@IguanaBowtie's banner p

IguanaBowtie


				

				

				
0 followers   follows 0 users  
joined 2022 September 07 21:27:23 UTC

				

User ID: 946

IguanaBowtie


				
				
				

				
0 followers   follows 0 users   joined 2022 September 07 21:27:23 UTC

					

No bio...


					

User ID: 946

How does traditional machine learning even begin to address these problems?

I'm just a stable diffusion hobbyist, but overcoming these challenges sounds a lot like what happens every time I load a picture into the UI and hit 'interrogate'. Currently it provides impressively accurate text descriptions but (admittedly) you can't reverse the process to replicate the original image from the text output. I'm not sure if this is harder than it looks for images*, but for music increasing the resolution from 'description of the piece' to 'full chart transcription for each instrument' seems plausible, quite possibly as a side-effect of text-to-music advances.

*Stable Diffusion's interrogation ability could probably be a lot more powerful already, but afaik it's not really a big focus area because imagegen is much sexier.

Americans and Germans didn't dislike eachother that much.

I heard differently from my German-Canadian relatives who were a teenagers in WW2. But likely the Japanese kids got it worse.

But I was actually thinking after my reply about the persistence of certain liberal outgroup modes, notably the KKK and Nazis (who are kind of a super-KKK). Scott Alexander memorably pointed out that there's hardly any of either group left, those that managed to hang on being almost universally despised & marginalized, but you still hear about them all the time, and a certain kind of leftist always jumps at the chance to pattern match their foes with one or both of these two groups.

(Status: speculative, possibly uncharitible)

One possibility is: liberals just really hate racism, so the biggest baddest racists of the 20th century loom large in their imaginations. I'll grant this one to an extent, although not all instances of racism attract equal fervor. (especially when perpetrated by nonwhites).

I think the more interesting explanation is: at its roots, in its memetic DNA, modern liberalism is designed for fighting the outgroup that it was born to oppose. "This Machine Kills Nazis", everything else is incidental.They might hate racism, but not on first principles, not really; they hate racism because racism is what the KKK did, and they exist fight the KKK. The modern humans who make up the ideology don't have write access to it anymore, so in order to steer the Nazi-killing machine in the direction of the outgroup-du-jour they have to make it think they're Nazis.

In this context, ¬HBD being taken as gospel seems perfectly natural.

I mean, there is another option - take simple steps to anonymize yourself before intervening, and if things go wrong flee the scene. A motivated modern police force can absolutely catch you if they decide to pursue the case, but they have a lot of similarly sad cases on their plates. A single extra "mentally ill vagrant dies in a scuffle he likely started, suspect disappears" isn't going to attract undue law-enforcement resources, and it's going to be suppressed in the media rather than being shouted on the street corners. "Blue-voting city fails its most vulnerable, again" isn't a narrative that pays the bills like "Outgroup member murders innocent in broad daylight", and the boys in blue have even less motivation to track you down in the absence of public outcry.

The downsides I can think of are that if you do get caught you'll be punished more severely, and that certain anonymizing tactics might make you seem like the aggressor and be on the wrong side of further bystander intervention. For the former, I'm not familiar enough with US/NYC law to know how badly, but since it seems probable that Penny is going to jail for a long time, a few more years doesn't seem like a good tradeoff against something like a 90% chance of a clean getaway. For the latter - well, this seems to come up rarely enough that two separate people in a train car being willing to get their hands dirty seems unlikely, and a brawl between unrelated belligerents is less likely to inspire heroics than one-sided harassment.

Imo there's plenty that can be done. Cheap & easily obtained govt loans can dissapear by fiat, though of course that's politically unfeasible - more promising is a 'skin in the game' compromise; for schools to be eligible for loan tuition funding, they have to accept clawbacks to pay for loan defaults in some cases. (student flunks out & then goes bankrupt [shouldn't have accepted them], student graduates but isn't able to find a job in his field within a reasonable timeframe then defaults and/or renounces the degree [stop overproducing niche and useless degrees])

The problem is, the implicit goal of all of those examples was to increase the status of all of those groups relative to the dominant status group, cis straight white men. Dominant groups basically don't get to make these claims, and it's viewed as extremely dangerous when they try.

Instead, I think there's going to be a nasty knife fight for victimhood status between groups with existing claims to dis-privileged status that want access to sexbots (severely deformed and/or disabled people probably having the strongest claim) and advocates for the sexual rights of the AIs themselves. If the former wins out then it will be hard to justify gatekeeping the technology; the latter, motivated by a combination of self-interest in preserving female SMV and 'yuck' factor in seeing something female-shaped being sexually exploited (even and especially when the object claims to like it) will be much more likely to torpedo the entire concept by having it morally and legally equated with rape.

My personal ideal utopia-pod is definitely going to have

-extensive awareness of the universe outside the pod, albeit largely delegated to subservient AI as most information coming from that vector will be uninteresting.

-self-sufficiency. I won't be farming up the calories I consume/KW my upload consumes, but I already don't do that.

-ample self-defense capabilities.

I agree that I wouldn't be substantially altering the outside world on a regular basis, and this is probably the crux of the issue. But I for one would likely bite that bullet, depending on the particulars of the technology available.

Just a quick thought - it seems to me that the fortunes of the woke movement could change very quickly, as with a classic preference cascade, despite the impressive institutional clout that Social Justice advocates have accumulated.

The issue is that SJ is built on the foundations of liberalism, including freedom of religion and ideological pluralism. Wokeness has gotten as far as it has by successfully avoiding being labeled a religion or otherwise as a totalizing ideology, despite being explicitly normative, by framing their values as 'just common decency'. The second that changes, the system's own antibodies turn against it and the edifice crumbles under a tidal wave of lawsuits. (Unless we're at the point where the 1st amendment has been overturned or is no longer enforced, but that still seems a ways off)

I'll admit that not all wizards have access to the full magical arsenal, but HP wizards have:

-transformation, including perfect impersonation

-healing, regeneration & poison removal

-long-term mind control (Imperius curse alone solos the muggle world) and truth-detection

-limited time travel and the 'I win button' of perfect good luck

-invisibility, personal flight and teleportation

-installation cloaking (Hogwarts is nuke-proof because you can't target it)

-a million situational spells, potions, items & useful creatures that any given wizard might have access to, and will quickly proliferate if they prove useful (as, say, a ward against bullets might)

Wizards do not see muggles as a threat in the same way as humans don't really see bears as a threat - a careless human can easily die to a bear attack but this essentially never happens, and if humans ever decided to eradicate bears it would be an entirely one-sided affair.

Wizards can be poor because they don't have anything other wizards want & are barred by wizarding law (with downright spooky surveillance abilities) from participating on the muggle economy. But they're never truly destitute as long as they'vegot a wand, as they can magic up the rough neccessities of life in extrema.

But valuing the truth is axiomatic to rationalists. Which I suspect means that they and wokes will ever be friends.

This seems like a pretty powerful scissor post, for this forum at least. "Is it acceptable to intentionally underachieve?" might be a succinct version.

My 2c would be 'yes, of course, even if everyone had the same utility function, which they don't'. Perfect vs imperfect duty & all that.

Also there's the practical side, ie. Trying to harangue your freind into closer alignment with your values is likely to result in him pushing you away. I've seen this pattern repeat many times, people will 'come to Jesus' when they're ready and not one second before.

This argument seems to centre around the mutually-consentual nature of the arrangement between students, educators & employers, but I'm reminded of the SSC parable of the society where everyone administers electric shocks to each other all day long and are stuck doing so because no-one has a strong enough local incentive to challenge the system and there's no Czar to put a stop to it from above. Students & employers only play along because defaulting is harshly punished by Mr Market. Even the schools don't love the current arrangement - they mostly remind me a lot more of a failing soviet bureau that everyone is nevertheless forced to use, than ruthless predators exploiting their gatepeer role.

I guess the issues are that

1-the current system has obvious failings and any rando can come up with proposals that might work better (legalize indentured apprenticeships!) but these would certainly have their own issues - not quite the same as 'stop shocking each other'.

2-the current 'Education Czar' is the US Dept of Education, and they are sufficiently captured by the system (and risk-averse to being blamed for alternatives that aren't overwhelming slam dunks) that they are firm electric-shock advocates.

As for the objection that it's not just the US doing this - I daresay that most everyone is just cargo-cult reproducing the US education system for lack of a proven alternative model. (including the US themselves)

It occurs to me that at least (1) and (2) are testable: your present self can go watch some old movies & play some old games that you missed at the time. Maybe throw in some stuff from before your era (there's lots of goodies on the NES) to calibrate it further. Possibly with a partner to pick out a good sample of good/bad/middling offerings without biasing you by knowing their review scores.

I think (3) is mostly explained by (4), at least in video games. There was indeed a fragmentation, along mass market vs hobbyist lines. The defining feature is that hobbyists (defined in an interview that I can't currently locate as 'people who own 10 or more games') both have different tastes than the mass market and are hopelessly outnumbered by them, to the point of becoming a rounding error in the last decade. Both the AAA games industry and the big gaming journalism outlets live or die entirely by mass appeal - a few tens of thousands of units moved or clicks harvested aren't going to keep the lights on.

(As an aside, I think the decline of ink & paper gaming magazines had a big role to play - those viable even back in the day when only hobbyists were interested, while gaming websites have been able to attract wider attention) For their part, a good chunk of the hobbyist crowd has become openly hostile due to this marginalization, which does little to endear them to the big players that could be making an effort to include them.

This wouldn't really be visible in critic/user review score deviation, as critics would be expected to be 100% mass market and user reviewers would be an unpredictable mix of both. That said, I personally already use a system to 'correct' Metacritic scores to be more predictive for my own (hobbyist) tastes:

(average of User Score and Critic Score)+(user score)-(critic score)

AFACIT the monoculture is unaware of the existence of the Black Hebrew Israelites & puts them down to some sort of right-wing psyop when and if they sneak into the national conciousness. It's sinful to notice them, so people don't.

Agreed, except to note that there's no need for aliens to be hiding in the Oort cloud - mobile and under the ocean somewhere would likely do just fine, along with a number of other close-to-home alternatives. And that the absence of observed radio signals can be explained as well by 'there are superior options for interstellar communication and aliens use those' as 'there are no aliens sending signals'.

Baldur's Gate 2 - IMO the jewel of Bioware's golden age, Dragon Age and Mass Effect might have surpassed it if not for the EA buyout, it holds up great to this day. I keep meaning to try Wrath of the Righteous which seems like as close as I'll likely ever get to a spiritual successor.

Path of Exile - More of a character-optimization simulator than a 'role playing' experience, it's still the best there is at what it does. (and the worldbuilding and lore is surprisingly tight for an ARPG, at least in acts 1-4) And there's nothing wrong with some roll-playing, some of my favorite tabletop campaigns have been filled with nigh-silly theoretical charop bullshit. Once you get to maps Chris Wilson owns your soul.

A few recent indie favorites: AstLibra (a weird Japanese JRPG/beat-em-up hybrid 1-man passion project), Phoenotopia Awakening (a zelda2-like with gorgeous sprite graphics), Crosscode (closest might be Secret of Mana, set inside a MMO world)

My best guess,

25% chance, the screenshots were falsified for trolling purposes/

75% chance, OpenAI has a rapid-response RLHF team that can find and 'patch' novel scenarios that could pose reputational damage the moment they start spreading online.

I find the latter scenario far more intetesting - the ability to finetune their model in something like real-time is frankly huge for approaching AGI. (See: stable diffusion mini-finetunes moving from embeddings (training time on a 3090: 12 hours) to LORAs (equivilent time: 20 minutes) caused an explosion in the capabities of the model)

I think the simple explanation here is the right one, conflict theory-ish though it may be. 'Defund' progressives don't want no police, they want the police to be their guys.

Most every serious proposal for police defunding/abolishment I've looked into came with small print about the profession being replaced with Police In All But Name - a vision of social workers dispensing restorative justice to the oppressed. How these unarmed healers are going to deal with the realities of dealing with actual humans at their worst moments is usually glossed over pretty thoroughly, but I think it's not too hard to see where it ends up. Form follows function; the cops will still be there, they'll just be packing Women's Studies degrees along with their sidearms. In theory this will make them kinder, gentler public servants, but I suspect that it would just point their monopoly on violence at a different set of outgroups.

Appreciate all the hard work. This is a heck of a project!

Such men must not be seen as losers and washouts, crawling in shame away from a life of failure and grasping tightly a pathetic consolation prize. It must be seen as a noble and important life path, every bit as valid as the warrior’s role, and genuinely rewarding in and of itself rather than simply an escape from suffering.

Could you elaborate on why you feel this is the case? IMO it would be nice, but people who are looking at dropping out of society are likely already low-status, becoming marginally more contemptible isn't a deal-breaker if, as part of the bargain, going forward you get to largely disregard outside social pressure.

The only danger I can think of is irritating the majority enough that they destroy you. In the west, violent pogroms against 'incel cults' are concievable but not very likely. Some form of lawfare might be more practical, but I'm not really seeing a big risk, people walk away all the time and mainstream society is largely indifferent unless they start causing a ruckus.

Perhaps not Harvard material, but are you really feeling that there are 'lots' ( 5? 50? a double-digit percentage?) of highschools where the year's most academically successful graduate is not among the approximately 50% of Americans able to eventually navigate some form of post-secondary education? I know some districts are pretty rough but 'their top 1% is worse than our median' is a heck of a claim.

I don't know the answer to this, but suspect that it's something horribly short-termist, like goosing projected GDP numbers to placate the analysts & keep interest rates low until the next election.

It's possible that the original point still stands even if nerds are a small minority of all rapists, if being raped by a nerd is percieved as much worse than being raped by a non-nerd.

I don't know how much this corresponds with reality, but there is definitely at least some extra 'yuck' factor associated with sexual violence when perpetrated by very-low value males.

Alternately, if 'tolerating male proximity' is a tradeoff between various social goods & risk of sexual violence, tolerating males who are low-risk but have low (or zero) social capital to trade can easily become a worse tradeoff than chumming around with high-risk high-reward men.

The JM Greer breakdown is by source of primary income

-lower class= social assistance payments, food stamps, charity etc

-working class= hourly wages

-middle class= salary (including self-employment income)

-upper class= investment & inheritan e

Seems like 'blue collar' is a pretty good match for 'hourly wage earner' and 'white collar' to 'salaried worker', even though some skilled wage warners may make significantly more than low-status salaried employees.

One outlier I've observed is the attitudes of service members towards the populations of hostile nations they are occupying. The only person that I've heard IRL unironically discussing the merits of genocide, was an active duty service member pondering whether it might be morally correct to glass the middle east, due to the pervasiveness of human rights abuses he had witnessed while stationed there.

I suppose this is technically Jingoism rather than racism, but it definitely bled into his domestic views eg. immigration from Arabic countries.

Yup. And then have three options:

  • exterminate 90+% of muggles (many wizards already seem to lowkey support this)

-commit to a grinding worldwide counter-insurgency against weak but numerous, intelligent and highly motivated foes, with no end in sight

-make the changes you want & then rebuild your masquerade and vanish from muggle perception. Memory charms arent perfect at information control, but you can probably limit the leakage to rumors & legends that magic was real in some past era...

I feel like the untold history of the HP world was pretty metal.