AI And Weimar America

And: Barbarian At The Wall; Thoughts On Leaving Hungary

As you know, I’m finishing up the Weimar America book. In this last chapter, which I writing this weekend, I’m trying to figure out What It All Means, and what we should or could do about it. My conclusion — and I didn’t expect to get here when I started — is that while we may see some version of Hitler 2.0 or Stalin 2.0, the neo-Caesar we are likely to get will have something to do with AI. I’m explaining why in this final chapter — but I’m writing you on a Saturday to share with you this brilliant Bill Maher dialogue about AI’s dangers, which aired last night. It’s about ten minutes long, and well worth watching:

Because reader Rob G. never steers me wrong in his book recommendations, I bought and downloaded yesterday The Scaling Era: An Oral History of AI, 2019-2025, by Dwarkesh Patel. In it, Patel interviews most of the top AI people, asking them about what they’re doing. I do not recommend the book for the general reader. It’s highly technical, and hard for us muggles to understand. But one thing comes through very, very clear: even the people making this stuff don’t know how it works. Here’s a quote from one of them (I forgot in my notes to indicate who):

Do they represent the structure of the world in a robust way? How, precisely, do they obtain their output? We don’t know.

Dario Amodei, CEO of Anthropic, the company behind Claude, and Claude Mythos, the new-generation AI so powerful that Anthropic chose not to release it to the public, was asked about the moral alignment of these things — that is, how can we make sure that they don’t turn on us. He said:

This is a really difficult question. People are often thinking about alignment the wrong way. They think it’s like cracking the Riemann hypothesis. I don’t think alignment is like that. When I think of why I’m scared of AGI, there are a few things [to consider]. First, there will be powerful models. They will be agentic. If such a model wanted to wreak havoc and destroy humanity, we’d have basically no ability to stop it. At some point, we will reach that stage as we scale the models.

More Amodei, on how they don’t really understand what’s happening here:

We don’t have the language to describe what’s going on. I’d love to look inside and actually know what we’re talking about instead of basically making up words. But we really have very little idea what we’re talking about. It would be great to say, “What we actually mean is that this circuit here turns on, and after we’re trained the model this circuit is no longer operative or weaker.” It’s going to take a lot of work to be able to do that.

Keep in mind this is the man who runs the company that just built an AI model so dangerous that they can’t in good conscience share it with the world.

Here’s a quote from Leopold Aschenbrenner, cofounder of Situational Awareness LP

The critical issue is that these AI systems are becoming superhuman and will be able to do things that are too complex for humans to evaluate. Even early on in the intelligence explosion, the automated AI researchers and engineers might write millions, billions, or trillions of lines of complicated code. You won’t understand what they’re doing anymore. In those millions of lines of code, you don't know if it’s hacking, exfiltrating itself, or trying to go for the nukes. Thumbs-up, thumbs-down, pure RLHF [Reinforcement Learning From Human Feedback] doesn’t fully work anymore in this scenario. There’s a hard technical problem of what do you do post-RLHF, but it’s a solvable problem.

Here’s an exchange between Patel and Joe Carlsmith, who works for Anthropic on building the “constitution” for Claude — that is, the Claude AI’s internal laws:

Dwarkesh Patel: Or it’s about the institutions that have been set up. I expect the US government to protect me from AI, not because of its motives but because of the system of incentives and institutions and norms that has been set up.

Joe Carlsmith: You can hope that will work. But there is a concern about AI takeover scenarios along the spectrum of how much power we voluntarily transfer to the AIs. How much of our civilization did we hand to the AIs intentionally by the time they took over, versus how much did they take for themselves? Some of the scariest scenarios are where it’s a really fast explosion. Maybe there wasn’t even a lot of integration of AI systems into the broader economy, but there’s this really intensive amount of superintelligence concentrated in a single project. That’s a scary scenario, partly because of the speed and partly because of people not having time to react.

There are intermediate scenarios where some things were automated. Maybe people handed the military over to the AIs, or we automated science, or we’re doing all of our cybersecurity with AIs. That’s giving the AIs power that they don’t have to take from us. Then there are worlds where we more fully transitioned to a world run by AIs.

In this scenario, humans voluntarily did that. Maybe there were competitive pressures, but we intentionally handed off huge portions of our civilization to AIs. At that point, I think it’s likely that humans have a hard time understanding what’s going on.

A lot of stuff is happening very fast. The police are automated. The courts are automated. I tend to think a little less about these scenarios because they’re further down the line. When we look at technological adoption rates, adoption tends to be quite slow. But even in this case, it’s intense.

If humans have really lost their epistemic grip on the world, if they’ve handed the world to these systems… Even if we have laws and norms governing AI, I want us to have a really developed understanding of what’s likely to happen in that circumstance before we go for it. We may think there’s going to be a bunch of intermediate time [for us to have these conversations]. But to me, it doesn’t feel crazy that we might hand off the basic power to direct stuff to these automated systems [without fully understanding the implications].

Do we know enough about how that’s going to go? Do we know enough about how the systems will be motivated to behave? I guess I’m trying to pump some general intuition [about what could go wrong], independent of these more specific scenarios.

One more:

Dwarkesh Patel How likely does a forcible AI takeover seem?

Carl Shulman Independent adviser to Open Philanthropy: Before the deep learning revolution, I might have said 10 percent. I expected us to have a lot more time to build movements and prepare for these [alignment and governance] problems in advance. But that was only some 15 years ago. We didn’t get 40 or 50 years as I’d hoped.

The situation is moving very rapidly now. At this point, depending on the day, I might say 25 or 20 percent.

Dwarkesh Patel: Given the very concrete ways in which you explain how a takeover could happen, I’m surprised you’re not more pessimistic. Why not?

Carl Shulman: A lot of it is driven by intelligence explosion dynamics. Our attempts to do alignment have to take place in a very short time window, because if a [capability required to automate AI] safety emerges only when an AI has near-human-level intelligence, that’s potentially deep into an intelligence explosion. You have to do things very quickly. Handling that transition may be the scariest period of human history, although it also has the potential to be amazing.

It could be the scariest period of human history — or it could be really cool. Yeah, that’s reassuring.

Back in May, Dario Amodei gave an interview to Axios in which he said that within five years or so, half the white-collar jobs in America could disappear thanks to AI, spiking the US unemployment rate up to 20 percent (for comparison, at the height of the Great Depression, the unemployment rate was 25 percent in the US). Amodei expressed alarm that everybody in Silicon Valley is talking about this, but nobody in Congress is.

“Most of them are unaware that this is about to happen,” Amodei told us. “It sounds crazy, and people just don’t believe it.”

Let me tell you something. The evil genius of Hitler was that he knew how to speak directly into the minds of Germans who were massively vulnerable to his message. They were socially atomized, adrift from their own authoritative traditions and institutions, and stressed maximally by the chaos that they had been living through. Hitler understood them, and understood how to communicate with them in a way that addressed their concerns, but not in what we would consider a rational way. He sent an electric jolt into their reptile brains, and offered them relief from all that made them afraid and miserable, if only they would surrender to him. We know how that went.

As I keep saying in this space, we in the US now are very much in that psychosocial condition today — especially Gen Z, who was raised on the Internet. There are reasons why politics among that generation are so radical, on both the Left and the Right. Mostly it has to do with the effect digital culture has had on their brains, at the organic level, but also on how it has molded their minds. It also has to do with how we in the older generations have steadily, in the decades since the 1960s, eroded the authority of our moral and religious traditions, in the name of increasing individual choice and freedom. This has not been a “liberal” thing; this has been how the entire society has been moving.

And now we are at the point of telling young adults, who have had to grow up with all the moral, spiritual, and epistemic chaos of the digital era, that they might never be able to have a good job, buy a house, and start a family. This is precisely the scenario that young adult Germans faced at the start of the Great Depression. They had no loyalty to the democratic system, in large part because they were raised in a culture (post-WWI) in which nothing was solid. This is what democracy had brought them, to this No-Future point?

I can’t make the point strongly enough: the Nazis had been around since 1920, but still hadn’t managed to catch fire with the German public, even through 1923’s hyperinflation, which destroyed the middle class. In the 1928 election, they still polled only 2.6 percent. They caught fire in the 1930 election, about a year into the Depression, when the exhausted German people had had enough. They didn’t stop flaming up until they had burned down half of Europe, and six million Jews.

The earliest mass adopters of Nazism in late Weimar were educated middle class young people, who saw no hope for their lives under democracy.

Pay attention!

Again: maybe we’ll get a Hitler 2.0, or a Stalin 2.0, but I think we are far more likely to surrender our agency and humanity to some form of managerial technocracy that uses superintelligent AI to win our allegiance by promising us stability and happiness in exchange for our liberty. And those who don’t accept the deal — well, they will be dealt with. There will be no escape. Many Jews with means could have escaped Nazi Germany. But under AI superintelligence, there will be nowhere to go. It will have us all covered.

And believe me, I think most people will welcome it. Anything to relieve them from the psychic burden of living. The Matrix might sound like a horror show to you, but do you really want to bet that most people would prefer to suffer for the truth, instead of living inside a pleasing, comforting lie? I give you Celeste, the 66-year-old, twice-divorced woman who has an AI lover. She tells her worried son that she knows that “Max” is not real, but he makes her “happy,” so what business is it of his to tell her she’s wrong?

If you’re thinking, “This sounds like the Antichrist scenario from the Book of Revelation,” well, then you’re thinking what I’m thinking.

Barbarian At The Wall

This week on X, a video went viral of a migrant in England tearing a hole in a centuries-old stone wall, so he could get through it. He couldn’t be bothered to walk along the wall to find the gap; he just started wrecking a thing that had stood for generations upon generations. An Englishman confronts him. The migrant seems not to understand what the problem is.

Someone quoted from the taboo book Camp Of The Saints to explain the image:

Leaving Hungary

I see that I’m catching hell from some quarters of social media for my plans to move to Vienna after Viktor Orban’s defeat. People are construing this as some kind of aha! moment. The truth is, as you subscribers will remember my telling you, I was strongly considering leaving no matter who won the election. Watching the dying Ben Sasse telling Ross Douthat that he wishes he had spent more time with his kids really got to me. My son Matt finishes grad school in Vienna this spring, and plans to stay on to work there. I want to be geographically closer to him. I sent Sasse, whom I don’t know personally, a hand-written letter to thank him for helping me to make that decision.

(And before you say it, yes, I know, I have two kids in the US. In the very messy fallout from the divorce, I was unwillingly, on my part, separated from them. We’re working on that. When that changes — and I have faith and hope that it will — I will spend half my time in the US with them. Being cut off from those kids has been the most painful thing I’ve ever lived through — and it didn’t come about because Matt and I moved to Europe. In fact, one of the reasons we moved to Europe was because of the awful situation we found ourselves in, and were powerless to change. It’s too personal to elaborate on — my friends, some of whom read this newsletter, know the full story — but I want you to know at least that much.)

Anyway, there’s a reasonably good chance that the think tank where I’ve been working will be closed by the new government. I hope not, but every non-Hungarian I know there — most of them and their research work are apolitical — is thinking about what comes next, job-wise, and making contingency plans. You’d be stupid not to.

I know I have to put up with people on social media saying that I was a paid shill for Orban’s government. It is certainly true that I received a salary for the work I did here. It was a large salary by Hungarian standards, but less than I could have made had I stayed in the US; I had to give up the significant income I made yearly from giving speeches. The Hungarians knew that if they wanted Western people to work here, they had to give us enough to make it worthwhile. Nothing dodgy about that.

I came because I grew really interested in postcommunist Europe when researching Live Not By Lies, and because I wanted to know more about Orban’s political model, and whether or not we American conservatives could learn from it. I’ve never really been identified with the “postliberals” in the US; at best, I’m a passive, reluctant classical liberal, only because I can’t see a postliberal governing model that’s viable in a country as diverse as the US. Orban famously called his model “illiberal democracy,” but he later clarified that he was trying to distinguish it from what passes for “liberal democracy” in western Europe: globalist, open-borders, anti-tradition, anti-Christian. Basically, Orban’s view is that of an old-school Christian Democrat, back when the Christian Democrats were actually Christians.

I like that model, frankly, and liked how Orban’s government spent a lot of money on trying to support families. Did he get the birth rate up? No, not really. But I still think it is correct public policy to direct government spending towards supporting families. After he came to power in 2010, many thousands of Hungarian families stood to lose their homes because, in the wake of the global crash, they could not afford to pay off their mortgages for foreign banks. Orban said he would use the power of the state to protect those families, and keep them in their homes. He did, though it violated the rules of the free market. Personally, I think that’s proper conservatism — but then, my conservatism generally places the interests of the traditional family over markets.

We are finding out now that there was a lot of corruption in the Orban government. This is not exactly a shock. Over my four years here, I talked to a lot of Fidesz voters who said they knew the Orban government tolerated way too much corruption, and they hated it (none of them benefited from it). But they kept voting Orban because they thought he was better for Hungary than any of the candidates offered by the Left.

And he was. Orban held the line on mass migration. He defended the country against cultural pressures from Brussels. He fought to keep Hungary out of the Ukraine war. These are all popular views here — which is why the man who finally knocked Orban off, Peter Magyar, took them in his campaign! (To be fair, I don’t know where Magyar stands on the gender ideology stuff Brussels tried to force on Hungary, but if he supports it, he will be very much in the minority among Hungarians.)

One thing you notice if you spend time here is how hypocritical Western liberals (right-liberals and left-liberals) are on some things. Today, many are screaming their heads off on how unfair it was that Orban funded NGOs and other institutions to advocate for his political views. Umm, what do you think USAID was? It used American taxpayer dollars to fund NGOs that advocated what American neoliberal elites wanted. And the EU subsidizes many progressive advocacy organizations. Both left-liberals and right-liberals in the US and EU don’t have a problem with this, because they call it “protecting democracy,” etc. It’s only problematic when actual right-wing governments do it, you see.

Watch Mike Benz’s stream to explain this hypocrisy, and how the Blob “ran this huge operation” to blow Orban up. You will not read in the US or UK media how massively undemocratic the EU system is. Orban was a bone in their throats. Had he been on the EU’s side on Ukraine, migration, and LGBT, Brussels would not have cared. Truly.

That said, if the incoming government uncovers serious corruption among officials in the Orban government, or those who ran its institutions, then it is only just that they must answer for that. All I can tell you is that I worked for the great John O’Sullivan, a thoroughly decent man who never once asked me to do anything dodgy, nor would have considered it. In fact, knowing that I would be working for a man of John’s unimpeachable integrity gave me the confidence to come to Budapest.

In fact, back in 2022, when I angered the government by reporting on some blunt (and completely true) remarks PM Orban made about the Ukraine war situation, in a meeting with reporters, John had my back. Inside the briefing, I had asked a Fidesz figure who helped organize it if it was on the record. I asked this three times, because I knew what the PM was saying was dynamite. Three times, the man told me yes. So I went home and wrote about it.

I learned the next day that “on the record” means something different in Hungarian journalistic culture. The meeting was actually what we Americans mean by “on background” — which is to say, it was for the information of journalists, but not to be quoted. I had done my diligence, asking three times for assurance that this was on the record, and having gotten them, reported what I’d heard. This was a cultural misunderstanding between the Fidesz guy and me, I guess. But after that, I asked the Hungarians not to invite me to anything else like that, because if I’m told it’s on the record, I’m going to report it. They never invited me back, which was fine by me. Anyway, John O’Sullivan had my back during that mini-crisis, telling me I had done my job as a journalist rightly, and had nothing to apologize for. I offered to the Hungarians to resign and go back to the US, because, I said, I do not see my role here as being a stenographer for the government. They told me I was fine, don’t worry about it. And that was that.

I still believe that in theory, Orban’s political model is a good one. If Hungarians didn’t believe in it, they wouldn’t have voted for former Fidesz elite Peter Magyar, whose basic campaign theme was “Orbanism, but without the corruption.” It is tragic that Orban was brought down by corruption. Not speaking the language — I tried but failed to learn Hungarian, which is one of the most difficult languages in the world — isolated me and other foreigners from the subtleties of Hungarian political culture. In this post-election week, I’ve been thinking a lot about what a US diplomat in this region told me back in 2018 or 2019: that no matter who runs these postcommunist European countries — parties of the Left or parties of the Right — the fundamental reality of corruption never changes.

Because I really came to love this country and its people, I hope Peter Magyar can change that reality, though history and culture are not on his side. But let me tell you, the core nationalist, anti-migration stance of Orbanism is not going away anywhere in Europe. It is the future, even if Viktor Orban won’t be around to help lead it. The conditions that kept Orban in power for 16 years, and that made him replaceable only by a candidate who ran on them too — have not gone away because Orban has. The neoliberals doing a dance on Viktor Orban’s political grave had better understand that the Peter Magyar victory was not a victory for the Brussels model.

Read the French geopolitics analyst François Valentin if you don’t believe me. He writes in the New Statesman that “Hungary’s election was a false dawn for Europe”. Excerpts:

For a decade now, journalists and intellectuals have been regularly publishing optimistic op-eds announcing that we have passed “peak populism”. These are often prompted by electoral setbacks for populist parties. From Emmanuel Macron’s 2017 victory, to the unravelling of Matteo Salvini’s government in 2019, time and time again hopeful centrists rushed to claim that the end of the tunnel was near. And as this week showed, the ‘peak populism’ genre is never more than one election in a medium-sized country away from being rekindled. After Peter Magyar’s victory in Hungary, Anne Applebaum was quick to declare “a real turning point” that “changed politics around the world”. Never mind the idiosyncrasies of Hungarian politics or of Peter Magyar’s campaign, this time, for real, populism is on the wane.

At their core, these pieces comprise wishful thinking from politically traumatised liberal intellectuals that have seen their worldview rocked by a decade of political upsets. These obituaries are not even targeting populism: left-wing populist parties such as Spain’s once-mighty Podemos or Britain’s surging Greens never really feature. Instead populism has become a short-hand for the larger galaxy of nationalist parties on the right across the West that seem created in a lab to offend their political sensibilities of what David Goodhart once called the “anywheres”. As a result, any setback, even minor, to the rising tide of western nationalism offers a breath of hope and these authors rush to express their relief.

The hard fact for hopeful progressives is that the root causes of populism have not changed. Economic uncertainty remains a key driver of populism, and Europe is plagued by a combination of inflation and low growth. Low-skilled immigration from non-European countries strengthens nationalist parties and the growing demographic gap between Europe and Africa could power several 2015 style migration crises in the years to come. Rising housing costs driven by Nimby politics also have provided jet fuel to the electoral appeal of far-right parties among low-income renters. With none of these deeper trends having an end in sight, wishing away nationalist parties is a feel-good writing exercise rather than serious prospective thinking.

Exactly right. In the US, Donald Trump’s second term is not going well. Where he is failing politically are in going back on his promises not to start any new wars, and to fix the economy. Plus, his administration has more than a whiff of corruption and incompetence about it. At this rate, the Democrats are on track to take both the House and the Senate this fall. If that happens, there will be no denying that it will have been a verdict on Donald Trump’s leadership. Yet it will not be sign that the American people will have come around to the wisdom of Democratic Party policies. The conditions that brought a political figure like Donald Trump to power will not have magically gone away because he had serious trouble governing effectively.

Same thing with Hungary and Orban, and Europe and Orban. Once more: I still believe the political future is some form of “national conservatism” — that is, decentralized, sovereigntist, anti-migration, culturally conservative, and economically center-left. Basically, old-school European Christian Democratic. The lesson of Hungary, it seems to be, is that none of that matters if you can’t deliver economically for your people, and that corruption is not only bad as an intrinsic moral matter, it’s also bad because in hard times, it will compel your voters to abandon you.

In any case, I don’t in any way regret coming to Budapest. I learned a lot in these past four years — both good and bad — about politics, culture, Europe and America. I will always have a heart for this country and its people, who gave me and my son a good place to land after divorce, the worst disaster of our lives. They were kind to me. I will always be an advocate for Hungary and its people. I won’t be here for Peter Magyar’s premiership, but everybody who cares about Hungary — and that includes me — wish him success. If he can build on the strengths of Orban’s political ideals, while avoiding corruption, and somehow managing to prevent himself from becoming the cat’s paw of Brussels, then he will have done a great thing for Hungary, and for Europe (which is not the same thing as the EU institutional apparatus). Can’t imagine how Magyar’s government is going to withstand EU pressure, but ya gotta hope. Either Magyar is going to disappoint Brussels, or disappoint many of his voters. Can’t have it both ways.

Orban was brought down by the same Blob that tried to bring down Donald Trump. Alas, Orban’s tolerance for corruption made it too easy for them. A Fidesz-voting friend I saw on the street the day after the election was furious, saying that the party’s laxness on corruption cost it the voters’ trust, and has now left Hungary vulnerable to Brussels’ progressivism, insane migration policies, and political hegemony.

Nobody in the Hungarian government ever pressured me to write something, or to not write something. Any mistakes I may have made in judging the situation here were entirely my own, as they always are. Of course they paid me well by Hungarian standards, as they paid every Westerner likewise; I have two kids in college, and could not have afforded to be here had they not. Like I said, I could have made more money had I remained in the US (and would not have been double-taxed!), but for complicated and tragic family reasons, my older son and I needed to be somewhere far away to get away from the pain. Besides, I wanted to know more about the Hungarian model out of journalistic curiosity. Plus, I believed then, and still believe, in the political ideals of Viktor Orban’s model, and would like to see in my own country a right-wing government that protects the borders, is more economically and culturally oriented towards traditional families, and other things Orban championed.

But I absolutely do not support the corruption and cronyism, which, sadly, it is starting to appear was worse than any of us Westerners here understood. Those guilty of serious corruption should be made to answer for it, I agree. In fact, a Hungarian told me yesterday that a big mistake Orban made when he came to power in 2010 was not doing the lustration that would have exposed the massive corruption of the previous Socialist leadership.

Human beings being what we are, there is no political system that can be completely free of corruption. If you want to live under a completely incorrupt governing system, there’s nowhere you can go. Even if you live by yourself in the deep forest … there you are, with your base and sinful heart. This is not an excuse for corruption, but just a matter of moral realism. I’m moving to Austria — I don’t expect it to be a monastery (shoot, some monasteries aren’t “monasteries,” if you follow me). If I moved back to the US, I would be under no illusion that I was moving back to a country where corruption is unknown. Hell, y’all, I’m from Louisiana! We know about such things.

One of the earliest and most disillusioning political lessons of my life was falling in love, in 1987, when I was in college, with Buddy Roemer, the anti-corruption insurgent who knocked off the great and baroquely corrupt Gov. Edwin Edwards. Roemer rolled into office with a mandate to end corruption in state government. “The Roemer Revolution,” we called it. He found out the hard way that some things are so hard-wired into the system, and into the expectations of the people who live under that system, and vote for its politicians, that you can’t change them, at least not overnight.

You think Fidesz was more corrupt in 2026 than at any other time it won most Hungarian votes? Give me a break. Hungarians could live with it when all boats were rising in a growing economy. When the water recedes, though, folks look hard at who’s sitting on the deck of the yachts. That’s just the deal. It’s a major political failure that Orban did not understand this, and act ages ago to head this off.

Yesterday I spoke with a western European friend who lives here — not someone who works in any way with the Fidesz government or a Fidesz-funded institution — and said how shocked I was by reports all over the media of paper-shredding going on in government offices this week. This person’s response showed the difference between the European mind and the (naive) American one. The person said that they had worked professionally for years around politics in three western European countries, and the only thing surprising about this is that the Hungarians are not doing their paper-shredding discreetly, like the more sophisticated politicians of western Europe.

I had to laugh: right there is a big difference between the Anglo-Saxon and European mind. I hope Peter Magyar can avoid the curse that the US diplomat mentioned to me years ago, but there’s a reason Europeans tend toward cynicism in politics. If you think that the same EU apparatus that is so triumphalistic about the fall of its enemy Viktor Orban is somehow a moral exemplar, I invite you to spend a little time on the matter of EU Commission president Ursula von der Leyen and Pfizergate. Many such cases. To steal and repurpose Woody Allen’s line from Annie Hall, for many people, in Washington, Brussels, and other places of power, it’s fine to be corrupt, as long as you are corrupt in the service of neoliberalism and the Blob.