official_techsupport
who/whom
No bio...
User ID: 122
official_techsupport
who/whom
3mo ago
A limitation of usual Bayesian reasoning.
Scott is doing his annual subscription drive and I was reminded of a (still) private post of his I disagree with: https://www.astralcodexten.com/p/but-seriously-are-bloxors-greeblic
In my post on uncertainty around AI, I wrote:
If you have total uncertainty about a statement (“are bloxors greeblic?”), you should assign it a probability of 50%. If you have any other estimate, you can’t claim you’re just working off how radically uncertain it is. You need to present a specific case.
Commenters were skeptical! I agree this important topic needs more discussion:
And then he proceeded to list some of the objections and his objections to objections. The objection I'm personally most partial to was not listed, so I assume it's a sort of novel idea, at least in that (and this) community.
Suppose that in your travels you encounter a shady guy who offers you an opportunity to bet on the outcome of a coin flip. Nearby stands a yudkowsky, who tells you that according to his observations the coin is biased and the next flip is about 66% likely to land on heads. You know that yudkowskis are honest and good Bayesians, so you trust his assessment.
The shady guy flips the coin and it lands on tails. Now consider two possible worlds: in one the yudkowsky says that his new estimate is 50% heads, in another he says that he has updated to 65% heads. That's two very different worlds! It turns out that the yudkowsky has an important parameter: how many coinflips he has observed so far, and therefore how much of his estimation comes from the observations and how much from the prior, and for some reason he doesn't tell you its value!
Scott's assertion is correct in a narrow technical sense: in a world where the shady stranger forces you to make a bet at gunpoint, you are forced to use the yudkowsky's estimation and the yudkowsky is forced to use a symmetric prior that gives him a 50% probability of heads when he has not seen any flips at all yet.
However in the real world there's almost always an option to wait and collect more data, and whether you want to exercise it critically depends on the difference between "it's a 50/50 chance based on observing 100 coinflips" and "it's a 50/50 chance based solely on the prior I pulled out of my ass".
So what's going on I think is that people intuitively understand that there's this important difference and suspect that when Scott says that normally they should start with a 50/50 prior, he's trying to swindle them into accepting Bayesians' estimations without asking how sure they are about them. And rightfully so, because that's a valid and important question to ask and honestly Bayesians ought to get a habit of volunteering this information unprompted, instead of making incorrect technical arguments insinuating that the estimated probability alone should be enough for everyone.
official_techsupport
who/whom
8mo ago
I totally disagree with the conclusion. First of all, we are literally living in the time where one man's vision is about to revolutionize space travel by making a rocket that can lift 100 tons of payload to LEO. Yeah it's interplanetary for now, but why not interstellar next, maybe by the next man with an itch for it?
And second, why do you need to persuade the whole society to migrate? Most of the old world people didn't migrate to America and it was their loss. The few people who did migrate multiplied and prospered. "Indirect evidence of extrasolar planets will never be enough" -- for whom? So we will have bootlicking statists like the author waiting for the government to give them credible evidence and orders to go, while adventurous types will be populating the galaxy.
official_techsupport
who/whom
10mo ago
To be fair complaining about lawn aesthetics is just as bad as adoring lawn aesthetics. Same goes for reading all that shallowness and unwarranted feeling of superiority into people you never met, vs being that shallow etc. A house like in the picture is perfectly suited for reading Dostoyevsky in, and that's the important part, no?
I also don't think that there's a problem that warrants a solution, much less a centralized solution. If someone figures out how to use resources more efficiently without compromising much more important aspects of life with their "part and parcel of hustle and bustle", let them try it somewhere! If it works, people will come and other places will emulate it! If on the other hand you operate under an assumption that people are deluded about what is best for them and must be forced into correct living conditions with an iron hand, it's overwhelmingly likely that it's you who are wrong.
On the latter note, in my experience there's an inverse relationship between the quantity and quality of interactions with neighbors and population density, in terms of inviting neighbors to your birthday party or a random bbq, vs not knowing who even lives in the apartment next door. Like, you might think that people living in separate houses naturally become a sort of haughty recluses shunning human contact, while people forced into a sort of human hive naturally form vibrant local communities--nope, for some reason it's the exact opposite in my experience.
I don't know why, maybe it's because a separate house on its plot of land is much more self-sufficient in certain senses, you don't just sleep there, you hang out there in the evenings and on weekends, your kids play there and around there, stuff like that, so naturally you interact with the neighbors all the time. While if you live in a pod and have to go to do all other activities in other designated areas, you just don't get many opportunities to interact with your pod-neighbors.
official_techsupport
who/whom
11mo ago
There's a thing however: the most vocal Holocaust deniers are also very very stupid. I remember arguing with one on this forum and he genuinely didn't understand how fire works, like he couldn't understand that it might be hard to ignite something but after you got it burning it keeps burning if that releases much more energy than is required to evaporate the stuff. That was in the context of whether you could burn a bunch of human bodies in open air with a minimum of external fuel or do you just multiply the amount of fuel modern crematoriums use to burn a single body to ash by the number of bodies. For the record, IIRC my back of the napkin calculations produced like 10MJ/kg released and 2MJ/kg required to evaporate the water when burning a corpse. The holocaust denier never engaged with these numbers.
If you could be teleported to the past and talk to Neanderthals about fire, they would understand it better than a modern day Holocaust "revisionist". Idk, maybe Australopithecs and Denisovans too. Holocaust deniers are inferior to literal subhumans intelligence-wise, and trying to discuss their arguments with them is a waste of time. I tried it again and again just to make sure and no: they are all very very stupid, that's all there is to it.
official_techsupport
who/whom
10mo ago
Why do mosquito bites itch?
Is it entirely accidental, as in, evolution only cared about whatever stuff mosquito inject acting as an effective anesthetic for the duration of the bite, not about what happens next? Or maybe it's beneficial for humans (makes us much more alert and aggressive towards further mosquitos) or maybe even individual mosquitos due to intra-species competition?
official_techsupport
who/whom
11mo ago
For the record, IIRC my back of the napkin calculations produced like 10MJ/kg released and 2MJ/kg required to evaporate the water when burning a corpse. The holocaust denier never engaged with these numbers.
To expand on how I see these things. If a flat earther comes here and not only makes arguments but engages with counter-arguments in good faith, that's all good. If they flat out ignore the strongest counter-argument, because they are literally too stupid to understand how, for example, fire works (which neandertals understood), or maybe even pretend to be that stupid (which is also stupid in its own way), then I think that:
-
It's good for everyone else on the forum to be made aware that the person in question is very very stupid (or pretends to be) so arguing with them is a waste of time.
-
The person in question forfeited their right to be taken seriously by not taking counter-arguments remotely seriously.
-
The forum would benefit from such people being named and shamed for their real or pretend stupidity and be driven away and the land of the heathen consume them.
official_techsupport
who/whom
11mo ago
My cats clearly don't understand how doors work.
They don't even understand that I open doors. One has learned that if she stands with paws on the door and looks at me and meows then the door opens sometimes, but I'm really not sure that she understands that it's I who opens the door, she just knows that "the door's open" tends to happen after "meow near the door". She doesn't understand "push on the door" leads to "door opens", I checked.
How would something like that work for humans? I mean, is it possible to imagine beings that are so above us as we are above cats and who affect our lives in meaningful (open doors etc) but incomprehensible ways? Like I have a literal guardian angel, I pray to it (like my cat meows to me), and then somehow things work out better for me, in an incomprehensible way. How could that work? How could an angel "opening a door" for you appear to a human?
official_techsupport
who/whom
8mo ago
Oh, oh, you had drama with Hlynka, and one that sounds like it involved some very stupid things said?! Spill the tea pls!
official_techsupport
who/whom
8mo ago
The real question we are interested in: "we can have an intervention that would make this black man a productive member of society that you don't even have to pay for, or you can pay $30k/year for decades until he grows too old to do crime".
official_techsupport
who/whom
11mo ago
You gotta admit though, it's a fun contrast between how you diagnosed someone who doesn't pre-match their socks with depression, autism, and a laundry list of other possible disorders, but then admit that you don't understand how someone can have all their underwear in the same color instead of matching it with their visible clothing.
official_techsupport
who/whom
3mo ago
That was one of the objections listed in the post, Scott's response was that you should only be neutral about elementary propositions, not about compound ones ("bloxors are greeblic AND bloxors are grue").
I personally think that this entire kind of objections can be dismissed by pointing out that Bayesian math works correctly and without contradictions, and when looking at actual priors there's not much disagreement about how to choose them either, in practice. Nobody actually has arguments against assigning a symmetric prior to a coin bias, or even can muster a lot of enthusiasm to argue that you should use a gaussian instead of a uniform prior.
People get hot and bothered when they feel that someone tries to hide how much information they have actually updated on and how much is their prior.
That's the difference between socialist and libertarian approaches I guess. A Libertarian seeks to reduce the scope of consequences of decisions to the maker of them. A Socialist seeks to increase the scope until everyone is affected including the people in power so they are forced to make decisions that are good for everyone else too. Or like everyone is forced to talk about it and make decisions that are best for everyone, because everyone's in the scope.
official_techsupport
who/whom
8mo ago
How would this get any good outcome? Does the George Soros fund care about helping prisoners, or do they want to advance their political goals?
They were very successful at electing people like Kim Foxx and Chesa Boudin and George Gascón who kept and keep doing exactly what was expected of them, so if that part were somehow removed and they were forced to select for people who are good at overseeing rehabilitation initiatives because that's the only way recidivists don't get back to prison, the people the Soros Foundation would choose for that would be pretty good at it.
It's, like, I'm saying what I would do if I were the Czar of the US prison system. I'd set some inviolable rules but then let Soros and friends do their best within the rules instead of trying to micromanage everything.
official_techsupport
who/whom
8mo ago
One thing unique to DNA/RNA is that they can be used in two distinct ways: directly copied or interpreted as instructions for building stuff. This is a pretty fundamental property because it allows constructing https://en.wikipedia.org/wiki/Quine_(computing) , with the cell corresponding to the program and the DNA corresponding to the text constant in the program that the cell uses to construct a copy of the cell and also copies and inserts directly. Also it avoids the issue of, how do you replicated a hammer without using a stronger hammer to disassemble it -- instructions for building a hammer don't possess the strength of a hammer and can be examined and replicated easily.
I'm not sure that something that doesn't have this duality can be a somewhat general purpose replicator, not by default at least, and I'd expect any good paper proposing some replicator mechanism to be aware of this.
official_techsupport
who/whom
10mo ago
O, tick bites are a great example because they do not itch pretty much at all (and don't swell at first either), while mosquito bites start itching within minutes. So it is possible to anesthetize the bite location without immediately causing an immune response: why don't mosquitos do that, do they simply not care (evolutionarily speaking) or maybe there's some non-obvious benefit to it?
official_techsupport
who/whom
6mo ago
The brutality and cynical tactics that Hamas uses do lead to them having lower support than they would if they were less sociopathic though.
That's the weird thing though: their cynical tactics used to be launching rockets from hospital rooftops and parading the inevitable Palestinian corpses, or having Palestinian kids shot for throwing stones, etc etc. The grift has always been provoking Israel to violence and posing as an underdog.
But this, parading enemy civilian corpses around, is a diametrically opposite thing. It's something you do when you have several thousands of tanks ready to roll over the enemy capital, you expect to win, and you want to demoralize the enemy to win easier.
So I don't know, either Hamas expects Iran to nuke Israel, or the old guard that understood the nature of the grift all retired or something and the new leadership got terminally high on their own supply.
official_techsupport
who/whom
7mo ago
I'm reading the webnovel A Practical Guide to Evil on @official_techsupport's recommendation.
For the record while I can recommend several webnovels, I've never read this one and also have a very low tolerance for badly written stuff.
official_techsupport
who/whom
8mo ago
I think I can. I'd also ask like half of the people I play Blood on the Clocktower with regularly and they will probably agree and we will make it.
I think that if/when there's a call for it, it won't be the average ability of some particular ethnicity/faith/whatever, but self-selected (but also gatekept) people that are multiple deviations above the average in whatever traits you think are good for manning a generation ship. A random 1000 mormons have nothing on the top 1000 <insert the ethnic group you most despise> most excited about colonizing Alpha Centauri.
official_techsupport
who/whom
8mo ago
By intervention I meant that hypothetical Soros-funded anti-recidivism experiment that funds all sorts of activists trying various ideas.
The alternative to that is the current situation when violent recidivists are in fact locked up for a long time on taxpayer's dime.
My point is that I'm sure that pretty much nobody, including KKK Grand Dragons, hates black people in a sense that they would actively pay to harm them. So we shouldn't worry that our hypothetical program would receive a pushback from the nonexistent group of people that prefers more black criminals around.
official_techsupport
who/whom
8mo ago
Your hypothesis is stupid. It’s needlessly complicated, and paints your enemies as both naive and devious. Why should activists expect results from foisting “the problem” onto an unwilling majority?
Because then the majority has skin in the game and has to deal with the problem. It's naive and usually doesn't work, but if you totally expect it from deBoer, why not from other activists?
Is desegregation really worthless except as a means to an end?
I never said or implied that. But you're conflating two very different things: all the stuff that MLK mentioned, let's call it "negative desegregation", meaning that black people are no longer prevented from being in white spaces, and "positive desegregation" that actively tries to mix up communities, such as busing (read https://en.wikipedia.org/wiki/Desegregation_busing btw, it was an unmitigated disaster opposed by everyone involved expect the actual rich and affluent whites). Can you give me your best pro-busing steelman? Because I'd put making whites have skin in the game wrt education quality on the first place, tolerance through familiarity on the second, and aesthetic preference for de facto desegregation as a sort of a strawman (though no doubt real) justification.
Same for the Soros DAs. I'm not saying that there were devious conspirators planning to have a crime wave, I'm sure that they hoped that there wouldn't be a crime wave! But don't you think that it was very weird to do things in the opposite order to what I proposed? Like, first you figure how to prevent criminals from reoffending, then you go soft on reoffenders because they won't reoffend yet again? If you go soft on reoffenders first, what exactly do you expect? Is "uh I hope someone also figures how to prevent recidivism now that they are forced to" painting my enemies as more stupid and evil than "idk I don't give a fuck lol"?
official_techsupport
who/whom
8mo ago
Because then it doesn't have a moral authority, and you're no better than Yazidis, who believe that Satan doused the fires of hell with his tears and escaped and now rules the world, so you'd better worship him if you know what's good for you.
official_techsupport
who/whom
10mo ago
Scrapes and cuts, especially scabs, itch too.
It's not comparable at all. A cut doesn't begin to itch until after several days, and don't itch at anywhere the intensity proportional to the affected area. Needle pricks don't itch at all and they are tens or hundreds of times larger by area than mosquito bites. So no, it's evidently a reaction to the anesthetic stuff they inject.
official_techsupport
who/whom
11mo ago
If I were trying to invent a leftism capable of handling a workerless future I think I'd go along the lines of some of the recent questions as to how copyright interacts with the output of LLMs. Namely that these things are sampling out entire culture and to some degree we all have some claim to the output, even if it's only through the influence we exert on someone who exerts influence on someone who exerts influence on someone who directly produces lots of the training data.
Longposters shall inherit the earth, yay!
official_techsupport
who/whom
3mo ago
How do I know that "bloxor-1 is greeblic" is elementary, if I am totally uncertain about this proposition, and I don't even understand the terms?
Skill issue.
What do you mean "correctly"?
That I, doing Bayesian math about some bets against you, will leave you poor and destitute in the long run, unless you're using Bayes too. What do you want to use instead of Bayes for the record?
the Allais paradox
My point is not that the poors are always instinctively right. My point is that they have well-honed instincts for when someone is trying to take advantage of them, and the usual Bayesian reasoning like the above rightfully triggers it, even if they don't have the concepts or the introspection to communicate to us what was that, that triggered them.
My point is that a Bayesian megamind is entirely justified in asking the yudkowsky what fraction of his prediction came from the data, and basing his bet amount on that, and grumbling about the yudkowsky being useless if he refuses to answer.
Nobody actually has arguments against assigning a symmetric prior to a coin bias
How many of the arguments in probability theory have you read to come to this judgement? Because I can think of large parts of the literature dedicated to exactly this point.
Huh?
Certain social policies are misguided attempts to ensure skin in the game.
Aristotle I think claimed that most vices correspond to the absence or perversion of some particular virtue. I reread Scott Alexander's review of Freddie deBoer's "The Cult of Smart" and it helped crystalize an observation that a certain kind of approaches to solving social ills is likely to be a perversion of the concept of having skin in the game.
Skin in the game is, basically, the idea that things tend to work out much better when people making decisions are also the people reaping the consequences than when they are not, because otherwise you get https://en.wikipedia.org/wiki/Principal–agent_problem and doctors doing lobotomies on the involuntarily committed and people forcing their politics on culturally different communities and the government spending taxes on catastrophically wasteful projects instead of letting people buy private services with their own money.
Now, Scott REALLY HATES public schools. Literally in CAPS LOCK. So it was kinda funny how he nodded sagely along with Freddie explaining how public schools don't really teach anything, commended his analysis that Montessori schools maybe aren't much better at teaching but at least they aren't DYSTOPIAN CHILD PRISONS, and so on and so forth, until the last part of the review where he COMPLETELY LOST HIS SHIT upon realizing that Freddie's solution is making everyone go to public schools and forbidding all alternatives.
(this was when Scott already mostly calmed down by the way)
But Freddie's approach actually makes total sense as an attempt to force people to have skin in the game. The society is largely controlled by the rich and affluent. As long as they can escape to charter schools, catholic schools, or hire private tutors, they will do that instead of using their wealth and affluence to fix public schools, so those being horrible will remain solely poor people's problem. If we close all alternatives then it becomes everyone's problem and everyone has to solve it.
You can also see this approach in what is currently happening with the US justice system. America has a huge prison population and high rates of recidivism, which maybe could be solved by adopting the Nordic model of rehabilitative justice. But it's hard, it's much easier to lock up recidivists for decades, so that's what the system had been doing until roughly 2018, when a coordinated campaign had elected a bunch of progressive DAs in all major cities, who simply refused to prosecute a lot of crimes. Now with the crime wave affecting everyone people have no choice but to take rehabilitative justice seriously.
Most of desegregation initiatives also had this component to them (in addition to tolerance through familiarity I guess). As long as black people live in their own ghettoes and send their children to their own schools, rich and affluent people by and large don't care what happens there. But if you have gangs selling drugs in your kid's school and a crack den next door, you'll have to care about and fix the problem, right? Right?
Of course all such approaches range from simply not working to greatly increasing the harm they were meant to prevent plus causing other catastrophic consequences. Here's some reasons why:
Just because you incentivized someone to solve a problem doesn't mean they will be able to figure how. Some problems are very hard and you have to try to solve them purposefully instead of setting up incentives and hoping for the best.
Unless you want to live in a North Korea (and can bring it on), it's really hard to incentivize wealthy people to solve problems like that. They'll look at it, admit that they have no idea what to do about it, shrug, and move to another place. So attempting to overmilk that cow will leave you without milk at all.
You are not incentivizing the actual rich and affluent people, you're incentivizing middle class, which is not affluent enough to solve much. Or more precisely, it's the actual rich and affluent people who are doing the incentivizing and they sure weren't busing their own children to mixed schools etc.
Affluent people who end up in charge of solving social ills are usually ideologically incapable of solving them. For example, a school that has problems with drugs and discipline should punish and eventually expel troublemakers, but that's precisely the kids the progressive school board cares about the most, so it would demand that the parents solve the problem with discipline without disciplining anyone, at which point the wealthy parents will shrug and move elsewhere.
Or regarding crime: let's be real, most criminals aren't Jean Valjeans stealing a loaf of bread to feed their younger siblings, they pick $1000 worth of Gucci bags and go do drugs and have fun because it sure beats working a week at Walmart, and that's the truth. If you tell them otherwise they will laugh you in the face. If you ask them to think about the poor Gucci shareholders they will laugh you in the face. The only way to fix them is to promise them a reasonably long stint in prison, at which point our prison abolitionist decides that Gucci shareholders deserve it and secretly gives up on rehabilitation.
Is it possible to force people to have skin in the game in a way that works? Yes, you have to make sure that you're forcing the right people and they can't wiggle out of it. So regarding prison reform again: first we go to the nearest KKK Grand Dragon and ask him if he's willing to pay $30k in taxes to have an innocent black man imprisoned for a year. Having thus established that the vast majority of the society doesn't have real incentives against rehabilitative justice, we greenlight anyone who wants to test their theories about how to rehabilitate criminals--more particularly, we ask the same George Soros fund that elected progressive DAs to bankroll and vet these initiatives, to make sure that the obvious grifters are excluded.
The most important part is that we also pass relatively strict laws against recidivism, say, doubling the term every time. This really incentivizes the anti-prison activists to do their best job trying to rehabilitate their charges. That doesn't mean that they will succeed--that any of them will meaningfully succeed--but they will try their best, and what more can we ask for?
This way instead of making the society hostage to criminals and hoping that someone figures out how to rehabilitate them, we take the criminals hostage and incentivize them and their rehabilitators to succeed.
More options
Context Copy link