@Lykurg's banner p

Lykurg

We're all living in Amerika

1 follower   follows 0 users  
joined 2022 December 29 10:51:01 UTC

Hello back frens

Verified Email

				

User ID: 2022

Lykurg

We're all living in Amerika

1 follower   follows 0 users   joined 2022 December 29 10:51:01 UTC

					

Hello back frens


					

User ID: 2022

Verified Email

Of course it's not exactly hard to figure out why that might be

And if there were a lot of overlap, it would not be hard to figure out its because the revisionists are far-rightists and so obviously carry water for Putin.

If left and right are different versions of enlightenment philosophy, this leaves open the possibility of an opposition from outside the enlightenment. But you dont seem to think theres anything in that box. Anytime you argue that someone is not a real rightist, they are placed firmly in the "left" box. Why do you think that is?

The joke is saying theres this opinion that conservatives have thats getting censored, and they dont want to admit that opinion. Getting censored applies to all social conservatism pretty evenly. Progressives accusing conservatives of secretly holding some opinion can apply to all of it, but applies most commonly to race. The last line is saying that we should have enough information now to know which one was meant. So the author thinks only one thing meets the previous conditions sufficiently. And race meets them the most, so it can only be that.

I’m reminded of a tweet:

[Image not reproduced in comment]

If youre familiar with brain-teasers, theres a certain gimmick where a puzzle isnt uniquely solvable until you get told that it is, usually by someone being said to know something or through a subtle use of the definite article. The joke in the tweet has this, too. It could be any sort of social conservatism, until the last line tells you its about race.

This is a great question, and most of the disagreement youre getting is just insisting on the Enlightenments self-presentation with little argument. Sad.

That said, I dont think "We know how to solve all our problems." is a plausible candidate. First its a very simple idea, thats easy for lots of people to stumble upon in lots of situations. Secondly, confidence can grow quickly, and shrink quickly, too. So if that was what the enlightenment was about, it would not be a historical trend with a definite starting point, it would be something like "Cannibalism in crisis" or "Wars of succession" that pops up occasionally and peters out again.

I think that, at least as far as politics is concerned, a good summary of the Enlightenment is Cartesian dualism. It leads to things like the original position, our definition of authentic desire, people apparently appearing ex nihilo as fully formed adults, "What if you had been born an X", and most of the other driving arguments of "progress".

Living out one's ideals is a costly signal of sincerity, and achieving success and happiness by doing so is the least refutable argument. This is a big reason why religion is so persistent despite sounding batshit crazy from the outside — and I say this as a religious person.

Is it? AFAIK religion is negatively correlated with most measures of success before you correct for income and education and such. Now, it might be that those corrections are necessary to find the true causal effect, but its clearly not just "follow what the successful people are doing" any more.

In fact I think this works against religion. Far more people avoid religion because its associated with low-status people then would ever care about whats objectively reasonable.

IMO this is just people not believing AGI is possible, or only believing it in the sense the physicalism requires them to say so.

There's been some research to check for transfer to offline environments

Well yes, if we believe in reinforcement or some other mechanism like that, that can carry the short-term consequences into the long term. But there the proteus effect is not an alternative way that the character can take over long term. All the stuff about the mechanism of it suggests it doesnt have an independent long-term effect.

isn't it just going to end up with everyone in an echo chamber?

I think its less of an echo chamber than sorting by upvotes. And as I said:

Personally I would like to see a replacement for sorting by new. There are fewer deep-in-the-tree discussions and more first-order replies without further replies than there used to, and my impression from memory is that this came gradually after the change to sorting by new.

So clustering could hopefully solve that with less of an echo chamber effect. But if you weigh the risks of partisanship vs declining quality differently, maybe you do want to stick with things as they are.

as long as you're reasonably experienced with Python. You do get to learn Python if you don't know it already :V

If you think thats the time-consuming part, then the whole thing doesnt sound too bad. I hope Ill get to it after finals season.

So I'm currently making no attempt whatsoever to cluster users :V

Well, Im glad I gave a new idea. Feel free to ask me about math details.

You could have made that argument about things that actually happened as well though. It needs to get a foot in the door somehow first, unassited, doesnt need to be very big or mainstream. Then I think liberals would support it. Not to increase minorities, or because they believe being gay is better, but in the same sort of way they do with trans now.

Interesting topic in a thread below

Consider: a society of just downies and Henrys* wouldn't even be a society, while a society of Enron, Google, and AXA is just ... our society.

A society of my aunt and Henrys would necessarily devolve into hunter gatherers who would be in a precarious position.

A rival hunter gatherer society of entirely Enron, Google and AXA professionals would be a tribe that my retarded aunt and Henry with comparable numbers of similar nature would probably subjugate easily, eventually integrating violent strong men or wise old women, humiliating the rest in servitude.

I think this is probably a point of disagreement for many here and so worth discussing on its own. I see two larger topics that this could become a test for: One is the model of "general competence"/IQ maximalism, expecting successful people to be successful at ~everything, vs a tradeoff between abstract thinking and practical or social skills. Second, whether our current elites are in some sense a paper tiger - bullshit jobs, Overcredentialism, etc.

*"I had a patient, let’s call him ‘Henry’ for reasons that are to become clear, who came to hospital after being picked up for police for beating up his fifth wife."

There is but it's not particularly relevant to this discussion

I wasnt particularly disagreeing with you; I genuinely would like to know what you think is in there.

secular progressivism with at least two-scoops of Marx and Hegel

How much do you know of Hegel? My impression is that while he caused a lot of brainrot, you are closer to his object-level positions than to Hobbses.

Also Holy thread necromancy Batman.

Im here so rarely now, I pick out the pearls.

Having just returned after a while, I notice that theres no easy way to find recent quality contributions posts from the front page. I know to look on reddit, but maybe we would want them more prominent for new users?

The higher IQ applies to the ashkenazim and is thought to be from selection in the late middle ages and after, but the pattern of concentrating in certain elite professions and the majority getting mad about it applies to jews much more generally, and so is presumably not explained by it.

When the ADL puts enormous pressure at the highest levels of power to "Stop Hate", is that progressivism masquerading as ethnic power, or is it ethnic power masquerading as progressive morality?

But I think you will agree that the ADL didnt get its power from "fierce advocacy". The advocacy and the being-persuaded-by-it are fake. My point is that "doing identity politics" suggests a pretty specific plan of action: You want to be very loud about how your group is treated badly, maybe have an organisation dedicated to that, make an ethnic voting block, etc. But those parts are kayfabe, they dont actually make you win. Now, maybe you mean something else by it, but if so its pretty prone to misunderstandings, because I still cant tell what it would be after rereading your comments with the assumption that its there.

This is very much in accordance with how average German, British or Hungarian liberal will be obsessed with black people but ignore or even just dislike the (sometimes much larger) Turkish Arab or Albanian populations in their midst.

The turks do have a kind of minority politics. Its a lot less intense than the US with blacks, but they have their highly credentialed representatives that get a good bit of stage time and diversity-grants, and theyre a topic in political discussion. The mindset you describe exists and is something you might filter into as a visitor from anglostan, but its pretty niche. So I dont think the atlantic fully explains the situation with the gypsies.

The philosophy may still be sound, we don't judge the art by the artist

An important part of the philosophy here is the claim that you can improve rationality in a domain-general way. That you could learn to avoid e.g. motivated reasoning in a way that would work on all topics simultanuously, so that your preformance in even the weakest field that a critic might adversarially pick will be ok (and that he has done this, obviously). Claiming to have a metabolic defect that would be lethal in the ancestral environment is strong evidence against that.

people do seem to behave differently based on their avatar

As I understand it, this would be in effect only while you wear the avatar. I interpreted the sentence I responded to

Outside of the more out-there therians and actors, though, this can be hard to notice from the outside, and harder still to distinguish from normal personality changes from simply being in these environments.

as being about long-term effects. The short-term effects are interesting, but I dont see how they would lead to the character taking over in the off-time.

Closer to the central claim, though, I think there is some difference between, for example, playing a character that is foo and doing foo, for wide varieties of characteristics. The latter probably is better at encouraging that specific action! But the former makes you think about the broader characteristics and motivations and how all those things would interact. Which, to be fair, is still a new behavior that's established and getting reinforcement. Just a different one.

I somewhat agree, depending on what you imagine for "just doing foo". If you get told what to do over earbuds, thats less dangerous than "playing a character" normally. I would say this is because in the latter case youre figuring out what to do, and that way of figuring out can be reinforced. I dont think its essential for that figuring out to involve thinking about some character.

And I think this is essentially the same way normal behaviour changes in an environment: You go in with somewhat different mood/disposition each day, and some of them get more positive reinforcement than others.

BTW, I think often doing a specific action is not the best way of encouraging it. Many actions lie at a point within the decision tree that youd never normally get to, and training that last step more wont help.

Well, obvious part is getting the regressions for (user, judgement) against various mod decisions. Getting interactions between different users is propably not feasible, but we could try clustering them. (Which reminds me, is there still any interest in this?) This could let us consider interactions between those, and help assign comments to volunteers to get a representative quorum faster. Do you think you can afford to double-check all decisions that went against the user? And there should propably be some report number over which to check approvals.

How would you document actions the system took? You wanted people to not know they had an impact, but the modlog would obviously show the action as not taken by a human, and however the modhat comments are handled would show it too.

I checked the modlog for this and have some suggestions there as well: First, bans arent robustly connected to their modhat comments. If you view all actions you can tell chronologically, but filtering for bans you dont get it. Also there apparently arent modhats for removed comments, or did I just not find them? No list of currently banned users either. I think it would also be nice if the list where you select a mod action to filter for had little numbers showing how many of each occured in the last month or so.

Also: I know how to programm in principle (coming from mathematics), but I dont have experience with git, interacting with databases, etc. How much of a time investment do you think it would be before I could contribute to dev?

I mostly agree.

nor any interesting non-additive effects

I dont know what people have with non-additive effects. In a highly polygenic trait, non-additive effects of genes are hard to detect because theyre almost certainly irrelevant even if real.

Would they be better off – in the expected direction of less dysfunction – than the baseline, or rather, than random implanted embryos? You can bet on it.

That is mostly what I expect as well. Im just saying that theres a lot more evidence that it "might as well be" IQ/genetics, than that it actually is.

Re the first part, I think your reasoning here depends on the directions orthogonal to beauty still corresponding relatively closely to terms in which we normally think about art.

the SF Federal Building, the Toronto subway sketches and the MLK Embrace statue all achieve their hideousness in unique ways, and all seem to strive toward various other indicia of elite art

Do I read correctly that you think its possible to make something thats clearly art of our current elite and also beautiful?

We also hear that beauty is consumerist, looks cheap, is reactionary, means embracing an aesthetic of a white supremacist past, etc

What did you have in mind with "looks cheap"? Are there really people who would say e.g. the Lincoln memorial looks cheap?

"Reactionary" here means basically the same thing I did with "fascist", and the association with bad old times is somewhere between made up and self-fulfilling, so it cant be the cause of the dislike.

I think the shared core of the argument is: An account of whats good for an animal based on what that animal itself pursues, in terms that are causally relatively close to perception and behaviour, and independent of their actual environment, and a claim that we have a responsibility to individual animals to do whats good for them.

Its not important for my argument whether these actually are wrong.

When was the vaccine mandate rational? I remember when the debate got big here in Austria, there were already multible countries with 90%+ vax rates that had new flareups.

It depends. Presumably you can also regain the capacity by practicing it again, for example, and in that case the longer time-horizons wouldnt care it went away. And if you set it up in a way where it did matter, then probably your capacity to slavishly obey someone would matter in a similar way. The formalism youve found just isnt particularly related to your problem, and if you find a way to make it do what you want it will be mostly your additions that are doing that work.

I guess you can just make a spectrum argument where the further away from our thought and society the harder it is to evaluate.

I think this is clearer than it may seem. The field itself tells you what is an improvement, and an improvement is within the field. It may look very different afterwards, much like the output of a programm may look very different after youve corrected an off-by-one error. This can be difficult to judge, in that it requires deep understanding of the field, but its not unclear. Of course, minor drifts which are not improvements happen all the time, but while a tradition is alive and well, they are eventually corrected and do not accumulate into larger drifts the way a random walk would. When it no longer does this there is simply no capital-H Historical period for a while. They dont need to be everywhere, just like not all of nature is alive.