curious_straight_ca
No bio...
User ID: 1845
curious_straight_ca
1yr ago
It's not like we can't talk about other things, we just don't as much.
This is why I'd like a BLR back, maybe just with those topics outright banned, as more seeds for discussion on a variety of topics.
curious_straight_ca
1yr ago
The board badly fucked up their shot. I think they had a chance at persuading at least a fraction of OpenAI that sam did some things wrong, if indeed he had. A number of OpenAI employees care about AI safety, and some even expressed online that, initially after Altman's firing, they were open to the possibility that it was deserved. But when the board said nothing for days, even to the first and second interim CEOs, that left the employees with nothing to think about but concerns about their future income and whatever Sam told them.
They also could've spent the preceding months building up a case against sam, maybe picking a replacement CEO who the openai team would be amenable to, etc. But they didn't!
curious_straight_ca
1yr ago
What's the obvious deal that America should dictate to Israel that'd solve the I/P conflict? The only thing I can think of is 'actually occupy, subdue, and govern Palestine', but that's less pro-palestine than US policy currently is.
curious_straight_ca
1yr ago
That you disagreed highlights how Sam's position isn't so implausible that it must be dishonest on his part.
But those who are claiming it was a pretext for Sam's power play have a point imo. The paper wasn't widely read or reported on, even in AI safety nobody had heard about it until this incident. Why would Sam care then? If it was a NYT op ed sure.
curious_straight_ca
1yr ago
Yeah, but, that just isn't true? I've written fiction before, and grand statements like that are very compelling in fiction. In real life, though, I don't really see how the three month old fetuses, or slaves, or even farm animals are all going to be wicked enough to deserve instant annhilation!
In logic, there's the 'principle of explosion'. With a false statement, you can prove anything. We're seeing something - different, but related - here. It's the principle that, to cover for a lie, if you aren't skilled, you end up inventing more and more complicated and implausible lies. God destroyed a city - okay, sure, whatever. But God is benevolent and fair, too! Hm. Okay, A->B and B->C so A->c, everyone in the city deserved it. And now we run into problems - the babies deserved it? The deaf, blind, and mute deserved it?
Spoilers for a 3000 year old story, but Abraham manages to talk God down from 50 to 10 and is then only able to find 1. At which point God basically tells them "RUN, and don't look back".
And isn't that just beautiful! Exactly one! Poetic. How likely is it that that, like, actually happened though? Versus just being poetic?
curious_straight_ca
1yr ago
What I (probably obviously) meant with 'uniformly' is that humans are very varied and it's entirely implausible that a nation wouldn't have over 1 in 20 people who, by your standards, it'd be obviously an unforgivable and titanic evil to murder for their 'wickedness'. The sick child, the quiet boy who took the abuse and did nothing to anyone else, the street beggar, the elder who did his best to guide his society towards a better place within its constraints. The man who does his job and doesn't bother anyone.
Again, consider the example of China nuking one of their provinces because it was 'wicked'. Does this strike you as reasonable, by your moral standards?
We all have it coming.
This just ... highlights ... the absurdity. Okay, God, right now, glasses America. Is this good? Probably not, right? But we all have it coming, so...
You're just giving a billion times as much deference to things in the Bible as you would anything else.
curious_straight_ca
1yr ago
That's all very well written, but I think it dodges the point that you'd consider it the worst evil of the century if the world population, or even a city, or even every firstborn in a city, were suddenly killed for wickedness or immorality. Yet that's what God did, right? Seems pretty cruel. Also pretty arbitrary, there have been a lot of wicked people and most don't get struck down halfway through their lives to make points about morality.
How do we know, if we're not fit to say? What good is praise from creatures who don't understand you? It's like when I worked as a sign language interpreter, and hearing people would come up to me after a job and tell me, "You were great!" I'd say, "Oh, do you speak ASL?" and they'd say, "No."
Notably, Islam also uses this 'mysterious ways' excuse to justify an entirely different set of god-atrocities, so it probably proves a bit too much.
curious_straight_ca
1yr ago
There are plenty of people in history who were wicked who weren't killed by god, though.
curious_straight_ca
1yr ago
Again, it is obviously untrue that the whole world was uniformly wicked and deserved to die, or that an entire city was uniformly wicked and deserved to die. Everywhere has some upstanding people.
Again, compare it to: Imagine if China gassed everyone who played video games, or smoked, or had premarital sex. That'd be pretty capricious, right? How is that different than what God did?
also: this has nothing to do with consequentialism at all. boxing shadows.
People throughout the world grow extremely wicked, God destroys them as punishment while protecting a righteous man and his family. Not capricious.
I find it unlikely that everyone was uniformly wicked, there were probably at least 10% of the population who - by your standards - shouldn't be slaughtered just because they're adjacent to wickedness. Like, how would you feel if China decided one of their provinces was really degenerate and gassed everyone in the province? It's clearly capricious, and you're clearly using a form of reasoning that provides infinite deference to religious claims.
curious_straight_ca
1yr ago
maybe get a new gimmick?
curious_straight_ca
1yr ago
I think people, maybe incorrectly, have a much higher activation energy for three paragraphs of characterization than for just copypasting a link. And that's a good filter for wokes r bad, 500th edition, but here if it would've prevented this whole discussion it's dumb.
curious_straight_ca
1yr ago
FWIW @jewdefender I appreciated you posting that article here and think the discussion it spawned was fine, not sure why you deleted it!
curious_straight_ca
1yr ago
Having the models not say certain things isn't going to stop OpenAI from building AGI. They just finetune the model once it's trained to not say like .01% of all facts. It still has the other 99.99%. It's annoying, but it's not at all fatal.
curious_straight_ca
1yr ago
It was the example you gave. I'm still not sure this is true though? I think it's directionally true - see how many subreddits are banned - but it's not that true. Even pre-elon, there were a lot of nazis on twitter. There were and are a lot of nazis on tiktok. Communists, too. Lots of weird fetish stuff too. Youtube is still mostly native creators, BBC and CNN are a very small fraction of the views.
curious_straight_ca
1yr ago
Where in the article does it say you can't be friends with them?
Friends and coworkers, often people that you or your white friends marry, people you integrate your life with. If white identity means anything at all, it means caring more or doing more with white than nonwhites.
Well, it hasn't worked yet, and it's chances of working sense white identity politics decreases with every million illegals that cross the border.
I don't understand your point here? You're arguing that in the status quo you're losing on demographics. And if all whites suddenly became wignats, then you'd start winning on demographics. That's true. It's also true that if everyone became HBD liberals, you'd stop losing on demographics. I don't see why this privileges white nationalism over HBD liberalism. There needs to be some other argument.
curious_straight_ca
1yr ago
Asians create different civilizations from Europeans who create different civilizations from Africans and so on.
I mean. Do they? I feel like modern China and Europe and America are more similar to each other than they are to any of China/Europe/America 2k years ago. Recently, both Russia and China had relatively similar communist revolutions, and it's China that managed to crawl their way into a functioning capitalist system.
White identity also gives people meaning. We are a tribal species. We find meaning in groups.
Even assuming we lump Jews into whites, 'white identity' still cuts out like 1/5 of my friend group and people I like online. Sure, those are mostly Asians and Indians as opposed to hispanics/blacks.
There certainly doesn't seem to be anything stopping Indians or Asians (or Jews) from mastering online dissident-right ideas and twitter reach, if you go by demographics. If anything they're overrepresented, as one would expect from IQ.
Why should I care about whiteness individually, when Indians, Chinese, Japanese, some south asians and middle easterners, seem to be perfectly productive and intelligent citizens?
The costs of not embracing white identity are already colossal. Whites are losing control of their own civilization; they are a quickly dwindling and despised minority in the countries they created
To the significant extent that that's true, it's entirely of their own desire and creation. Embracing white identity here is a blunt proxy for rejecting wokeness, liberalism, whatever you call it. I think Philo would admit he'd have no reason to embrace white identity if white/asian/jewish/indian elites had no interest in uplifting blacks or mass hispanic immigration or w/e. But ... why not just ... directly advocate for that? And if you can't do the that, you can't do the much tougher task of embracing white identity anyway.
So, uh, how did Yglesias's clearly uncharitable interpretation of things get blasted to the top of the most popular conservative leaning news aggregator? It feels like the Gina Carano situation all over again.
I mean the first tweet in the thread was
To the cowards hiding behind the anonymity of the internet and posting "Hitler was right": You got something you want to say? Why dont you say it to our faces…
And elon replied to a response.
It's not that uncharitable.
But when you look at other tweets by the guy who Elon agreed with, he's very explicit that he thinks antisemitism is bad, and is just criticizing a specific trend in jewish activism.
I think it's more Elon just having the kind incoherent idiosyncratic political opinions that normal people have than being coherently a cryptonazi. Later in the day he approvingly replied to a tweet about how racism is always bad no matter what.
curious_straight_ca
1yr ago
If you don't understand it, there is no hope of coding for it.
Huh? We don't currently understand how GPT-4 works. Evolution didn't understand how biology worked either, it just randomly mutated and permuted stuff. That's the template here.
evolution is not a person, it's a process, one which we cannot replicate in a practical capacity with LLM.
Why? Why can't we do a ton of FLOPS to do evolution on neural nets? That is, kind of, what current ML already does.
the thing you don't understand about regulation is that, more often than not, it is used by the incumbent actors in a space as a barrier to entry for new and more agile competitors. There is a reason Altman is all for it and in the same month there was a leaked Google memo that basically said that OAI, google, facebook et all didn't have a moat.
"more often than not" isn't confident! I think over the long term there's strong enough incentive for AGI someone will get there.
curious_straight_ca
1yr ago
-
... yeah, and yet we still manage to have agency and reshape the world? I don't understand your point. Current AI methods are more 'evolve a huge complicated weird thing' than 'understand how it works and design it'.
-
evolution did it, why can't we do it? Even if some new thing above neural nets is necessary, we're going to work very hard on it.
-
this is the same thing as 'a cop killed a black guy wrongly once so all cops are racism'. you're comically overgeneralizing a newsworthy culture-war-adjacent event to everything. Regulation has opponents, opponents that care more about big piles of money and power than saying bad words online.
curious_straight_ca
1yr ago
Middle schoolers still share weird shit with each other in private discord servers/dms. I'm not sure if the prevalence of that over the whole population has increased or decreased, but it's still there.
Here's how I think about this:
Just in my personal experience, HEPA filters seem to help with allergies, and visibly help filter out smoke when it's detectable by smell.
But let's ask The Science. What do we see? The first question to ask - does it remove microscopic particles from the air? If it doesn't, we can just stop. This is something I'd expect good data on, it's a (relatively) hard-sciencey question. Looking at this and this, HEPA filters seem to reduce particle concentration (of pm2.5 and bigger particles, which google says is how big allergy-related particles are) by around 50%. Which is substantial, but intuitively seems like less than you'd want. Of course, though, this depends on exactly what HEPA filters they used, how much ventilation there was, what kind of particle, etc. The first study says "ventilation was through the door" - does this mean through gaps in the door, or was the door open? idk. The second study also compares having multiple HEPA filters to one, and three filters on 'medium flow' seems to (eyeballing a very poorly constructed graph) reach 75% reduction in particles. Just intuitively, by having a good model on a (somewhat noisy) high setting, and either having it in a small room without too much air exchange or having multiple for your home, you'd get meaningful (80%+) reductions.
The second study does note some increases in ion concentration (not particle concentration) in HEPA on scenarios. But the increases are only present for some ions, in some particle size categories, so I think it's just noise.
I'm not really sure how much to trust random studies like this anyway. One has authors from "Dr. Bhimrao Ambedkar University" and "Savitribai Phule Pune University, Pune" and the other from Taiwan. The Indian study mislabeled several of their tables to seemingly show that air purifiers double particle concentrations,.
Okay, just from personal experience I'm pretty sure it removes particles, and the science seems to confirm it removes the right ones. But inferring that it stops allergies from that isn't particularly justified. What about studies on allergy symptoms? Well, there are a lot of positive results on google scholar. But they're all positive results on some measure but not others, and they're all either n=20-40 RCTs or meta-analyses of 10 n=20-40 RCTs. I don't think one should believe any of these, either as evidence for or against.
So, uh, my guess is if you pick the right ones and use it correctly, they're effective, mostly because some people I know claim they are. It looks like good ones are $100-$150 on amazon, so if it doesn't work it won't hurt too much. I'm not sure how to match the airflow rate or w/e to how big your room/home is, there's probably a guide on that somewhere.
But if you need increased airflow, that'll significantly reduce the effectiveness of the air filters. Certainly airflow from outside, and if the AC is introducing new particles from itself or outside, as opposed to just recirculating them, that'll hurt too. I guess it'll probably still help, but idk either way.
curious_straight_ca
1yr ago
It's a lot more reasonable to demand clear boundaries and communication in BDSM. You're combining acts that can cause strong negative feelings, physical restraints or intentionally caused pain that can cause permanent damage if you do it wrong, with roleplay where you intentionally ignore the usual ways of judging if something's gone wrong (ignoring physical resistance and asking to stop, enjoying pain, ...).
Whereas in a TTRPG it's literally just words.
curious_straight_ca
1yr ago
For example, it's become very common to put question marks at the ends of statements to indicate uncertainty
I do this in text conversations. It takes the place of a certain tone of voice, and is just useful. It's just language changing, 2000s english has a number of errors by the standards of 1900s english, which has errors by the standards of 1800s english, and so on and so on in an often continuous process until eventually it's mutually incomprehensible.
No one seems to know how to spell led, no one, all right, or its
I think people, especially dumber people, have made these sorts of mistakes for a long time, but 1) we see a lot more writing by average people than we did in the past due to the internet, and 2) typing is a lot faster and more casual than writing with a pen so people care less about little mistakes.
Also, though, english spelling conventions are dumb. Why not just have a simpler, understandable correspondence between phonemes and spelling? Like, chinese characters were horrendously complicated until they were Simplified, and there's still a lot more complexity there than is necessary, but it was a net benefit.
Compare a newspaper article or even worse a scientific journal article from today versus 70 years
Not sure about the news, but there is a lot more science being done today than there was 70 years ago, so the % that are high human capital necessarily decreases.
Concretely, I thought the Aporia article he posted was a perfectly reasonable thing to post.
More options
Context Copy link