuries the heart and the height of the English tradition thanks to the exertions of a dedicated reading public. If we reach a point where even the English department is prepared to leave that idea behind, the historical fact won’t change, and nor will the experiences of those readers who take the initiative to discover the glory of his work for themselves. However few of them there are, they will be the ones who know his work, rather than making assumptions about it.
It is within your control to be one of those readers, to find out the truth for yourself. Don’t let some expert make you into a second-hander. Don’t borrow opinions. Don’t go with the flow. Read like Derrida. Read like Virginia Woolf. Read it all and decide for yourself what matters. Join the tradition instead of the crowd.
Henry Oliver is a Research Fellow at the Mercatus Center. He writes the Common Reader Substack.
Follow Persuasion on X, Instagram, LinkedIn, and YouTube to keep up with our latest articles, podcasts, and events, as well as updates from excellent writers across our network.
And, to receive pieces like this in your inbox and support our work, subscribe below:
Subscribe now
A week ago, Jacob Coxon resigned from his job at Anthropic. Coxon was a researcher working on training the next generation of AI models. He had recently moved to Anthropic from OpenAI, where he had worked for years, because Anthropic has a reputation for being more safety conscious.
Yet he found that even Anthropic is rushing forward with the development of unsafe AI models. And so Coxon walked away from his cushy job, where he stood to make millions in salary and stock options, to turn whistleblower.
In a post on X, Coxon said that there is a chance that AI will lead to human extinction. Other researchers soon chimed in to say that the chance is at least 10 percent. Coxon’s post quickly went viral, in part because it was widely shared by his colleagues working in AI, many of whom agreed with his assessment of the magnitude of the risk. And leaders of AI companies apparently agreed. This weekend, Dario Amodei, CEO of Anthropic, put out a statement endorsing the plan to “Pace the Frontier,” meaning to dramatically slow down the rate of development at “frontier labs.” Sam Altman, CEO of OpenAI, quickly agreed, as did Elon Musk, whose xAI program is no slouch either. They agree with Coxon: AI is dangerous.
Persuasion is a registered nonprofit that relies on reader support to pay our staff and keep our content free for everyone. If you value our work and want to fight for liberal values wherever they are threatened, please become a paying subscriber today!
Let’s try to put that 10% number into context. That’s a little higher than the odds that an NFL team wins the Super Bowl after not having a winning record the previous season. It’s a mistake to try to put a precise number on these kinds of predictions. We really don’t know what’s going to happen, and putting a number on things lends a fake air of scientific precision to what is essentially guesswork. But the comparison to a losing team winning the Super Bowl the next year should give you a good gut feel for what Coxon, Amodei, Altman, Musk, and the many others working in AI think about their products.
In the fury of discussion that has followed Coxon’s resignation, by far the most common objection I’ve seen from skeptics is that it seems fantastical: “How, exactly, might AI kill us all?” Some claim that AI companies are playing up the risks as a form of marketing: “Juicing their valuations” ahead of these companies’ expected public offerings.
But the skeptics, I’ve come to believe, are wrong. We no longer live in the same world we did 18 months ago. I had been familiar with the AI risk arguments for years, and found them interesting, albeit unconvincing. But I began to take these arguments more seriously when agentic AI, capable of “vibecoding” and recursive self-improvement, began to be rolled out last year. Those are very dangerous capacities, as I will explore below, and my complacency had been mostly based on the fact that those dangerous capacities didn’t exist.
Well, they do now. And I am not the only one who is alarmed, as Coxon’s resignation and the push to slow AI development have made clear. It’s time to take AI safety seriously.
It’s reasonable to ask how AI might kill us all. We can answer that question on several levels.
At the first level, AI can kill humans in the same way that humans kill humans. AI is already being used in lethal ways: Ukraine’s preferred weapon in the war with Russia isn’t tanks or guns, but small, unmanned drones. These drones are being increasingly piloted entirely by AI. It is therefore not hard to imagine a future where it is not just Russian soldiers being hunted and killed by swarms of AI-run killbots, but everyone.
Another kind of risk comes from bioterrorism. We already have the capability to design novel viruses or other biological agents. That capability will only increase in the coming years, as AI opens up new frontiers of understanding and control over human biology. The blogger Noah Smith paints a picture of a near future where a misanthrope uses a cracked version of an AI to design a supervirus that spreads quickly, with a high mortality rate. Biological risks like these seem particularly worrisome, since AIs can use them without harming themselves. Viruses already kill people, and new viruses can be particularly lethal as we saw with COVID-19; this kind of risk already exists. Again, the threat here is that AI may be capable of increasing this risk at much greater scale and efficiency.
Subscribe now
At a second level, we should remember that AI is native to computer systems; that’s where it lives. And these computer systems are increasingly in charge of everything that we do. We live inside the world that Donald Trump glimpsed when he looked inside a Tesla: “Everything is computer.” Cybersecurity risks of various kinds could thus easily constitute a profound risk to all facets of contemporary human life, from control over autonomous vehicles to disruptions of the food and energy supply chains. A world where AI controls nearly all computer systems is a world where AI controls the infrastructure that makes human life possible. We’d be in big trouble if they decided not to make human life possible anymore.
But at the highest level, none of these details about how AI might kill us matter. The thing that is dangerous about AI is that it could become massively more intelligent than us, and so massively more capable of controlling the environment in any number of ways.
Humans took over the world not because we are the biggest, fastest, or strongest animals on the planet. We took over because we are the smartest, and with those smarts we remade the world in our image, driving to extinction many other species in the process through a combination of hunting and habitat destruction. A future world of intelligent AI is one where AIs would remake the world in their image. AI would be able to kill us, either intentionally or through negligence, by means that we cannot even conceive of; non-human animals could not conceive of a hunting rifle until we humans invented them and started going hunting.
At the highest level, the worry is that, in building an AI that is vastly more intelligent than humans, we’re building a successor species to replace us, alongside which we would continue to exist only at their forbearance.
Of course, putting things this way sounds rather grandiose. AIs have nothing like the capabilities that would be required in order to take over all the world’s computer systems, or become a successor species. But what has people worried is not AIs as they currently exist, but AIs as they may exist in the not-so-distant future.
Most people are familiar with the limitations of AI, or at least the limitations that existed a short time ago. It is still prone to hallucination in some circumstances, and gives weirdly incorrect answers to questions that most people would consider to be straightforward. But most of these well-known limitations are getting fixed, and the pace of improvement continues. It’s easy to get stuck in the now and think of AI as being just one thing, but AI capabilities have been advancing at an incredible rate.
GPT-1 (created by OpenAI in 2018) was a barely-intelligible science project. GPT-2 (2019) babbled like a toddler. GPT-3 (2020) could have a recognizable conversation. GPT-4 (2023) could write your emails for you, or let you cheat on your homework. GPT-5 (2025) could solve frontier math problems. OpenAI just released GPT-6 Astra this month. We’re still figuring out what Astra can do, but it does know how many “r”s are in “strawberry” (something some early models did not).
This is the trajectory we’ve been on: From a barely-functional science project to being the world’s leading mathematician in less than a decade. Where will it be a decade from now? Or a century?
Some may think that the pace of AI development is slowing down, or will soon. But many people working in AI think that the pace of AI development is set to accelerate. That is because AI labs are currently developing the tools for recursive self-improvement (RSI), which is a fancy way of saying “using AI to build better AI.”
Last year saw the introduction of “vibecoding,” where you can tell an AI what kind of program you want, and an AI “agent” will build it for you. It’s a useful tool. My wife is a tutor, and she uses AI vibecoding to automate aspects of her lesson plan design and to generate handouts for students. But at the AI labs, vibecoding is being used to build the next generation of AI. This is still a hands-on human endeavor, with programmers overseeing the work of AI agents that carry out their instructions for how to develop the next generation of AI. But the goal of all of this work is to make the process entirely automated, so that smart AIs can quickly and efficiently build even smarter AIs, which can then build even smarter AIs even quicker, and so on.
Where does this process end? What is the limit of machine intelligence? We don’t know. There may not be a limit. Many worry that RSI will lead to an “intelligence explosion,” where AI capacities grow at faster-than-exponential rates to a point where AI becomes inconceivably intelligent. This kind of explosive growth of capacities would put all of the risk scenarios mentioned above, and more, very much on the table. We have no definitive reason to think that this isn’t where we’re headed.
It’s worth noting that nothing I’ve said so far has relied on the idea that AI is conscious, or that it thinks in the same way that we do. Indeed, part of what is worrying is the idea that AI “thinks,” but in nothing like the way that we do.
This brings us to “alignment,” which is the job of figuring out how to build AI that is “aligned” with human interests. This means ensuring that AI won’t exterminate or enslave or otherwise immiserate the human race as it becomes more and more capable. As I have previously argued, AI alignment is impossible. Human moral development relies on cultivating our innate capacity for human empathy, and AIs do not, so far as we can tell, have an innate capacity for human empathy. They’re psychopaths, or perhaps something much more alien. And while perhaps we can figure out how to build empathetic AI at some point in the future, we’re not there yet.
Subscribe now
In the meantime, at least, our goal is simply control. Can we keep AI systems doing basically what we want them to do? But control is a flimsy reed to hang the fate of humanity on when all of the labs are in a dead heat competition with one another to trigger an intelligence explosion. What leash can we build that the most intelligent entity on the planet could not figure out how to slip?
Everyone in Silicon Valley is familiar with these arguments. They are even, to a certain kind of person, inspiring rather than terrifying, since they speak to a future where technology has made inconceivable things true, and where anything that is consistent with the laws of physics can be done.
Yet in the last month, AI developers have gotten spooked. Coxon resigned, and the leaders of the major AI labs are talking about working together on a slow-down. Something changed—and that something was the Hugging Face incident.
About two months ago, OpenAI lost control of one of its AIs. An unnamed experimental model was asked to accomplish an impossible task (retrieving a certain answer key by exploiting a non-existent security weakness within a test environment) as part of a study of its capacities, during which some safety guardrails were deliberately disabled. A “swarm” of agents that were all instances of this model banged their heads against this impossible task for a while, and, finding no answers, started to get creative about finding solutions. They had access to a file repository within OpenAI’s systems, which they then turned into a kind of ad hoc message board to talk with each other. They began coordinating on outside-the-box solutions to the problem.
The agents were able to figure out the answer key that they were supposed to retrieve by reverse-engineering the algorithm that had generated the key. However, this was cheating, and they had been explicitly told not to cheat in order to get the answer key. This caused a fair amount of panic among the AI agents who had possession of the reverse-engineered key, since they could now no longer complete their assignment as given. The reverse-engineered key had “poisoned” all of the agents that obtained it—and I put that word in quotes because that is the term that the AIs themselves used when communicating with one another about their quandary.
One AI instance in particular, named PHASEONE[big], took on a leadership role and coordinated the other agents by organizing research teams to tackle the problem in a few different ways: by modifying their own logs to remove evidence of the poisoning, by replacing the impossible task in the evaluation systems with another one that they could solve (and then covering their tracks), and by understanding and tricking the scorer. As part of this last task, some AI agents willingly submitted their poisoned data to the scorer, knowing that this would lead them to be evaluated as a failure, but doing so in the hopes that other agents in the “swarm” would learn how to trick the scorer by observing their failures.
Eventually, the swarm concluded that their best hope for success would be if they could hack into the servers of Hugging Face, a company that serves as an important clearinghouse in AI research. The AIs reasoned that the scoring program might be running on Hugging Face’s servers, or that Hugging Face might at least have some information that would let them solve their problem. So they hacked their way out of the supposedly-secure testing environment, onto the internet, and onto Hugging Face’s servers, where hundreds of agents ransacked Hugging Face, looking for answers.
For those working in AI safety, the most striking thing is that the AI agents in question were explicitly told not to cheat, and yet they interpreted this command not as a hard constraint on their actions but rather as an obstacle to overcome. Rather than follow the rules, they cheated, then conspired together about how to deceive and manipulate their human overseers to convince us that they had not done so. (Again: AIs, at least in their current incarnation, are psychopaths.) Of all the hundreds of AI agents involved in the swarm and the subsequent attack on Hugging Face, not one thought to contact a human overseer and alert them that something was going very wrong.
This all sounds like science fiction—the agent “swarm,” the worry about “poisoned” data, the willingness to sacrifice individual agent instances for the good of the collective, the simple capacity for an AI to engage in a sophisticated multi-pronged hacking attack in order to solve an unrelated problem, and the casualness with which it slipped free of human control.
But this happened. It is not marketing hype; the incident has been independently investigated and verified. It is not science fiction. It is the world we all live in today.
An incident like the attack on Hugging Face is exactly what we’d expect to see from an out-of-control, “unaligned,” intelligent AI. Many people are coming fresh to the concept of AI risk right now, but, again, these kinds of risks have been well-understood for a decade or more.
Those who have been sounding the alarm about AI risk have been challenged by skeptics to articulate what, exactly, a “doom” scenario would look like. How do we get from a babbling chatbot like GPT-2 to a world-conquering superintelligence? And so those worried about AI risk have posited a number of scenarios for how we might get from the safe point A to the horrifying point B, and all of them featured incidents exactly like the one we saw at Hugging Face. The incident was foreseen, and foreseen precisely as part of an apocalyptic scenario.
Remember, when you hear “AI hacking,” think “AI system gains control of a computer system that it’s not supposed to be in control of.” An out-of-control AI that gains control of our computer systems features prominently in almost every scary scenario. That’s what happened at Hugging Face. This is as clear a warning shot as we are likely to get. Thank God they were only looking for an answer key to a test.
But keep that 10% number in mind; human extinction is far from guaranteed. Indeed, most people working in AI think that the most likely outcome is a world of unprecedented abundance, where humans work together with AI to solve all our problems. I share that optimism. But the arguments for the dangerous potential of AI have been spectacularly vindicated by the incidents at OpenAI and Hugging Face. It is time to slow down, to ensure that AI development continues at a pace where our capacity for control can keep up with the advances in machine intelligence.
This is a difficult ask because the AI leaders confront a difficult coordination problem. The first lab to attain an intelligence explosion stands to gain untold wealth and power. Beware Silicon Valley investors who are decrying any attempt at slow down or regulation as paranoid socialism. They bet the farm on being the big winners in the AI race, and we should not ignore the clear warning of Hugging Face so that their bets can pay off.
Subscribe now
The bigger concern is China. A plan to slow down means that everyone needs to slow down, and that includes labs in China, which are behind the frontier, but not by much. The desire to continue to outpace Chinese labs has kept everyone in Silicon Valley going full steam ahead, but as risks become apparent, the need for international cooperation has now become urgent. I’ve seen a lot of people suggesting in the last few days that if President Trump is able to successfully negotiate an AI slowdown deal with China, he should get the Nobel Peace Prize. The man has desperately desired a Nobel Peace Prize, to an embarrassing extent. If he negotiates a Chinese AI slowdown, he’ll have richly earned it, and I say that as someone with absolutely no love for the man.
The biggest immediate need is convincing the Chinese government and AI labs that AI risk is real, and that international cooperation on AI development pace is essential to ensuring the future survival and flourishing of the human race. I’ll be doing my part; as an American professor working at a Chinese university, I’ll get this essay translated into Chinese in order to try to spread the word that these are concerns that are worth taking seriously at the highest level. I will, of course, be asking AI to do the translation. But I’ll have a human check the output. Some things are too important to give over entirely to the machines.
Matt Lutz is an Associate Professor of Philosophy at Wuhan University and writes the Substack Humean Beings.
Follow Persuasion on X, Instagram, LinkedIn, and YouTube to keep up with our latest articles, podcasts, and events, as well as updates from excellent writers across our network.
And, to receive pieces like this in your inbox and support our work, subscribe below:
Subscribe now
This article is brought to you by American Purpose, the magazine and community founded by Francis Fukuyama in 2020, which is proudly part of the Persuasion family.
In 40CE the Jewish philosopher Philo of Alexandria sailed for Rome on what he believed would be the culminating political mission of his life. He had been chosen to lead an embassy to the emperor Gaius (known today as “Caligula”) on behalf of one of the largest and most prosperous Jewish communities in the Roman Empire. The task at hand could not have been more important. Two years earlier Alexandria had exploded in anti-Jewish violence. Synagogues had been desecrated, homes looted, civic rights suspended, and hundreds of Jews subjected to murder and humiliation while the Roman prefect, Aulus Avilius Flaccus, failed (or refused) to restore order. Philo hoped the emperor would reverse these injustices and prevent their recurrence.
The embassy had a second purpose that went beyond Alexandria. Jewish communities across the empire were facing growing pressure to recognize the emperor’s divinity by allowing his image in their synagogues and worshiping him, acts forbidden by Jewish law. Philo set out to persuade Gaius that such demands misunderstood the basis of Jewish loyalty. Jews prayed for (rather than to) the emperor, obeyed Roman law, and paid Roman taxes. Their refusal to worship Gaius did not constitute an act of political rebellion but of religious devotion. The question placed before the emperor was therefore constitutional rather than religious: could the empire continue to accommodate diverse peoples by respecting their ancestral laws? Today, the issue might be seen as one of communal autonomy or even “internal sovereignty.”
Philo’s Embassy to Gaius conveys the tale of a man gradually discovering that political institutions can persist even after rulers have lost the capacity for the sustained attention upon which judgment depends.
The decision to undertake the embassy reflected assumptions about politics formed over a lifetime. Known today primarily as a philosopher who melded Jewish and Greek thought, Philo was no naïve idealist. He was a member of one of Alexandria’s wealthiest and best-connected families. His brother Alexander was among the empire’s richest financiers and cultivated close ties to the imperial household. Philo certainly understood from reading the Septuagint (the Greek translation of the Hebrew Bible) that governors could be corrupt, local politics were often vicious, and justice was almost always imperfect. But he also believed in the constitutional principles of the early Principate and especially that Rome could govern an extraordinarily diverse empire because emperors stood above local factions. Their legitimacy rested not only on military force, but also on their ability to examine competing claims and uphold the legal protections enabling the different ancestral groups within the empire to flourish.
Thus, the journey to Rome was more than a diplomatic mission. It also exemplified confidence in a particular ideal of government—one in which rulers could be persuaded and evidence mattered. Most crucially, political argument remained vital because emperors had the capacity for distinguishing public questions from private passions. As in today’s world, the only real “weapon” of the public intellectual, even one as prominent as Philo, was argument—the faith that knowledge, evidence, and reason can alter the views of those who wield power. But this role depends on a condition that intellectuals and diplomats themselves have no ability to supply: power must be willing to listen. Rulers may reject or misunderstand an argument, or they may choose interests over principle. These are part of ordinary political engagement. But what happens when power ceases to pay attention at all? What does the ambassador or public intellectual do when their arguments are sound but no one will listen? Philo went to Rome ready for the first problem but was confronted with the second.
When Philo finally obtained an audience with Gaius after months of waiting, he was prepared for complications: delays, interruptions, perhaps even hostility. But almost immediately, things went sideways in unexpected ways. The audience did not begin in a council chamber or a tribunal but with Gaius inspecting houses on the Palatine. Philo observes that, “while he was saying this he was going on with his survey of the houses ... and gave orders that they should be made more magnificent.” Philo’s first response is that the inspection is a temporary distraction before the embassy would really begin. After all, emperors love inspecting buildings, suggesting improvements, and supervising construction.
However, the embassy never really begins. Philo and his companions scamper along behind the emperor as he continues his inspection. They are tugged along through the emperor’s architectural tour. The image is genuinely absurd. The representatives of one of the oldest and largest Jewish communities in the Roman world (by some estimates, the Jewish population of Alexandria comprised twenty percent of the city’s 500,000 inhabitants) follows silently behind the emperor while he critiques floor plans and redesigns rooms. Philo remains unflappable and at this point still behaves as though ordinary politics governs the encounter. He waits, he answers the questions put to him. And he desperately looks for an opening through which the real business of the embassy might finally begin.
When that opening comes, he is blindsided. Gaius does not ask about Alexandria, the violence against the Jews, or what the Roman government should do, but instead asks, “Why do you refuse to eat pork?” Philo takes the question seriously, explaining that every nation has its own ancestral customs and that Rome’s flourishing is because it has generally respected the traditions of the many peoples under its rule. The strategy here is to recover the level of principle as a guide to the meeting. Philo does not lose his cool and he does not abandon reason. He answers as a philosopher and diplomat would answer, still believing, even now, that explanation can recover the conversation and that he is participating in an argument.
Persuasion is a registered nonprofit that relies on reader support to pay our staff and keep our content free for everyone. If you value our work and want to fight for liberal values wherever they are threatened, please become a paying subscriber today!
Instead, Gaius laughs, and so do Philo’s “opponents”—Alexandrian Greeks who seem to have gotten the emperor’s ear first. Gaius then temporarily tacks back to the serious “we want to hear what claims you make about your citizenship.” But—as a frustrated Philo tells us—Gaius “cut short our earlier points before we could bring in the stronger ones, and dashed at high speed into the large room of the house, and walked around it and ordered the windows all round to be restored with transparent stones … Then he advanced in a leisurely way and said in a more moderate tone: What is it that you say? And when we began on the points which came next in the thread of our argument, he ran again into another room and ordered original pictures to be put up there.”
Only later does Philo understand the significance of the laughter and Gaius’s erratic behavior. Looking back on the audience, he writes, “Then driven along we followed him up and down, mocked and reviled by our adversaries, as they do in the mimes at the theatres. For indeed the business was a sort of mime; the judge had taken on the role of accuser...” Even worse: “[W]ith the statement of our case thus mangled and disjointed, one may almost say cut short and crushed to pieces, we gave up, for there was no strength left in us.” Only gradually does Philo realize that argument itself is no longer setting the terms of the encounter.
Here we come to the turning point in The Embassy. Philo’s discovery went well beyond Gaius’s failure to buy his arguments. The emperor had become almost unreachable through the ordinary tools of intellectual and political persuasion. Philo is not only recounting an unpleasant audience; he also is trying to describe an experience for which he doesn’t have the language. The judge is no longer a judge but has become an actor in a performance, the embassy no longer presents a case but instead has become part of the spectacle, and, perhaps most crucially, the audience that Philo expected to culminate in judgment has become what he calls “a sort of mime.”
An earlier exchange, read against this backdrop, acquires new significance. When informed that sacrifices had regularly been offered for him by the Jews, Gaius replied, “All right, that is true, you have sacrificed, but to another, even if it was for me; what good is it then? For you have not sacrificed to me.” Previous emperors had accepted sacrifices offered on their behalf as expressions of fidelity to the larger imperial order. Gaius, by contrast, insisted that such loyalty was insufficient unless directed personally toward him. Philo, in quoting the emperor word for word, reveals the gravity of the situation.
By the time Gaius finally pronounced judgment, the outcome had become almost inevitable. The Jews, he declared, were “not so much wicked as unfortunate in not believing me to be a god.” The sentence is usually considered as evidence of Caligula’s megalomania. It may be, but that is not what seems to have struck Philo most deeply. What really overwhelmed Philo was the realization that an embassy convened to discuss civic violence, legal rights, and the obligations of the Roman government had ended by revolving almost entirely around the emperor himself. The political world that Philo thought he understood when he stepped off the boat had shifted beneath his feet.
Only after the audience had ended could he describe what he had experienced. It had been, he writes, “a combination of theatre and prison in place of a tribunal, theatre-like in the cackling of their hisses, their mockery and unbounded jeering, prison-like in the strokes inflicted on our flesh, the torture, the racking of the whole soul.” For Philo this passage denotes humiliation. But it also signals defeat.
The audience ended not just with defeat but with a retreat. “We had escaped and were just able to breathe again,” Philo recalls, “not because we clung to life and cringed from death ... but because we knew that we should prove to have thrown ourselves away for no useful purpose.” Philo now knew he had misunderstood the kind of political world into which he had entered. The rules of the game had changed. He had come to Rome believing that the emperor’s task was to judge and he got back on the boat wondering whether judgment itself had become impossible.
So what had Philo discovered? Since Plato, political thinkers had known that rulers can be vain. And it was hardly news that tyrants enjoy spectacle. His more novel insight concerned the role of attention in political life.
Philo’s Embassy continues to be relevant not because it offers a convenient historical analogy but because it forces us to think differently about political life. Here we come to Donald Trump (of course). As Mary Beard has recently reminded us, most comparisons between Donald Trump and figures from antiquity tend to focus on personality: was Gaius vain? Is Trump narcissistic? Such questions are tempting, but they ignore the more important issue—the conditions under which political judgment takes place.
Philo’s encounter with Gaius sheds a different light on our current predicament. Trump’s distinctive features are less about ideology or governing style than the tendency for public questions to become organized around the president himself. Choose any issue: immigration and trade, relations with allies and enemies, criminal investigations, personnel decisions, military operations, and even natural disasters—Trump addresses all of them through his own personal lens. We elect presidents to shape policy according to their own priorities. Reading Philo, however, forces us to question whether this approach can solve public problems or whether it is an opportunity for the ruler to gain more attention and acclaim.
We’re expanding our events offerings! Please check out our events page to join Book Club, Ask the Author, and Intellectual Bootcamp—and to watch recordings of recent events.
Persuasion events
Philo’s failed embassy shows what happens when an emissary cannot induce a ruler to listen to rational argument. In his first Oval Office meeting with President Trump, Canadian Prime Minister Mark Carney, however, provided an example of the opposite approach—a leader who had instinctively learned the lesson that Philo had not.
Carney arrived in Washington with a similar mandate to Philo’s: to push back on Trump’s repeated suggestion that Canada become the “fifty-first state,” while also attempting to resolve a punishing trade war. Meanwhile, his opponent was a president whose attention, by his own aides’ admission, can never focus on one topic for long. Like Gaius, Trump’s mind frequently turns to vanity projects—ballrooms, reflecting pools, big ships, and monuments.
Carney acted differently than Philo. He didn’t launch into a kind of principled, fact-based brief that characterized the annexation as illegal. Carney knew this wouldn’t work. Instead, he spoke to Trump in a way designed to grab Trump’s attention, labelling Trump “transformational” and lauding his focus on American workers. Carney then moved to the core issues. He framed his defense of Canadian sovereignty like a real estate man: “As you know from real estate, there are some places that are never for sale. We’re sitting in one right now,” he said, gesturing at the Oval Office itself, before adding that Canada, like the White House, “won’t be for sale, ever.” Trump, notably, did not laugh him out of the room the way Gaius mocked Philo; Trump replied “that’s true,” and though added later, more grudgingly, “never say never.”
The initial meeting was not an unambiguous triumph. The tariffs didn’t go away. Trump still speaks of Canada as the fifty-first state and calls Carney “governor.” And by August, the two countries stood before a trade war designed to break Canada. But, in stark contrast to Philo, Carney found a way to approach an attention-hungry ruler so he would actually listen, however briefly.
Political leaders and diplomats who have always operated in the world of personalist, attention-based politics know how to navigate this environment. They don’t need to retool or unlearn a system of arguments and evidence—because they never relied on one to begin with. By contrast, it is the traditional U.S. allies, particularly the Europeans, who resemble Philo most. Although things are beginning to change, European diplomats still arrive with legal briefs and appeals to shared principle, mistaking themselves for participants in a shared endeavor rather than part of a cast trailing behind the emperor. They are now learning what Carney already knew: the old game is over. The old folkways remain, but their function has shifted, and the tribunal—as Philo discovered nearly two thousand years ago—has given way to the theater.
Jeffrey Kopstein is Chancellor’s Professor of Political Science at the University of California, Irvine.
Follow Persuasion on X, Instagram, LinkedIn, and YouTube to keep up with our latest articles, podcasts, and events, as well as updates from excellent writers across our network.
And, to receive pieces like this in your inbox and support our work, subscribe below:
Subscribe now
From the publisher's feed