Astral Codex Ten

Viktor Orban

20 min read

Viktor Orban

Article

Viktor Orban is a recurring person in the Astral Codex Ten archive, appearing 6 times across 6 issues between November 04, 2021 and September 18, 2025. The archive places it in contexts such as “Viktor Orban founded an extra-curricular society at his college called The Alliance Of Young Democrats”; “the pig is Viktor Orban”; “So Viktor Orban got everyone from his liberal democratic party together”. It most often appears alongside Middle East, Angela Merkel, China.

Metadata

  • Category: People
  • Mention count: 6
  • Issue count: 6
  • First seen: November 04, 2021
  • Last seen: September 18, 2025

Appears In

Source Context

Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.

Dictator Book Club: Orban
November 04, 2021 · Original source
Orban: Europe's New Strongman and Orbanland, my two sources for this installment of our Dictator Book Club, tell the story of a man who spent the last eleven years taking over Hungary and distributing it to guys he knew in college. Janos Ader, President of Hungary. Laszlo Kover, Speaker of the National Assembly. Joszef Szajer, drafter of the Hungarian constitution. All of them have something in common: they were Viktor Orban's college chums. Gabor Fodor, former Minister of Education, and Lajos Simicska, former media baron, were both literally his roommates. The rank order of how rich and powerful you are in today’s Hungary, and the rank order of how close you sat to Viktor Orban in the cafeteria of Istvan Bibo College, are more similar than anyone has a right to expect.
Inline links: Orban: Europe's New Strongman, Orbanland
Our story begins on March 30 1988, when young Viktor Orban founded an extra-curricular society at his college called The Alliance Of Young Democrats (Hungarian abbreviation: FiDeSz). Thirty-seven students met in a college common room and agreed to start a youth organization. Orban's two roommates were there, along with a couple of other guys they knew. Orban gave the pitch: the Soviet Union was crumbling. A potential post-Soviet Hungary would need fresh blood, new politicians who could navigate the democratic environment. They could get in on the ground floor.
Still, there was something about him. To call it "a competitive streak" would be an understatement. He loved fighting. The dirtier, the better. He had been kicked out of school after school for violent behavior as a child. As a teen, he'd gone into football, and despite having little natural talent he'd worked his way up to the semi-professional leagues through sheer practice and determination. During his mandatory military service, he'd beaten up one of his commanding officers. Throughout his life, people would keep underestimating how long, how dirty, and how intensely he was willing to fight for something he wanted. In the proverb "never mud-wrestle a pig, you'll both get dirty but the pig will like it", the pig is Viktor Orban.
Highlights From The Comments On Orban
November 11, 2021 · Original source
There were a lot of conservatives in the comment who made basically this point, or who compared various things Viktor Orban did to various things Joe Biden (actually or imaginarily) did.
Another is to be Viktor Orban. Go against elite opinion, and when the elites try to stop you, crush them. Crush the judiciary and replace it with your college friends. Crush the media and replace it with your college friends. Crush the intelligentsia and replace them with your college friends. Then do whatever you want, and the judiciary, media, and intelligentsia will take your side!
Viktor Orban is, alas, not George Washington. He became powerful enough to get whatever policies he wanted (and if you’re a conservative, you can really appreciate the policies he enacted), but then used the power to make it really hard for the voters to remove him.
Who Gets Self-Determination?
March 30, 2022 · Original source
AP has the take that Visegrad shows the way. Integrating with the West to enjoy its security guarantees and material benefits, but developing your own civilization instead of destroying it. Press X for doubt. Viktor Orban might go down in the next election, and Polish conservatives appear to be doubling down on all of the dumbest mistakes of American Republicans.
Deceptively Aligned Mesa-Optimizers: It's Not Funny If I Have To Explain It
April 11, 2022 · Original source
Prosaic alignment is hard… “Prosaic alignment” (see this article for more) means alignment of normal AIs like the ones we use today. For a while, people thought those AIs couldn’t reach dangerous levels, and that AIs that reached dangerous levels would have so many exotic new discoveries that we couldn’t even begin to speculate on what they would be like or how to align them. After GPT-2, DALL-E, and the rest, alignment researchers got more concerned that AIs kind of like current models could be dangerous. Prosaic alignment - trying to align AIs like the ones we have now - has become the dominant (though not unchallenged) paradigm in alignment research. “Prosaic” doesn’t necessarily mean the AI cannot write poetry; see Gwern’s AI generated poetry for examples. … because OOD behavior is unpredictable “OOD” stands for “out of distribution”. All AIs are trained in a certain environment. Then they get deployed in some other environment. If it’s like the training environment, presumably their training is pretty relevant and helpful. If it’s not like the training environment, anything can happen. Returning to our stock example, the “training environment” where evolution designed humans didn’t involve contraceptives. In that environment, the base optimizer’s goal (pass on genes) and the mesa-optimizer’s goal (get genital friction) were very well-aligned - doing one often led to the other - so there wasn’t much pressure on evolution to look for a better proxy. Then 1957, boom, the FDA approves the oral contraceptive pill, and suddenly the deployment environment looks really really different from the training environment and the proxy collapses so humiliatingly that people start doing crazy things like electing Viktor Orban prime minister. So: suppose we train a robot to pick strawberries. We let it flail around in a strawberry patch, and reinforce it whenever strawberries end up in a bucket. Eventually it learns to pick strawberries very well indeed. But maybe all the training was done on a sunny day. And maybe what it actually learned was to identify the metal bucket by the way it gleamed in the sunlight. Later we ask it to pick strawberries in the evening, where a local streetlight is the brightest thing around, and it throws the strawberries at the streetlight instead. So fine. We train it in a variety of different lighting conditions, until we’re sure that, no matter what the lighting situation, the strawberries go in the bucket. Then one day someone with a big bulbous red nose wanders on to the field, and the robot tears his nose off and pulls it into the bucket. If only there had been someone with a nose that big and red in the training distribution, so we could have told it not to do that! The point is, just because it’s learned “strawberries into bucket” in one environment, doesn’t mean it’s safe or effective in another. And we can never be sure we’ve caught all the ways the environment can vary. …and deception is more dangerous than Goodharting. To “Goodhart” is to take advantage of Goodhart’s Law: to follow the letter of your reward function, rather than the spirit. The ordinary-life equivalent is “teaching to the test”. The system’s programmers (eg the Department of Education) have an objective (children should learn). They delegate that objective to mesa-optimizers (the teachers) via a proxy objective (children should do well on the standardized test) and a correlated reward function (teachers get paid more if their students get higher test scores). The teachers can either pursue the base objective for less reward (teach children useful skills), or pursue their mesa-level objective for more reward (teach them how to do well on the test). An alignment failure! This sucks, but it’s a bounded problem. We already know that some teachers teach to the test, and the Department of Education has accepted this as a reasonable cost of having the incentive system at all. We might imagine our strawberry-picker cutting strawberries into little pieces, so that it counts as having picked more strawberries. Again, it sucks, but once a programmer notices it can be fixed pretty quickly (as long as the AI is still weak and under control). What about deception? Suppose the strawberry-picker happens to land on some goal function other than the intended one. Maybe, as before, it wants to toss strawberries at light sources, in a way that works when the nearest light source is a metal bucket, but fails when it’s a streetlight. Our programmers are (somewhat) smart and careful, so during training, they test it at night, next to a streetlight. What happens? If it’s just a dumb collection of reflexes trained by gradient descent, it throws the strawberry at the streetlight and this is easily caught and fixed. If it’s a very smart mesa-optimizer, it might think “If I throw the strawberry at the streetlight, I will be caught and trained to have different goals. This totally fails to achieve my goal of having strawberries near light sources. So throwing the strawberry at the light source this time, in the training environment, fails to achieve my overall goal of having strawberries thrown at light sources in general. I’ll do what the humans want - put the strawberry in the bucket - for now.” So it puts the strawberry in the bucket and doesn’t get caught. Then, as soon as the humans stop looking, it throws strawberries at the streetlight again. Deception is more dangerous than Goodharting because Goodharting will get caught and trained away, and deception might not. I might not be explaining this well, see also Deceptively Aligned Mesa-Optimizers? It’s More Likely Than You Think: We prevent OOD behavior by detecting OOD and obtaining more human labels when we detect it… If you’re (somewhat) careful, you can run your strawberry-picking AI at night, see it throw strawberries at streetlights, and train it out of this behavior (ie have a human programmer label it “bad” so the AI gradient-descends away from it) …and we eliminate the incentive for deception by ensuring that the base optimizer is myopic A myopic optimizer is one that reinforces programs based only on their performance within a short time horizon. So for example, the outside gradient descent loop might grade a strawberry picker only on how well it did picking strawberries for the first hour it was deployed. If this worked perfectly, it would create an optimizer with a short time horizon. When it considered deceiving its programmers in order to get a payoff a few days later when they stopped watching it, it wouldn’t bother, since a few days later is outside the time horizon. …and implements a decision theory incapable of acausal trade. You don’t want to know about this one, really. Just pretend it never mentioned this, sorry for the inconvenience. There are deceptively-aligned non-myopic mesa-optimizers even for a myopic base objective. Even if the base optimizer is myopic, the mesa-optimizer might not be. Evolution designed humans myopically, in the sense that we live some number of years, and nothing that happens after that can reward or punish us further. But we still “build for posterity” anyway, presumably as a spandrel of having working planning software at all. Infinite optimization power might be able to evolve this out of us, but infinite optimization power could do lots of stuff, and real evolution remains stubbornly finite. Maybe it would be helpful if we could make the mesa-optimizer itself myopic (though this would severely limit its utility). But so far there is no way to make a mesa-optimizer anything. You just run the gradient descent and cross your fingers. The most likely outcome: you run myopic gradient descent to create a strawberry picker. It creates a mesa-optimizer with some kind of proxy goal which corresponds very well to strawberry picking in the training optimization, like flinging red things at lights (realistically it will be weirder and more exotic than this). The mesa-optimizer is not incentivized to think about anything more than an hour out, but does so anyway, for the same reason I’m not incentivized to speculate about the far future but I’m doing so anyway. While speculating about the far future, it realizes that failing to pick strawberries correctly now will thwart its goal of throwing red things at light sources later. It picks strawberries correctly in the training distribution, and then, when training is over and nobody is watching, throws strawberries at streetlights. (Then it realizes it could throw lots more red things at light sources if it was more powerful, achieves superintelligence somehow, and converts the mass of the Earth into red things it can throw at the sun. The end.) III. You’re still here? But we already finished explaining the meme! Okay, fine. Is any of this relevant to the real world? As far as we know, there are no existing full mesa-optimizers. AlphaGo is kind of a mesa-optimizer. You could approximate it as a gradient descent loop creating a good-Go-move optimizer. But this would only be an approximation: DeepMind hard-coded some parts of AlphaGo, then gradient-descended other parts. Its objective function is “win games of Go”, which is hard-coded and pretty clear. Whether or not you choose to call it a mesa-optimizer, it’s not a very scary one. Will we get scary mesa-optimizers in the future? This ties into one of the longest-running debates in AI alignment - see eg my review of Reframing Superintelligence, or the Eliezer Yudkowsky/Richard Ngo dialogue. Optimists say: “Since a goal-seeking AI might kill everyone, I would simply not create one”. They speculate about mechanical/instinctual superintelligences that would be comparatively easy to align, and might help us figure out how to deal with their scarier cousins. But the mesa-optimizer literature argues: we have limited to no control over what kind of AIs we get. We can hope and pray for mechanical instinctual AIs all we want. We can avoid specifically designing goal-seeking AIs. But really, all we’re doing here is setting up a gradient descent loop and pressing ‘go’. Then the loop evolves whatever kind of AI best minimizes our loss function. Will that be a mesa-optimizer? Well, I benefit from considering my actions and then choosing the one that best achieves my goal. Do you benefit from this? It sure does seem like this helps in a broad class of situations. So it would be surprising if planning agents weren’t an effective AI design. And if they are, we should expect gradient descent to stumble across them eventually. This is the scenario that a lot of AI alignment research focuses on. When we create the first true planning agent - on purpose or by accident - the process will probably start with us running a gradient descent loop with some objective function. That will produce a mesa-optimizer with some other, potentially different, objective function. Making sure you actually like the objective function that you gave the original gradient descent loop on purpose is called outer alignment. Carrying that objective function over to the mesa-optimizer you actually get is called inner alignment. Outer alignment problems tend to sound like Sorcerer’s Apprentice. We tell the AI to pick strawberries, but we forgot to include caveats and stop signals. The AI becomes superintelligent and converts the whole world into strawberries so it can pick as many as possible. Inner alignment problems tend to sound like the AI tiling the universe with some crazy thing which, to humans, might not look like picking strawberries at all, even though in the AI’s exotic ontology it served as some useful proxy for strawberries in the training distribution. My stand-in for this is “converts the whole world into red things and throws them into the sun”, but whatever the AI that kills us really does will probably be weirder than that. They’re not ironic Sorcerer’s Apprentice-style comeuppance. They’re just “what?” If you wrote a book about a wizard who created a strawberry-picking golem, and it converted the entire earth into ferrous microspheres and hurled them into the sun, it wouldn’t become iconic the way Sorcerer’s Apprentice did. Inner alignment problems happen “first”, so we won’t even make it to the good-story outer alignment kind unless we solve a lot of issues we don’t currently know how to solve. For more information, you can read: Rob Miles’ video above, direct link here, channel here.
Inline links: this article, Gwern’s AI generated poetry, electing Viktor Orban prime minister, Goodhart’s Law, Deceptively Aligned Mesa-Optimizers? It’s More Likely Than You Think, my review of, Eliezer Yudkowsky/Richard Ngo dialogue, here, here
Book Review: Cyropaedia
January 10, 2024 · Original source
Along with teaching self-control, the Persian education system turned Cyrus and his fellows into a band of close friends who trusted each other with their lives. This was a big advantage in a world where everyone else was betraying each other. Cyrus made his school friends into officers and satraps, and was always confident in their competence and loyalty. This sounds similar to the story of Alexander the Great and his generals, making me think this really was a big advantage in those days. It is cliques of schoolboys who conquer the world (for a more modern version of this trope, see Viktor Orban).
Inline links: Viktor Orban
Defining Defending Democracy: Contra The Election Winner Argument
September 18, 2025 · Original source
Sources: Babylon Bee (yes I know it’s satire; notice the direction), Spiked, WSJ, MacIver Institute The most common response is to say that fine, democracy is about who wins votes, but we also like liberalism, liberalism is under threat, it’s too hard to talk about “liberalism” because in the US it sometimes means being left-wing, and so we use the related concept “democracy” as a stand-in. This is reasonable, and some accused-democracy-destroyers like Viktor Orban even accept it for themselves, calling their brand of government “illiberal democracy”. But I think there’s an even stronger response that doesn’t require admitting to a bait-and-switch: democracy isn’t just about having an election. It’s about having more than one election. Imagine a system where the winner of a fair election gets unlimited authority during his term. What forces this person to ever hold another fair election? Why can’t he ban the media from reporting on his missteps? Or confiscate opposition parties’ treasuries? Or order the police to murder any candidate who runs against him? The preparations for the next election, and the election itself, occur while it is still his term; if he can do whatever he wants during his term, there is nothing guaranteeing a fair election besides his personal goodwill. When we adjust for this - when we consider how to accord a leader enough power to do anything except rig the next election in his favor - we find that this is such a hard problem that it already requires most of the checks, balances, and civil society that we call liberalism. For example, the simplest way to win an election is to murder opposing candidates. We cannot merely constitutionally ban the leader from murdering people; if the leader controls the judiciary, he can pack it with sympathetic judges who will find him innocent of murder even when he does it in broad daylight (for some reason, no Russian judge has ever convicted Vladmir Putin of any of the assassinations that so many Western sources are sure he committed). So in order to give teeth to even the most basic ban on murdering rival candidates, you need an independent judiciary. (and although having “unelected bureaucrats” sounds bad, it’s important that these people not be directly elected at exactly the same time as the leader, because if the same electorate that puts the leader in power puts the checks on the leader in power, they’re likely to come from the same party. In the US, we solve this in a variety of ways, especially by staggering appointments - some officials are appointed by the previous leader, or the one before that.) But an independent judiciary is useless if the leader can ignore it without penalty. And the penalty cannot be purely legal, because legal penalties are levied by a judiciary, ie the organ that such a leader is ignoring. So this penalty must bottom out in extra-legal consequences: either the public relations consequences of the populace realizing that their leader has become a dictator, or - in the worst-case scenario - the military realizing this and taking direct action. But these extra-legal consequences require a well-informed populace (or at least a well-informed military). Now we also need freedom of the press. And a token freedom of the press, only sufficient to print the single line “the leader has defied the judiciary”, won’t be enough. People need context: is there an emergency? Was the judiciary actually trying to overstep? Is this part of a pattern? Is the leader generally a bad enough actor that this should tip people over the edge to vote against him, or to protest him? Many people will be reluctant to protest if the economy is strong and the borders are peaceful; is the economy actually strong, and the border actually peaceful, or is this just state propaganda? Answering these questions requires a flourishing journalistic ecosystem, including investigative reporters. A well-informed populace is useless without the ability to act on its information. Consider what might happen in a flourishing democracy if a leader tried to fire all the election monitors and replace them with toadies who would stuff the ballot boxes in his favor. Someone at the election office notices and informs the media (this step goes better if you have whistleblower protections enshrined in law, which may require an independent legislature).
Inline links: Babylon Bee, Spiked, WSJ, MacIver Institute, “illiberal democracy”