Overton Window is a recurring concept in the Astral Codex Ten archive, appearing 9 times across 9 issues between November 11, 2021 and November 26, 2025. The archive places it in contexts such as "there’s a tiny Overton Window of things you can actually get done"; "no act powerful enough and gameboard-flipping enough to qualify is inside the Overton Window of politics"; "the real Overton Window isn't centered around that point". It most often appears alongside United States, America, Christianity.
- Article page
- Overton Window
- Mention count
- 9
- Issue count
- 9
- First seen
- November 11, 2021
- Last seen
- November 26, 2025
Outbound links
- https://abc7news.com/post/graffiti-in-san-francisco-tagging-vandalism-street/13801629/
- https://slatestarcodex.com/2019/06/10/book-review-the-prices-are-too-dmn-high/
- https://www.amazon.com/Closing-American-Mind-Education-Impoverished/dp/1451683200/ref=sr_1_1?crid=SAPRLAFUD8ML&keywords=Closing+of+the+American+Mind&qid=1680632698&s=books&sprefix=closing+of+the+american+mind%2Cstripbooks%2C221&sr=1-1
- https://www.astralcodexten.com/p/details-that-you-should-include-in/comment/61387082
- https://www.astralcodexten.com/p/secrets-of-the-great-families
- https://www.statista.com/chart/21581/public-opinion-on-easing-stay-at-home-restrictions/
- https://x.com/RichardHanania/status/1867335298491068769
Issue trail
- Highlights From The Comments On Orban
- Practically-A-Book Review: Yudkowsky Contra Ngo On Agents
- Against "There Are Two X-Wing Parties"
- Your Book Review: Njal’s Saga
- Highlights From The Comments On Polyamory
- Highlights From The Comments On Nietzsche
- Prison And Crime: Much More Than You Wanted To Know
- Open Thread 359
- Why AI Safety Won't Make America Lose The Race With China
Related pages
Source context
Okay, fine. So regardless of your politics there’s a tiny Overton Window of things you can actually get done, things that you can sneak past the gauntlet of various liberal and conservative elites trying to frustrate you at every turn. You can either do some minimal thing within this window - Trump’s tax cut, Biden’s spending-bill-lite - or you can crush all those people.
Parenthetically, no act powerful enough and gameboard-flipping enough to qualify is inside the Overton Window of politics, or possibly even of effective altruism, which presents a separate social problem. I usually dodge around this problem by picking an exemplar act which is powerful enough to actually flip the gameboard, but not the most alignable act because it would require way too many aligned details: Build self-replicating open-air nanosystems and use them (only) to melt all GPUs.
Taken as an absolute claim, it's meaningless. Both US parties are on the same side of center? What center? By the standards of the Soviet Union, both US political parties are extremely far right; by the standards of Pharaonic Egypt, they're incomprehensibly far left. Whose standards for center are you using? The objective standard? Are you sure that exists? Are you sure you're not just taking your own personal beliefs about what seems reasonable, declaring the middle of that the objectively correct center, and then getting angry when the real Overton Window isn't centered around that point? People act as if you should just be able to take the leftmost thing imaginable, the rightmost thing imaginable, draw a line between them, find the middle, and then get angry if both US parties are on the same side of that line. But maybe they have poor imaginations. The leftmost thing I can imagine is an insectoid hive-mind; the rightmost thing I can imagine is a rapidly expanding cloud of profit-maximizing nanobots. Are we sure that a line drawn exactly midway between those two things lands on Joe Biden? What if it lands on anarcho-capitalism? Does that mean every existing human is left-wing?
We are many centuries of domestication removed from Ragnar of the Bloody Axe. Literal murder isn’t on most of our radars. Still, anyone on Twitter can sympathize with the ancient Viking feeling of getting insulted and debating how strong a response is warranted. On one side of the modern Overton Window, you have Elon Musk, who will ban people who offend him from Twitter, or sue them, or spread rumors about them being pedophiles. On the other side, you have - I don’t know, turning the other cheek doesn’t tend to generate a lot of news articles. But when I am in these situations, I try to think of Njal, kindest and most forbearing of men.
The question isn’t whether the success mode of mono is worse than the failure mode of mono. It’s whether moving the marginal couple from mono to poly or vice versa makes success more vs. less likely. I’m not sure what the answer is there. If you’re very optimistic, you could imagine that the people in the story above, instead of divorcing to be with another partner, could have stayed married, stayed with their children, and had another partner on the side. I’m not that optimistic. The people described don’t seem like they would be very good at any form of relationship. If you tell them to be monogamous, they will cheat. If you tell them they can have as many relationships as they want as long as they do so ethically and honestly, then they’ll do it unethically and dishonestly. But occasionally, I do meet some people who I feel would be better served by polyamory. Here’s a conversation I sometimes imagine having with certain friends and/or patients (this is partly stitched together from a few real conversations with different people, and partly imaginary): ME: So, you’re having a messy divorce. THEM: Yeah. ME: Because you cheated on your fourth husband. THEM: Yeah. ME: And now you’re dating a new guy. Are you worried that this might also end with you cheating on him? THEM: No. ME: Remember how we talked about the secret ancient rationalist technique of thinking about an answer for five seconds before you give it? I want you to try that now. Are you worried that this new relationship will end with you cheating on your partner? THEM: . . . . . yes. ME: Okay. Why do you think that is? THEM: I guess I’m just a bad person! ME: Can you be more specific? THEM: I guess I’m just really impulsive, and sometimes I see someone and can’t hold myself back! I’m too flighty and horny to ever be happy staying with the same guy for too long. ME: Okay. Can you think of ways that you could potentially address that in your next relationship? THEM: No, I’ve already tried therapy, and I’m too extraverted to be happy never going to any places where I could meet new men. I don’t know what else to try! I guess I’ll just never be able to be in a relationship without destroying it. ME: How would you feel about talking to this new guy you’re dating and telling him all this? Maybe he would be willing to agree to some kind of open relationship, so that if you felt something like this again, you could have a safe outlet that wouldn’t destroy the relationship. THEM: No that would be unethical. I’ve also met some people in the same situation as the “them” above who did switch to open relationships and it did seem to let them have a stable life / marriage / family in a way that they weren’t able to do before. I don’t think this is right for everyone, but the people who need it, need it. And here’s the kind of thing I see in relationship advice forums: I used to hate when my wife went out to parties or conferences because I always worried she was meeting other men. I told her she couldn’t leave the house unless she texted me exactly where she was going and explained why it was necessary, and it made her mad, but she eventually agreed. But she still talks on the phone to this one guy she’s been friends with since grade school. They’ve never had sex or anything, and he’s married to someone else, but they chat on the phone a couple of times a week and I always hear her laughing. I think it counts as an emotional affair and I told her that if she didn’t cut off all contact right away, I was going to end the marriage. Now she’s all mad at me. What should I do? This just seems like a fundamentally unhealthy outlook on relationships and life. People with better emotional skills can figure out some middle ground where they still agree not to have “emotional affairs” or whatever but don’t freak out every time their partner has contact with another human, but seeing the version where this goes bad has kind of radicalized me. I think there’s something really attractive about being poly even if you never get around to having any other relationships, just so you don’t have to constantly be getting angry at your partner for having normal human desires. TGGP writes: The real reason to prefer monogamy is that most cultures/societies have been polygynous, but the smaller number of monogamous ones were the winners in the contest of cultural group selection. I’m nervous about cultural selection arguments because they seem to justify anything, and to rapidly switch what they justify. Three thousand years ago, a cultural selectionist could have said that most societies in the world were polygynous, so we should avoid monogamy. A hundred years ago, they could have said that most societies in the world were monarchies, so we should avoid democracy. The problem is that nobody ever deploys cultural selection arguments against a failed trend that everyone hates. Nobody ever says “throughout history, most societies haven’t made people marry rocks, so we shouldn’t force people to marry rocks now”. Nobody wants to marry rocks now, so nobody needs to argue against it! Cultural selection arguments only get deployed against things that were once unpopular, but are in the process of becoming more popular - ie alleles that are currently being selected for! This isn’t necessarily hypocritical. You could say that certain institutions are spreading for good reasons, and others are spreading for bad ones (and TGGP links my Competing Selectors post which makes exactly this point). But now you’re not really making a cultural selection argument, you’re making an “I did some armchair reasoning and decided this was bad” argument. I think TGGP would answer that it doesn’t require any selective pressure to explain why people would want to have more romantic partners, since this is inherently fun for everyone. But first of all, no it isn’t - you can read the responses to this thread to see how viscerally unhappy some people are about this. And second of all, however fun it is now, it was probably equally fun a hundred years ago, so we still need an explanation for why it’s happening now and not earlier. Another possible argument is that polyamory isn’t spreading by monogamous cultures dying out and being replaced by polyamorous ones. It’s spreading by word of mouth. But this is also how almost every cultural trend spreads. Monogamy didn’t spread to Scandinavia because the Vikings died out and the Romans colonized their land, it spread because missionaries converted them to Christianity. If some memoir converts somebody to polyamory, that seems like the closest modern-day equivalent. I don’t really know how to rescue cultural selection arguments from these kinds of considerations. Ascend writes: Okay, I have two basic objections to polyamory. The first is a broader but weaker point, and the second is narrower and stronger. 1. If you actually love someone, that person should be enough for you. Especially, if most people who claim to love someone do, in fact, find that that person is enough for them. I struggle to see how it doesn't almost follow by definition that if Aaron wants to be in a monogamous relationship with Brenda, and Carl wants to be in a polyamorous relationship with Diana, then Carl loves Diana less than Aaron loves Brenda. Now maybe this assumption can be refuted somehow, by reference to irreducible personality types or something, but it certainly seems like the prima facie assumption, and requires an affirmative defence. Splitting this up into two paragraphs so I can answer one at a time. The traditional response is to ask: Do you think it’s bad for someone to have a second child? Surely if they really love their first child, one should be enough for them! More than one friend? In fact, why do they need to have a friend at all? Shouldn’t their partner be enough? I think you can think of this in one of two ways. One is to imagine the original objection in the mouth of a Dickens villain: “We can’t invite other people to our Christmas party! They would consume our limited supply of Christmas cheer, and then there wouldn’t be enough left for us!” Obviously at the end of the Dickens book, the character discovers that cheer isn’t a limited resource, and multiplies rather than divides when shared with others. Some people genuinely think this way. This isn’t really how things work for me. I want my wife to definitely be the most important person in my life and vice versa. But I find I can carve out a category “secondary partner” that doesn’t interfere with this, any more than her having friends , hobbies, children, etc interferes with this. Probably other people’s psychology doesn’t work this way, and those people wouldn’t enjoy being poly. 2. There seem to be two types of polyamory: the "hippie" kind where a group of people are all in a single demarcated "relationship" with each other, all know each other as either friends or lovers, and having sex with anyone outside that group would be condemned as cheating. And the "open relationship" kind where two people are dating or married but are "free" to sleep with other people (but still have each other as a "primary partner" or some such). The first, while I'm not endorsing it, seems to have decent case for being a form of actual love. The second, unless I'm missing something, looks like despicable pure hedonism. First, because the "primary" aspect shows it really is all about sex, not about sharing your love with someone else. Second, because it's basically legitimised cheating, and for all the talk of mutality what's to stop someone pressuring their spouse to "consent" to "opening" their relationship? (I've seen a number of online stories of this happening, though with beautiful poetic justice where the pressured partner ends up finding someone who actually values him or her and wants a true relationship, and the other partner ending up entirely alone and certainly not finding the harem they were expecting). And third, because it creates *competition* between the two spouses over who can get more partners, and for fuck's sake a marriage is the ONE place such toxic sexual competition should not exist! Re: 2 - As I mentioned above, polyamory seems more about the romance angle than the sex angle. One interesting demonstration of this is how many asexual people are poly. In the 2017 SSC survey (the same one cited above), about 5% of polyamorous people described themselves as asexual (having no sex drive) compared to about 3% of monogamous people. These people are probably in relationships for the emotional benefits. I also checked sex drive among people who were in monogamous relationships, stable polycules, and open relationships (this is the 2019 SSC survey, which went into more depth on sex and romance). There were no significant differences in sex drive among these three categories. Piotr Pachota (blog) writes: Invectives about polyamory are inevitable now, in the current phase of polyamory Overton window shift. This is the only acceptable way polyamory can be currently portrayed and discussed. As Overton windows move for new social norms like LGBT and polyamory, the mainstream media narratives evolve along the axis of omission/taboo -> negative/critique -> ambiguity -> struggle/oppression ->positive -> new normal. We have seen LBGT cover almost all these steps during recent 40 years, currently I think we are between positive and new normal. With polyamory, we have just recently moved from omission/taboo to negative/critique. Portraying polyamory as positive or struggling/oppressed in mainstream media is still out of the Overton window, showing ambiguity is now at the edge, only negative critique is in the window. This is also why almost all movies and shows about polyamory are telling a story of a failure of a polyamorous relationship. Positive polyamory testimonials exist on social media, but usually the comments below are a shitstorm. This also proves how positive portrayal is unacceptable. At the same time, positive polyamory testimonial + shitstorm comments = negative meta-content bundle that itself fits well within the Overton window. Note that these negative portrayals still promote polyamory somehow, as they at least put it on the map. 10 years ago there were no polyamory movies, shows or social media content - it would have been unthinkable, as we were in the omission/taboo phase back then. Yeah, one pattern I see pretty often goes something like: Weird people, who maybe weren’t able to cope with normal institutions, find some weird new thing that works for them, okay, cool, whatever.
Inline links: writes, Competing Selectors, writes, Some people genuinely think this way, blog, writes
I think the way I wrote it is one end of the Overton Window, the way 10240 writes it is the other, and to me the window seems small enough that I don’t see much difference between our two formulations - but I appreciate the criticism.
Roodman 2017. Written by a researcher at the effective altruist nonprofit Open Philanthropy Project, which was trying to determine whether to continue to advocate decarceration. They (Roodman’s bosses) had already started some advocacy, so Roodman admits he is at risk of a soft-on crime bias. This was extraordinarily well-written, and is one of the most careful meta-analyses I have seen on any topic - Roodman tried to get data from every study and replicate it himself and test all of the assumptions - but I am still interpreting it as the anti-incarceration faction presenting their side of the case. Each of these reviews gets deep into the statistical weeds to bolster or discredit their favorite/least favorite studies, and in most cases I wasn’t able to have a strong opinion on who was right. My strategy was to use the pro-incarceration and anti-incarceration reviews to determine a sort of “Overton window” of what smart people thought they could get away with claiming. Then, if the window was very wide and I couldn’t figure out where I landed, I used the apparently-unbiased Nagin as a tiebreaker. All three reviews separate the possible effects of long prison sentences into three bins: Deterrence. Would-be-criminals can consider the severity of punishment before committing their crimes, and, if the punishment is too high, decide it’s not worth it.
Inline links: Roodman 2017, whether to continue to advocate
So we expected Three Strikes to decrease crime by 80%, but in fact it decreased it by 0-7%. Why? Because California’s Three Strikes law was weaker than it sounds: it only applied to a small fraction of criminals with three convictions. Only a few of the most severe crimes (eg armed robberies) were considered “strikes”, and even then, there was a lot of leeway for lenient judges and prosecutors to downgrade charges. Even though ~80% of criminals had been arrested three times or more, only 1-4% of criminals arrested in California were punished under the Three Strikes law. Why can’t we have a real Three Strikes Law? For the same reason we saw in Sweden earlier. I can’t find any graphs of the US population broken down by number of past offenses, but we know that about 8% of Americans have at least one felony. Let’s say that about half of those have at least three felonies. That means a real Three Strikes law would require increasing the incarceration rate from its current 0.75% up to 4%, ie quintupling it. We’d need to build 6,000 new prisons and 10,000 new jails, locking up an additional 5-10 million people, and spending somewhere between $400 billion and $1 trillion per year (ie around the same as the entire military budget) on prison-related costs. This is light-years outside the Overton Window and I’ve never heard anyone seriously propose it. Still, it would decrease crime by 80%. Again, the lesson is that - despite power laws - small increases in incarceration cause small decreases in crime, and only very large increases in incarceration are capable of causing very large decreases in crime. Why Did A Ten Strikes Law In The Netherlands Massively Decrease Crime? Given the ambiguous results of three strikes in California, I was surprised to learn that a ten-strikes law in the Netherlands did reduce property crime by 25% (Volaard et al). In 2001, the Dutch government passed a law allowing longer sentences for criminals with at least ten previous offense who were not good targets for rehabilitation (eg rejected or had already failed drug treatment). The law allowed judges to increase the typical sentence for petty theft (2 months) to a longer sentence (2 years). A quasi-experimental study found that property crime, though not violent crime, decreased by 25%. It’s not surprising that violent crime didn’t go down since the law was almost entirely deployed against thieves. Vollaard found that the population affected was extremely criminal; they had an average of 31 past offenses, and on surveys they admitted to committing an average of 256 crimes per year (mostly shoplifting). Before the law was passed, they spent an average of four months per year in jail (probably 2 x 2 month sentences); afterwards, they spent two years in jail per crime. It’s hardly surprising that imprisoning these people decreased crime. So why didn’t the California law do the same? I think there are two reasons. First, Europe imprisons far fewer people than any US state. The incarceration rate in the Netherlands is only an eighth that of California. The pre-ten-strikes-law Dutch penalty for shoplifting was two months; the equivalent California penalty alternates between six and thirty-six months depending on which set of propositions won the last election. Vollaard found that the ten strikes law had diminishing returns: On average, we find the benefits of the policy to exceed the costs by a large margin. We find the benefits to go down rapidly with a more intensive use of the law, however. The marginal crime-reducing effect of convicting another prolific offender to an enhanced prison sentence declines by some 25% when going from the 25th to the 75th percentile in the rate of application of the law during 2001–7. The benefits of the policy remained higher than the costs, however, even for the cities which used the law most intensively. So I think it’s plausible that California was already much more punitive than the Netherlands even before the California Three Strikes Law was passed, and there were diminishing returns from becoming even more punitive than that. Second, the California law only applied to certain serious crimes (eg armed robbery). People don’t commit these as often as shoplifting. The average person affected by the Dutch law shoplifted 256 times per year. Even the most energetic criminal would struggle to commit 256 armed robberies per year, and they would probably get killed (or murder a victim) long before reaching that point. So petty theft is longer-tailed than serious felonies, and it’s easier to decrease the petty theft rate by imprisoning the few worst shoplifters compared to decreasing the armed robbery rate by imprisoning the few worst armed robbers. Why Can’t We Just Incarcerate Those 327 Shoplifters In New York City? This time we just suck. Remember, each of these shoplifters was arrested 20 times per year. So they can’t be going to prison in any substantial way. Even if they got a one-month sentence for each arrest, they’d run out of months to shoplift in after twelve! So the question here isn’t “why are prison sentences for shoplifting so short”, but rather “why can’t New York City incarcerate repeat shoplifters at all?” I’ll come back to this question later. What About El Salvador? They famously solved crime by more imprisonment - how? By doing a lot of it. I said above that we might be able to decrease US crime rates by 80% if we quintupled our incarceration rate. Between 2014 and today, El Salvador quadrupled their incarceration rate: (source) They now have by far the highest incarceration rate in the world, 2-3x that of America (which is itself the fifth highest, after various dictatorships). How did they afford this? Through a combination of lots of funding, not being too picky about human rights and prison conditions, and not being too picky about whether the people they imprison were guilty of any specific crime or just kind of gang-adjacent. As incarceration rate quadrupled, homicide went down by a factor of twenty: We previously predicted a similar increase in incarceration would lead to an 80% decrease in crime in the US, but El Salvador got a 95% decrease in crime. Why did they do so much better than our prediction? I think because they started with half our incarceration rate and ten times our murder rate. When you’re starting from someplace terrible, without any of the low-hanging fruit picked, it’s easy to make progress! I can’t find good statistics on other crimes like theft, but the crappy statistics I find say it hasn’t budged (1, 2). Why not? Either my statistics are bad, or the gangs that the government cracked down on weren’t in the theft business.4 Incapacitation Fine, so despite power laws there’s no way to easily solve crime just by imprisoning a small number of people. How much bang for the buck do we get by incapacitating criminals? You would think this would be easy to figure out: just determine how many crimes the marginal prisoner commits per year. Then that’s how many crimes incapacitation prevents per year. But although it’s easy to see how many times the marginal prisoner has been arrested, most crimes don’t result in arrest. How do you know how many crimes they really committed? Some bold scientists have tried asking them - giving prisoners surveys about their criminal histories - but obviously these should be greeted with heavy skepticism. The method criminologists have settled on is to wait for big shocks to incarceration - big enough to affect the general crime rate - then see how much the crime rate goes up or down. The best study here is probably Levitt 1996 (you may know Steven Levitt from Freakonomics). In the 1970s, US prisons were overcrowded. The ACLU argued the overcrowding was a rights violation - a form of “cruel and unusual punishment” - and sued a dozen states. They won all their lawsuits, and judges in all states said the government had to free prisoners until prison crowding returned to a non-cruel, usual level. So at a slightly different time in each state, many prisoners got released all at once. By examining the effects of this sudden release on the crime rate, we can determine how much crime the incarceration of those prisoners was preventing. Levitt does a lot of fancy statistics, and Roodman reanalyzes with even more fancy statistics, but the good news is they both agree and get numbers somewhat contrary to Roodman’s biases, which make me trust them more. Each year of imprisoning the type of prisoner who got released under the ACLU lawsuits prevented 6 property crimes and 1 violent crime. This suggests the average criminal commits ~7 crimes per year, which I think matches well with the data above showing that the median prisoner has 10 past arrests and some have 30+. Other studies on incapacitation, mostly taken from Roodman, that I trust less than Levitt: Owens (2009) investigated a Maryland law that caused some criminals to get released early. They found a crime increase corresponding to about 3 crimes per prisoner per year. This is lower than Levitt’s estimate of 7, but crime rates went down in general between Levitt’s study period (the 70s) and Owens’ (the 2000s), so they might both be right.
Inline links: 1, 4, 8% of Americans have at least one felony, Volaard et al, https://substackcdn.com/image/fetch/$s_!9L-h!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F29a25928-8620-45d1-b677-34447be88ad6_778x528.png, source, fifth highest, https://substackcdn.com/image/fetch/$s_!ww3Y!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F028e701c-5f95-436d-b1ff-15742274e221_2000x2000.png, 1, 2, 4, Levitt 1996
And here’s Bentham’s Bulldog trying to convince you to donate to the Shrimp Welfare Project. “I’d be surprised if we got to heaven, asked God what the highest [moral] impact thing that we could have done is, and his answer was ‘oh, something very normal and within the Overton window.’”
Narratives about regulation stifling progress are attractive because they are often true. A time may come when the Overton Window shifts to a set of AI safety regulations strong enough to substantially slow American AI. Perhaps this will happen at the same time that China finally solves its own chip shortage - likely sometime in the 2030s - and America can no longer rely on its compute advantage for breathing room. Then the threat of Chinese ascendancy will be a relevant response to concerns about safety. Perhaps people raising these arguments now believe that they are protecting themselves against that future - better to cut the safety movement out at its root, before it starts to really matter. Be that as it may, their public communications present the case that AI safety regulation is already a big threat. This is false, and should be called out as such.