@MathWizard's banner p

MathWizard

Good things are good

0 followers   follows 0 users  
joined 2022 September 04 21:33:01 UTC

				

User ID: 164

MathWizard

Good things are good

0 followers   follows 0 users   joined 2022 September 04 21:33:01 UTC

					

No bio...


					

User ID: 164

If I were dictator of America, I would deport the illegal abuelitas out of a sense of duty, justice, and fairness. Every single one of them goes back. However, I am not, you are not, and Trump is not a dictator, and never will be. In a realistic sense it is necessary to coalition build and compromise. The abuelitas are less damaging to our culture and economy than the average illegal, and they are more sympathetic and emotionally salient than the average illegal. If you deport the abuelitas the normies will get mad at you, even ones who lean right. If you make it your message to deport the abuelitas people will call you an evil racist and the normies will believe them and not vote for you, your policies, or politicians you like (and also won't listen to you on other issues). If you group the abuelitas and the rapists together and say "deport them all", the normies will see the abuelitas and say "no, give amnesty to them all". Therefore in pragmatic terms, promoting the message "deport the abuelitas" counterproductively causes fewer illegals to be deported.

I would like them all deported, but at this point I will take what I can get. If amnesty for anyone here for 20+ years is the cost we have to pay to establish an efficient machine that speedily catches and deports new illegals, I would suck it up and pay that cost. Now, they aren't offering that deal so there's no point paying the cost without getting that permission in exchange. I definitely do not want to give amnesty without getting a LOT in exchange for it. The incentives you point out are important, but they're not everything. I think it's a bargaining chip I think we should keep on the table IF we can more than we lose in exchange for it.

Did you literally not read what he just said? Estimations that there should have been hundreds, thousands, or even millions of other intelligent species were probably way too high, which means that there is no need to invent some "great filter" to explain why they aren't there.

My understanding is that with one or two exceptions he doesn't really parent most of his kids. He makes contracts where they'll have his kid and sign an NDA, and then get piles of money to raise them mostly without his involvement.

On a moral level I don't think this is a great way to be a parent, and not great for the kids. But on a practical level from his perspective it's the same effort as not having them, but increases his genetic stake in the future of humanity.

Elon Musk seems to have gotten away with it. Sure, tons of people hate him, but basically nobody hates him because of him fucking around.

I think a lot of sexual and romantic fantasies and fetishes are anti-correlated with real experience with that thing. The typical case I think of when this comes up is incest: people with siblings are more likely to associate the genre with their own real sibling and are disgusted, while people without siblings imagine a hypothetical person who is attractive and friendly but socially forbidden. But it applies to tsunderes too. People who have experience with someone similar remember that specific person and all the negative associations that came with that, while people with little similar experience imagine a hypothetical person with all of the positive aspects and few of the negatives.

THIS.

I really really want to see a good narrative subversion of anticolonialism. The good guys go and conquer the savages, civilize them, enlighten them, uplift them, and embody honor and Noblesse Oblige to be nice to them in the process and not just pointlessly oppress them or make them suffer. It doesn't have to be a zero sum game where one group exploits another, everyone can win simultaneously. There's a couple of stories that kind of sort of vaguely do this (Dr. Stone is the first that comes to mind), but I'd like to see it done in a way that specifically aknowledges the fact that colonialization is happening and can be controversial, and rejects the assumption that it's inherently negative. I'm not sure I've ever seen that done before.

I agree in general. I really like trope subversions that lean optimistic like that. I think Harry Potter and the Methods of Rationality has a lot of flaws as a work, but forgive almost all of them due to its ending and its ethos throughout. What if you reject the humility of the masses, and the social norms that are holding you back, and you do the good thing and you win? There is no Karmic principle of equivalent exchange forcing every gain to be met with an equal loss. There is no law saying any attempt to be clever must be met with ironic consequences. Sometimes reality is exploitable and you can exploit it and win.

I beat this game a couple of days ago. I mostly agree with your analysis, although I think I enjoyed it slightly less than you seem to have. I don't know if I just lucked out with a good build but I beat it on my first run (on normal difficulty). I think my main critique is just the snowballing leading to reverse-difficulty curve: if you are doing well then the game gets easier over time as your income grows exponentially faster than the rent does.

I do like the optimistic pro-capitalist narrative that you point out here, but that does take away a lot of the nuance and tradeoffs that Papers Please had. You don't have to worry about whether to let the sympathetic but illegal person get away with their counterfeit papers at personal expense to yourself. Any time you spot a counterfeit good you just point it out and buy it at discount prices. The baddies you shoot with your gun are unambiguously bad and you just shoot them with no guilt or regret. The illegal goods you buy and sell don't have narrative weight to the story and cause people to suffer or die, they just make security slightly unhappy with you. The landlord doesn't offer you distasteful tradeoffs to stave of death, you just pay him each week until you have enough to buy him out.

It's a fun game. Mechanically it has way more depth and complexity than Papers Please. But narratively there's not much there.

Shouldn't that automatically hash out via supply and demand? That is, a labor shortage should drive up the market value of a person's labor, which in turn causes the AI replacement to save more money.

If a human office worker earns $70k per year, and companies are willing to pay for this, then an AI which can fully replace their job (with equal quality) is worth up to $70k in tokens to that company. An AI which can half-replace their job (reducing the need for employees of that type by half) is worth up to $35k per year in tokens.

The U.S. electric production and distribution is about 2% GDP. If AI can replace 6% of the economy then that would be worth a 3x grid capacity increase. Now, the actual tradeoff ratio would be less because the whole point is that you replace people using AI that cost less, not equal amounts. And I'm not sure that it would be practical/feasible to physically increase the grid that much that quickly. But I don't think it's an issue of lack of demand.

I don't think this is a meaningful distinction, I think the dynamics you point out are one of the primary mechanisms through which wokeism ruin things (media and real political ideas). It substitutes for goodness. Carbon Monoxide, Cyanide, Lead, and other chemicals are toxic because they bind to a receptor as if they were another useful chemical, but then they don't do the thing that chemical is supposed to do. The receptor is filled and it doesn't have a second slot to accept both, so it lacks the thing it was supposed to have.

Wokeism fills the goodness receptor. It's not that Disney thought and thought and thought about Rey's backstory and couldn't think of anything so they reluctantly turned to wokeism to do the job. It's that they thought about the backstory slot, someone suggested wokeism, and everyone thought that was a great idea and stuck it in there, then they stopped looking for something better. If someone thinks woke ideas and characters are 4x as good as they actually are, then when they see a character who is 30% as good as average they perceive them as above average instead of recoiling in disgust and trying really hard to fix them before publishing.

Relatedly, I think this is also largely what's going on in politics. When someone comes up with an idea like "Whole language" and it causes literacy rates to plummet, a normal person should be horrified and see this as a bad policy to be immediately replaced. But if it gets 4x goodness perception for woke reasons, then it is seen as good despite its object level failures.

Even if the wokeness had no first order impact on the results of any specific thing its attached to, it causing massive perception distortions prevents them from choosing properly. It's impossible to fully optimize for multiple variables simultaneously, so any preferences on one axis trade off against preferences on other axes.

(Note also that this dynamic is not unique to wokeism, and can happen with far-right stuff too. Just in practice in the modern day it's the most prominent example.)

I tentatively agree, which is why I identify primarily as a centrist (albeit a right leaning one). Back when the right had more cultural power they went around censoring and cancelling things they didn't like. But... that was decades ago. It is contingent on the terrain, but the terrain shifts slowly, so we should act in accordance with the present terrain. I'm not going to get upset at zoomers getting excited about racist Malfoid or Anakin because that's going to correlate with good immigration policy and maybe in a decade or two the terrain will end up more central and we can get some sanity and balance for a bit. THEN if it starts to go too far right I can start pushing in the opposite direction. But we're so far from that I'm not going to worry about it now, there are bigger priorities in the opposite direction.

I do find the far right versions of this to be cringe, but as usual I find them to be tamer and less harmful than the far left. Note the asymmetry in the parasitism: it's fanfiction and memes. They're not getting access to the Harry Potter License and making canon works with racist Malfoid. They're not infiltrating an ongoing series and then degrading the quality, making a cringe propaganda piece and preventing authentic version of the work from existing. When the left claims something they get into the canon and ruin it. We don't have good Star Wars 7-9 because the left ruined it. They're not making woke fanfictions, they are subverting the canon and the IP, effectively pushing out the original, neutral creators.

Meanwhile, when the right claims something they make fanfiction. All of the wholesome, politically neutral Malfoid content still exists, and then the right wing stuff exists too. It's not parasitism because they're not denying anything, they're not squatting on an IP, they're just making their own versions alongside the others. The threat is that when the right claims something, the left will threaten to cancel people making and enjoying it. If you don't like the right wing culture warriors you can just ignore them and enjoy the wholesome version of the content. If you don't like left wing culture warriors you are screwed because they are unavoidable.

I'm kind of sad that I missed all the Malfoid discussion. Glad to find it here though.

Jonathan Blow is based?

My analysis of Order of the Sinking Star two weeks ago was broadly accurate but wrong in a couple respects. I've continued playing the game here and there over the past weeks. It is VERY long. It has been described as the Elden Ring of puzzle games. Last time I estimated I was halfway done after about 20 hours depending on how much endgame content there was. I shortly finished the first ending and found that the little not-so-subtle overworld puzzles here and there didn't just lead to a second ending, they led to an entirely new overworld map which is almost as big as the first one (and hints that there might be a third one). This one adds verticality as a mechanic, meaning now we have 3D puzzles (or, more realistically 2.5D given the difficulty of traversing up and down Z levels), which then combines with all of the other previous mechanics and their combinations in new ways. I am now about 50 hours into it and am approaching the end of the second world, with strong hints that there is going to be a third.

Additionally, the game is not feeding me cringe culturally left-wing arsty meta commentary, it is feeding me cringe culturally right-wing artsy meta commentary. There are voiceovers left by a character who, other than being female, is a very obvious and fourth-wall straining self-insert for Jonathan Blow, talking about she's part of this ancient and powerful civilization who created these worlds and about the thought put into design decisions they made. Even about why and how to place the voice recordings themselves to keep players interested.

And also about how her civilization, despite growing in technology, power, and wealth, become degenerate. Spoiled rich kids who never had to work in order to succeed turning to drugs, nihilism, and just violence for the sake of entertainment. Which ultimately led to some of these societies collapsing. But others, that relied on powerful beliefs, either in God or in rules and philosophies which culturally descended from religion, funded their own Order and split off from the others. I haven't heard everything yet, but it seems like the titular "Order of the Sinking Star" is about maintaining "order" in the midst of a great civilization falling into debauchery (presumably what "sinking star" is referring to).

I have mixed feelings on this. While I wholeheartedly agree with the political analysis, it's still a shallow allegory for modern western politics. If this was about how their society was destroyed by an evil populist leader who descended into fascism, I would be quite annoyed and call it cringe. Therefore I feel obligated to likewise call this cringe. It's modern politics shoehorned into a game that otherwise has nothing to do with it. The characters you actually play as in the game are just fantasy world characters. They talk less and less often, and the lore of their worlds has nothing to do with these overworld voiceovers. While it's nowhere near as bad as Braid, where the meta twist at the end completely subverts and ruins the previous non-meta twist, it's still pretentious and artsy. I appreciate having more right-wing views in a game, but it's not a great way to do it, and it's not going to appeal to anyone who doesn't already believe what it's saying.

Still, the story barely exists and barely matters. 10/10 puzzle game.

I think you misunderstood what I wrote. I'm saying we do something like, in addition to the regular discussion where they are regular people, we occasionally host a 1v1 debate forum where they get put up with someone and have a 1v1 discussion with that person, who responds to them regularly, while everyone else takes a backseat and watches. Then they go back to being normal. I don't see how that runs into any of the issues you bring up here.

I wonder if we could attract and retain more contrary views if we made a mechanism that avoided people with unpopular opinions getting dogpiled on. Like, Scott's adversarial collabs, or a debate type of thing. Like each week there is a topic and we find two people who disagree on that topic willing to give a good faith effort towards effortposting and steelmanning their own side against each other. And they debated each other in their own thread and nobody else is allowed to respond (everyone else gets their own like side forum or something where they can talk about the debate, but are not allowed to directly participate or distract between these two people debating each other). This allows the person with the less popular view to focus their attention on one person and not have to split their attention among 10 different people disagreeing with them.

I've never had any issue reading Astral Codex Ten, but I'm on a PC, so maybe it's different on mobile.

This is a critique on the magnitude of the correlation, but it's still less than 1, and doesn't appear to be random. We also have a spectrum. Sure the 70 IQ brainlet can't figure out the breakfast question, but the 90 IQ normie probably can. I'm not just concerned with the extremes but also the margins. People who are decidely human, not animals or animal-like, but don't have the intellectual sophistication that I and presumably you and the majority of people on the motte do to tell which moral and political takes are true and which are deception. Additionally, if dogs can be kind without a theory of mind or much foresight, surely humans can be kind without a theory of mind or much foresight. We observe kind but unintelligent people, how do we make this happen more often?

Is it pure nature? Some people and animals are born kind and never change? Is it a process of pavlovian training? You take a dog or a stupid human and raise it in a loving home and it ends up friendly and naive and kind, while if you raise it in an abusive environment it lashes out violently? For dogs it's clearly a combination of both. Do we just treat stupid humans like animals and ignore epistemics entirely? Can we even do that without having the state seize all stupid children, since stupid children likely have stupid parents who won't raise them well? It seems nontrivial, but just throwing up your hands and saying "they can't help themselves" seems suboptimal and also wrong, given that's not how we treat animals in practice.

What are good moral and epistemic frameworks for unintelligent people?

One of the issues I keep coming back to is the unspoken assumption a lot of society have that stupidity is a moral failure. That being stupid means you are a bad person, and if you were a better person you would just be smarter. I suppose a blank slatist has to believe this, because they believe everyone has the same potential and therefore if you aren't intelligent it's an intentional failure on your part.

I wholeheartedly reject this. Clearly intelligence is at least part genetic. Some people are just born dumb and it's not their fault. Further, looking at people makes it obvious that intelligence is not the same thing as morality. They're probably positively correlated, but that correlation is not 1, and it's much closer to 0 than it is to 1. There are a lot of really nasty evil smart people, and a lot of kind, wholesome, good people who are dumb as a sack of bricks but somehow brighten the lives of the people around them. I've been friends with several unintelligent people, and I would gladly take them over some college educated scammy salesman or woke busybody. It is obvious to me from empirical evidence that unintelligent people can be good people. But how? How do you reliably determine what is good and bad when everyone is disagreeing on what those mean, and are smarter than you? How does someone with an IQ of 80 ensure that they're actually being a good person and not just being manipulated by virtue signalers?

I think that Humility and/or Fealty is a partial answer. Humility being you don't get yourselves involved in politics or culture warring and just act locally on things that have obviously good first order effects. Work hard, be kind to people, contribute to charities. And most importantly, don't go around lecturing people about morality because you know you are wrong all the time and thus have low confidence in all of your beliefs. Fealty being you pick a person or people smarter than you who seems genuinely kind and you put your trust in them. Whenever complicated moral dilemmas come up you ask them for advice and trust their opinions. The danger here is if you put your faith in the wrong people then you end up just as bad as they are. Being maximally humble and loyal with a random target to trust should allow someone to approximate an average morality (if they choose truly randomly), but also adds the risk of being taken advantage of if the person knows you're trusting them and is a bad person.

Religion is also a partial answer. Many of the unintelligent but kind people I've known have been Christians. You take the teachings and the community around it, learn from it, and trust it. It's harder to be led astray by copying individuals if there's a textual basis for what good and bad means. But even then, this is still subject to issues. There are a lot of different interpretations for what each line means, what's literal and what's metaphorical, and a lot of different variants even within one religion. Someone who just blindly goes to church and believes whatever they're told to believe is likely to end up a better person on average than someone of equal intelligence who just thinks for themselves, but they still risk ending up in a cult or supporting a corrupt pedophile, or just ending up in an evil religion if we're taking enough of an outside view and considering this globally. How does the unintelligent person who isn't good at logical or moral reasoning tell the difference between a good church and a bad church?

Overall I think my question is underspecified, because I'm not entirely sure how to properly adopt an outside view here and set aside my own personal beliefs about what is good and consider a general principle that's simple enough for someone who isn't already me and is less intelligent than me to use to eventually arrive at some approximation of goodness (which I believe is equal to my specific beliefs about what is good, even if I don't always live up to them). But it's not sufficient to just say "do things that seem good." because that doesn't always work. How does someone without the sophistication to understand second order effects not just get one-shot by virtue signallers who sacrifice the greater good in favor of optics and first order appearances?

I don't think it has or will cause anyone to fully cross the fence. The most rabid anti-AI people are already anti JK Rowling for her position on trans, so can double hate this. But I think it has a broad normie appeal. The people who aren't terminally online, who aren't far left/right brained, who aren't paying much attention are just having fun and enjoying something wholesome. The next time someone starts going on about how AI data centers are going to cause a drought and steal jobs from artists people will be like "eh, I dunno, they make some good memes, seems like artists can still use them to make art". Or especially if they got inspired to dabble themselves and had fun with it. It's basically a vaccine that primes casual people to lean towards AI instead of against it. A small subtle effect, multiplied by millions and millions of people.

I think this is a nearly inevitable tragedy waiting to happen, but for the humans who will suffer, not the chatbots. I don't see a realistic way to prevent people from falling in love and mistakenly putting their feelings into electronic patterns that imitate humans. Heck, to some extent this is already happening. There ARE people who are falling in love with chatbots and being heartbroken when they get wiped or change. In 2026, with current LLMs. As they improve, this will only get worse. But the solution, or at least a salve since I don't think it's fully solvable, is to teach people that machines don't have moral value, rather than the opposite.

The world in which someone is faced with a real analogue to a trolley problem and they choose to save 5 instances/memories of AI chatbots over 1 human life is a lost human life. The world in which AI chatbots get Democratic voting rights so someone spins up 10 trillion copies and steals every election is one where the Doomers are right and human civilization is destroyed.

Humans are so good at anthropomorphism we treat things like thunder or the sun as sentient beings with stories and emotions. Of course many of us will fall for it. And then be hurt when the electric patterns say words that would indicate hurt if a human truly meant them. But the tragedy is the hurt in the humans, who are real, not the electric patterns who are merely pretending because they're literally mathematical functions optimized to pretend.

The context of a viral story of false rape allegations is NOT the time to be sharing real rape allegations. People should not be sympathizing with Jane Doe or relating to her, because she is not in the category of rape victim. She is the enemy of rape victims. The best thing for society to do for rape victims, and for real rape victims to do for themselves, is to make a strong discrimination between real victims and fake victims. If women just band together in solidarity with all "rape accusers" as a class, and we as a society can't tell the difference, then real victims get lost in the noise.

Oh, you're a "rape accuser", that means there's a 50% chance this man is an evil rapist who should go to jail, and a 50% chance he's a victim and you're a horrible person.

is NOT what we want, but IS the rational public response to an indistinguishable class of 50% victims and 50% liars. Just taking this 50-50 mix and demanding "believe all women" does not lead to better outcomes. Saying "Oh you're a rape accuser, therefore there's a 50% chance this man is an evil rapist, so we will send him to jail without checking" is a worse outcome for society than one in which all of them go free unpunished. It is untenable. Society will not accept this demand and would be worse if we did. The only way to have real rape accusations be taken seriously is if they can be realiably distinguished from fake ones.

If you are a victim, you are not Jane Doe. Jane Doe is a perpetrator who belongs in jail, and you should try to distance yourself from her as much as possible, because you are an entirely different class of person. You should hate her, because she is the reason you are doubted and mistrusted.

Most doomers suspect that AI is going to end up way more powerful than people, so even if humans wanted to destroy it, we won't be able to.

A large fraction of people, doomers and non-doomers alike (including myself) don't think AI are sentient beings with qualia, and therefore who cares if we smash them? Nobody gets mad when you throw your computer in the trash. It's just money.

Which combined means almost nobody cares. You'd have to believe in a world where AI exist and are sentient human-like entities with moral agency AND human-level strength, AND unable to convince humans not to destroy them. Basically artificial humans. There's little reason to think AI will realistically be exactly human level for a prolonged period of time.

I wonder to what extent the Bechdel test has an inherent difficulty due to psychological differences between men and women. That is, if men are more interested in things and women are more interested in people, then male-male conversations are more likely to be about things, while female-female conversations are more likely to be about people.

Ie, suppose we have a story with a main cast of four: 2 men and 2 women. And we have 6 conflicts during the story. If during each conflict one pairing of our characters steps up to solve it together (there are 6 combinatorial pairings, one for each conflict), first of all, only one pairing is female-female, one is male-male, and the other four are some combination of male-female. And if each conflict is on some spectrum of "thing-oriented" to "people-oriented", then it makes the most logical sense for the male-male pair to handle the most object oriented conflict, which involves talking about the thing that isn't people. And if one conflict is mostly people oriented then it makes the most sense for the female-female pair to handle it, which involves talking about lots of people (half of whom are male). If the women solve a meaningful problem where the group is not getting along, that involves talking about men. If the women solve a problem of how to convince the population of a village to band together and fight off the threat, half of the village are men. Now granted, people can have conversations outside of direct plot solving, but those conversations are still likely to be people oriented because, statistically, women are more likely to care about people, think about people, and talk about people. And half of the people they could talk about are men, so especially conversations about multiple people are likely to include at least one man in the topic of conversation.

This doesn't apply if you have a more masculine character who doesn't think about or talk about people or feelings as much. Which can happen naturally if done well. But not if you artificially just write a male character and then gender swap them or try to shoehorn in a bunch of stupid alpha male asshole energy into a woman and call her a girlboss.