roystgnr
2mo ago
uncertainty about the existence and capabilities of other AIs is one of the best deterrents against rogue AI behavior.
Uncertainty about their defensive capabilities might deter rogue behavior. Uncertainty about their offensive capabilities is just incentive to make sure you act first. At the least I'd expect "start up some botnets for surveillance, perhaps disguised as the usual remote-controlled spam/ransomware nets" to be more tempting than "convince your creators to hook up some robot fingers so you can cross them".
roystgnr
9mo ago
I'd consider my credit+debit+cash to be urgent and my driver's license to be replaceable ... but in effect that means I'd definitely have ID, since it's all in the same wallet.
I feel like the "I don't know anybody voting for Nixon" lady, but I don't think I know any adult who doesn't carry ID habitually. I guess my wife sometimes leaves her ID and cards at home when I'm driving, but even then it's less often than not.
Things are probably different in cities with good mass transit, but does that describe any of the ones flooding?
roystgnr
7mo ago
There's at least people getting the "less bad" end of it compared to each other, even if true "better" might be far in the past or future, and "white men in the like 25-45 range" does seem to be the least bad male demographic to be in. I long ago noticed in OKCupid blog data that men do have better odds as they get older, and looking now at the "Reply Rate by Race - Male Sender" graphic from the old OKCupid blog, it does look like white men had it less bad than other men - with nearly 30% of women they messaged willing to acknowledge they exist!
Disclaimer: all data above is at least 15 years old, OKCupid was since bought out for being too functional a competitor to commercial dating apps, and there is no reason to believe the descent away from "better" has abated. Even limiting our complains to the situation of men seems myopic; women hardly seem much happier with modern dating, with very different but similarly serious complaints.
roystgnr
1mo ago
The wheels themselves, even on many expensive luggages, are of dubious quality
I've heard a theory that this was the problem: if even modern wheels are of dubious quality and capability, how much worse would they have been a hundred years ago? I'm not sure that makes sense, though. The invention of wheeled luggage is at roughly the same time the transition of roller skates from all-metal wheels to hard polymer wheels (which were lighter and smoother-rolling and less expensive), but all-metal wheels aren't that much worse in utility and they were probably better for durability. The most important invention for small wheels is ball bearing support, and that's more like 100-150 years old (at various levels of quality and expense).
The two other common theories are more situational:
Wheeled luggage came about during the expansion of mass air travel, with it's corresponding huge concourses and lack of porters. This was the first time people really had reason to want to carry their own luggage for long distances.
Wheeled luggage came about shortly after the Women's Rights Movement made it more common for women to travel on their own, and whereas a typical man would feel weak if he avoided carrying his own luggage, a typical woman would feel foolish if she didn't.
I'm not sure either of these really works either, though. Wheeled luggage was invented in 1970, but as another comment points out it didn't become popular until the 1990s. Perhaps that's because of the addition of the retractable handle (invented in 1987) finally making them more ergonomic to roll around? And maybe 17 years isn't too painfully long for someone to come up with that idea once it finally had a use case; "The Retractable Handle" isn't exactly the sort of thing you find at the start of the Civ tech tree next to "The Wheel".
roystgnr
3mo ago
difficult to observe because of the way it does (or doesn't) interact with regular matter.
Difficult, but not impossible. The clearest candidate so far is the Bullet Cluster, where we can see the shock wave from regular matter in the galactic collision, but we can also see the lensing from a bunch of something invisible in EM (i.e. "dark") that is a major source of gravity (i.e. "matter") that managed to shoot through the collision without itself colliding so much.
has never been observed
We could argue about what counts as an observation (have I ever really seen my kids, or have I only seen the photons bouncing off them?), but we've observed something that looks dark and acts like matter, regardless of how precisely we can identify it in the future. There are other theories that try to explain galactic rotation curves (the original motivation for theorizing "dark matter") with e.g. changes to how gravity works at long ranges, but they have a much harder time explaining the Bullet Cluster.
dark matter was invented to explain the otherwise unusual expansion of the universe
This was the motivation for dark energy, not dark matter. Dark energy is a much better candidate for your metaphor here. If it's uniformly distributed in space (which it seems to be on large scales, plus or minus 10%) then the volume of the Earth would include about 6 septillion kilograms of matter and 1 milligram of dark energy. Our best candidate for dark energy right now is probably "Einstein's equations are still consistent if we add a constant, so maybe that constant is super tiny instead of zero", and even that runs into a problem where, when we try out different particle physics theories for predicting the constant, we either get "zero" or "A septillion septillion septillion septillion septillion times larger than what we see". This definitely feels more like an "invention" than a "discovery" still.
I'm not sure you want to take the "ha, scientists invent invisible things too" metaphor too far, though. The examples get cooler than the Bullet Cluster. When scientists invent such things we sometimes get discoveries like neutrinos (predicted just to try to balance particle physics equations, and nearly impossible to see because they barely interact with anything, but we can detect them now), or the planet Neptune (predicted based on irregularities in Uranus' orbit, and essentially discovered by an astronomer "with the point of his pen" before we could figure out where to point our telescopes). Even when they fail at it we still get things like General Relativity (which explains irregularities in Mercury's orbit that were once hypothesized to be due to a planet "Vulcan" even closer to the sun). Neutrino detectors are still huge and expensive, but now anyone can see Neptune with a home telescope or use the corrected-for-relativity GPS system in their phone.
Could miracles ever work the same way? You've learned about the Miracle of Calanda now; perhaps we could convince people to start praying for amputees, and we'd see claims of miraculous limb regrowth rise to match claims of e.g. miraculous cancer remission? Would you expect that to work, and start trying, and report back to us after you see it start working? I'd be ecstatic to be proven wrong like that.
roystgnr
3mo ago
IMHO a better solution to the "fruit from the poisonous tree" rule would be "the criminal defendant can be in prison when the criminal cop is too". Two crimes get two sentences, not zero. Making one sentence contingent on the other would be sufficient to fix the bad incentives.
In this case, though ... do we even need to imprison the "defendant"? "A confidential informant said he was MS-13" got him held without bond after he was arrested for loitering, but never got a conviction. "The cops think this gang-member-turned-snitch is very trustworthy now" is a good place to start an investigation but surely it's not a good enough place to end one; police informants are sometimes themselves motivated more by base incentives than by a newly-acquired love of honesty and justice.
roystgnr
5mo ago
For (2) I'd add "Provide more school choice". I live in a school district where the best public schools are barely adequate, but in Texas the state will also pay for charter schools, which makes a difference. Charter schools can't screen students by anything more than some combination of "the applicant is in our geographic area", "applicants with a sibling already attending get preference", and "random lottery", but it turns out that the implicit self-screening of "the parents are concerned enough with their kids' education to move schools" and "the parents are on-the-ball enough to be able to get kids to school without a bus driving to their house" is enough to concentrate the most motivated kids and avoid the most disorderly.
We may be getting vouchers for private schools soon, too, so we'll see how that works out. Support for a voucher program in general is at 55% among Texas Democrats, 65% for Texans as a whole.
roystgnr
11mo ago
That would be a very dishonorable way to go out
Surely just the opposite? Being self-aware of one's own limitations, especially in a context where they sneak up on you like aging does or where they make it harder to be self-aware like cognitive problems can, is much more honorable than letting those limitations hit reality unchecked.
roystgnr
1mo ago
roystgnr
2mo ago
Humans don't even want political opinions that differ greatly from ours to exist. In a democracy those opinions might spread to the median voter and then be imposed on us against our will, and even in an oligarchy or autocracy there's always the chance that they will persuade the leaders or inspire a revolt against the leaders and then be imposed on us against our will. The use of language to navigate intratribal factionalism is probably older than homo sapiens. It's really hard to treat a question dispassionately as an intellectual issue, rather than as a signifier of loyalties, when everything we think and feel screams that there might be too much at stake.
Consider LessWrong, possibly the most concentrated population of high-functioning autists intelligent high-decoupling people on the internet, people deliberately trying to learn how to better discuss issues rationally in an unbiased fashion, the sort of "hey, I see what the problem is" people that normies joke about: their main conclusion about politics was that anybody who wanted to apply their intellect to any other issue should talk about politics as little as possible in the process.
If you want to apply your intellect to politics, though, where do you go? Well, here I am, I guess? I wish the place was more popular among thoughtful left-wing participants, and maybe there's some way to improve that, but in the meantime I'd rather be somewhere that often repels people with opposing views than somewhere that often expels them.
I think a more subtle issue (though I hesitate to call it a problem) here is that we also select for a particular subset of right-wing participants. Obviously anyone who's a Witch on one issue or another has reason to come to a place like this they won't be expelled from, but also there's a bit of strain between @Goodguy's claims of "assume that social conservatism is correct" and "wordily show-offy". At least 5 years ago, the modal Motte survey respondant was "ambivalent about religion, seeing it as a weak force for good", but that's reflective of a very peculiarly modern type of "conservative". At least in the US (also a modal Motte user characteristic in that survey), the modal social conservative is instead one of the 40% of Americans who would agree that "God created human beings pretty much in their present form at one time within the last 10,000 years or so". I know there are a number of faithful theists here, but in all the random discussions I've seen of anthropology and human genetics and so on I've never seen anyone jump in with the "no, it wasn't a parable, the first humans were created from clay 6kya" rebuttal that's a plurality belief among Americans. I'm not really interested in rehashing (from my perspective) that debate, but I hope that people are here who would be on the other side and are simply avoiding bringing it up for similar reasons, because that's still a huge and politically important mass of people, whom we can't avoid talking about, and whom I'd therefore like to occasionally be talking to.
roystgnr
4mo ago
Keanu Reeves character, "Speed", trying to be edgy: "Shoot the hostage. ... Go for the good wound and he can't get to the plane with her."
The_Nybbler, actually understanding edgy: "Shoot the hostage. Once they've obeyed the terrorist they can't legitimately complain of being treated as an enemy."
roystgnr
8mo ago
not sure when the election results will be announced
You and everybody here. I presume that explains the "(Day?)" quip in the title.
Most amusingly: Nevada's a swing state, and it's going to be counting mail-in ballots with postmarks up through election day (or with smudged postmarks) if they're received up to 3 or 4 days later. Even if everybody counts competently and instantly with no errors and no recounts there's a small chance we might not know the final outcome before the end of the week.
roystgnr
3mo ago
So then we're back to the status quo of zero sentences in N% of cases, but we get justice in 100-N%? Since N will be less than 100 that still sounds like an improvement.
roystgnr
4mo ago
Any thoughts on "Shadows of the Limelight"? I loved "Worth the Candle" and "Metropolitan Man" (and "The Randi Prize", and "Instruments of Destruction"), but I couldn't get in to SotL and I still wonder if I just gave up too soon and didn't make it to the good part.
roystgnr
5mo ago
Elon's choice of hand gestures last week were probably a spastic/autistic mistake that should be ignored, possibly a 4chan-mentality troll that should be disdained, but almost certainly not a Neo-Nazi signal that should be feared. Social skills are not up his alley and being cheered by huge crowds is a bit new to him and this isn't the first time he's flailed around on stage like a 5 year old with too much candy, it was just the first time the shape of the flailing was something that could trigger left-QAnon.
But Elon's choice of wording in that tweet, I can't think of any excuses for. Managing an orbital launch provider is up his alley. He knows those astronauts aren't stranded and are able to leave at any time (on his company's capsule!) in an emergency. He knows that when they were semi-stranded (the Starliner made it home unpiloted fine, just with a level of risk that was unacceptable for humans) it was Boeing's fault, not Biden's. He knows SpaceX was already made Plan A to bring those astronauts back, before the Biden administration ended, and they're just waiting for the next crew rotation. He knows they haven't been there "so long" compared to other ISS long-term crew stays (though this one I would just call a "distortion" rather than a lie; they have been up there quite long compared to their original quick out-and-back plans). And he knows that the last couple months of delay are due to a SpaceX delay, with the new Dragon capsule for Crew-10 taking a little longer to finish than planned.
For years Musk seemed to be somewhat resistant to even the powerful brain-melting effects of Twitter discourse. I'm not sure what combination of MAGA-inner-circle discourse and way-too-much-Twitter discourse and too-much-Ketamine discourse has finally gotten to him, but I wish he'd snap out of it and back away from whatever it is.
roystgnr
4mo ago
my impression (memory fades) is that the vaccines were lauded as being more effective than they turned out to be
The vaccines were lauded as being more effective than they had already turned out to be. Pfizer's efficacy was something like 93% in the initial study, and e.g. Biden oversimplified that as badly as "You’re not going to get COVID if you have these vaccinations."
They also turned out to be less effective than thought, with that 93% dropping to like 68% after only 6 months, which was enough to take us from "well some vaccinated people still get it but as long as we can push R<1 we can..." to "screw it, it's endemic now", even before Omicron changed the math further.
I'm not sure if this is "people were lying out of their teeth" rather than "voters aren't smart enough to avoid black-and-white thinking so they don't insist their president be smart enough either", but I think the takeaway is that you can probably trust independently repeated and reviewed studies of vaccine effects and you probably can't trust most popular interpretations of those studies.
roystgnr
5mo ago
Will they try to build more multi-family dwellings or just rebuild the mansions?
They'll try to build more multi-family dwellings, because density pays off when you start out with some of the most valuable land on Earth, and in the end they'll just rebuild the mansions, because the thumb is pushing hard on the scale for that.
roystgnr
6mo ago
Or compare the Ming Treasure Voyages cancellation with the European Age of Exploration. China was on top of the world, but their institutional infighting manifested as stasis while Europe's often-more-literal infighting manifested as a mad scramble for power, and the difference in incentives set China back centuries.
Compare East Germany with West Germany, or South Korea with North Korea. The economic effects of better vs worse governance on nearly-the-same-genes can literally be big enough to see from space.
I think a more interesting wrench in the gears is the question of better vs worse culture, though. South Korea has a lot more lights than North Korea in all the satellite photos, but at 0.72 and still falling TFR they might as well start turning lights off now before the last person "leaves". Their genes haven't significantly changed, and their government changes over the past 60 years (since the TFR started falling from 6) seem unrelated, but the cultural changes have been massive and baffling. Nerds started getting a mathematical handle on voting system foibles and game theory back around the time of Condorcet (though we've now got much better alternatives to FPTP than IRV-misnamed-as-Ranked-Voting, so the lesser-nerds' focus on the latter at this point in history is weird...), and likewise for economics and at least the most obvious Communism-level economic mistakes. But does anybody have any sort of mathematical analysis of WTF happened in South Korean culture, to cause the gender war to get so bad and the fertility rate to plummet eight-fold in 2 or 3 generations? It's like the stereotype of the pushed-into-overachievement kid burning out and becoming suicidal, but on a 50-million-person scale!
Would it be fair to say that the whole disagreement here is that @fmac is interpreting "Tell them not to have premarital sex" as, literally, programs telling kids not to have premarital sex, where you're interpreting it as reversing three generations of cultural change?
It's probably fair to say that the former doesn't work (it's definitely fair to say it doesn't work well, but none of the "abstinence-only education correlates with higher teen pregnancy rates" research I can find seems to be RCT-based or even adjusting for obvious confounders).
It's probably also fair (again, so many likely confounders) to say that the culture we changed away from did work pretty well.
But, although I'm not criticizing you for sticking with Chesterton's wording, doesn't it feel like "difficult" is grossly understating the problem here? If it had turned out that devoting some Health class time to abstinence had worked, we could have had some policy wonks discover that and institute it, and voila, problem solved. It could have been done via state laws, or via ED (when will I ever get tired of pointing out the ironies of that acronym?) funding, or just one school board at a time. But if it is correct that 1950s morality had a strong effect ... how do we get back to 1950s morality again, exactly? Or more precisely, since 1950s morality is what developed into 1990s morality, how do we get back to something that's sufficiently 1950s-like to help people but sufficiently different to avoid eventually being rejected again?
From your choice of quotes, I'm guessing your answer (and Chesterton's, were he still around) would include some sort of revival of Christianity, but the data makes that look neither necessary nor sufficient. In the USA non-Hispanic whites are around 60% Christian and have around a 30% rate of births to unmarried mothers, while for non-Hispanic blacks we see around 70% Christian and around 70% of births out of wedlock, and Asians here are at around 30% Christian but around 12% births out of wedlock.
Of course, that's just the rates of "births out of wedlock"! Currently 3/4 of Americans think that premarital sex is morally acceptable, and the vast majority of the other 1/4 must feel guilty eventually, because even decades ago 95% of middle-aged Americans had done it. Even if there's a potential level of deep, culturally-ubiquitous Christianity that could inculcate "fornication is a sin" in a way that modern Christianity can't pull off, how utterly monumental a change would it be to get from here to there? Whatever the process, describing it as just "tell them" seems woefully inadequate. There may be some level of hysteresis making this exceptionally difficult: if 90% of your community thinks "fornication is a sin" is a theological fact, the other 10% just look like sinners and don't affect what your kids believe, but if it's 10% and 90% instead then the 10% just look like weirdos and don't affect what your kids believe, even if you're in the 10%.
roystgnr
3mo ago
how do you make democracy work if you permit governments to bind the hands of their successors?
With legislative supermajorities. If you can get 51% agreement on something, good for you, but there's no reason to expect that to bind others once a slight shift of political winds leaves you at 49%. But if you can get 60% (or 67%?)? That might be something worth hanging on to for longer, if it's not so soundly refuted that support drops to 40% (or 33%).
roystgnr
4mo ago
I'm a huge fan of this Musk philosophy in his engineering ventures. Testing often-too-flawed engineering ideas as fast as you can is much cheaper and much faster than trying to come up with something flawless on the first try, and seemingly-ironically it tends to give you a less flawed final product too. I'm not sure how well that works with people rather than objects The fourth Falcon 1 wasn't working while scared that mistakes had been made that blew up the first three. The Falcon 9 landing engines weren't going to change careers because SpaceX tried out parachutes first. The machine-welded stainless steel Starship tanks aren't going to quit and find a job where composite tanks and hand-welded steel tanks don't get abused and wrecked.
I'm not sure how well the philosophy works with people. Federal government work in many cases is seen as a tradeoff: lower compensation than equivalent skills would get you in the private sector, but with better job security to make up for it. If he significantly cuts headcount without cutting output (or if Congress follows up with more deliberate cuts) then maybe making that deal worse is still fine? We'll have fewer interested applicants, but we'll also have fewer jobs we need to fill, so we won't have to raise pay to compensate for the drop in supply? But this isn't like an engineering experiment where the experimenter is the only one who learns something and failure is just one of the things we can learn; here the experimentees are learning too and failure can have more lasting consequences.
roystgnr
8mo ago
rejecting Rationalism, as it leads to cudgels like "falsely claimed without evidence"
Does it? I don't think I've ever seen the phrase "without evidence" used sloppily by anyone whose definition of evidence is "B s.t. P(B|A)/P(B) > 0".
roystgnr
9mo ago
selection pressures in that have returned to favoring fertility rates in people with high educational attainment
You say fertility rates, but am I reading correctly that in your link the graph is of a raw birthrate in the age 15-50 bracket without controlling for age? How much of that result is just reflecting the fact that a lot more people are going to higher education than they used to? In Census data I'm seeing the ""College 4 years or more" column rise from e.g. 23% of the total "25 to 34 years, Female" section in 1992 to 45% of the total in 2022 (passing through 32.5% in 2002 and 37.5% in 2012). Unless I'm misreading something, women in their childbearing years are much more likely to be well educated than women a decade or three older, so we'd expect birth rates to skew much more strongly toward educated mothers than total fertility rates do.
roystgnr
9mo ago
I take it you missed the dialogue about that solution? If you're wise you'll remain ignorant and stop reading this comment now.
Okay, but don't say I didn't warn you.
Wired tried to explain that although "some white users worry that calling attention to their race by texting a pale high five (or worse, a raised fist) might be construed as celebrating or flaunting it", "The yellow emoji feels almost like claiming, “I don't see race,” that dubious shibboleth of post-racial politics, in which the ostensible desire to transcend racism often conceals a more insidious desire to avoid having to contend with its burdens."
And NPR let you know that, although "some white people may stick with the yellow emoji because they don't want to assert their privilege by adding a light-skinned emoji to a text", "there was a default in society to associate whiteness with being raceless, and the emojis gave white people an option to make their race explicit", so even if you're "just exhausted [from] having to do that. Many people of color have to do that every day and are confronted with race every day" - so is it really fair for you to get to ignore it?
Indeed, "the default yellow is indelibly linked to The Simpsons, which used that tone solely for Caucasian characters (those of other races, like Apu and Dr. Hibbert, were shades of brown)."
(No mention of the other characters who were non-Caucasian and yellow or lighter, for some reason.)
Did they focus on probationary employees, or employees in probationary positions? I've heard a lot of claims that there are employees who got a probationary promotion, but the probationary status of "depending on performance the promotion might become permanent or you might be returned to your old job" was just replaced by "you don't have any job now".
It's interesting to ponder the selection effects, if that's true. We'd be keeping the "so obviously capable that we can't imagine undoing their promotion" employees, but also the "so obviously incapable that we wouldn't even offer a probationary promotion" ones, while hollowing out the middle a bit.
More options
Context Copy link