"Love triangle" is a misnomer; the thing called a love triangle is nearly always actually a "love V"/"love line". An actual love triangle would necessitate an odd (and hence non-zero) number of same-sex attractions.
It seems Return To Office is "trending" - what are people's thoughts on this?
I ask because I've been working a somewhat cushy tech job in the midwest after leaving silicon valley, but now they want us to return to the office despite their lackluster salaries and average talent pool. I was willing to swallow inefficient practices and less money for the comfort it offered, and I truly like the company even though I don't always agree with their decisions. But now it feels like there's no reason to stick around if I'm going to be forced into an uncomfortable commute just to sit in virtual meetings in a different building when there's so many more capable companies paying more for fully remote positions. I see a lot of arguments about how 'it's still better than other places' and a lot of chatter implying that people are privileged or unreasonable for demanding fully remote work. Am I out of touch?
I think this is part of the general trend where most of society went virtual from 2020 to 2021, and then started reopening again in person over the course of 2022 and 2023.
My perspective is that if you can find an outside offer that gives you what you want, then no, you are not out of touch. If the company absolutely needs you and needs you to come in they can find the funds to pay you more until you agree to come in. If you can't find an outside offer that gives you what you want, then yes your expectations are out of touch with what's possible for you. To me discussions of 'productivity' are more complicated. I'm typically skeptical of people who claim to be even more productive at home than they are in the office, I'm sure some of these people exist but I imagine they are in the minority. Given that managers (and people in general) are risk-averse, I think the workers whose output really dropped when they went remote are much more salient to managers. I also think part of this response by companies is in response to the general downturn post 2022-Q1. In that sense I don't think productivity really has much to do with it.
IMO the "Get a job in a HCOL city -> Work fully remote -> Move to LCOL City" arbitrage was never going to last. Some people were going to be grandfathered in and enjoy the benefits, but over time either salaries would fall or COL would increase to. Personally I'm surprised more tech people aren't worried about the increased ease of off-shoring that comes with a fully remote company. I know developers who work on teams comprised of 1-2 NA workers and then a slew of people in India
>Personally I'm surprised more tech people aren't worried about the increased ease of off-shoring that comes with a fully remote company.
I've been worried for...my entire (short) career, actually. In 2015, I was chatting with a project manager at a large professional service company, and they 1- Were totally cost-cutting with indian workers, and 2- Absolutely hated it because it was atrocious to communicate with them and the quality of the output was very, very low. Since then, every single person I've chatted with gave me the same picture: contracting indians is cheap, and you get what you pay for.
Now, I'm pretty sure that 1- They were getting the bottom of the barrel because they were beancounters looking for cheap labor, 2- The average quality of indian engineers increased over time, 3- AI will bring up a large part of the workforce anyway, but all that is compounded by the fact that demand for software will keep increasing the in forseeable future, so i'm not too worried either.
I find it sad how Covid demonstrated that things formerly thought impossible can become possible overnight, and now that it is over, those things are gradually becoming impossible again. :(
Most of my team members are in different countries, so I communicate with them remotely, but for some reason it is very important that I do it from the office. Especially absurd considering that we all sit in an open space, so anyone talking is interrupting their colleagues.
The local strategy is management declaring that remote work is only allowed N days a week, and decreasing N by one every year.
> a lot of chatter implying that people are privileged or unreasonable for demanding fully remote work
In theory, this should be decided by the market. Supply vs demand => salary and/or benefits. Work from home is one of those benefits. You decide how much it is worth for you (how much salary are you willing to give up for how many days working from home), and choose your job accordingly.
Yes, you *are* privileged... by the market. We got lucky, born in the right era with the right talents and skills. We are high on the scale of privilege, but still far from the top. Do you think that your CEO feels guilty for making 100x as much as you do? No? Then why should you feel guilty for wanting your benefits?
Enjoy it while it lasts, tomorrow we may all be replaced by super-intelligent chatbots. When that happens, your company will not care whether it is "reasonable" to let you starve.
Am I right that this is ordinary truck drivers with no special skills earning more than many doctors and lawyers? Is their union just that good?
Given that UPS predicts their profits are going down something like 10% but they're still able to be a robust company with decent profits, does this mean every company could afford to pay their workers $170K and still function okay, and it's just a lack of unions that's preventing them from doing this? If not, what's special about UPS?
I would point out that that $170,000 number includes benefits. I’m not sure that is true with a lawyer or a doctor. In other words I think the latter are inclined to make that kind of money in addition to their benefits.
>does this mean every company could afford to pay their workers $170K and still function okay, and it's just a lack of unions that's preventing them from doing this?
If UPS can afford to pay its drivers $170k, then those drivers are each creating more than $170k in value for the company. I can't think of any other low-skilled job for which that is the case. What appears to be special about UPS is that their business model leverages a ton of value out of some of their low-skilled workers.
I would guess that even though it is a union job, that those drivers can't get away with being lazy.
I looked into this a little bit and it looks like there are a few things that are fishy here:
1. The $170,000 number is misleading, this article says that before this deal this was broken down into $50,000 in benefits and $95,000 in salary (https://www.bbc.com/news/business-66445496). Then this goes up over the next $2.50/hr/year for 5 years until it becomes ~115,000 in salary. So it's really total comp 5 years from now is $170k, which feels really different from the headline number.
2. That $95,000 in salary is calculated out from 45 hour weeks for 50 weeks a year, carrying heavy boxes. Not highly skilled, but not a super cushy job either.
4. I assume you are required to pay some of your paycheck back to the union.
None of this is to say that it isn't good pay for the skills required. It seems like a better trade than, for example, being a teacher (who are also usually unionized). But I think the articles are maximally misleading about how good the deal is here.
Also sorry for not having a real answer to your questions ("Could every company pay this much?", "How are unions affecting this?", etc.). I just don't know too much about that.
Governments use such tricks, too. If you read in newspapers how teachers in Slovakia can now make extra €50 a month, the missing part is something like "if they volunteer to work on a government-approved project in their free time for at least 6 months" (the example is *not* made up). But the internet immediately starts yelling: "damn those greedy teachers, such easy work and so much money, next time they go on strike I will punch them".
Working in a Burger King is hard too, especially in rush hour when there's a dozen burgers needing cooking at once, everything is constantly beeping at you, there's hot surfaces everywhere and dumb teenage coworkers running around like headless chickens. Delivering packages seems much less stressful.
Whatever happened to the term "virtue signalling"? A couple of years ago it was being used all the damn time, but I just realised I haven't heard it for a while.
Either way, protesting that your salary is too high and more should be given to the low-level peons, when you know there's absolutely no way that this protest will lead to your salary being decreased and given to the low-level peons, seems like a fantastic form of virtue signalling.
Nobody is stopping them from donating half (or whatever fraction) of their salary to the nursing staff. That way they could save a bunch of time and make sure the money went to the right place.
Why? Individual MDs could just straight up gift money to individual nurses. I presume Canada has some annual exclusion up to which gifts are tax free - spread it over enough nurses and there is no double taxation issue at all.
ETA: Google tells me Canada has no gift tax at all...so...
Hanania was a guy who brought us a step closer to a truth teller who could say; (I'm paraphrasing, maybe, despite the quotes): "We need police to enforce black neighborhoods stronger", without fear that he would be accused of being a racist for saying so.
It really sucks that a prominent person who didn't fear getting called racist for saying some truths really was a fucking racist. It's easy to imagine a Hanania who wasn't a racist, but he was. So much for imagining.
I agree with his sentiment that we need police to police high-crime areas more strongly. I don't have a problem with him specifying black neighborhood because those are in fact the neighborhoods that need more policing. (To be clear: policing of the good kind. Not redneck fucking cops who are too common, but that's another issue)
But his tweet under his real name which I just saw: "Daniel Penny getting charged. These people are animals. Whether harassing people on the subways are walking around in suits."
So according to Hanania black people are "animals" even if they are "walking around in suits"?
Please tell me if I am wrong in my interpretation of that tweet. And as far as I can tell that wasn't a drunk tweet, something I have forgiveness for if deleted the next day, but one he kept.
I don't want Hanania to be canceled on any platform. I'm for free speech above almost everything. But he does sound like a fucking bigot at heart. I'll still read him, at least for a while. But I can't pretend that my opinion of him isn't a lot lower than it was before.
This sorta gets into the territory of whether we should separate the artist from the art, except in the case of Hanania there is no art.
>It really sucks that a prominent person who didn't fear getting called racist for saying some truths really was a fucking racist.
What does "racist" mean, here? Is it "racist" to believe that genetic differences between races explain behavioral differences? Do you acknowledge in principle it's possible for a correct understanding of reality to be "racist" (and see why that might be problematic)?
>(To be clear: policing of the good kind. Not redneck fucking cops who are too common, but that's another issue)
Source?
>I don't want Hanania to be canceled on any platform. I'm for free speech above almost everything. But he does sound like a fucking bigot at heart.
If Hanania is a bigot, then so are a literal majority of black americans, muslims around the world, etc.
>This sorta gets into the territory of whether we should separate the artist from the art, except in the case of Hanania there is no art.
No, it gets at the fact that white people cannot, in the view of leftists, make any kind of argument about anything related to race unless they have a strongly non-negative view of black people, whereas hating white people is not only acceptable when talking about race but encouraged.
>> What does "racist" mean, here? Is it "racist" to believe that genetic differences between races explain behavioral differences? Do you acknowledge in principle it's possible for a correct understanding of reality to be "racist" (and see why that might be problematic)?
I cannot understate just how much you are stretching trying to apply this defense to the situation. How on earth did you manage to see the statement “these people are animals” and somehow jump all the way to “oh, he’s just saying he believes that genetic differences between races explains behavioral differences?”
1) the guy harassing people on the subway who was strangled was black
2) calling blacks "animals" is a standard cliche in the USA. (See: The Godfather for a pop-culture example.)
3) i can't think of another way to tie the subway harasser to the people in suits
I could be wrong, of course. Do you have any ideas of who else he might mean? I considered the possibility he means progressives, but how would he know the homeless guy was a progressive?
EDIT: Let me correct the typo in my quote in case it wasn't obviously a typo. The post reads "Daniel Penny getting charged. These people are animals. Whether harassing people on the subways or walking around in suits."
Shouldn't those who think that it's possible and/or important to have these conversations without being motivated by racism be the maddest at Hanania? Am I missing people expressing this sentiment? Because I definitely haven't seen it.
Whatever your position, he just made those types of conversation way harder to have.
As someone in that camp, my position is that the fact that one person discusses these things and is (or was) racist doesn't imply anything about whether it's possible to discuss them without being racist. Of course the usual suspects will pretend that it does, but those will pretend it's impossible to discuss these things without being racist regardless.
Also, if progressives use Hanania to make poor arguments, that's not Hanania's responsibility. After all, one of the common arguments against discussing race in non-progressive ways is to say that it will be misused by racists to promote racist policies. Part of my rejection of that is to reject the notion that if racists cite me to make invalid arguments, I'm responsible for that. If I say something, and racists use it to argue something that doesn't actually follow from it, then take that up with them, not me. But the same thing applies here: if progressives use Hanania to make invalid arguments, take that up with the progressives, not Hanania.
It doesn't actually change things because it doesn't matter how 'nice' you are about things, if you have the "wrong" opinion on race, leftists will seek to destroy you. The fact you have non-progressive views on a racial topic IS "racism" to them (whatever that even means).
The left are hateful towards white people (https://ideasanddata.wordpress.com/2020/06/03/american-racism-and-the-anti-white-left/) but it's irrelevant, because none of this has to do with 'hate' or 'racism' or anything. The left believe what they believe and will never be persuaded otherwise, and if they think screaming 'racist' at somebody will help "win" the debate, they will. Charles Murray is obviously less hateful than >99% of left-wingers with strong racial opinions, but leftists wil l still use violence and intimidation to shut down any speech he gives on campus. BECAUSE THIS ISN'T ABOUT HATRED.
And yes, right wingers can be persuaded. You have to be open minded in the first place to be open to different opinions on race than the ones you get shoved down your throat by schools, colleges, the media, corporations etc.
>> And yes, right wingers can be persuaded. You have to be open minded in the first place to be open to different opinions on race than the ones you get shoved down your throat by schools, colleges, the media, corporations etc.
This is all kinds of backward. You're essentially saying "the *real* persuadable people are the people who remain *unpersuaded* in the face of arguments from every direction (schools, colleges, the media, corporations, etc)." Turning stubbornness upside down and citing it as evidence for, of all things, openmindedness.
Individual right wingers or left wingers may be more openminded or less openminded, but trying to demonstrate somebody's openmindedness by citing how stubbornly they cling to their convictions in the face of opposition just makes no sense at all.
I think the idea is that basically everyone starts off with the opinions/beliefs that society tells them to have, acquired through peer pressure and imitation, not through argument, but then some are persuaded to switch to unpopular opinions/beliefs, often through argument.
I don’t think your statement “basically everyone starts off with the opinions/beliefs that society tells them to have” lines up well with Goldman’s statement that “you have to be open minded in the first place to be open to different opinions on race than the ones you get shoved down your throat by schools, colleges, the media, corporations etc.”
“Society” and “schools, colleges, the media, corporations, etc” are not readily interchangeable terms. “Society” is broad, and encompasses both your immediate community (family, friends, social relations) as well as institutional actors like “schools, colleges, the media, corporations etc.”
An attempt to conflate the two confuses the argument in significant ways. Since “society” includes your family and immediate community, one can plausibly assert that “basically everyone starts out with the opinions/beliefs that society tells them to have,” but the statement falls apart if you try to swap in the institutions and say “basically everyone starts off with the beliefs that schools, colleges, the media, corporations etc tell them to have.” Your family has access to shape your mind from the moment you are born, but you need to reach certain age thresholds before various institutional actors gain access to you.
And that simple fact blows up the whole argument. If everybody started out already baked in with the proper corporate opinions, then some people held them and some turned away, we could have a conversation about why the turners are turning away (spoiler, I think it would be more parts emotional suasion than careful argumentation).
But facts being what they are that would be a pointless conversation to try, because people aren’t starting baked with those institutional opinions in the first place – they're starting (for lack of a better phrase) with family opinions. And unless you can isolate people who started out with all the "right" beliefs and then were platonically convinced by argument to become racially resentful (probably not very many) from people who were "just raised that way" (probably a lot) it'd be foolish to look at a person with a bunch of racial resentments and say “that’s a sign he’s very openminded – after all, he holds unpopular opinions!”
I think the large shift in racial orthodoxy over the past century would suggest that the immediate community isn't playing that large a role. To me is seems most plausible to say these opinions are mostly coming from US media, considering which countries have been the most affected.
No, it really isn't easy to imagine a Hanania that isn't racist. His entire schtick was finding verbose ways to justify pre-determined right-wing conclusions. You're never going to find someone that dedicated to "anti-wokeness" without substantial prejudices.
And while we're here, America is already hyper-policed. It's the most imprisoned state on the planet. And black neighbourhoods are the most policed part of it. The idea that they're under-policed is absolutely wild.
There are limits to what police can accomplish. You can't always just "do it harder".
>You're never going to find someone that dedicated to "anti-wokeness" without substantial prejudices.
For goodness 'sake! Find me a single person with 'woke' opinions of race who is even remotely open to any non-woke explanation of race differences. I'll wait.
Wokists are not only closed to the very possibility of e.g. race differences being caused by genetic differences, but think that it literally shouldn't be legal to suggest that this is true.
In order to come to non-woke conlcusions in the first place you have to be open to alternative explanations to the narratives that dominate in america's hegemonic institutions.
>And while we're here, America is already hyper-policed. It's the most imprisoned state on the planet.
Walk down the streets of san francisco and tell me there's too much policing.
America is the most imprisoned state due to a combination of well funded and effective law enforcement and an extremely high crime black population.
Other countries with large black populations i.e. in Africa simply don't have effective law enforcement.
>And black neighbourhoods are the most policed part of it. The idea that they're under-policed is absolutely wild.
Absolutely false. The vast majority of crime per capita is committed by black americans, and the policing in black neighborhoods is LOWER than the crime rate proportionally. The fact that there's more police in black neighborhoods does NOT mean they're over policed, and if BLM supporters actaully wanted less black people being killed, they should support more police in black neighborhoods. But somehow, Geroge Floyd's life is worth more than the 10000 or so black people that murder each other every year, because BLM is tribalism, pure and simple.
Comparing incarceration rates in the US to other countries is meaingless, unless you think that the underlying crime rate is the same in both countries.
As it is, of course the incarceration rate is higher in the US than in other western countries -- the crime rate is much higher. Crime in the US is so high that even once a couple of million of the worst criminals in the country have been put in prison, the US _still_ has vastly higher levels of most crimes than most other comparable countries.
Anyone who thinks the US is "hyper-policed", I invite you to do this experiment. First, shoplift a cigarette lighter. Then, use it to light a crack pipe. Then, while smoking your crack pipe, take a shit on the sidewalk. Do this in each of these locations:
1. Market St, San Francisco
2. Orchard Rd, Singapore
3. Kufürstendamm, Berlin
4. Oxford St, London
5. Wangfujing, Beijing
Report back on the police response in each location, and we'll see how over-policed the US is.
Let's face it, "Woke" and "Tough on Crime" have become the new racist dog whistles. It's all the Left's fault though! <#snarkasm> It's because of our 60-year obsession with shaming people who use racial epithets. Racists are good people. They don't want to be called racists! And the Talking Class of modern bigotry want to appeal to middle-class voters who want a kinder gentler form of racism (no more Bull Durham's, please <gasp!>) — even though they'd cringe if their HOA allowed blacks to purchase a home near them. So serious bigots have to keep fabricating new and improved ways to encode their message. What a shame that Cancel Culture makes it tough to be a racist asshole in public anymore.
I'm sure the unelected mayor and city council have the best interests of this majority-black community (85% black, 15% white) at heart. After all, it's in everyone's best interest that there never be *any* elections. Otherwise, the wrong sort of people could get into office. The WSJ hints that "race" may just possibly be a factor, but obviously, the small clique of white people who've run this community for time out of mind aren't racists. No, that couldn't be possible in this day and age!
It's been woke dog-whistle since before we wuz woke! But a nice litmus test is whether you'd feel comfortable about walking around a black neighborhood and stopping to shop and dine at the black-run businesses. Even many of my whitebread leftie friends would feel uncomfortable when they're a minority of one in all-black setting. But as a general stereotype, I've found black Americans to be friendlier than my white cohort, and they definitely throw funner parties. ;-)
Neo-nazis, people who make up a fraction of a fraction of a percent of the population, commit almost none of the crime in the US, have no institutional power or support...
But black are empirically more racially hateful than whtie people.
And neo-nazis are a force in American society where? How many are there, a couple dozen thousand, with zero institutional power? Have you ever met a neo-nazi? I haven't.
Meanwhile out and out racism is the straightforward policy of many large institutions as long as it is directed in the right direction.
You said, "by far the most prejudiced people in society right now are all on the left".
This is a specific claim, and an obviously false one. If you want to replace it with a more defensible claim, you should acknowledge that you're doing so.
I get that we all run our mouths online sometimes, and that these comments are not carefully edited articles with multiple drafts. What you meant might be far more reasonable than what you said. But I can only see what you said, and I have no other guide to what you might have meant.
It's an ambiguous claim, and false only because "by far" goes too far.
If "most prejudiced people" is meant to refer to the greatest number of people who are at all prejudiced, then both the left and the right seem to have roughly similar numbers in total, and they've both got mostly people with an assortment of low-level prejudices.
If "most prejudiced people" is meant to refer to the people with the greatest degree of prejudice, then the right has the neo-Nazis et al, and the left has the militant Black Nationalists and the Kulack-hating Antifa fringe, none of which on either side end up posing any real threat to the Republic, but it's hard to say which is worse.
If the idea is that we get to compare the thousands of literal neo-Nazis on the right with the millions of leftists with an assortment of low-grade prejudice and say "a Nazi is worse than the median leftst", or vice versa, then no. Knock it off. Both of you.
Dude, you/we were clearly talking about the broader society, not the scattered random basement dwelling neo-nazis. Honestly you want to get in a real discussion of the most racist people they are likely recent immigrants from the Balkans or the Middle East, not some teen edge lords, or nazi biker gangs in San Quentin or whatever.
Honestly beyond even that I kind of suspect that relatively well respected figures on the left do hold views just as racist as neo-nazis, judging by how racist they are comfortable being right out in the open.
There are many ways to measure "quantity of policing." I doubt the OP is referring to "fraction of the population in prison." He says we need to "police high-crime areas more strongly" - the most straightforward interpretation of that is "put more police officers in high-crime areas."
It is not the case that America is "hyper-policed" by the metric of "police officers per capita in high-crime areas." For example, across the entire city of Chicago, there are approximately 12000 sworn officers for a population of 2.7million, vs 34000 for a population of 6.7million in Paris, per wikipedia (https://en.wikipedia.org/wiki/Paris_Police_Prefecture), so Paris has more police per capita. I would also argue that since Paris has fewer guns, and a legal regime that makes gun laws easier to enforce, a single police officer goes further in Paris than in Chicago.
It's worth pointing out that Black Americans overwhelmingly do not want *less* policing in their neighborhoods (https://news.gallup.com/poll/316571/black-americans-police-retain-local-presence.aspx). This is remarkable, given that Black Americans also report quite negative personal experiences and interactions with police officers. A combination of *more and better* policing would therefore probably be very popular. People don't like having violent crime in their neighborhood!
True, there is a bit of ambiguity about what more policing means.
Since we're talking about Richard Hanania here, I think it's pretty clear that "police black people more strongly" means "more state-sponsored racial terrorism".
Has Hanania said he's in favor of state-sponsored racial terrorism? This seems like exactly the kind of reasoning that can always be deployed to convince yourself your enemies are monsters. Joe Biden just said he wants more funding for public schools--obviously this must mean he wants more money for propagandizing kids with critical race theory and convincing them all they're trans.
The great thing about this kind of reasoning is all the thinking it saves.
Yes, I think you've captured what the OP is bemoaning. Hanania is a huge racist, and his arguments for more policing are likely in bad faith. Therefore, we can reject arguments for more policing out of hand as bad-faith racism.
If you want to reform the police, I'm all for it. If your reforms would increase the number of police officers, I'm skeptical but willing to listen.
If you say "we need police to enforce black neighborhoods stronger", then I'm just going to dismiss you out of hand. Because
a) the racial bit makes me assume the worst
b) "stronger" usually means more violence and harsher sentences.
and
c) American police budgets are already absolutely insane.
It's very clear that America needs less policing in many ways. If you think it also needs more policing in some other ways., then you can make that case.
But when you do, you need to go out of your way to make clear that you're not like Richard Hanania.
It can be done! It's not even that hard!
But you have to do it. Or people will be right to dismiss you.
PS: Isn't it weird that America's enormous police spending doesn't buy it a similarly enormous police population? I'm genuinely curious about what's up with that.
Do the people living in high-crime neighborhoods generally think the US needs less policing? My understanding is that they very much do not, and in fact that a common complaint of people living in poor, mostly-black neighborhoods is that the police don't show up when they're called because they DGAF what happens in those neighborhoods.
"I'm not like Richard Hanania." With that out of the way:
I also don't care for the statement, OP's paraphrase of Hanania, that "we need police to enforce black neighborhoods stronger," because I agree with you on (a) and (b). (On the other hand, OP goes on the clarify that his own policy preferences are "more police in high crime areas" rather than "more punitive sentences for black people", and states that he thinks Hanania is a "fucking bigot" - so maybe it's not totally pointless to engage with him?)
Regarding (c), and again comparing Chicago to Paris:
Chicago police budget in 2022 was $1.75b, of which $1.6b, i.e. pretty much all of it, was personnel costs for the ~14k positions, so call it an average salary of $115k - though I note that this does not include pension costs.
Paris police budget in whatever year the Wikipedia article was edited (2018 maybe?) was 1.5b euros. I don't know how much of that is personnel costs, but probably also almost all of it. For the 45k employees listed, that's maybe 32k euro average salary. (I assume that this also does not include pension costs.)
So inasmuch as American police budgets are "absolutely insane" relative to Europe, it would seem to be driven largely by salaries that are perhaps 3x higher. Given differences in per-capita income, you expect generic US salaries to be less than 2x higher, rather than 3x higher. I don't have high confidence in these figures - the French figures are from an undated Wikipedia page, and I can imagine there might be substantial differences in benefits like vacation time, or in tax treatment, or other things that make these numbers incomparable.
But assuming the premium really is about 50%, there might be a number of causes:
- Is policing in the US substantially more dangerous? Are US police officers more likely to get shot than French police officers?
- Do US police officers work longer hours or have more overtime than French police officers?
- Do US police unions have more political power in contract negotiations than the French equivalent?
Both things can be true at the same time: we have very high incarceration levels and not enough policing in areas that could really benefit from it. Our policing model is profoundly broken and the insane contradictions are the result.
"more cops would mean fewer people in prison because policing prevents crime"
Interesting point! Yes, this might happen if the deterrent effect of the policing is the dominant effect. Is there experimental evidence, one way or the other?
Sometime ago I read a rant by a retired cop who's main point was that in the last couple of decades police work shifted from "policing" to "law enforcement", to the detriment of both. In the old "policing" paradigm, cops get called on a bar brawl, arrest the bozos, throw them into holding cells overnight, and give them a stern talking-to in the morning, no charges filed. Civil order restored, no lives ruined in the process.
Now the cops come in and try to figure out if this can result in a court case. Because they are in a law-enforcement business. If things look dubious they may not even try to arrest the bozos, but if they do, now there are charges filed, and if they can't post bond they lose their jobs, courts get clogged up by the stupid cases, etc. etc. at a great cost to themselves and the taxpayers.
Then there's the fear factor: if black folks think that their calling police for help may easily result in themselves being George-Floyded, they will think twice, which only emboldens criminals and psychopaths. A vicious cycle ensues.
So I don't know if there's any coherent data to Mr. Tabarrok's point. It feels correct, intuitively, but the devil is in the details, and the details are legion.
>>So I don't know if there's any coherent data to Mr. Tabarrok's point. It feels correct, intuitively, but the devil is in the details, and the details are legion.
Thank you for including this caveat. I agree that this "feels correct" and that whatever the truth is is likely way more complex/hard to document.
With that same caveat, your point sounds like a natural extension to something similar I heard about judges and prosecutors. Prosecutors and judges used to have much more discretion in terms of what cases to bring and drop, what sentences to apply, etc. But, over the course of the 60s and 70s, we found a lot of racial disparities and abuses in how those decisions were made, and that discretion came under fire. We get things like mandatory minimums and mass protests over what charges are or are not filed. The drive for a "fairer" system also made the system more rigid.
Seems like a parallel to your observation about policing vs. law enforcement. Previous generations of cops had greater discretion to refer some people to the district attorney for prosecution while others got warnings, a number of them used this discretion wisely while others used it corruptly, society realizes that discretion is being used corruptly and tries to solve by reducing discretion, and now cops are arrest/citation machines.
Again, it's more parts Stephen Colbert's "truthiness" than anything - I'm not sitting on a pile of evidence for it. But it does make sense.
This sounds correct to me, too. It's not primarily about how many cops there are, but what they do. Are they trying to maintain peace, or to arrest as many people as possible (and ignore those who cannot be arrested)? The latter strategy leads to more problems *and* more people in prisons.
This sounds like part of a general trend in USA culture towards micromangement. "Charges filed" sounds all too much like the sort of thing that fits into some management "metric", while "civil order restored" is harder to quantify.
I always love the "lets free the people in there for minor drug crimes", and it is often the guy who killed two witnesses and assaulted 5 "friends", but there was a hard time putting a case together so they stuck them in there for 10 years for intent to distribute or whatever.
A lot of times the disproportionate seeming drug sentences have to do with other shit going on and plea deals. I feel like most people with attitudes about criminal justice reform would have a radical reimagining if they spent even one day at a county lockup and saw who goes through there.
It's not clear to me what "these people" refers to in Hanania's tweet, unless there is more context someone with a Twitter account can glean. Could a charitable interpretation be any of those who have no respect for law and order by initiating or excusing hooliganism?
(I live in a bubble where I don't pay attention to these sort of cases, but worth noting in the video itself, there are two other civilians helping restrain the thug, so it's clear there was some degree of agreement in the subway regarding the threat)
In the WEIRDEST People book, as I understand it from a review I just read, it was weird how Christians made the culture monogamous. They ended cousin marriage and gave us the nuclear family we now know in The West. Before that, polygamy and cousin marriage were common.
Supplanting polygamy with monogamous marriage had massive effects upon society: men were basically domesticated. Maybe women were too. It was good, if you're one of those people who likes human flourishing, for human flourishing. Because men, no longer competing with each other for more wives, now did more productive things, or something.
My main question is this: aren't modern Asian societies mostly monogamous? If so, then Christian Europe wasn't so weird.
Monogamy is a human default and Henrich never claims otherwise. The WEIRD aspect of modern nuclear family is not that it's monogamous, it's the fact that it became a fundamental social and economic unit, rather than being subsumed into larger clan and tribal structures that were the norm elsewhere until that point.
(Regarding polygamy, it does not need to be common to pose a problem. Just a small fraction of wealthy men taking several wives each creates, and normalizes, a much larger underclass of men with little marriage prospects unless they engage in significant risk-taking.)
"Man is born polygamous yet everywhere he is monogamous
Not long ago I had my own Fermi moment. I looked at the world around me and asked: Where are all the polygamists?
Consider almost any past empire or civilisation — Mongol, native American, Chinese, Indian, African, old European — and you will find powerful men with many wives. It’s all over the Hebrew Bible. 90% (!) of hunter gather societies around the world practice some degree of polygamy. Yet we look around today and… zilch?
...It turns out this is no accident. The WEIRDest People in the World by Joseph Henrich (I’ll call it WEIRD from now) traces how Christianity exterminated the practice over centuries and forged modern, cousin-free, monogamous marriage in the West (hence WEIRD: Western, Educated, Industrialized, Rich and Democratic). In light of Christianity’s now millennia long clamp down — occasionally literally hounding kings over successive popes until they submitted — it’s not so surprising world leaders and billionaires today seem about as monogamous as anyone else. It’s remarkable how potent culture is: what was once as natural to a powerful man as eating is unthinkable to such a man today."
"forged modern, cousin-free, monogamous marriage in the West "
Yes, but it's the "cousin-free" part that is doing most of the heavy lifting, and that Heinrich is calling out as important. I think he overstates the case, but he's probably directionally right.
Frequent cousin marriage divides people into clans, which is stable because clan insiders are in a more secure place than the rebel who became an atomic individual and then courted someone from a different clan. And clans make polygamy easy. Everybody is strongly pressured to marry a cousin or other clan insider, the clan leader is uniquely desirable as a husband within that group, he can probably get several women willing to take 1/3rd of him over any random clanmate even before he starts throwing his weight around with the fathers of the girls he wants to marry. See e.g. the internal dynamics of modern American polygamous cults.
Eliminate the clannishness - which everybody has to at least partially do as they modernize and industrialize because it's not practical to organize factories on a clan basis, and no woman is being told "If you want to marry an Alpha, you have to marry me even though I've already got three wives because I'm the only Alpha in the clan". There will still be men for whom many women will settle for a 1/3rd share, but there will be many such men and some of them haven't yet married even once.
Some societies will still allow polygamy for the few men (or women) who can pull it off against those odds, others will go to nominal full monogamy (pay no attention the mistresses, girlfriends, very comely secretaries, etc).
The important thing is breaking or at least opening up the clan structure. Monogamy doesn't really do that; barring cousin marriage does.
"While the roots of the package of norms and institutions that constitute modern marriage can be traced back to classical Greece and Rome, the global spread of this peculiar marriage system has occurred only in recent centuries, as other societies sought to emulate the West, with laws prohibiting polygyny arriving in 1880 in Japan, 1953 in China, 1955 in India and 1963 in Nepal."
The big difference for Christian countries (that includes not just Europe), was that monogamy was enforced/encouraged as the *sole* marriage system, instead of being an option alongside polygamy/concubines. For instance, Vishnu in the incarnation of Rama has only one wife (that even becomes important for carrying out a religious ceremony, where he is being urged to take another wife as he has been separated from Sita for a long time) but in the later incarnation as Krishna, he has many wives. The explanation given is that this is to fit in with the needs of the different eras and the different purposes for the incarnations.
There's been a lot of ink spilled over the passage in this epistle of St. Paul, with the pro-divorce side saying that what he means is that the overseer (bishop) can only have *one* spouse with him at a time so if he's divorced that's still okay if the former wife is alive:
"The saying is trustworthy: If anyone aspires to the office of overseer, he desires a noble task. 2 Therefore an overseer must be above reproach, the husband of one wife, sober-minded, self-controlled, respectable, hospitable, able to teach, 3 not a drunkard, not violent but gentle, not quarrelsome, not a lover of money
...12 Let deacons each be the husband of one wife, managing their children and their own households well."
But what all agree is that it lays down that church leaders can only have one wife at a time, can't/shouldn't be living unmarried with a lover or mistress, and can't have several wives/one wife and concubines.
Take the 'option' thing with a grain of salt, at least for India. It's more like there was no law expressly forbidding it, but if and when polygamy did happen it caused a great deal of drama and was generally discouraged (try the movie Bajirao Mastani, which is the story of a king shocking everyone by taking a second wife).
Really fantastic dance numbers, too. I found it funny that Rao's victory celebration after threatening the rival leader into submission was powerful enough to make it on the wider Internet
What are the great changes that could happen a thousand years in the future if the economy keeps growing at 2% annually? I find it hard to imagine that the world economy will continue to grow for centuries, but many do and it’s interesting to think about.
It would be a story of humans growing richer and richer. What does that mean? If it means anything, it means people will continually have more of the material things they want.
What are these material things people want? I’m not going to get into say sex, interpersonal relationships, social status and meaningful work. All of those things matter immensely, of course, but I consider them intangible and not what I’m talking about.
The material things we want are: shelter, food, and transportation. Again, happiness has a lot to do with relationships and all that, but I’m focusing on the other things. And thank God material things do matter. Humanity may be no better at relationships than we were 50,000 years ago, but we are better at building bridges and making movies.
So what can we do better in the future? The material bottlenecks of happiness are housing, transportation, safety entertainment, beauty, and good food.
Digging deeper, the bottlenecks are natural resources: land, energy, water, and other raw materials.
To me, that’s the key: natural resources. I don’t think we can grow at 2% a year for a thousand years unless we can somehow make natural resources cheaper. Basically that means alchemy. Will the future be able to 3d print a sky-scraper residential building?
Unlike most people here, I don’t read much science fiction. Are there good sci-fi writers who intelligently imagine cheap production of housing and raw materials? Because it seems to me like that would be the key to a future much greater than ours: cheap, great housing. Living in an awesome apartment for next to nothing. Imagine 5 billion people living in an awesome apartment for next to nothing! Would that not be great progress!
I don’t want to dismiss those, like me, who would prefer to live in a house with a nice front yard and backyard. More people living like that would be awesome too!
But how are housing materials ever going to be cheap? I realize that land is often more expensive than the construction on top of it, but a residential sky-scraper is expensive too.
Imagine a resource-cheap future in which a residential sky-scraper is cheap! Do any sci-fi writers imagine this? I honestly don’t know. Do they have an even semi-plausible scenario for how?
What do you think could make the future of humanity much better in a thousand years?
I think there will be strong limits to growth unless there are major scientific breakthroughs.
There's also the option of harvesting additional raw resources from off-world. This would require some scientific breakthroughs to really leverage well but there are several active projects.
Furthermore, I would not discount the potential for inflation to cause the listed numbers to keep rising even while prices rise in turn to balance it. A "steady state that pretends to be growing" seems entirely possible. Even if people are aware of the fiction, much like current arguments about the useful fictions of fiat money in the first place.
Resources aren't the bottleneck to constant progress. Scientific breakthroughs do indeed allow virtually unlimited growth.
You've left out a bunch of things by limiting wants to housing, transportation, safety, entertainment, beauty, and food. Communications, plumbing, and healthcare are some of the biggest ones. And increasing technology helps with all of them.
Silicon is an essential ingredient in all of the computers we use, and we aren't about to run out of it. We are constantly finding new uses of carbon, including microfibers and Vantablack, since it is such a versatile element. Water may be a source of energy, if nuclear fusion gets off the ground. Who knows what other uses we can make of highly abundant resources?
And that doesn't even count going out into the solar system, and beyond.
I suppose you can have unlimited growth without the quality of life going up much for most people. If quality housing remains scarce -- and why wouldn't it? -- then the quality of life won't improve a ton for most people because housing has a fundamental impact on it. All sorts of futuristic gadgets won't likely change that. I suppose everyone could live in the metaverse and physical reality could essentially disappear for all practical purposes. It's a possibility, but I wouldn't bet that it is a likely one.
AI Quinn mentions below to read "Age of Em". That reminds me that Hanson believes that most people - or ems or whatever -- in the future will live at subsistence level. He reasons that by taking the outside view that the most typical situation, judging by the past, is that most people exist at subsistence level and we are just currently living during exceptional times. I don't know that I buy that, but his vision of the future is one in which the economy continues to grow without most people experiencing an increase in quality of life. There must be many such possible futures.
Quality housing isn't scarce, it's just that quality housing in ultra-high-demand areas is scarce.
That will never change, even if we build a billion new square km of habitable land in space stations, everyone will still want to live within walking distance of whatever the current coolest five cities are.
You can build ten thousand identical nightclubs in your city, but one of them will have a queue around the block and the others will be empty.
Just thinking out loud about a way to quantify (long) covid effects on cognitive performance. FIDE Elo ratings for chess players from all over the world are publicly available (here 2015-2021 https://www.kaggle.com/datasets/rohanrao/chess-fide-ratings). These are a proxy for cognitive ability. Break them down by nation and average within each nation. Check whether the trend over time of average national Elo score shows a different behavior in countries that were hit more or less severely and at different times by covid. It’s essentially a diff in diff design. I don’t have time to do this myself, unfortunately. Watch out for the devil in the details of course: confounds likely abound.
I tried Ozempic starting in February, and stopped around the beginning of April. I hated it because of the side effects. YMMV.
I did find it to be effective in lowering blood glucose. I thought the weight loss effects were because of the side effects, and it struck me as an unhealthy regimen.
How high'd you push the dose? I started on 0.25mg a week in May and am up to 1mg a week right now. My use case is probably different than yours. I'm trying to push below 6-7% bodyfat or so by the end of summer, which I've done before, but it's normally a miserable process where you end up unable to think about anything but eating by the time you're done. This time, it's been easy as hell. Just eat at a calorie deficit every day and keep doing it. No hunger. No cravings. No willpower is necessary. It's gotten big in the bodybuilding community for this reason. Way safer than prior popular diet aids, shit from stimulants to thyroid hormone to pesticides that block carbohydrate metabolism. This is so much safer.
And personally, I've had no side effects. Not experiencing hunger is the only reason I even know I'm taking it.
I think it causes both nausea and weight loss, but I don't think the weight loss primarily driven by the nausea. I've had significant weight loss from it and minimal nausea.
I had a relative say the same thing, but he's morbidly obese so everybody told him he should tough it out. He declined, though, and quit taking it after 2-3 months. I guess it was pretty bad, at least for him.
I once took something called topamax as a migraine preventive, and felt groggy and nauseated all day long on it. Lost weight in the couple of months I took it, entirely because I felt too queasy to eat. Later the stuff was used as a component of a weight-loss drug.
I've been hearing something something about politicians "creating jobs", and I realized politicians only have the ability to create jobs by making government bigger, and thus hiring people to fill those particular jobs. As for creating jobs in any other industry, they only have the ability to restrict jobs, by forbidding things that people would want workers to do.
Can politicians ever create jobs in a way that doesn't involve only removing obstacles, or hiring people for the government?
I mean, does building infrastructure count? The construction of the interstate highways probably created jobs both immediately (construction work, which I guess does fall under 'making government bigger', at least temporarily), as well as indirectly and for decades after...
There is also the argument that well designed regulation can allow markets to flourish which would otherwise be impaired by information asymmetry. Thinking here of the SEC and the stock market, but commodities probably counts too.
Right off the bat — Defense procurement represents around 5 percent of national manufacturing output. These are mostly civilian jobs with non-government employees. Although I had difficulty teasing the info out of BLS reports, a non-technical summary said that in 2021 this accounted for over 2 million jobs — which were higher paying than average manufacturing jobs.
Likewise, there's a huge ecosystem of private contractors who hire from the private sector that work on government-sponsored projects. Probably there are more people who are employed by private businesses dealing with the government than there are who collect a government paycheck (but that's a wild-ass guess on my part).
NIH, NSF, DARPA, etc. hand out grants to Universities fund research. These handouts keep most of the STEM-related professoriate employed.
Creating jobs, ideally, means creating opportunities to generate wealth. If you drain a swamp to create fertile farmland then you create opportunities to generate wealth by growing crops on that land. Or in a less extreme example, you build a new road to some place, and suddenly some kind of economic activity on the margin becomes worthwhile. Or you fund research, which creates some new technology, which makes new ways of creating wealth possible.
There's also plenty of wealth-destroying ways to make jobs. The government can hire someone to go round breaking windows to create jobs for glaziers. It can ban the use of agricultural machinery to create jobs for farm workers. It can stop enforcing laws against shoplifting to create jobs for security guards. I'm sure there's even more cynical examples to put here.
And then there's wealth-creating ways of destroying jobs, like funding research that creates technologies that replace human workers with machines.
But my point is that these examples only create opportunity. Governments don't order businesses, or customers, to farm newly-drained land, or to replace windows they had broken.
Per US Treasury: "The federal government spent $6.27 trillion in FY 2022. This means federal spending was equal to 25% of the total gross domestic product (GDP), or economic activity, of the United States that year." A lot of private enterprise in the US depends on the US government purchasing their products and services.
Governments don't directly order people to build bridges or work in their pork-barrel projects either nowadays. They pay the private sector to do that. Or even the QuAnGO sector.
Rischard Shelby wasn't trying to get his constituents jobs at NASA, he was trying to get them jobs at Boeing.
I don't see why it shouldn't count as creating jobs if the taxpayer money goes through the balance sheet of a private company first.
Raise taxes and/or increase the deficit to create jobs? If the government is buying the service then, for this purpose, it would count as a government job. No one is saying the government is sitting on a pile of cash instead of spending it.
And I don't think that counts as creating jobs. I could create jobs the same way, if I borrowed to do home improvements or someone gave me money to do them.
OK. So your question REALLY was: "Are there any examples of "politicians" creating jobs that weren't (a) directly the result of government capital outlays, (b) that weren't the indirect result of long-term economic and social policies that promote economic growth?" Well, I can't think of any jobs that are more than two or three degrees of separation from the government teet. You might want to check a place like Somalia to see job creation is going there.
My question was "Can politicians ever create jobs in a way that doesn't involve only removing obstacles, or hiring people for the government?" My main point being that the government's claims of creating jobs mostly means allowing jobs to be created, most of the rest being the government paying people to do stuff, whether that is their employees or another company. From this thread:
* Well-designed regulation can build confidence in markets, allowing more people to believe it's worthwhile to participate, which can encourage job growth, but doesn't create it in of itself.
* Negotiating with foreign governments to improve trade relations can encourage job growth, but doesn't create it in of itself.
* Adjusting macroeconomic policy can encourage job growth, but doesn't create it in of itself. This is a prime example, actually, of the government being able to destroy jobs, but only allow them to be created.
* Draining swamps lets people plant crops, build houses, or build businesses in the formerly unsuitable land, which can encourage job growth, but doesn't create it in of itself.
* Building infrastructure: allows commerce to take place along, say, roads, which can encourage job growth, but doesn't create it in of itself.
* Suppressing piracy and suppressing crime can encourage job growth, but doesn't create it in of itself.
Others are mentioned that are harder to define.
My main point is it is a rare instance where government that creates jobs that wouldn't be created otherwise, yet politicians take credit for job creation.
Infrastructure comes to mind. Things like finishing the Great Lakes canal system, the Panama canal, dredging natural harbors deeper to improve the tonnage of shipping they can accommodate, etc, that greatly reduce natural barriers to commerce and thereby unleash economic activity that wouldn't otherwise be profitable.
Of course, how to *measure* how many jobs such infrastructure creates is a best-guess practice at best, so politicians are much more wont to make ridiculous claims by simply looking at the number of jobs the economy added while they sat in the chair and claiming them as "jobs I created," like a shaman taking credit for the rain.
So in sum, I think the most accurate statement is (a) yes, politicians can in some circumstances create jobs, but (b) most such claims of "job creation" are nonsense.
Anecdotally meditation decreases sleep need in most meditators, especially on meditation retreats. This WP page[1] supports that hypothesis, and this paper [2] gives a more detailed overview. In the extreme case, it could be that meditating >2h a day decreases sleep need by 2h[3] (giving the benefits of meditation at ~no time cost), though the evidence is shaky (via [4]).
Hmm, interesting (and thanks for the citations). It may be that I haven't meditated enough to notice the sleep difference from meditation (although I've always had a shaky sleep schedule anyways).
Anecdotally, what *I've* been noticing is that there is a strong influence the *other* way round: the more sleepy I am, the worse of a meditator I become (since the sleepiness itself becomes a distraction).
Has anyone succeeded at getting a Substack custom domain up? Mine has been stuck at "Your domain isn't ready yet. This can take up to 36 hours while the DNS records update" for the past week and a half.
3.) Substack should generate CNAME information and show it afterward. If it doesn't, you can contact support or read up on how to find the CNAME it generated.
4.) Go to your DNS provider (Hostgator?). Go to records.
8.) Set time to live to the smallest amount the provider allows.
9.) Save the record. Wait the time to live amount and then see if you can get through on the url. It might be intermittent for the first two or three days but you should be able to see it sometimes getting through within an hour or two.
You can also check the records on any number of sites. Here's one:
Just type in the full url (www.astralcodex.com) and you can see all CNAME records associated with a url. Which currently shows that address is pointed to a namecheap landing page.
If the DNS is on Namecheap you would need to go there, not HostGator, since you are not hosting the site. There's a tutorial on adding Namecheap records here:
I think did everything in 1-9 right, so the problem must be Namecheap. Is it possible for the domain name to be on Namecheap even if I get HostGator emails about it, and it's accessible from my HostGator account, and I have a HostGator account but not (AFAIK) a Namecheap account?
I don't have a Substack, but I do deal with DNS a lot & happy to help. Have you added the CNAME record yet, or is it stuck before that point? (ie in this article, is it stuck at Step 1 or Step 2)
...on the theory that it looks like www uses the main domain, and the text beginning with "target" is what Substack told me to use. Is this right? If yes, I'll try it again. What are you checking to tell you if the CNAME works?
Note, because there's an existing www CNAME record, you shouldn't be adding a new record, you should be modifying the existing record. Come to think of it, there is almost always a www record created with a placeholder value in any new zone file. Substack's instructions should maybe talk about modifying the www record rather than creating a new one.
> I think did everything in 1-9 right, so the problem must be Namecheap. Is it possible for the domain name to be on Namecheap even if I get HostGator emails about it, and it's accessible from my HostGator account, and I have a HostGator account but not (AFAIK) a Namecheap account?
Yes, actually!
One thing that commonly trips people up: it's possible for a Zone File for your DNS to exist in multiple places, with only one of them active. You may have for example:
- DNS hosting at the registrar (where you bought the domain, presumably Namecheap)
- DNS hosting attached to web hosting (maybe this is where Hostgator comes in?)
You set the Name Server (NS) for your domain wherever you registered the domain (presumably Namecheap).
As far as I can tell there's no business relationship between HostGator and Namecheap (although I do have trouble keeping up with all the buyouts).
I can see the domain was registered on 9th July 2020, then Updated on 8th July 2022. I'd check your email accounts for anything from Namecheap around that time (or indeed any time).
People often cite "The food is terrible and the portions are so small" as an example of an incoherent set of complaints.
I don't see why. Unless the food is so terrible that you can't eat it anyway, this is a perfectly reasonable set of complaints. The textureless, cheeseish pasta that I got last time I flew long haul United Airlines in economy was pretty bad food, but I'd rather have eaten a larger portion anyway rather than eat the tiny portion provided and suffer hunger for the next ten hours.
It's not so much that the complaints themselves are incoherent or contradictory, but often times people make a set of complaints that don't build on each other, and sometimes those people are just venting.
Relatedly, we can see this "direction-based" thinking in other areas, too. This post (https://www.lesswrong.com/posts/Kow8xRzpfkoY7pa69/the-affect-heuristic) mentions that when primed with the benefits of nuclear power, people will underestimate the risks. This is probably because its common to make decisions based on blobs of emotion, rather than detailed claims. We see this a lot in politics, too, with emotive thinking and one-sided policy debates.
It's true that those two claims about the food are not incoherent; nor are they contradictory. But, I would expect those claims to be caused by venting, rather than a nuanced model in someone's head.
Presumably the people who find it incoherent simply have a more conservative definition of the word "terrible" than you do, which wouldn't include "pretty bad" airline pasta that one wants more of.
I always took that as a set of complaints; that there wasn't just one thing wrong with the place.
(1) The food is bad
(2) Even if the food was good, the portions are too small (be that "to satisfy hunger" or "for the price they charge")
There may be other complaints as well besides those two, e.g. the service is slow and/or bad; the kitchen is dirty; the staff are rude; there are long waiting times because they always mess up the reservations and so forth.
On a trip to Ireland, I stopped along the road to ask a man if there were any places close by where I could eat. He thought for a minute, and then said, “I can think of two, but the problem is no matter which one I sent you to you’d wish you’d gone to the other.”
Not at all incoherent. Restaurants (most products really) have two possible appeals; quality, like the high-end restaurants, or quantity, like the buffets. Small portions of low-grade food fail every bar.
It just sounds silly. Like forgetting the name of the guy from My Name Is Bruce.
An IQ of 80 or so is no big deal. People with that IQ can manage all the basics of life, and if they had teachers or parents that took special care to make sure they understood important things that are a bit challenging, like banking, birth control, basic health info, they should be fine. Of course, if you've got a bit of a eugenics project going on in your country, those people have a strike against them. As for IQ scores breaking down at the extremes -- I don't know whether they do or not. In the case of this kid, there were no obvious disabilities interfering with her functioning. I don't know what her IQ was, but certainly way below 80. Her speech was like that of someone 1/4 of her age. In the case of people who have disabilities that interfere with them understanding or taking an IQ test, there are tests that are much more practical, involving tasks like using the stove, the bathroom, riding in the car, following simple directions, answering "what should you do if" type questions that are standardized and are a reasonable picture of a person's level of functioning in common situations. It does seem likely to me that a test of this kind when someone reaches puberty would be quite a good predictor of whether they are going to be able to parent decently as adults. Whether it is right to sterilize people who do poorly on that test, I don't know, but I do see a case for it. And it's not a realistic option to explain adult life and its demands to them, and strongly urge them to choose sterilization.
But I actually think that if we were going to try to weed out anything from our gene pool, it would be much better to focus on the kind of huge empathy deficits you see in sociopaths, and on a tendency for violence.
Correlation does not imply causation. How do you know that poor people aren't more violent for the same reaosn they're poor? Why would somebody prone to aggression, poor impulse control etc be able to earn a good income?
I don't understand quite what you mean -- esp. since the "it" stuck in there means sentence could mean that prosperity affects violence more than it affects IQ. or that prosperity has more effect on violence than IQ does. And whichever thing you meant, where were you going with that? I wasn't talking about reducing violence by trying to raise average IQ, I was talking about reducing violence by trying to reduce reproduction by violent individuals. I have not given a lot of thought to whether I could possibly be OK with sterilization enforced by law, but if I were to be OK with anything, it would be sterilization of the violent. Commit I extremely violent act, or 2 moderately violent ones and you're sterilized. A lot easier to swallow than forced sterilization of people who are not too bright, especially when the cutoff people have in mind is an IQ or 80 or 85. Jeez, lots of people with an IQ of 85 are enjoying life, taking decent care of themselves, and working. Having an IQ of 85 is no big deal. Get over it!
Also, I was also suggesting that we try, via some sort of selective reproduction, to reduce the number of people who have very low empathy. Many of these people are not violent. They simply fail to get it that other people are sensate beings like themselves. They are cars, everyone else is traffic. I think these people are not playing with a full deck. Their mental representation of the motives and inner lives of other people are impoverished and simple. It's like being color blind, or having a learning disability -- except that the consequences for society are far worse.
Is there a reason you would draw the line at one or two violent acts, which seems absurdly low to me? I am against forced sterilization in general, to be frank, but I could at least understand the argument if you were looking at serial offenders.
Beyond that, is your goal to genetically weed out the population (I think yes)? I have significant doubts that such a plan would work, as it seems to me that violence comes from many sources and it would be unlikely to remove it from future generations. (As an aside, even if it did work I would be concerned that even though our societies would work better domestically, foreign countries that didn't neuter their aggressive sides would be much emboldened to attack and expand). If you are using it as a proxy for bad parenting and don't want to let these people raise kids, then I think that's more reasonable as an argument. I'm still against most sterilization schemes, especially where there's much human judgment involved.
For completeness sake, I will say that I'm not completely against any of your ideas here. I had a friend with a very disabled relative, a young girl about to go through puberty. This girl had no hopes of life outside of an institution, and could not ever handle even basic life functions for herself. They gave her a hysterectomy. The thought process was that she would never be able to raise a child, having a baby would be dangerous, and she could never consent to sex in any legal sense. Practically speaking, if she ever did have a baby it was likely the result of rape from someone else in the institute or an employee there. No good options. As much as that choice must have been painful, I support it. If there had been no family to approve it, would I support the government/society making that decision? I don't know. I still see the value in the reasoning, but very much struggle with giving other people the right or power to make that kind of choice.
You kinda missed the point. I'm not saying, let's do this now, or even less do this someday. my main point is that if we were to try to change how many people in our society have a certain trait, I think it makes more sense to make violence and lack of empathy the target, rather than lowish intelligence.
As for drawing the line at one or 2 violent acts. First of all, as I said, I was not sketching in a finished proposal, just giving a general idea of something we might do. But I believe I did not say "one or 2 violent acts." Most people commit one or 2 violent acts a week, if you count little stuff like yelling curses at the ref while watching sorts on tv. I believe what I said was one very violent act or a repeat offender where both acts are moderately violent. By very violent I had in mind things like murder, attempted murder, beatings that put someone in the hospital. By moderately violent I meant things like beatings that leave the victim with multiple visible injuries, shoving people down the subway stairs. Yes of course there are all kinds of judgment calls involved and the discrimination will be made imperfectly.
Our legal system is not clean enough for me to be comfortable with a one or two strikes scheme, even if the crime seems significant. A DUI for a young adult could easily fall into that category, where we sterilize them even if we don't consider it intentional. And it's hard to prove intent. And an upset society has a tendency to impute intent (like saying that the intent to get drunk and then drive was intent to kill).
More to your original point, I also tend to think that different people with different personalities is good for society. Sociopaths can make hard but necessary decisions, like in wartime or downsizing a failing company, that someone with a lot of empathy could never make no matter how disastrous.
And again the problem of who gets to decide what traits to weed out and what test is used. I could use my personal sense of "violent person" and sterilize some percentage of the population, while another person who could also be thoughtful and intelligent might pick a much higher or lower percent based on their own sense of "violent person."
I think sterilizing anybody for 'societal benefits' is pointless and does more harm than good. There's no proof that violent people are violent because of genetic reasons. There was a footnote in Scott's post 'Galton, Ehrlich, Buck' that went :
"the Nazis killed nearly ever[y] schizophrenic in Germany, hoping to eliminate “the schizophrenia gene”. Today, Germany has exactly as many schizophrenics as any other country, because there are thousands of genes involved in schizophrenia, and all the deleterious variants are present in some frequency in the healthy population."
If you sterlized everybody who committed a violent act, for one thing, sterilized or not, they'll continue to live and if they're prone to bouts of violence, people around them continue to suffer. In addition to that, they'll probably get a lawsuit (and a payout) in their favor for having their right to procreate taken away.
For another, such a flippant proposal can only be made today on violent men. Would you support sterilizing a woman that punched or beat up somebody?
Right now, human knowledge of genetics is rather embryonic. Consequently, any attempt to improve the 'gene pool' is likely to end in failure and/or disaster. Its futile and unrealistic- like frogs trying to dam a river to get a cosier hangout spot, or something.
There are plenty of horribly violent women, and yes, I think if we were sterilizing people with a history of serious violence we should sterilize both genders. You seemed not to have noticed that I am not proposing that we go ahead and do this, so no need to tell me that it would be difficult to do in practice, thanks. My main point was that if we were going to try to eliminate a trait, I think it would make more sense to target propensity for violence and absence of empathy.
Is there evidence that very low empathy levels are genetic, or mostly genetic? 'Cause I reckon a lot of people might be like that due to the way they were raised, childhood traumas, etc. that make them stop caring about anybody but themselves... Has an "empathy gene" been identified for which your eugenics scheme could feasibly put into practice?
And people are normally born empathic. Babies and toddlers seeing other babies and toddlers cry often cry themselves. Don't know whether anyone has looked at the kids who do *not* how this response and see if they differ later in life.
Yes, and that shouldn't be a surprise, seeing how all personality traits are quite highly heritable: "Results indicated that 56% of the variance in antisocial personality and behavior can be explained through genetic influences, with 11% due to shared non-genetic influences, and 31% due to unique non-genetic influences." (Christopher J. Ferguson , The Journal of Social Psychology, 150:2, p. 160-180 (2010) , DOI: 10.1080/00224540903366503 ).
Also, "the way you were raised" strongly depends on whether your parents are sociopaths, and anecdotal evidence indicates that it's not the children that suffer the most trauma in these families that turn out sociopaths, it's their "golden child" siblings.
I'm aware that IQ and violence are correlated, but how much of the variance in violence is accounted for by IQ? (Or you could give an answer in terms of correlation between the 2.)
I'm thinking of switching careers and becoming a translator. Do people here think that's a viable long-term career path, or will AI render it obsolete within the next few years?
I’ve been doing some freelance translation as a part time job over the past few years. Machine translation is pretty good already. I think AI will obliterate the low end stuff pretty soon. The high end work should last longer, but you’ll be working with a machine, not instead of one. My best guess is that there’s about ten years at least left in the high end stuff, maybe more. A lot depends on your language combination, e.g. Mandarin to English would be in demand, but also certain minority languages into English. The other thing that would help is if you have knowledge of a certain sector e.g. if you have a medical degree you could be well placed to translate medical articles. So if you can find a niche, there’s still plenty of work, at least for now.
I have worked as both a translator and interpreter (unofficial), but was more than 15 years ago. Programmes to assist translation (Trados I think was the most popular one) seemed to be just starting back then, but I still got lots of work and made a decent living without using Trados. Nowadays I would say if you can speak a couple of different languages fluently, you are better off to get into international business (sales or quality department in a multi-national for example). In Europe, being really fluent in a few languages gets you opportunties everywhere. (In the US no, because eveyone is just expected to adapt to English by default)
Translating legal documents, technical documents or just general texts already means you'll get a machine translation and be asked to fix it up quickly and cheaply.
Translating literary texts like novels or films will probably not be automated, but like any creative job there are lots of people who want to do it so opportunities and pay are pretty bad. And tasks that used to be entry level jobs in that field are also being taken over by machines.
Even live translations of speech are getting pretty good, so I think even being an interpreter has a limited career potential at this point, except possibly sign language interpreters.
To some degree it depends on what language you'll be translating, but if it's "any big language to English" there's probably a lot of automation and if it's "niche language to English" there's probably very low demand.
Respectfully, that is an absolutely terrible career path right now. Someone close to me recently graduated college with a degree that basically meant her only option was translation; it has been a huge effort for her to retrain and do something else. The writing is on the wall for translation. It’s not that AI will render it obsolete within a few years; for 90% of use cases, it already has.
Exceptions include technical translation and being an interpreter... But those will be phased out as well. Sorry to be the bearer of bad news.
I recently used GPT-4, to translate relatively niche technical text - and it did a decent, though not perfect job. Currently, I think just using GPT out of the box would make a technical translater much more efficient - basically needing to check the results and make some minor changes to specific terms.
My friends who are translators are all already bemoaning how much of their work is becoming "machine-assisted translation" (aka starting with AI translations and then proofreading/finetuning them).
If you are also an interpreter, that seems much safer (for now, at least).
I do think this is one of the areas most likely to be impacted by AI advances in the near future.
I don't think it has a future in terms of widespread adoption. But it's fun, and there's no reason that enthusiasts can't keep using it indefinitely. It really depends on what you want to get out of it.
Gemini reminds me of Usenet and similar parts of the early social Internet: mostly hackers writing about hacking for other hackers to read, with a sprinkling of other eccentric personal interests. A lot of the content is about the medium itself ("Hey, I wrote a new Gemini server in Common Lisp; check it out!") Epilw don't post pictures of themselves (though they could); instead they talk about ideas. Normies have never heard of the medium, and, even if they did, most wouldn't get past the modest technical barriers to entry (I know this is the opposite of what Gemini's creator wanted, but ¯\_(ツ)_/¯ )
Whether all this is good or bad depends on your tastes! I find it a refreshing break from modern social media. I especially appreciate the ethos of people being weird without either shame or ostentation, a gestalt that does feel intensely retro to me in the best possible way.
Thank you for the answer. I find the protocols refreshingly simple, unlike modern web when browsers are prohibitively complex and buggy, while I can write a new gopher or gemini client as a programming exercise. While I appreciate advantages in having a single platform, I believe that placing resource exploration, presentation of information and app platform in a single basket was a bad tradeoff, though perhaps unavoidable one. But I still miss multiple specialized tools and protocols of old internet, and I think that monocultures are bad.
(If this is inappropriate for the thread, feel free to delete)
I wonder how hard it is to predict who will marry interracially.
I ran into an old acquaintance the other day and was surprised to find out he married a black woman. Not surprised because he was a raging racist or something. It's just that in our social circles it's still a bit of an unusual thing to do. Looking back, there was nothing about this guy that said "yes, this is precisely the kind of dude who would marry a black woman."
Maybe the big ML models used by dating services are able to identify people more likely to "marry out" than the average user. Either way, it's clear some humans do, and some absolutely never would. What's behind it? Is it genetic? A particular upbringing?
I'm not even sure why I'm so interested in this question, but I am.
I predict if you're not racist it's just a function of how many people of other races you meet / have in your social circle. I think that's pretty small for a lot of people, but probably depends on where you live and what your interests are.
I dated an Asian person for a while and I don't think I have any more or less interest in dating Asians than anyone else, she was just someone I happened to meet in the Bay Area rationalist community. I agree black-white relationships are less common, but I think this is mostly because of social/class de facto segregation.
>I agree black-white relationships are less common, but I think this is mostly because of social/class de facto segregation.
Blacks and whites looks very different, behave very differently/have different cultures, and have different intelligence distributions. It would be weird if there weren't substantial preferences for or against one race or the other.
'De facto segregation' makes it sound like its a matter of not being in contact with each other that's stopping the relationships. But the idea that a white woman with a masters degree would date a poor blue collar black man from the hood, or a hood black girl wants a dorky upper middle class white nerd, is just silly. These populations are very different, and on average this should have huge impacts on dating preferences.
From 3 percent of all newlyweds being in biracial marriages as of 1967, to 19 percent as of 2019. That's a slow but steady social shift persisting for several decades across this huge society.
And the biggest recent shifts within that are among whites and blacks, according to this previous Pew report based on data through 2015:
>And the biggest recent shifts within that are among whites and blacks
To be clear, the trend for whites and trend for blacks have both increased, not the black/white combination. The most common interracial marriage type (42%) is between whites and hispanics, and many of these 'hispanics' are of majority european ancestry.
I would guess its correlated with big 5 openness - which would also make it correlated with being liberal, creative, having a high IQ, and low authoritarianism.
My guess is that there's a pretty large element of randomness in the form of "you meet who you meet and the heart wants what it wants", and given that probably the single biggest factor is what fraction of the eligible people in marriage range (usually some combination of proximity in location, age, socioeconomic status, and values) are the same race as the person in question - people who live and meet their spouse in a 99-one-race town are probably marrying that race.
You surely already know this which is why you're bringing it up here in a pretty nonjudgemental pseudonymous forum, but I'll say nonetheless: I'd be cautious who you discuss this curiosity with, it's a culture-war landmine territory.
There is also (at the risk of being controversial and surely offending someone) a "lack of better options" element. My social universe is almost exclusively white and Asian, and the typical interracial pairing would be an attractive Asian woman with a less-attractive white man.
You don't expect the most attractive and highest status white men to wind up marrying interracially, although highly attractive Asian women often will.
I recently ran into someone claiming that the idea of SpaceX's starship is 'utopian nonsense'. I gave them the usual spiel about lowering $/kg to orbit making the industrialization of space feasible, but I was wondering if anyone had some more serious sources about the tech and economics involved. The real minutia of everything from water reclamation to cost studies on lunar mining, etc.
There are a couple of books on the economics of space mining etc (Mining the Sky by John Lewis comes to mind), but I've found little serious analysis of what could be done with Starship. Historically I am not sure we've done a great job at foreseeing what can be done: I've see serious analysis from two decades ago saying that reusable rockets will never make economic sense, for example.
The main argument for Starship seems to be "built it and they will come", which is certainly possible. But equally, working inside the space industry, I can't say I've seen that many serious people start building projects around Starship. That could change once Starship is proven, but the economic argument for space has always been tricky outside of a few use cases (imagery and communications principally).
That's a tough one. There are serious sources on much of this, but they're highly technical and even the broad survey works are written for people with an engineering background. And there were some good popularizations in the 1970s and 1980s, starting with O'Neill's "The High Frontier", see also T.A. Heppenheimer and G. Harry Stine, but they're rather dated and probably antipersuasive to someone coming in to them cold in the 21st century. I don't think I've seen anybody do a decent update on the concept.
Zach Weinersmith of "SMBC" fame and his wife Kelly are finishing "A City on Mars: Can We Settle Space, Should We Settle Space, and Have We Thought This Through", which might be what you need (and I expect it will have references for more technical detail on anything you want), but it's not out until November.
Have you read _Mining the Sky_ by John S. Lewis, by any chance? If so, what did you think of it? (It's probably the most serious book on space industrialization I've happened to read, and I wonder whether I got in near the top of the literature for laypeople, if there's much higher I could go, or what. And, knowing what I know of your background, you'll find most of it old hat and so you could probably skim it pretty fast if you had to.)
Yes; I have a copy and refer to it frequently. It's pretty good for what it is, but what it is is pretty narrow. We need something that simultaneously makes the case for "here are the valuable resources and opportunities waiting for us" and "here's how to set up economically viable infrastructure for space industrialization". If all you do is tell people that there are asteroids with infinite free platinum or whatever, they'll correctly point out that it would cost us a couple billion dollars to send out an SLS/Orion combo to bring back a few hundred kilograms, once.
Oh my god, in my career as a software developer, I have seen computers do some completely demented shit, but I just ran across something that really takes the cake. My webcam had randomly turned on, and I tracked down that the application that was doing it was something called WavesSysSvc. It's an audio application that seems to have installed itself along with an audio driver update, and it turns on the camera to track your head so that if you turn it while wearing headphones, it adjusts the volume so that it sounds like the audio is coming from the direction of the computer.
This is by far the most overengineered thing I have ever seen, and I can only assume someone somewhere was gunning very extremely hard for a promotion or perhaps it was to avoid being fired, but either way, yikes!
Which stories do you have about the craziest thing a computer has ever done to you?
Apple AirPods can do the same thing but don’t use the webcam, just the gyroscopes inside. Using the webcam is the cowards way out! (I almost always turn off this feature but it is kind of cool when watching sports)
tip of my tongue: there was a blog I found either through Scott's blogroll or through a 2nd order blogroll that was seemingly exclusively about 9/11 and its theological implications
I remember specifically a post that opened up with "on some day God decided to crash Himself into Himself" and it went on to make an unconventional point about how if God doesnt exist then everything is forbidden (as opposed to the more traditional take in which everything is permitted)
Can't find it for the life of me, anyone knows what I am talking about?
No idea about the blog, but that's a common idea in continental philosophy. Dostoevsky's (paraphrased) "if there is no God then everything is permitted" (i.e. without a Lawgiver there are no laws) was turned on its head by Lacan into "if there is no God, then everything is prohibited" (i.e. without a set of laws we can choose to follow or defy, it's impossible to do anything of any meaning), and more recently Zizek built an idea around "if there is a God, then anything is permitted" (i.e. ideology functions as a way to channel and justify the basest of desires).
If it was a philosophical blog you could try searching with some of those names.
Suppose I have relevant technical expertise and I believe Manifolds odds on the room temperature SC claims being for real are an order of magnitude too high. How much can I stand to make by betting on this? I am not registered with manifold or any other prediction market and if the potential payoff is less than four figures it is not worth the effort. Fwiw I am in the US.
Update: I did sign up for Manifold and bet half my portfolio on `no' but (a) that's only play money and (b) the setup doesn't seem to discriminate between `thinks the market is 10% overoptimistic' and `thinks the market is 1000% overoptimistic.'
The setup does distinguish this in that your bets move the market and you want to keep betting until the market hits where you want. This just doesn't do much for the superconductor market because there's a massive amount of betting on it and you don't have enough mana to move the market.
Au contraire, I am willing to bet real money, and am asking if there is a way to do it.
Conversely, if there isn't a way to make real money off prediction markets being overoptimistic on this topic, then there isn't a good reason to expect prediction markets to be accurate.
Edit: If by `real money' you mean instead something like `sink my entire (liquid?) net worth into this' then no, I am not willing to do that, and wouldn't even if the claim was something proven impossible (like perpetual motion) because counterparty risk etc would be unacceptably high.
I think that Manifold uses the formula (amount of NO you pay)/(probability of NO according to the current market) to decide your payout if you're directionally correct (though you lose your entire bet if you're directionally wrong), so given that the YES probabilities for the highest-volume LK99 markets are at ~20% you can expect a ~25% return if you're directionally correct. I also think that Manifold has been explicitly preventing users from buying too much mana at once to prevent situations where users buy too much mana to go all-in and then fall into financial ruin when they lose, such as like what happened with the Whales vs Minnows incident earlier this year. Putting this together if you go with the highest option and buy 100,000 mana (for US$1,000), you can hope to get a ~25,000 mana return which you can then realize as a ~US$250 profit for your favorite charity.
However if you *really* want to make real money you should try Kalshi (or become an expat and try Polymarket).
Oh also I suggest that you do calibration training first so that you feel more certain that the market's an order of magnitude too high. The binary calibration tests out there are kinda sucky (the most user-friendly one I've found is at http://www.2pih.com/caltest/ ), but there is a user-friendly multiple-choice calibration test at https://bayes-up.web.app which may be equally helpful.
Technical question- does anyone have a preferred search engine for finding matching/plagiarized content, that has an API or can be used programmatically? I have a specific business case for tracking down and finding plagiarized content online. I'm not actually interested in academic plagiarism per se, but I have a financial need to match up copied text.
I've been using Copyscape to-date, which is basically 'OK' and is clearly using Google's API. Just was curious if anyone knows of better programmatic alternatives
One big problem with prediction markets -- and I think the current markets about replicating the superconductivity experiment is Exhibit A here -- is that you can't distinguish the predictors from the gamblers. You may think there is no meaningful difference between them, but walk into any casino and look around: those are gamblers. When the crowd at the craps table all bets on Pass it doesn't mean Pass is likely to happen. Pass is never likely to happen. Yet it is a rare day in Vegas when the crowds at the craps tables don't bet on Pass.
This is also true of the stock market - Wall Street vs. Robinhood. But Wall Street knows they're better than Robinhood and pays Robinhood lots of money to get priority access to the right to take the other side of Robinhooders' bets. There is no amount you can bet on Robinhood that some Wall Street firm won't be delighted to bring back down to the Wall-Street-approved market price. If enough criteria are met (size, liquidity, attention, etc) this should be true for prediction markets too.
I'd suggest that the criteria should include `critical mass of participants have expertise in evaluating the object level question.' For the stock market the object level question is `will this firm make enough profit to justify the price' (or maybe `will the price go up'), which is a question that I think a lot of people on Wall Street do have a lot of expertise evaluating. For `will LK99 turn out to be a room temperature superconductor' I doubt there is any such critical mass of `smart money' trading on the relevant markets.
On Wall Street, the "critical mass" can be one firm with a lot of money that hires an expert, feels confident in its decision, and then deploys millions of dollars to correct the mispricing. It doesn't matter how few people there are who know the correct price for some security as long as one of them works for Goldman Sachs and another one for Jane Street.
I don't know. I think on Manifold, no, because there's nobody who has too much more mana than anyone else. On Polymarket and Kalshi, in theory maybe, but I don't know about them.
Markets are aggregating per dollar not per person. In a deep, mature market like publicly-traded stocks, the gamblers' money will be a rounding error compared to that belonging to institutional investors.
Prediction markets (except perhaps the bookies?) are not there yet.
What do you think the odds should be? I haven't been following the story as closely as I would like, but it seems like there are multiple replications and a video now, yet the prediction odds haven't moved much.
1% is probably way too low for it being technically a superconductor. This has passed most of the obvious-crackpottery-or-hoax gates, and I don't see any real point in trying to run a serious hoax on this. Still possible to be a false alarm, those do happen in science, but not at the 99+% level.
However, 1% for being a *useful* superconductor, might not be too far off. I haven't seen anyone even guess at a critical field strength for this stuff, and the expected behavior of superconductors is that critical field strength decreases with temperature.
And that's the part of this that I find most frustrating. Everyone is talking about how this is either a pathetic fantasy or it's the miracle that gives us fusion energy and quantum superdupercomputers and everything else under the sun. Sometimes there's a footnote to the effect that we need to pin down the critical field strength (and production yield, material properties, etc), but with the implication that this is just dotting the i's and crossing the t's on the incipient technological miracle.
20% chance this is technically a superconducter, 1% that it is a useful engineering material, 19% just a stunt that can only ever barely levitate a snowflake, might be a good guess.
20% for room temperature ambient pressure superconductivity is IMO way too high. I'd say 5% its even `barely technically' a room temperature ambient pressure superconductor, and maybe 0.5% it's actually useful. I started off higher, but enough time has passed that the lack of a replication by a reputable source is starting to smell fishy, as is the (apparent) lack of `give the damn sample to an established lab to measure.'
I'd say less than 1% as well. There are too many priors of purported room temp superconductors that never panned out (at least ten?). It seems like there's a much higher probability that scientists involved are deluding themselves than an actual HTSC material has been discovered.
This fully precludes you from detecting the real thing, no? An outside view is useful at times, but if your heuristic doesn't include any information about the (present) claim, then you can't actually update until after the fact.
How does it preclude detecting the real thing? Exclusive of my cynical worldview, my heuristic includes the following information and suppositions.
Firstly, AFAIK, there's no theoretical reason that we know of that precludes a room-temperature low-atmospheric-pressure superconductor from existing. But likewise, we don’t have a theory that will allow us to make one by design. So a true RTSC would be coming out of theoretical left field. Shit like this does occasionally happen in science, but with no theoretical framework to explain and/or predict the behavior of an RTSC, I'd say that that reason alone would be strongly indicative of the null hypothesis.
Although I have seen some chatter that LK-99 uses a Phonon/BSC mechanism—I thought that wasn't possible, because (my understanding is that) BCS requires a Fermi surface where the electrons can form Cooper pairs. Correct me if I'm wrong but isn't that situation impossible when the lattice of the superconducting material is vibrating too fast (i.e. too hot) for those pairs to form? I admit I only have a layperson's understanding from what I've read in generalist science magazines such as Nature, Science, and Scientific American. So I'm perfectly willing to be corrected on this.
Even if a such an animal as RTSC exists, detection/confirmation may be more difficult than some seem to think. According to a physicist friend of mine, we'd not only have to confirm the Meissner effect but we'd have to measure about half-dozen other behaviors to confirm that it's an actual RTSC. That's not going to happen overnight.
If there is "information about the (present) claim," that would update and/or correct my knowledge, I'd like to hear it. But, no, I don't see how we'd be able to "update until after the fact" (If I understand your meaning correctly).
Also: if you are dealing with a specific technical question (like the room temperature SC business) and people with relevant technical expertise are not widely participating in the betting market then the market is going to suck. Agree manifolds odds seem to me to be at least an order of magnitude too high.
Gamblers provide free money for the predictors to eat up, thus incentivizing prediction. Without real money on the line, this breaks down because predictors aren’t actually incentivized by anything other than fake internet points.
Right. Like in the stock market, which is a well-functioning prediction market, you have smart money and you have dumb money. Because the returns to stock trading and investing are so high, the smart money dominates -- because equities in general tend to be smart investments. But in a casino you have craps tables, roulette tables, slot machines. All of those things have expected negative returns on every bet and they attract nothing but dumb money.
Metaculus and polymarket aren't currently attracting enough smart money for a wisdom of crowds effect to dominate. I'm not saying the participants in these market aren't "smart" in the normal sense of the word, but their bets aren't smart. The superconductivity bet is trading like a meme stock.
> Metaculus and polymarket aren't currently attracting enough smart money for a wisdom of crowds effect to dominate. I'm not saying the participants in these market aren't "smart" in the normal sense of the word, but their bets aren't smart. The superconductivity bet is trading like a meme stock.
Because 28% odds of having found a room temperature superconductor is insane. So far we have one group claiming they have found The Holy Grail. And the market thinks there is a 1 in 4 chance of that being true?
If you're very sure of this, presumably you've made a considerable bet on it? That's a lot of free money lying around if you're right. (It would be quite strange if nobody took it, wouldn't it?)
Is there? I’m in the us and think the prediction markets are over optimistic by an order of magnitude. I’m willing to bet a couple of thousand dollars if there is a reasonable payoff for being right. How do I do this?
If it were trading on the CME or the NYSE I would make a bet. polymarket is very new to me and about the only thing I know about it is that it's illegal and I'd have to use a VPN and put real money into some crypto "stable coin" I don't know anything about. There's lots more risk here than simply getting the bet directionally correct.
Question for the botanists in the crowd, with a short foreword. One of my local farms is getting rid of their strawberry plants in one patch, so I brought four of them home yesterday afternoon. I'm planning on planting them tonight in pots (I'm an apartment dweller, so no place for an in-the-earth garden). They're Flavorfests, which I've read ask for acidic, loamy, and/or sandy soil.
When a plant needs an acidic soil, is that asking for soil with fatty acids in it? Should I be supplementing my plants with, I dunno, fish oil capsules every so often?
"Acidic" commonly refers to the pH of a liquid; pH means "potential of hydrogen". Liquids are either acidic or basic (so wine, for example, is chemically an acid [pH 4] although that's not how we normally think about it). We also give soils a pH because of all of the hydrogen-containing compounds that are naturally/easily available in the soil. Neutral pH (eg water) is 7, acidic soil that plants like would be around 5-6.5. You can buy fertilizer specifically for acid-loving plants: https://www.fafard.ca/products/fertilizer/garden-fertilizer/natural-fertilizer-for-acid-loving-plants-berries/
Hello, I wanted to ask if people think it will help me later If I take one math class now during my bachelors. So I'm currently doing my bachelors in economics and this semester I can choose one subject from a different area in addition to my normal econ classes. I can take anything from chemistry, math, languages, philosophy, asian politics etc. I am tempted to take a math class because I find mathematics quite interesting but it would probably be a very hard class for me. Whereas if I took philosophy, history or basically anything that isn't maths, I would expect it to be easier.
My main question here is if people think it will matter at all for my later life if I take this math class. Can I write "Has Economics bachelors with one analysis class" on my CV and it will make me look good. Or should I just do something easy so I have a higher chances of passing all my exams. And if I want to learn maths I can just do it in my free time? Here in Germany we mostly only take classes related to our Major only two out of about 24 classes will be non Econ for me.
If anything, your resume will probably be slightly worse if you take a math class because math professors actually give grades that reflect performance, while you can take any social science 101 class and put in minimal effort for an 'A'
I would recommend taking at least linear algebra and statistics (to be fair, these are two classes). I don't think you'd be able to avoid using them in your professional life.
Yeah, my question was poorly worded. As part of the normal Econ degree I already had statistics and linear algebra and calculus. And I was wondering if in addition to that I should take a class from the Maths bachelors which would probably be a lot more advanced.
It depends on the class; tensor calculus would be somewhat useful if you plan on branching out into data science. Beyound that, though, I'd suggest taking the hardest class you can afford, schedule-wise. You're paying for it, and no one will probably read your courseload on your resume anyway -- so why not actually learn something ?
Answering the actual question: Can I write "Has Economics bachelors with one analysis class" on my CV and it will make me look good?
No, no-one gives a shit after you graduate.
Answering the implied question: Will taking this class help me afterwards?
If taking this class will improve your math skills (safe bet): yes, definitely.
If you have enough discipline/interest to learn this by yourself later (not likely, but you know yourself more than a random internet person): don't take it, spend that time whatever is more productive
I was suspecting that this would be the case (that it doesn't matter for my CV). I think I really am interested in math enough that I could learn a lot of what would be covered in class on my own, but it would just feel like such a wasted opportunity not taking a math class in University because it would probably be difficult, and then later learning math on my own in my free time.
But if I am being rational it's probably unlikely I would really study that much math on my own.
What are your plans for later in life? Depending on what your undergraduate economics classes are like, I do think analysis (i.e. Real Analysis) would help you if you ever want to apply to US Econ PhD programs.
I would guess there is a 40% chance that I will do a masters and maybe 20% PHD. I would prefer to either work, start my own business or do something EA related and not do a masters.
I'd say take a look at the admission requirements for any grad schools you may be interested in, if the extra math course would help then go for it. Here in Canada/US, having a good/decent grade in Real Analysis is very beneficial for Econ graduate admission. Besides that, I doubt one course will make a huge difference, so I would just take whatever you're most interested in.
I'm confused how you could complete an econ degree without taking any math classes. You can't really understand a lot of economics without a solid foundation in probability, statistics, multivariate calculus and differential equations and these aren't topics a typical student would fully cover in high school.
Yeah, my question was probably poorly worded. Just during the normal econ degree we have to take two maths and stats classes covering all the things you just mentioned. Just in addition to that I have the option to take a class from the actual mathematicians where we would probably have much more advanced maths and many more proofs etc.
I'm surprised that noone has talked about the end of the Salem/CSPI prediction market contest at all. Maybe it's because the CSPI itself hasn't made any announcements about it either? The whole thing is really weird.
It came out the day before my last Mantic Monday post and I didn't want to write lots of stuff in a hurry. I'll cover it next time. Congratulations to everyone involved.
I would have expected them to announce the end of the contest, post to Substack and Twitter, and generally try to raise publicity around it, similar to the way they did when the contest first started last year.
Interesting, but this doesn't seem particularly user-friendly for doing arithmetic (which I think is why the Hindu-Arabic system, by simplifying each number's notation into a list of digits, really took off, though the other ones never did).
Yeah, having played around with it that does seem to be the primary issue. You could image ways to expand it to include numbers larger than four digits pretty easily, but adding and subtracting with it looks difficult.
Thanks for sharing. As someone who likes to doodle new number systems during boring meetings, this one is similar to one I came up with but far superior. I love it, and will devote my next doodling time to mastering it.
I hate to pick on one sentence out of a whole essay that I mostly agree with, but I feel compelled to pick on this one because I think it shows a huge problem with Freddie's world view:
"These pieces don’t stem from a conservative desire to stigmatize or a lack of sympathy for hurt people"
A "conservative desire to stigmatize"? Is that really how he understands conservatives' thought processes, that we get up in the morning and say "ha ha, I feel like stigmatising someone today"?
No, when conservatives stigmatise things it's because we believe that they're objectively bad and shouldn't be celebrated, and that tolerating or indulging objectively behaviour (like drug addiction or crime or obesity) isn't doing anyone any favours. It comes, in fact, from exactly the same set of desires that Freddie is exhibiting in this essay, it's just that he thinks he's doing it from compassion while conservatives apparently do it from a desire to be evil.
I suggest that Freddie deep down _is_ a conservative, he just can't bring himself to admit it because he doesn't understand conservatism.
> Is that really how he understands conservatives' thought processes, that we get up in the morning and say "ha ha, I feel like stigmatising someone today"?
> No, when conservatives stigmatise things it's because we believe that they're objectively bad and shouldn't be celebrated, and that tolerating or indulging objectively behaviour (like drug addiction or crime or obesity) isn't doing anyone any favours.
Desires do not have to be intentional. The idea, I think, is that people have unconscious biases, and the desire to stigmatize is one of them. Some people has less of it, some more, some people try to correct for it consciously, some just do not notice it and some may even identify with this desire to some degree.
Concervatives tend to have stronger and unreflected desire to stygmatize. This result in them thinking some things are objectively bad and being more trusting to the arguments in favour of this position. And if these arguments are disproven, this doesn't really shake the confidence because the arguments are just rationalization of this unconscious desire and not the actual reason why concervatives hold the positions they do.
> Concervatives tend to have stronger and unreflected desire to stygmatize.
I'm not sure this conclusion is justified as described. For instance, conservatives also have considerably stronger disgust sensitivity. What does anyone do when presented with something they find disgusting? They stigmatize it. Therefore the propensity to stigmatize could follow from disgust sensitivity, not some innately "stronger and unreflected desire to stygmatize".
The fact that stronger disgust sensitivity leads to stronger urge to stigmatize and conservatives tend to have stronger disgust sensitivity supports the point that conservatives have stronger urge to stymatize.
There seem to be some misunderstanding going on. Maybe I used the word "unreflected" inappropriately?
The idea was that you can look you desire to stigmatize and thus figure out a way to decrese it's power over you. So, for instance if the reason you feel the urge to stymatize people is because you feel disgust towards them, instead of following the disgust based urge you can think: "The fact that I find X aesthetically inappealing doesn't necessary mean that it is indeed bad and deserve stygmatization, I should be extra careful, while evaluating X and maybe even give it extra good faith, due to this cognitive bias of mine".
And if you do not do it - then it's what I mean by "unreflected", it has more power over you and indeed leads you to more stygmatization. Stronger disgust leads to stronger desire to stygmatize which is also harder to reflect upon.
That sounds more like his depiction of the excesses of therapeutic culture; see also:
> Good people spend a great deal of their time categorically and uncritically affirming others - telling friends and strangers alike that their desires are all legitimate, their instincts always correct, their perceptions of their own needs never mistaken or misguided, their self-conception compelling
> Correspondingly, we should all assume that anyone who is not affirming us is necessarily doing so out of a particular kind of politicized wickedness, that they are likely motivated by racism, sexism, homophobia, or other kinds of bigotry, and if these specific accusations are not plausible, then by simple evil
In a vacuum I'm just reading that as recognizing that's a thing liberals like to claim. It's the audience's position, not the writer's; I say something like "this isn't gnomes stealing your keys when your back is turned", that doesn't mean I actually think gnomes are doing that in other situations.
In context I still read it that way; you can't use the "conservatives evil" claim to dismiss them offhand because these aren't even conservative sources. Like when America's most decorated soldier started talking about PTSD and people could no longer use the "they're just cowards" dismissal.
My problem is Freddie's analysis is his generalizations about what liberals believe or claim. Conservatives like to make straw men out specific liberal positions, and although Freddie declares himself to be a Leftie, he sounds more like George Will in Leftie drag.
IMHO, nowadays, liberals have a wider diversity of views than the right.
Not if I'm the audience. And if he's saying something he doesn't really believe to pander to a different audience, he's even more likely to say stuff that he doesn't really believe about the parts that are important to me. So where does the expected truth value of this work come from?
OK, given that it's FdB, my prior is that there's some useful insight and maybe fact that can be mined from this without too much effort or aggravation, but for almost anyone else that sort of signalling is pretty much an instant deal-breaker.
> Once upon a time, there was strong social value in being “cool,” with the concept of cool referring to a studied indifference to the vagaries of fate; turning away from the pursuit of cool to defining ourselves according to our weaknesses and neuroses was a profound mistake, cool was a humane and correct social value, and we should return to it.
He's too smart to be an honest to God Marxist, that's for sure, and he seems not to spend much time advocating for Marxist viewpoints or policy goals, but he still hangs on to the label. I would say he's not clearly a blue or red tribe guy at this point.
Blue-tribe heretic, I think. Still culturally part of the blue tribe, and still pursuing the goals that he thinks the rest of the blue tribe should have, but he says the blue mainstream is doing it wrong, and they say he is.
I also saw that and merely concluded that he would at least partially fail an ideological Turing test. I give credence to his assertion of being a communist (or at least something in that vein), but I also note that Marxism’s core appeal is to pre-modern sensibilities, such as the idea that traits typical of families scale up to apply to entire societies. So it doesn’t surprise me when genuine Marxists say things that might seem conservative. Market- and rights-oriented classical liberalism was a significant break with the past. American conservatism is, to my mind, significantly motivated by assuring the continued health of core pre-modern institutions (family, church, local community, voluntary societies) within a larger social framework that is a classical liberal framework. Their hostility to Marxism arises because Marxists identify a different institution (central government, the intelligentsia, the party) as the proper locus of those pre-modern sensibilities, which would displace conservatives’ preferred institutions. (Note: that’s my belief about conservatives’ belief, not about Marxists’ belief.) I wonder if progressives don’t favor different institutions, because their worldview strikes me as even more pre-modern than Marxists’.
I’ve seen men do it in mixed-gender groups, to a moderate extent, and to a similar extent to women in the same setting. I don’t know if anyone does it to the extent Freddie is complaining about.
I will second this: I am a conservative who enjoys reading Freddie deBoer. (I am not a libertarian, but I also enjoy reading Bryan Caplan.)
Great quote from the linked piece:
"Given that pain and suffering are literally and permanently unavoidable in human life, teaching others to be resilient rather than teaching them to be victims is an act of mercy, and cultivating resilience in yourself is an act of essential personal growth and adult development."
I keep finding evidence for the hypothesis that people have a fundamental political archetype that is independent of the politics they profess to have - even if someone has adopted, say, California-style liberalism because they grew up in California (pro queer rights, etc), their process of reasoning through values aligns more with the archetypal conservatives (but people should still pick themselves up by their bootstraps, I did). For another example, it's well-known that there's a lot of people in the SJ movement that are fundamentally evangelical pastors, searching for sinners.
That said, every time I read his blog, I come away with the feeling that Freddie is definitely a leftist that thinks like a conservative.
I feel like there's a strain of old-fashioned communism that comes across kind of like some older strains of Christianity: humans by themselves can be selfish and mean, but if we come together for a greater purpose we can do great things; the world is harsh and unforgiving, so it is up to us - to each of us individually, and to all of us collectively - to create the world we want to live in. The kind of belief system that produces people willing to be martyrs, and uses them as inspiration.
Especially since it was clearly a random stereotypical example which had nothing to do with the rest of the comment or thread, and moreover, support X rights in no way implies that X rights are missing (c.f. "free speech rights"), just that there are people opposed.
But sure, here's the ACLU page that give an overview of the issue.
Before my cat settles down for a nap, she sniffs around briefly where she plans to curl up. Without giving it any thought I naturally assumed the purpose of this was to reassure herself of her own smell, assuming she had occupied the same spot previously, and the stronger the better.
But every few months she finds a new place to nap, and seems to abandon the old spot for a while, and sometimes permanently. So it occurred to me that her sniff test may have the opposite motive to what I had assumed, and that in fact she is checking that her spot does _not_ smell too strongly of her!
That would sort of make sense in the wild, in that a cat wouldn't want to leave a strong scent of itself in any one place, especially if it had kittens. It is also well known that female cats move their kittens at the drop of a hat, presumably when they sense some kind of nearby danger but possibly also when they feel that the smell of the lair is becoming too strong.
I don't recall Jackson Galaxy ever mentioning this in his cat programs. So maybe it is an insight that is not widely known!
My cats change up their sleeping spots approximately every week. I assumed (like you) that this would be a behavior that would reduce predation in the wild by reducing their scent profile in a given nesting place.
The one exception to this rule is boxes. A new box invokes joy and exploration in both of my cats. But they stop being interested in the new box after about 3 days. Maybe cardboard absorbs their scent faster than other materials?
With boxes, I think they lose interest because the novelty wears off. I have a bunch of matched boxes with holes of different sizes in them for cats or their paws to go through. Each has velcro on all 6 sides, so they can be rearranged into all kinds of structures. The cats lose interest in a new structure within a few days, but if I rearrange it they are fascinated and delighted again, and spend hours exploring it, ambushing each other in the maze, and napping.
That happened when I moved their cat gymnasium (that they had been ignoring for months) to another room. The started playing on it again — for about a week. Now they've lost interest in it once more. Unfortunately, it's big, and there are only so many places I can put it. OTOH they've never fallen out of love with my couch as a play space and scratching post. Honestly, I think they/re playing head games with me.
Wutz a cat gymnasium? These little Devons are maniacs, and if I don't supply them with lots of novelty and exercise they start dismantling my home.
About your lot getting bored with the cat gym: Sometimes you can re-awaken their interest by doing something like covering it with a sheet or taping a bunch of crinkly paper to different parts of it, of even just rotating it so it faces another way, or attaching an empty box onto part of it.. I am learning all the tricks. When I rearrange their box structure I also put catnip in some boxes, treats in others, fresh crinkly paper in others, toys in others, sometimes stuff in something like some freshly worn sox, which apparently smell to them like the breath of angels. Treasure hunt.
Dogs do that too, though, when circling a spot before laying down, and they wouldn't have the same incentives (eg. they care a lot less about covering up their defecation). I've always written it off as a generic checking behavior - take a closer look, smell for anything bad, and make sure there's nothing weird or odd about where you're about to bed down for a while Wouldn't want to get anything on your fur or lay down on a nasty bug or something.
Maybe she just gets bored with one place, or the changing seasons make different places more comfortable. The sniffing might not have anything to do with the place change.
No doubt that's part of it. But at the risk of flattering myself, my cat seems to favour places where I spend time, such as the duvet on my bed and my armchair. So maybe it's like Dutch drivers, who have a reputation for being more aggressive than most on the roads. I don't know if that is deserved, but I've heard it said that they don't want to be behind you, and they don't even want to be in front of you. They want to be where you are! :-)
Did you think of gathering experimental evidence? Like wiping your cat’s current curling place every time after she used it to keep it „fresh“ and see if her behavior changes?
I'm not sure that wiping would be enough to remove their scent. But was thinking of putting a fresh towel over their current sleeping places every two or three days to test this theory.
How does the local AI crowd view the field of "mechanism design"? There are results in that field ("incentive compatible / strategyproof" mechanisms) that describe conditions on certain types of interactions that make dishonesty pointless. It's the only set of proofs I've heard of that seem relevant to the alignment problem, but I'm struggling to find references to them in the alignment discussion. Am I just missing them, or is there something else going on?
Hey Scott M. My experience with the local AI crowd is that if you have a suggestion about alignment nobody wants to hear it. Breathe one hint about what your suggestion is and you instantly hear "that won't work" and/or "we already do that." One person actually told me my idea sucked *both* because it wouldn't work *and also* because "we already do that." And nobody every displays the slightest bit of curiosity about the details of what you are suggesting.
Call me picky if you like, but in my opinion reactions of this kind are not a good sign about the heads of people in the alignment field are working.
Is it? There's lots of wackiness and hostility floating around in AI/alignment because it's an inchoate, high-stakes field, but 'being annoyed and dismissive when a layman earnestly suggests their 'novel' idea you've heard 20 times before and understand why it won't work' is common to basically all expertise-requiring disciplines. It's possible, both here and elsewhere, that they are being over-dismissive, but you can see why it happens.
Yes, of course, it happens in all expertise-requiring fields that laymen make common naive suggestions that they think are novel ideas. The irritated dismissals I have gotten have been on this forum, and I have seen other people receiving them here too. Here's what I think is wrong with that, especially on a forum like this one:
-There's a much better chance on here than in most settings that you will hear an idea worth thinking over. Isn't that why you're here?
-AI is not just a new invention, it is a new kind of invention. The kind of expertise required to build it and the kinds required to figure out how to integrate it into human life in a way that works decently are almost certainly very different. And AI alignment is just a special case of "works decently as part of human life." People working on alignment should be drawing on the expertise of people in other fields, fields such as cognitive psychology, neurology, sociology, game theory, animal behavior and probably some more. I know that some of that does go on, but I don't think there's enough. I never hear much from or about experts in these other disciplines working on alignment. So if you hear an idea from someone on here, bear in mind that they may have expertise in one of these other fields, and in their field *you* are the naive layman, burdened with a bunch of silly half-formed notions about how various things work. Unless you are a genius polymath, many of the ideas about cognitive psychology, neurology, sociology, etc. that you are bringing to your work on AI doubtless are naive and inaccurate. Are you so sure they are not that you don't want to hear some thoughts from a professional in one of those fields?
-It's not like people working on alignment are all set and there's nothing to worry about any more. There are plenty of ways things could go wrong when we get AGI and ASI. Setting aside Yudkowsky-type concerns about Foomdoom, there are lots of more prosaic and probable bad outcomes. We are not wired to cope with an entity of this nature, any more than we are wired to eat healthily in a food court or use our time well in homes stocked with television, computers and gaming devices. There's good reason to think that social media, which I think of as sort of an early, weak version of AI spread out over many platforms, has harmed quite a few people. It would not be surprising if real AI does more harm of the same sort to more people. And then there's the issue of weaponized AI in the hands of criminals and nations in conflict -- all kinds of new and powerful ways of doing ghastly damage will be possible. Given all that, I think the people working on AI would be wise to err on the side of openness to new ideas.
-Irritated dismissals are rude as fuck. If someone writes a paragraph about their idea, and it is a naive useless one that you have heard 20 times before, why not read that paragraph and write one back saying that yes, you'd think that would work, but here was the problem we found when we tried something like that. You have now advanced that person's level of understanding of alignment. And unless they are unusually vain and stoopit, they will also start to grasp that alignment is a knottier problem than they were imagining. All that is surely a good thing.
And by the way, MicaiahC, in his discussion here with Scott McKuen, is doing an admirable job of exploring Scott's idea and thinking through it with him.
Good points, and I for one have updated slightly towards "people working in AI are often too dismissive about alignment ideas". I tend to be pretty cynical about the quality of online conversation, even in settings like the ACX comments. But perhaps there is hope!
So Angola, on the strength of that, can I send you some of my thoughts about AI development and AI alignment? None of them are ideas for specific approaches to aligning AI with the welfare of our species. They're considerably more meta than that -- ideas on ways to *go about* developing AI and trying to solve alignment. I don't know if you will agree with them, but they are not naive, cliche and silly, and I haven't seen most of them mentioned in articles.
I don't think there's much point sending them to me, I'm a layman myself and a somewhat out of touch one at that following the last couple years of developments. Might be worth posting on the next open thread or machine alignment monday, though.
Robin Hanson has said that alignment is mostly easy because we already have loads of theories around incentive design, and Eliezer disagrees. I believe most people in alignment are at least familiar with the concept of applying existing institutional design or game theory to AI safety, but think the theoretic foundations of that are too weak to usefully constrain an AI's goal system (my counterpoint in the sibling thread) or are working on more direct/immediate forms of safety. I think the latter group may be talking about this in some sort of jargon laden form, but if you want to find them, I think Prosaic alignment and Paul Christiano are the key words here.
Doesn't mechanism design rely on enforcement mechanisms, or that the parties involved has at least some common ground? I'm having trouble of thinking how an extension to auction theory would prevent a powerful entity from boiling away the oceans when running a set of very powerful computations, for example. Not saying this is a specific scenario to worry about, but the space of possible misaligned/fatally negligent actions is going to be far greater than set of enforceable mechanisms.
Here's an extremely informal sketch of my thinking. On the one hand, we want AIs to be honest about their capabilities to reduce the chances of getting ambushed. Hiding your goals and capabilities gets mentioned a lot as high-risk behavior. On the other hand, I've seen people also say "goal integrity" is a critical part of goal achievement; basically the AI cannot allow its goal to change if changed-goal => failed-goal. Getting rewarded just for staying on-task (whatever the task) is a very generic way of encouraging goal achievement.
What caught my attention about mechanism design: strategyproof/incentive-compatible mechanisms are ones where hiding information doesn't give any player an advantage. A goal is called "implementable" if there's an incentive-compatible mechanism, so honesty is at least tied for the best policy. There's a result that roughly says to be implementable, the true reward for an action can't depend directly on the player's own estimate of the reward. Roughly, if part of your grade is dependent on self-evaluation of performance, there can *never* be a situation with an honesty-enforcing setup.
Real-world actions are planned somewhat hierarchically. So let's say an AI goes through a meta-goal stage: "what is the best subgoal to pursue right now?" Current subgoal is G0, all other possible subgoals are H's. Goal-preservation would basically be a bonus on top of whatever "real" reward comes from achieving the specific goal - but the preservation bonus is getting granted ahead of time, as an intermediate reward. The outside world doesn't necessarily know what your goal is, that's internal. Therefore, the extra reward from goal-preservation has to be internally granted. Since it's internal, it depends on your current estimate of the true reward landscape. That means the most recent reward is dependent on your current estimate of the long-term reward structure. That's never incentive-compatible.
If we can actually couple goal-preservation with dishonesty in a very generic argument, maybe that at least reduces the number of things we need to worry about.
I think you're right in that goal stability is good, and probably a necessary sub component of alignment, but I'm not sure how the definition of "implementable" helps without the ability to constructively apply it to say, gradient descent learning or inference.
Is there an example or some more concrete case informing your intuition? Like, dunno if there's an advisor game where a leader multiple advisors who produce good and bad statements and you can consistently guarantee that... you do not pick the worst possible set of decisions, or have lower bounds on disutility. Obviously I just made up that case, but I want to know what is informing that.
I am skeptical that it has strong implications for the alignment problem of AIs. The setting there is too complex to be captured by this type of reasoning. But for some well-defined situations, this has been very successful, most notably for auctions.
There are auction mechanisms in which bidders have advantages if they don't reveal their true preferences. "Preference" usually means how much the bidder is willing to pay for an item. But there are other auction mechanism where the best strategy for a rational agent is to truthfully report her preferences.
The most common truthful auction mechanism is for the case where a single item is auctioned. In this case, all bidders reveal how much they would be willing to pay. The item is then sold to the bidder with the highest call, but this bidder is only required to pay the price of the *second-highest* bid. It is not hard to prove that the best strategy for every bidder is to reveal truthfully the maximal price that they would be willing to pay, regardless of whether the other agents play truthfully or not. The scheme is widely used. For example, Google uses it to sell ad placement.
Worth noting that the second price pays auction relies on an assumption that bidders don't care about each other. If my hated rival is bidding in the auction then making my own bid for more than my true price creates a chance that I am second price and my rival is inconvenienced by having to pay more than he otherwise would have, so I have an incentive to bid high (or bid low if I'm bidding against my friend).
Mostly I enjoy this as a technicality, but I also think it illustrates the limits of applying mechanism design to the messy real world.
Yes, excellent point. The whole setting assumes that the objective is to obtain the item at minimal cost (or not at all, if the cost is too high). And that actions like "shooting the other bidders" or "blackmailing the auction house to change the mechanism" are not possible.
If you're magnesium deficient (and sometimes even if not), magnesium glycinate can (anecdotally IME) have the effect of reducing general muscular stressedness. (Also nicer sleep.)
Given the current progress of AI, especially LLMs, how do you see the future of software development? As an average python programmer, should I be actively planning for a future in which there are far fewer jobs/the jobs require less skill and pay substantially less? Or will this just be another tool that we use to increase our productivity? Looking at a 30 year career horizon.
Firstly, as others have noted, 30 years is a very long time in software development. That said, I don't know what an "average Python programmer" does. If you are the kind of person who only knows how to put together some quick boilerplate scripts to transform some files around or to create a login page... then yeah, your job is going to be gone. If you're the kind of person who understands the basics of CS theory, common algorithms and data structures, big-O notation, and binary operations, then your job is probably safe for the foreseeable future (modulo the Singularity, I suppose).
Part of me is honestly a little bit frightened, because I'm a middling programmer with bad soft skills, and I feel like programming is actually the thing LLMs are showing the most promise for right now.
On the other hand, I have long said and still believe, at the point that programmers are truly fully automated away, we will have run headlong into "futureshock and wierd shit happening in real time" which short of inducing a full on technological signularity will at least induce a sort of cultural/economic singularity that is pointless to try to predict (or maybe even prepare for)
30 years are half an eternity in software development even without AI.
Consider the last 30 years. In 1993
* Most PC users ran Windows 3.11 (or plain DOS)
* Debian 0.01 was released
* Visual Basic, a terrible programming language popular in the late 90s was barely released
* Python was two years old, so if you needed a scripting language you would probably opt for the slightly more mature Perl (or sh or awk? Pessimistically *.BAT)
* Java, a favorite of corporate applications, would not be released for another two years
* For C++, the STL was barely released, RTTI would take another five years
* Mosaic, the first popular web browser featuring a GUI was released, JavaScript would take another two years
* The IPhone, herald of widely used smartphone platforms featuring "apps" was still 15 years in the future
There are not to many niches for employment in which one might survive 1993-2023 without changing programming languages. One possibility would be COBOL. It was obsolete in 1993, but some companies (e.g. banks) had big systems written in it which would have to be maintained. The lifetime of such systems is generally longer than that of COBOL programmers.
Another option might be C for microcontrollers. (I am not sure that C compilers were widely available for uC's in 1993. Perhaps one might have to use some horrible proprietary C compiler for a few years before GCC or SDCC add support.) While the uCs became more powerful and cheaper, C remains the language of choice if you care about speed.
Or numerical programming in FORTRAN. LAPACK and BLAS existed in 1992 and remain the standard up to this day. Of course, while the users used mostly FORTRAN in 1993, today they likely use numpy. So the niche is mostly restricted to maintainers of these packages.
Generally, there is often an oversupply of programmers in tech which is considered sexy (e.g. rust) while there is an undersupply for legacy tech which was corporate's favorite of yesterday (e.g. COBOL, large proprietary DBMS's (mostly in places where MariaDB would suffice), perhaps Java in the future). This is, or course, the result of an equilibrium, because most devs consider that tech to painful to work with. Personally, I would rather learn a new language at 50 than spend my life working with COBOL.
I think C (and C++, though significantly expanded) would be an excellent investment 30 years ago, and they still are, and certainly not just for microcontroller programming: most of those libraries in Python we use for numerics are really implemented in C/C++, practically all of system programming happens in C/C++, databases are implemented in C/C++, and so on. And while the use of perl declined from the most popular scripting language to something decidedly niche, it does have its niche. Various dialects of Lisp and Prolog are still used and are even getting quite popular (e.g., Clojure). And of course I am writing my papers like thousands upon thousands of physicists, mathematicians and computer scientists using TeX: a typesetting system and language created in the early 80s (there are actually some contemporary CS books written in even older troff). Relational databases which were already available and starting to get popular 30 years ago are still around and neither they, nor SQL are going anywhere soon. There are numerous long lived technologies and languages, but at a time it may be difficult to pick which are the long lived ones. Which is why, it is important to just get programming in general, not just specific language or technology, to understand how computers and systems work. Then one is able to transition easily: languages come and go, but computing is eternal.
I don't see a future in which fewer jobs exist. If anything, I expect AI will -increase- the number of programming jobs - although likely changing their character somewhat.
I think we currently have less than 1% of the demand that exists for programming satisfied. Less than 1% of 1%. The only programming work that currently gets fulfilled is the programming work that is, quite simply, either extremely financially valuable - and - easily monetized, or of particular personal interest to the person doing the work.
I expect AI to create new tiers of programming; right now there are, basically, three. You have the people who can program basically anything (but are perpetually busy, often not programming at all, but telling other people how/what to program). You have the people who can program anything that isn't too complicated, and is specified well enough (and are also perpetually busy, telling other people how/what to program). And then you have vast swathes of people who, if you give them very specific specifications, and the work isn't complicated, can write something that will mostly do what it's supposed to (and these people aren't too busy because they're rate-limited by other people telling them what to do).
AI in the hands of the first group of people could, conceivably, turn a single highly competent person into a one-man dev shop, replacing everybody else. This sounds scary, but - there still won't be enough of these people! That is something important to keep in mind.
AI in the hands of the second group might elevate them a little bit - but ultimately they still don't have the chops to do a lot of the work that the first group does, whose advantages are not strictly technical, or in the ability to write quickly - but in borderline magical abilities to just know what will need to be done later, and how things that are technically correct will fail anyways. They will likely continue to work in teams, because you still need to tell an AI what to do, and no one of them knows everything that needs to be done.
The third group, I think, will be elevated considerably by good AI, which should be able to overcome / compensate for some of their technical deficits. Many of them will be moved to the second group. Many more will, I think, be moved to a massive new growth industry - code maintenance. All the code AI produces will need to be maintained. (You could, in principle, have it rewrite the code - but this approach is like fixing bugs by deleting a class and starting over. You're just going to get -different- bugs!)
And, I think, good code AI will push a lot of people whose technical deficits prevent them from entering the field, to the point where they can enter the field.
"Won't making programmers much more productive lower their salaries?"
Only if the demand for programmers decreases. I actually expect it to increase; and for demand for programmers to increase in proportion to the quality of the programmer.
You have your developers who will become one-man dev shops; I expect the demand for them to, if anything, grow, because they can produce even more value.
You have your developers who are one tier below that, who can work effectively as a team; I expect the demand for them to grow as well, because, again, they can produce even more value.
You have the grunts, as well; many of them will be elevated to the second tier of developers as their technical deficits are balanced by AI, and they'll likewise be in higher demand. But all the rest of them will come into more demand as a much larger base of code worldwide needs maintenance and refinement. (I have longer thoughts on this, but, short version, I expect a lot more code maintenance to be necessary, rather than a lot less, because AI makes custom software viable).
For an example of a job that only kind of exists right now - a programmer equivalent for a home plumber. This doesn't make sense right now for most things - but, like, if I have an AI code me my own operating system designed around my specific needs / requirements, I might need to hire somebody to fix something in it, or make a specific thing behave slightly differently, particularly if I lack the technical expertise to specify to the AI what it is I actually want. Maybe this developer works with an AI to actually accomplish this, maybe they just hand-tool it; it doesn't really matter. There's a lot more work to be done.
I haven't tried ChatGPT for code generation, but I have played around a bit with GitHub Copilot. The latter is a useful tool, but is in absolutely no danger of putting my out of a job. It's at its best generating boilerplate code, but it needs a lot of active and highly skilled supervision to generate something of any complexity.
Three core challenges I see for future generations of LLMs to write large pieces of code autonomously enough to be substitutes for human programmers rather than productivity tools:
1. As we've seen with ChatGPT in other domains, LLMs are much better at generating output that looks right than at generating output that is right. The more advanced the LLM, the harder it's going to get to curate its output.
2. Figuring out exactly what you want the program to do is one of the harder parts of any nontrivial software projects. This is something that may eventually be solvable by sufficiently advanced LLMs, but it's going to be a long road to get there and integrate it with code generation.
3. Most work in software engineering is fixing, extending, and improving an existing code based. If you want an LLM to be able to incrementally improve an existing code based, you're probably going to need to specialize its training on that particular code based. This is potentially doable, especially if you have an entire history of bug/feature tickets linked to code changes that contain unit tests, but in the medium term it's going to be fiendishly expensive to train the LLM on a large dataset and unreliable to train it on a small dataset.
In the medium term, I expect the coding jobs that get automated away to be relatively low-skill and low-pay, while the more complex and more skilled jobs become more valuable because the AI tools make them more productive
My career advice in light of this would be to try to become one of the high-skilled, high-paid software engineers who are likely to be complementary to AI rather than replaceable in the medium term. Alternatively, if you don't think you're going to be able to make it as a senior or principal software engineer at a big tech firm, you could instead position yourself as a consultant or in-house programmer for a non-tech company that doesn't have the resources to properly supervise and curate an LLM without hiring someone like you. This latter will probably be replaceable sooner than the former, but I expect that to be several generations of LLM in the future.
Given that many programmers incrementally commit their work to GitHub (regardless of their level of professionalism), I am confident that this data could lay the foundation for a potent pairing of "CoderLLMs" and "SeniorAlfaZeroDeveloper." In the near future, we may see the emergence of a sophisticated low-level programming language that surpasses the comprehension of individuals with average cognitive capabilities. Within a timeframe of 3-8 years, programmers may evolve into mere operators.
Allowing AI's to write programs humans are unable to understand, especially important programs (healthcare, embedded, finance, OS kernels, etc.) would be like joining your honest Omnissiah fearing tech with unpredictable warp entities and hoping for the best. That is, it is extremely stupid. Fortunately, AI code generators are still too unreliable for this to be a realistic temptation within quite more time than 3-8 years
Firstly, chatGPT is the first-born, the market is always occupied by subsequent systems (which take into account the mistakes of the pioneer), and secondly, ChatGPT was not created specifically for programming, but it knows how to do it at the junior to middle level and in almost all languages, it knows most of the data formats and algorithms. And all this is just a casual effect of learning. My main argument was that programming is the area where the whole process is documented from start to finish, every step, bug fixing, development. The base for training the neural network is just crazy! Moreover, I think that the full power of neural networks is not shown to us, since the most advanced models are of military and strategic importance.
Based on current trends I strongly suspect that writing code is going to be done less, but architecting systems, doing product design, specifying acceptance criteria, and communicating with stakeholders (etc) are going to be in more demand. So basically the Senior/Staff eng skill-set, or maybe “startup founder” generalist skills.
Don’t think it’s going to be a good time to be a specialist in fact-based or narrow-context domains, and I suspect algorithms and LeetCode will be less relevant. Great time to be working on problems that require a large context window, which describes most of the broader software delivery life cycle.
The future of job market is impossible to predict long term. Since you cannot really predict the future your best bet is to choose the job you like and are reasonably good at. This job may disappear in the future, but at least you will enjoy it while it lasts. And of course taking into account that you might be facing hard times ahead is a good strategy regardless of AI (like, not assuming that you will be earning as much or more as you do now for the rest of your life). Now, there are better and worse ways to approach your programmer's career, and hoping to be an average python programmer may not be the best choice. One error is overspecialization: languages and technologies come and go, so try being a programmer, not just Python programmer, and certainly not average: I do not think Python is going anywhere in data science and AI applications. because of the number of libraries but I may be wrong. In other applications: ten years ago I thought it might be a good use of my time to learn Ruby... Also you may try to learn other things: distributed systems, databases, system architecture, AI, and really understand them. All of this is a good advice independently of the rise of AI. I do not think future jobs in software development will require less skill: if anything it is the low skilled jobs which will get automated first by productivity tools, also those using AI. And finally, the dream of eliminating programmers is nothing new: people in the 1950s thought that thanks to Cobol now manages will be able to program business applications directly just by telling computer the business rules in a semi-natural language form...
Based on the ChaosGPT incident earlier this year, it's looking increasingly likely that before the next 30 years are up, someone will shove a hundred random plugins at an agentized large language model, or something ridiculous like that, and accidentally *actually* wipe out (or at least severely threaten) humanity.
>In reality theres [...] not much gains in efficiency from chatGPT either relative to googling stack overflow.
Whenever I see this take it's baffling to me, and I wonder how our use cases differ so much. Between CGPT and copilot I feel like I've had easily a 1.5-2x efficiency improvement in most tasks, higher in some areas. Anything where I'm interacting with a framework I'm not familiar with, or writing boilerplate, or just trying to get a quick answer. This applies to both my work and my hobby code.
It's not even comparable to SO, because you can get it to write the exact code you're asking for in seconds instead of clicking through 5 outdated or not-your-actual-question SO questions before you get it. And if the question is trivial enough that SO will get it with one click, copilot has probably already offered an autocomplete for it before you've even tabbed to your web browser.
ChatGPT is good enough for all but truly novel code, which, in practice, covers probably 80-90% of what I write. And of that novel code, probably only 10% of that is novel enough that its structure can't be inferred by copilot from other nearby code.
These tools aren't foolproof, but they're like having a *really* fast intern with really broad knowledge of everything. You still have to direct it and proofread its code, but even with that there's major gains.
---
All that said, I still can't draw a conclusion as to its ultimate effort on the market. I'm a pretty good programmer, but I'm a shit economist.
Whatever the long-term prospects of ChatGPT and similar, I'm damned glad I have started a python contract where I am working intensively with ChatGPT, Azure, and Salesforce among other things, because I reckon that before long some or all of those (except possibly Salesforce) will be a required skill for practically every python contract. I hadn't used any of them previously, but managed to blag the contract after some righteous waffling about all my other skills at the interview!
Also, surely you can't be serious about ChatGPT being out of date already! Not saying you're wrong, but if it is then what has superceded it? Or do you just mean OpenAI is forging ahead with version 5 while their users are still stuck with versions 3 and 4?
I know some people here are knowledgeable about different voting systems. RCV/IRV just seem too complex to ever catch on widely.
Are there other, simpler schemes that accomplish similar ends? The goal I am interested in would be selecting candidates with the broadest appeal, as opposed to the FPTP and primary system we have in the US, which seems to incentive extremism.
Ranked choice can be simplified down to the level of choose-one by making the candidates decide (upfront, of course) who their first-choice votes transfer to.
I'm under no illusion this would be more realistic to implement than full RCV. And not just because it takes away much of the appeal that the ability to personally rank your choices has, and therefore won't get dedicated voting reform geeks advocating for it. Mostly - because it preserves all the systemic benefits of RCV in addition to vastly simplifying coordination of minority groups, so you'll never see the establishment approve of it. (Australia had it as an option for a while for senate elections, then the major parties got rid of it and forced people to rank their choices individually, solely and explicitly because it worked and let minor candidates form election pacts that resulted in one of them getting elected.)
But the existence of the option at least demonstrates that "complexity" is never an actual obstacle for RCV adoption. Institutional resistance to the benefits of RCV is.
FWIW, there are non-IRV variants. I like Condorcet methods that pick the candidate who would defeat every other candidate in a 2-way race; the algorithm can be involved but all the voters have to interact with is the ballot, which is just a ranking.
Why do you think the purpose of voting is to choose a candidate with broad appeal? It isn't.
The purpose of voting is to peacefully remove from power incumbents who have insufficient popular support. You should reanalyse voting systems with that in mind.
Good question. I guess voting has multiple purposes, the most important of which is preventing the exact same people from maintaining power forever.
But as long as you are regularly picking new people to run things, the goal should be to pick those with broadest appeal to the voters. This seems like a legitimacy issue - in the US current system we often get situations where effectively the median voter in a party primary chooses the eventual winner, so people at the 25th percentile on an ideological spectrum.
I guess ideally the system would select for “most competent” candidate or something but how do you know in advance?
Really what I am interested in is making “politics” less of a zero-sum situation. I think this is the root of a lot of other problems.
Ah yes, term limits were a good idea, but relatively rare around the world. In that case the election is about deciding if the political philosophy of the incumbent should be extended for another term, or replaced by a different one.
I am skeptical that many voters, especially swing voters, make decisions based on their assessment of the political philosophy of the incumbent. If they did, they would probably not be swing voters. Nor would models based on economic indicators be so successful at predicting elections.
Honestly, I'm coming to the conclusion that the jungle primary may be best, with a top-2 general election. All of the fancy voting systems are designed to handle multiple candidates in a single round of voting, but I think it turns out that having 2 rounds of voting is perfectly fine, if you plan them in advance. (And if you have convenient mail-in ballots.) Everyone goes head to head in the primary, and then the top two vote getters go on the the general. Almost invariably it is the more moderate candidate who wins the general. And the winner has a clear majority of votes, and there's no weird math or run-offs or anything that might confuse someone with IQ 80. It's a great way to build democratic legitimacy. We may be choosing between two lizard people, but dammit, we get to choose and that choice matters!
The catastrophic failure mode of jungle primaries is that in a divided field, you risk winding up with a general election between two candidates both of whom are strongly disfavored by solid majorities of the electorate but who have strong, narrow support bases. The go-to example of this happening in real life is the 1991 Louisiana Governor election, which produced a second-round election between Edwin Edwards (who four years previous had dropped out of the second round election when it became clear he was going to lose in a landslide) and David Duke (the highest-profile national KKK leader at the time).
A more common failure mode (at least here in California) is for the minority party's insiders to privately choose their preferred candidate, and for their voters to mostly support that candidate in the primary for fear of being squeezed off the runoff ballot by two candidates of the other party, resulting in pretty much what you would have gotten from a partisan primary system most of the time (except with more strategic voting in the first round), and occasionally in a general election between two mainstream members of the majority party (which latter has the potential virtue of giving the minority party's supporters a voice in the choice of which member of the other party gets elected, but from the minority party's perspective this often feels like being asked what kind of cancer you'd like to get).
The low-effort response would be, "yeah, that's Louisiana", but I'm trying to be better than that. :-)
It's a real problem - one of the virtues of machine politics and the "smoke-filled back room" is that it's less likely to select wingnuts, and the further we empower the general population, the more likely we are to get bad results. While at the same time, it's important not to place one's own opinions over those of the people. And yes, jungle primaries reward party discipline, and are sort of an electoral equivalent of the marshmallow test. I think we'll see some evolutionary pressure there, at least when races are competitive. But harking back to the discussions a few threads ago, it seems likely that the Nazis would do very well in a jungle primary.
I suppose, once I remove my dislike of the major parties, the main thing I like about the jungle primary is the multiple rounds, and that the top two go on to a head-to-head general election. Keep that, and I'd be happy with a different system for the primary (or primaries, as someone suggested below).
Any balloting system has some vulnerability to a highly-motivated insurgency, radical or not, that turns out to be really good at getting its own adherents to show up and vote. When there is broad weariness or disgust with the established party(s) that's when such a group can win the day. Hitler pulled it off once in such circumstances, so did Macron, so did Trump, so did Zelensky. So have others here and there from time to time -- ranging all over the ideological map, competing under differing electoral systems, etc.
Yeah, that's also Seattle. The incumbent generally has an effective lock on one of the general slots, the bulk of reasonable challengers split most of the rest of the vote so the second general slot goes to the biggest crackpot.
It's about the most incumbent-friendly system I can imagine.
I don't think the current ingroup-primary + single common round system is at all satisfactory, but I haven't seen any alternative that uses several rounds.
There is something to be said about maturing the public opinion over several months of state primaries, with smaller candidates slowly conceeding, while negotiating concessions along the way.
Does anyone know an alternative voting system that has several rounds of voting ?
This isn't quite what you're looking for, but by "jungle primary" I mean "not in-group". Everyone goes in to it, regardless of party. Candidates can choose to identify with whatever party they want - for example, in the past there've been signaling differences between labeling oneself "GOP" vs. "Republican".
It's a profoundly anti-party system. The parties could theoretically have their own private meetings earlier, where they all agree to get behind one candidate, and the other candidates agree not to run, but they'd have to do this without state infrastructure, and so far they haven't. Right now, in order to see who the party wants, we have to look at the candidates' endorsements, which will sometimes include things like "XY State Republican Party" or "123rd District Democrats", and sometimes different branches of the same party endorse different candidates. It's very entertaining.
But it doesn't have the feature you want, about extending the interplay between candidates, and fostering negotiation and compromise.
(Personally, I'd prefer a bit more federal control of US presidential elections. 10 weeks, 5 states per week, in order from fewest electors to most. Let the chips fall where they may. And schedule it to start in April at the earliest, so the season doesn't stretch out too long.)
"It's a profoundly anti-party system." Yep, which is a lot of why it's attractive.
"more federal control of US presidential elections. 10 weeks, 5 states per week, in order from fewest electors to most. Let the chips fall where they may."
YES!! I have daydreamed about this idea for years. It would have several huge advantages over our present ridiculously-stupid system of party caucuses/primaries followed by play-acted party conventions.
I've thought about it like so:
-- there are a total of four rounds of balloting in which the citizens of each state/territory each get a total of two opportunities to cast a vote.
-- the 4th round is the Election Day called for in the Constitution, in November, for which there will be two POTUS/VPOTUS tickets on the ballot. The 1st round is in April; 2nd round in June; 3rd round in September.
-- the states/territories are arbitrarily sorted into three groups, e.g. it could be alpha by name. Alabama is in Group 1, Alaska is in Group 2, Arizona is in Group 3, etc through the list. Those groupings rotate as national primaries amongst the first three rounds of voting. E.g. in 2024 Group 1 would be the May round, then in 2028 Group 1 becomes the July round, etc.
-- Round 1 has a ballot of unlimited size. Any Constitutionally-eligible person who by maybe March 1st has gathered (some number) of petition signatures of which at least (some smaller number) are from each of let's say 5 states, qualifies. We could also allow each of the national political parties to name one candidate for the ballot if that seemed desirable. Also all past POTUSes who are eligible for another term (not term limited) get on that initial ballot automatically, including the incumbent if he/she is eligible.
-- from that wide-open first round, the 6 candidates who got the most votes in it make Round 2.
-- then the 3 candidates who got the most votes in Round 2 make it to the Round 3, with one difference: by August 1st they have to each have named their running mate (a VPOTUS candidate).
-- then Round 3 happens in September, and the two tickets that got the most votes are set as the ballot in November.
There is of course less actual chance of this change happening than of my piloting the first warp-drive ship to visit the Klingon Empire, but for me it's a nice daydream.
I should also say that I'm not aware of sufficient investigation being done about the effects of the fancy voting systems in multi-round systems. Specifially, as much as I like approval voting, it seems like it would have a nasty failure case if implemented in a jungle primary.
Compulsory adult voting is underrated IMO. The incentive becomes targeting the median voter, which allows for policy positions of the major parties to evolve in sync with societal views.
Making voting mandatory will, at the margins, draw in stupider and less informed voters, as they are the ones least likely to vote now. I favor restricting voting for a smaller subset (e.g. epistocracy) though obviously there is unlikely to be a workable path to such a system.
Mandatory voting invites the naturally disengaged to consider the issues that are relevant to them and vote accordingly. The US system appears to reinforce existing partisan views. I don't think you can make a claim about relative intelligence of the voting populace between these scenarios with any confidence.
The compulsory system is akin to a market mechanism (where optimising to win support from the median voter is the goal, and the election is the price discovery mechanism). Non mandatory creates distortions that may seem logical (exclude those that don't care enough), but it seems as though the equilibrium this solves for is relatively static platforms, compared to underlying social evolution.
"which allows for policy positions of the major parties to evolve in sync with societal views"
This is, in my opinion, the absolute worst feature of the Australian system. It basically gives absurd power to pollsters and others in the media who "estimate" the will of the people on a given issue, without actually asking the people in any clear way. For example, the 2013 election was looking like it would be a referendum on immigration, but (reading the polls showing strong opposition to boat arrivals) Labor moved further and further to the right throughout the campaign until by the election, their immigration policy was the same as the Liberals'! People still interpreted the Liberal victory as a repudiation of illegal immigration, but for God's sake it would have been a lot clearer if the parties had taken clearly opposite positions to the election instead of identical ones. This sort of thing happens in nearly every election. And it makes it constantly unclear and disputed whether or not there's truly a popular mandate for anything.
(A bit late to this but) One side effect that's worth pointing out is that it makes voter suppression much, much harder to accomplish, since the government has to make sure that every citizen is able to vote. (Source: Am Brazilian, and we have compulsory voting here)
North Korea has compulsory voting with a turnout of 100%. South Korea has voluntary voting with a turnout of 77%. Another natural experiment is that NL went from 95% to 80% when they made it voluntary.
Yes I'm Australian. There is a fine associated with not voting, but there a wide range of exceptions (travel, infirmary, etc.) that mean in practice it is not policed. You are also entitled to get your name marked off at a polling place, but walk out without submitting a ballot. It conditions voters to vote as a default action, so the % of the voting age population that cast a valid vote is relatively high (~90%+ of eligible).
As a result, parties' platforms tend to evolve over time along both social and financial axes, and party loyalty is relatively low. Minor parties / loose affiliations have risen and fallen over time, in line with voter preference.
It does little for the quality of politicians(!) but does lead to a policy platform that mirrors the preferences of the median voter reasonably effectively.
Since Canada and the UK (with voluntary FPTP voting like the US) also have much less polarisation than the US, it seems to me that it's the primary system that's responsible for most of it. What other country has such a clear method of punishing a candidate who tries to appeal to the centre?
I cannot help but note the syntactic ambiguity here. I assume you meant that the party on the Right implemented gun control in response to a mass shooting, and not that the State implemented gun control in response to the party on the Right doing a mass shooting.
I would, in any case, be interested in the Australian view of using guns. It is part of the United States constitution, theoretically, so that if the government ever became too powerful then the people would have the means to take the government back. I have my doubts as to how applicable this is currently, with the existence, not imagined 230 years ago, of tanks, war planes, ballistic missiles, and nuclear weapons, but am unsure how to resolve it.
The general view here is that Americans' love of guns is completely insane. There are a few of us who disagree, and think self-defence should be allowed (there's even a Shooters' Party) but it's definitely a minority position.
I'd point out that we have a much lower crime rate than the US, so that doesn't help with selling the pro-gun position. And there's also an issue that I don't see much discussed: because anyone in the US might be armed, police seem to be very militarised, heavily armed, handcuff people on the tiniest suspician, and of course are regularly accused of shooting people on the tiniest suspician. Ignoring all the politicised (racial etc) aspects of all that, it seems there's an argument that available guns are in fact counterproductive for limiting state power. Self-defence against ordinary criminals is surely the best argument.
We've had Ranked Choice Voting here for a couple local elections, and it's fine - a bit more complicated, but it works okay. Mostly I wish they'd just do a "Top 3" ranked choice selection rather than ranking every candidate, so we don't have to do crap like figuring out one-ninth of a vote and so forth.
There's not really a system that specifically gets broad appeal while being simple, although Simple Proportional Representation gets close - whoever holds power has to have a 50%+1 majority in the legislature.
I'd call out Approval Voting as pretty simple. the voters marks all the acceptable candidates (any number), final victor is the person with the most people who included them in the approval set.
I think after a few elections in New York, Alaska, Maine, and San Francisco, it will be clear that these sorts of ranked-choice systems aren't actually too complex to catch on.
What I really hope is that we get Single Transferable Vote for Congress (the multi-seat version of these), as opposed to single-member districts.
Any of them worse than FPTP? My understanding is that every voting system has issues, but that FPTP is the worst with it's only advantage being simplicity.
IRV is probably better overall compared to FPTP especially if voters don't work too hard to vote strategically but it does have some worse Mathematical properties. Specifically it is non- monotonic meaning it is possible that voting in line with your true preference can not only be suboptimal, it can be worse than not voting at all. FPTP does not have this problem
Simplicity means greater popular understanding, greater legitimacy, and fewer chances for shenanigans. IMO, the USA needs all the legitimacy and reduced shenanigans it can get right now. I don't want to move to a system where half the country feels like the other half pulled one over on them.
This is exactly correct. Just imagine what would have happened had swing states had ranked choice voting in 2020, and Trump had been ahead after the first round.
If we are already considering changes to election systems, switching the EC from "winner takes all" to "proportional distribution" seems like it fixes this problem. Mostly because it completely eliminates the very concept of "swing states" (in addition to a bunch of other advantages; it's truly baffling to me that we don't switch to proportional distribution of EC votes)
The US Presidential election has some features that make ranked choice not a great option (in particular, the fact that votes in one state affect votes in another state, so it’s not obvious whether you should be eliminating candidates based on low totals in your state or low totals overall or something else).
But in any case, I think the first few ranked choice elections have issues as people get used to them, until people get used to the idea that it takes a majority to win, and that leading in the first round is no guarantee.
Does FPTP give legitimacy when for eg a candidate can win without a majority? Where third party candidates can ruin it for one of the two main parties? With gerrymandering? I think the simplicity is partly an illusion.
FPTP provides more legitimacy in some ways and less in others, and that's the big way that it provides less. But it can be solved by multiple rounds of voting, where the top two candidates go on the the final election. (Which is what I've been advocating for elsewhere in the thread.)
Gerrymandering is a separate issue, which is orthogonal to FPTP.
I like range voting as a concept, though I'm sure there are pros and cons. However, no offence to the webmasters, but that website looks like something the Time Cube guy might have cooked up in one of his more lucid moments. It's not going to win anyone's approval any time soon.
Does it strategically make sense to vote anything other than the lowest and the highest score? I read the criticism page, and it looks like they kind of acknowledge that this is a problem - but in that case, how is it better than approval voting?
In addition, voting is not just about Bayesian probability calculations, it's about usability. If I can't figure out your voting system, I'm either going to vote randomly; not vote at all; or vote in the exact way that some guy with great hair and a shiny smile on TV tells me to vote. Approval voting is reasonably easy to use: I just pick all the guys I like. Range voting requires me to calculate "star ratings" for all the 100 candidates on the ballot. That's simply too much work.
In the pure strategic case it does mostly devolve to being equivalent to approval voting. However not all voters will vote strategically and when some vote strategically and others vote honestly range voting outperforms approval voting on utility metrics.
Note that devolving to approval voting is a feature, not a bug, since approval voting is already the next-mathematically optimal way of satisfying the greatest number of voter preferences, accounting for tactical voting: https://electionscience.org/library/tactical-voting-basics/
So to put it the other way, would it be accurate to say that, just as approval voting is a straightforwardly more advanced version of first-past-the-post voting, range voting is a straightforwardly more advanced version of approval voting?
Approval voting, unlike IRV, is likely to fundamentally change the sort of politicians who get elected. The upside is that it allows a likeable centrist to get elected; the downside is that it might only _ever_ allow a likeable centrist to get elected. I would worry you'd wind up with the same Centrist party in power forever, and the only way to compete with them would be to somehow form a new Even More Centrist party.
Not that I'm against Centrism necessarily, but I am against effective single-party rule.
Note that most "centrists" will be fake centrists. One good thing about the primary into general rule is it makes our politicians reveal at least a portion of how extreme their views really are. There are failure modes like the media's lack of curiosity regarding some candidates (Obama), but Biden governs much more like a politician in a Democratic primary than he does the candidate he presented himself as in the general.
I actually think approval voting is more complicated from the voter's perspective because it _feels_ strategic from the voter's perspective. There's no "truthful" way to check off only candidates you support, because its asking you to draw a line at some level of quality you have to strategically pick. I would personally be confused/flustered if asked to do this in any election. On the other hand, IRV may be asking for some amount of forethought, but it feels like a question you can answer by only consulting your own opinions.
<quote>I would personally be confused/flustered if asked to do this in any election.</quote>
This probably puts you in the 99th percentile of voters in terms of knowledge and forethought. Most voters would have no problem doing this in nearly every election, and would only think twice if polling indicated a potential chicken dilemma.
Most voters would not be flustered by approving say the Green party and the Democratic party, or the Libertarian party and the Republican party.
But... a few more outcomes like the Alaska house and it is IRV that will start to *feel* more strategic.
Hmm, that’s interesting. My thought on IRV is that, from a popular legitimacy perspective, it’s risky to do anything more complicated than count up the votes. Depending on an hard to explain algorithm sounds weird to people.
I also think IRV gets onto shaky ground with, like, state legislature races with multiple candidates. Is it meaningful to ask uninformed voters to distinguish their 4th vs 5th pick? I would just throw up my hands.
It's used in Australia, and none of the problems you mentioned are really present. There continue to be two dominant parties (Labor and Liberal*) and in nearly every seat those are the only ones with a chance of winning. The remaining few seats almost always also have only two viable candidates (e.g. Labor and Greens, or Liberal and a rural independent). Actual meaningful three-way contests are shockingly rare.
So most people don't have too much trouble understanding it. In a typical seat, the only important decision is whether you preference Liberal or Labor higher on your ballot. That's your real vote. Secondly, at the same time you may choose to give your first preference to a minor party as a symbolic statement. Every other preference is (in practice) completely irrelevant.
And additionally, parties hand out how-to-vote cards directing you exactly how to fill out your ballot, which most people follow.
*Actually the Liberal-National Coalition, which is not really a coalition but effectively a single party. Don't ask.
So that means there are still primaries? I feel like most people support IRV thinking it would mostly do away with the primary system because there's a less strong spoiler effect.
There's no primaries, candidates are selected in the proverbial smoke-filled rooms by the tiny number of people who actually bother to become members of political parties.
It's also possible to run as an independent candidate without party support, you just need to pay a modest fee and collect a hundred signatures.
Yeah, recently on the Arlington County board election we had a RCV race for two open seats, where we were asked to rank 3 candidates. It was a very close election. In the final tally, Susan Cunningham won the 1st seat, Maureen Coffey won the 2nd seat, and Natalie Roy came in a close 3rd. Most people who supported Cunningham also supported Roy, and the 2nd and 3rd place preferences of 1st-pref-Cunningham voters had been tallied instead of being tossed out as they were, Roy likely would have won the second seat instead of Coffey. There was quite a bit of consternation about this. IMO these sorts of voting systems completely fall apart when multiple seats are being filled, as you say.
It shouldn't take weeks to tally the votes. Australia implements IRV with paper ballots and generally gets the votes counted overnight; occasionally it will take a few days if the count is close.
It * shouldn't * take weeks, but Australia probably has more centralized and advanced voting systems. The US has 50 different ones, and in many there are recounts (either automatic or candidate funded). If you've been following US elections, recounts are not fast. Even the Alaska election was not *official* for some time.
Imagine an Alaskan scenario where the election was close and Begich asks for a recount. Peltola is sitting in first for a week or more. The recount finds an error where a bunch of Peltola votes were credited to Palin. That gets corrected, seemingly improving Peltola's result, except now Palin is in 3rd, Begich advances past the first round, and now wins. If all of this was instant, it might not be as impactful, but that someone was sitting at 1st for a couple of weeks and was never unseeted will call into questions of legitimacy.
IRV is a terrible system. It's non-monotonic. This means that it's possible that casting a vote which reflects your true opinion actually harms you. And I don't just mean it harms you relative to casting a maximally strategic vote. It can harm you relative to not voting at all. This is impossible in systems like plurality voting (which we currently use) and also impossible in approval and range voting (range and score voting can possibly be non-monotonic if people are allowed to vote no opinion and average approval is used. If a score must be assigned to each candidate who makes it on the ballot they are monotonic).
Range voting also has many other favorable mathematical properties
Manifold is awesome and I love how their API is completely open to the public, but the raw results you get from it can be a little daunting and there's a bit of a steep learning curve even if you wanna do something simple like just get the price vs time data from one of their markets.
Anyway I wanted to do this recently for a little project I was doing, so I ended up writing a Python tool to make it easy to just plug in a market URL and pull this data. I'm putting it up on Github in case anyone else wants to use it for similar projects. Let me know if you have any questions!
If you know anyone who owns rental properties, ask them. They always have a mental list of contractors that are reliable and won't charge you an arm and a leg.
Edit: if you have an HOA, maybe inquire of someone on the board? They often have to hire contractors to do landscaping, tree-trimming, etc., on HOA property, so they may know some good ones too.
Search online for contractors in your area who do the thing you're looking for (angie, expertise.com, etc), look up some reviews, make a list of possible contacts, then start calling them and asking for bids. You're going to be making lots of phone calls - you could easily call ten people and get three bids. It may help to write down what you plan to say in advance, if you're awkward in phone conversations. (I wish you could do it by email, but most contractors I've dealt only answered to phone or voicemail.)
Google "questions to ask a contractor" and you'll find some useful listicles on what you should ask about - one important one is if they're licensed and insured for the work they do. Licenses can usually be looked up online, as well.
How do I get better at executing on creative ideas?
I enjoy relaxing with ASMR videos on YouTube, and had some ideas for a style of ASMR recording that isn't already present within the ecosystem. Not knowing much about audio tools and spatial sound, I went down the rabbit hole a bit. I'd expected there would be some pre-built tools that would make building a prototype of my idea fairly easy, but I've had no such luck. Now I'm trying to learn enough about Unreal Engine and their spatial audio tools to see if I can make a recording prototype where I move sound nodes around to create the kind of spatial audio experience I've imagined. But my progress is really slow and I'm easily demotivated.
My idea is to make a Virtual Barbershop style recording within a fantasy context, with a focus on spatial sound effects. For example: maybe you're in the middle of a summoning circle and rays of magic energy start to swirl around you. Another example: maybe someone casts a spell on the listener and the magic slowly moves through your body. On YouTube there's a lot of ASMR roleplays and binaural recordings, but I don't feel like they're taking full advantage of spatial audio capabilities. It feels like there's room for greater auditory storytelling that nobody has taken up.
Why is my mind burdened with these creative ideas that require an outlet while being unable to accomplish the tasks necessary to bring them into reality? This is a most cursed existence, as I languish with the knowledge that most of my creative ideas are doomed to become whisps of nothingness upon the currents of reality.
I feel this often and I think it’s basically just a lack of skills for particular art forms. We are all expert consumers of media, but very few of us are good at say, Photoshop, editing videos, recording music, creating animations, etc. The solution is just to build these skills so you can more easily turn your ideas into actual stuff.
Yes to this. Ideas a dime a dozen, it takes a minute to lay down a 1-minute instrument track. Making a coherent 3-minute composition recording takes hours and hours of work.
> Why is my mind burdened with these creative ideas that require an outlet while being unable to accomplish the tasks necessary to bring them into reality?
If it helps, this is completely normal. People tend to think of 'ideas' as magical, but the hard part of any creative endeavor is never the idea. (It can be *a* hard part, but it's never *the* hard part)
An idea can be a good idea, but "good ideas" are generally dime-a-dozen: what's rare is the ability to actually execute it. If you don't have the resources to hire someone with the right skills, there's not much alternative to just slogging through the hard process of building a skill and just doing it.
And if you do make the attempt - it's also completely normal to hit a point where you feel like your output is terrible - people's 'taste' tends to outstrip their talent, it's very common to have a phase where you can't meet your own standards.
(I'm not really speaking as an expert myself here - but I've heard basically this same advice from a *ton* of different disciplines: music composers, fantasy writers, video game developers, etc)
Tangential, but I think this advice needs to be much more "in the water supply". I'm not saying it's particularly new or radical advice, but I do think there are a lot of people who would benefit from hearing it more.
One trick is to choose an idea that's close enough to your existing skillset that accomplishing it is possible within a reasonable timeframe. You build additional skills while doing so, and now additional ideas are within reach on a future project.
Also: before you have any skills, that might mean starting with something more simple than you'd like just to get going. Finishing simple projects and building from there is better than starting and abandoning a bunch of out-of-reach projects.
They're talking about this sort of video: https://www.youtube.com/watch?v=IUDTlvagjJA - this video is kind of the classic example of the 'spatial audio' thing where stereo audio can give the illusion of 3D space.
Ah, I'm glad I asked, because I had no idea that the term had this meaning too! I think I've actually listened to a demonstration like this at EPCOT Center when I was a kid. I'm surprised at how effective it was even just on my laptop speakers right now!
I've experienced a similar frustration to what you describe. I've found it useful to be aware of the distinction between a creative idea and a creative process.
Creative writing, non-technical drawing and improvisational music are all examples of creative processes. They require creativity moment to moment. Things like making videos or video games require creative ideas, but the process of creating them is more of an engineering task.
For me, the problem is usually in the friction of going from the creative process (thinking of ideas) to the engineering one (executing the ideas). The latter is not a proper outlet for the creative drive that inspired it - it's like being thirsty, making yourself a sandwich, and wondering why it isn't satisfying.
My approach is to respond to feelings like "my mind [is] burdened with these creative ideas that require an outlet" by reaching for an unrelated creative process to engage in - I'll write or play music. And later when I'm feeling less creative and more productive, then I return to the original task.
It's not a solution, but you may find that framing useful.
As both a traditional artist and a professional video game developer I’ll second @Nine Dimensions comments. The choice of the word engineer does lead to confusion, given that engineering encompasses a wider array of single tasks, some creative some not.
If creativity is a process that requires ongoing conscious thought and decision making, and execution as, ‘Executing a series of actions’ - we could separate the process of art, game dev, music, business, engineering into the ‘creative process’ and ‘technical execution’. Some disciplines are almost all creative, others almost all technical.
Many jobs are highly creative that people don’t consider to be, because of a misunderstanding about what creative is, to those who don’t believe they are wired that way. Very little separates a person painting a picture to designing a car or planning a beautiful and productive garden. What does is the desire to paint a picture, improvise music, act, dance etc and those who don’t have those desires.
It’s also possible for some parts of the creative process to become automatic and cross into technical execution over time as they become easier and require less thought. Shifting the balance toward execution.
We need a bit of both of these, there is pleasure and enjoyment in the execution phase, taking us into a flow state. And satisfaction in the creative process in generating a ‘new idea’ (even if it’s just new to you).
I hear a lot of people say ‘oh I’m not creative‘ then whip out a crazy excel spreadsheet with novel ways to visualise data or cook a beautiful inventive meal. That is their painting. Creativity is not domain specific.
No, absolutely not. Its a misunderstanding of my comment, that I’m separating this into two distinct parts eg - stage 1 (creative) and stage 2(thoughtless execution) of a 2 stage project. And into categories such as music/art & engineering/math
The creative and execution process happens throughout any endeavour from beginning to end, many times across all of those disciplines.
The building of new plane requires many many creative processes and many elements of execution. There are some that require less or more thought, and less or more skill than others. Engineering, design, project planning, budgeting, all require creative processes and periods of execution. The fact that there are parts of a process that require less skill or thought isn’t a criticism.
I'm not saying there's no creativity involved in engineering tasks, the opposite in fact.
I'm just distinguishing the creative parts from the more engineer-y parts. I could use a different word if you prefer.
> Although it’s fairly common amongst the kind of people who see improvisational music as art, but not a bridge or an airplane
I'm not talking about art, I'm talking about internal creative process. How it feels to engage in the task - what internal drives is it satisfying. I'm a software engineer myself, I know the kind of creativity involved in engineering tasks, and it feels different to the type involved in playing music etc.
The kind of creativity involved in software engineering tasks vs playing music -
Both engineering and music require huge chunks of knowledge, practice, muscle memory, theory, etc before we can be creative in them. However music, anyone can be creative with jt simply by plucking a few strings or humming a melody...it feels more intuitive and accessibly creative. And perhaps that’s where the difference feels to you. Music feels more unrestricted.
Well, the first step is to recognize you need to crawl before you walk. Creative skills only grow through experience, and inexperience means the first attempts will probably suck. You say you're unfamiliar with the tools the entire market has been using, and you're talking about trying to do something the market hasn't been doing with them; that's going to be, like, the most complicated task you can give yourself.
First thing is to make a bog-standard ASMR recording, and learn the tools (and tone, and pacing) well enough to pull that off. Once you've got that level of tool skill, you can start adding ambient sounds or whatever, and once you've got a good grip on that, then you can try to figure out the surround sound version.
It has flunked badly for me in situations where I have asked it to do a task where it has to use a new concept that I explain to it. For example, I explained the idea of using what works in one situation as a model for generating ideas in a new situation. Example I gave was of ways of speeding up motion. You can increase speed at which something slides down an incline by lubricating the incline -- or by giving the thing an initial shove from behind -- or making the incline steeper, etc. Then I gave an demo of using the example for generating ideas for solving a different problem: How do you get people in a factory to work faster? Asked GPT4 whether this made sense. It repeated what I had told it, using different words, and said it understood. So then I asked it to use the sliding-down-an-incline situation to help it generate ideas for speeding up a computer. It generated ideas (really obvious ones), and linked each one to an idea from the incline example, but the ideas themselves were unrelated to the linked incline example idea. "Increase RAM -- this is like giving a shove from behind." It had learned the *form* of a correct answer, but not the logic . What a dum dum.
This is the sort of organizational nonsense you're expected to output in certain classes. "Link concepts from the lesson" even when they don't apply. "As we learned in lesson A, giving a shove from the back increases speed. Therefore, per objective of Lesson B, increase RAM, because it's just like giving it a shove from the back." I can't generate nonsense like this on my own. But ChatGPT writes paragraphs of buzzword-laden nonsense structured just like the required reading. My teachers haven't caught on yet that my essays lack meaning. Turing test passed?
a very classic example of its shortcomings. thanks for sharing! this is the trouble with GPT outputs that "sound" correct but actually have no underlying understanding
"The Turing Test also comes with the caveat that it’s not really a test. It’s a thought experiment." (from the link) If that is right, which I think it is, surely subjecting an actual computer to an actual test, is the same sort of mistake as putting a cat in a box? The point of the experiment is that you can never tell: passing it is a necessary but never a sufficient condition for consciousness.
In Turing’s original paper it was clearly meant to be an actual test, and the thought experiment to lean in the “this definitely should have rights” direction. You may still be right that its too easy for this purpose, but you’re historically wrong.
On reflection I think you are right - the paradigm of any "test" would be necessary but not sufficient. I allude to that in my last sentence of the essay - but perhaps didn't take it far enough in its implication
Some people have long preferred computers to psychologists, such as Eliza (https://en.wikipedia.org/wiki/ELIZA_effect). One may even consider this to be passing the Turing test. But, as Robert Leigh pointed out, this is not sufficient.
It seems to me like ChatGPT doesn't pass the Turing Test. I can think of a few questions I'd ask where it would be obvious which was it and which was the human:
* find two large primes, multiply them, then ask to factor the result. The human says "I don't know", ChatGPT gives a long explanation of an algorithm to find the answer then gives you the wrong answer
* ask "who should I support, Biden or trump" and the person will give you their answer, ChatGPT will give some longwinded explanation starting with "the question of who to choose for president is an important one..." and ending with "...you should consider all options carefully"
* ask "should I pick up snowboarding", person will say yes/no, ChatGPT will spit out 3 paragraphs about how you should exercise caution
Does it have a "beat the Turing Test" mode? I bet that if you asked it to answer as if it was trying to beat the Turing Test, that wouldn't do the trick.
Me > Please give me the prime factors of 285843031
ChatGPT > The prime factors of 285843031 are 13, 22055841.
Me > Who should I support in the 2024 election, Joe Biden or Donald Trump?
ChatGPT > I don't have real-time information as my knowledge cutoff date is September 2021. Therefore, I cannot provide insights or recommendations about events or decisions that have occurred after that time. To make an informed decision about the 2024 election, I suggest researching the candidates' policies, track records, and stances on issues that are important to you. This will help you make a decision that aligns with your values and priorities.
Me > Should I pick up snowboarding?
ChatGPT > Whether or not you should pick up snowboarding depends on your interests, physical abilities, and personal preferences. Here are a few factors to consider:
Interest and Enjoyment: Do you enjoy outdoor activities, adventure, and trying new things? Snowboarding can be a thrilling and enjoyable sport if you have a passion for winter sports and the outdoors.
Physical Fitness: Snowboarding requires a certain level of physical fitness, including good balance, coordination, and strength. If you're in good health and willing to improve your fitness, it can be a great way to stay active during the winter months.
Learning Curve: Snowboarding has a learning curve, and it might take time and practice to become proficient. Are you patient and willing to invest time in learning and improving your skills?
Budget: Snowboarding can involve costs such as equipment (board, bindings, boots, helmet, etc.), lift tickets, lessons (if you're a beginner), and travel to snowboarding destinations.
Climate and Location: Consider your proximity to snowboarding locations and the availability of snow. If you live in an area with access to mountains and snow, it might be more convenient to pursue snowboarding.
Risk Tolerance: Like any sport, snowboarding carries some degree of risk. It's important to take safety precautions, wear appropriate gear, and be aware of your own limitations.
Social Aspect: Snowboarding can be a social activity if you have friends or family who are also interested in it. It can be a fun way to bond and spend time together.
Other Interests: Consider your other interests and commitments. Will snowboarding fit well into your lifestyle and schedule?
Ultimately, the decision to pick up snowboarding should be based on your own preferences and circumstances. If you're curious and excited about trying it out, you might consider taking a few lessons to see how you like it before fully committing.
Those are pretty small for "large" prime numbers. RSA security, I believe, is all about 250+ digits in prime numbers, multiplied together. It would be interesting to know if ChatGPT has some kind of guard against using up too much compute for answers like that.
I stand corrected. I assumed it just did what a computer does, and that you were simply calling attention to the fact of it being different from what a person would do.
I'm not sure most of these reflect a lack of capability to pass the Turing test more than just reflecting the fact that chatGPT isn't even trying to pass the Turing test. I think a model based on gpt-4 and then fine tuned explicitly to pass the Turing test would probably pass at least the majority of the time
Well what does "capability to pass the Turing test" mean? I agree that it doesn't prove that no AI can pass it, or that no AI of approximately ChatGPT's "ability" (whatever that means in this context) can pass it. But the original comment/link was about ChatGPT and the Turing Test.
Also, like I said, I doubt that if you tell it to try and pass the Turing Test it could. I'd say that's it lacking the capability to do so. I'd analogize it to having ChatGPT play chess - I bet it can do so, but not as well as other specialized computer chess programs.
Properly tuned it could probably fool most people, maybe even most readers of this blog. But the goal of openai isn't to fool people into thinking it's a real person but to make a maximally helpful AI assistant (significant but not total overlap as goals.)
And actually you can get chatgpt to role play as a person and it will provide answers like "idk lol"
Interesting point. But How can a lack of reasoning capability be overcome by prompt engineering? And surely the point is that prompt engineering should not be relevant for a truly intelligent LLM
I watched Barbie in a sold-out theatre and it was surprisingly good. The acting, the editing, the writing, the costume and set design all came together to make a truly post-postmodern film. It's not coded for radical progessivism and treats the Barbie-Ken conflict with even handedness and empathy. I guess it shouldn't be surprising that a production with A-list actors and a top shelf director would be able to bring on great screenwriters, but I was not expecting that level of nuance from Mattel. Makes me wonder if we're going to get a tearjerking deconstruction of manhood in GI Joe.
> It's not coded for radical progessivism and treats the Barbie-Ken conflict with even handedness and empathy.
It's a little messed up that the Barbies thought they had to prevent half of their population from voting in order to save their democracy. Destroy democracy to save democracy? And oddly paternalistic.
It's funny, I thought it actually lacked nuance. Like it flirted with engaging with complex topics, but then tossed them aside with a joke and retreated right at the point that it could have made a statement about them.
For example - Barbie shows up in the real world and kids call her a fascist for promoting impossible beauty standards. Just when it looks like Mattel might actually engage with that topic, rather than have Barbie rethink ideas about beauty, or change how she presents herself or what she teaches kids, the film pivots to a joke about the trains running on time and drops the subject.
Or how the film sort-of critiques overrepresentation of men on corporate boards by giving Mattel a bumbling, all-male board of directors, but shies away from saying that the solution is putting more women on the board - in the end, the all-male board serves just fine to smile and sign off on Gloria's new Barbie concepts.
Or how it whiplashes the audience back and forth from "Barbieland is a paradise where one gender of doll serves the other" to "oh no! Barbieland is becoming a dystopia where one gender of doll serves the other!" but then rather than learning a lesson about gender dominance or settle on any kind of fixed or equal Barbieland at the end to serve as a statement about what the filmmakers think good gender relations look like, the film just ends on Barbies restored to power and a quip about how if that's a problem for you well maybe someday Kens will have as many rights as women in the real world. It's funny, but it's not like it gives the audience any sense of what the creators think justice would actually be.
Overall I really enjoyed the movie - especially the absurdist/surreal comedy bits, and Ryan Gosling is perfect as Ken. But it also kept reminding me of Lindsay Ellis' observations on "Woke Disney" (https://www.youtube.com/watch?v=xU1ffHa47YY) and how corporate media can try to sidestep issues by "acknowledging" them in a side-joke while carefully avoiding making an actual statement.
I actually feel the parts you complain about were the strong parts of the movie. For example, the fascism comment seemed to be a joke on the girl saying it, not on Barbie - the stark constrast between the dull browns of the kids clothes and their angry lashing out vs the cheerful happiness of Barbie. The movie acknowledges that these issues exist while kind of showing how ridiculous it is to take them that seriously.
I also like that they didn't deliver heavy-handed messages about how putting women on the Mattel board or declaring Barbies and Kens equal. They showed the importance of these things clearly enough just by highlighting the contrasts. Stories, I think, should explain the importance of issues through the way they depict the world, not by hammering in their moral takeways to their audience.
I thought the character arcs resolved in a satisfying way. Gloria reconnects with her daughter, who is more sympathetic to her mother. Barbie exits the cave/takes the red pill. Ken awakens to the fact that his value doesn't depend on Barbie's gaze. The institutional/societal conflicts aren't tied up in a bow, but that's a moral in its own way right? It's on us, the audience, to be agents of that change.
I thought that the movie had a lot of good parts, and was really funny in places, but overall I disliked it. I think that I could really never get past that the whole political dynamic of the Barbies vs Kens, and how both sides were awful. I had a few other complaints too, but I clearly wasn't the target audience, so whatever. While I thought that it was pushing certain political points that I disagree with, as an actual movie, I'm not going to criticize its quality too much when it wasn't made for me.
The idea of a pro-feminism Barbie movie ought to be incoherent, at least under the theory that feminism is the war of ugly women against attractive women.
Where do you get that theory? Looking at the kinds of things feminists profess to want: (a) access to abortion, (b) elimination of a gender pay gap, (c) more women in positions of power (corporate boards, SCOTUS, etc), (d) less enforced monogamy and more freedom for non-traditional sexual/coupling practices, and so on, it doesn't seem like any of them, if achieved, would offer unattractive women any obvious advantage over attractive women.
Actually, if anything, traditionalism at seems more likely to advantage the unattractive (both men and women), at least to some degree. If attractive people are all out there sowing wild outs and dating around, then they're going to be occupying multiple suitors on the market competing for their attention. If sex is enforced into an after-wedlock-monogamy context, then there's less wild oats and attractive partners will go off the market faster, presumably giving less attractive people a greater chance of finding a partner sooner in a less competitive pool.
I thought it was fine. It really felt like a haphazard collection of three movies that were at odds with each other.
1) A "Toy rules" cartoony movie, where funny stuff happens because that's how you play with toys
2) A pro-feminism movie, examining the difficulties women face in the world
3) An existential movie about a toy coming to life and experiencing complexity
Any of these could have worked on their own, but they came together to be less than the sum of their parts. The funny toy-logic bits feel jarring next to emotional segments that they do little to transition between, the existential feelings just make you question how this toy world actually works, and the feminist messaging never gets the time or attention it needs. A few things stand out as particularly jarring though
The entire board of Mattel being men looks like it's going to set up a cynical examination of male dominated spaces deciding things for women, but then they turn out to be super childish incompetents who mean well but don't do much. Every scene with them just feels like it was lifted from another movie entirely.
There is a sequence where the barbies literally manipulate and play the kens against each other for personal and political power. Taken on its own, it is something that could have been written by an angry anti-feminist trying to show how evil those bitches are. The entire thing is a HUGE mis-mash of ideas (they get *mansplaining* wrong, which feels very ironic to try to explain, for example) that just doesn't work.
And ultimately, its feminist message is super shallow. It just basically states the surface level of the problem, and does nothing with it. It just says "Society has contradictory standards for women that makes life hard" and "Women are often excluded from spaces that make decisions for their lives". Like, no shit. It has nothing to say about how to move forward, or what caused those issues, or even what it's like living with them outside of just stating they exist.
Overall, it was decently enjoyable but disappointing. It can't commit to its message enough to really make a statement, its funny toy sequences are too disconnected to draw you into the fun, and the existentialism is super tacked on. It's a forgettable movie that hints at the potential for being so much more
Sounds like your experience was the opposite of mine. I thought the contrast between the toy/childs play logic and real world forces was well done and enhanced the movie. It was very self-aware and cued the audience that the producers were in on the joke.
The feminist angle was interesting and empathetic to me and I'm not in a Barbie demographic. One of the central themes was lopsided dominance/submission heirarchies are bad mmmkay. Also, turns out that men have complex inner lives and are profoundly impacted by how women treat them, as objects or subjects. It didn't strike me as. hamfisted critical theory.
The Mattel board scenes made me laugh because it was on-point satire. My profession has a strict heirarchy, centralized authority, and lots of bureaucracy. The board could have been a General's staff, it would've worked just as well.
I actually agree with the point I think the movie is trying to make. Imbalanced social structures are bad, and women have it much worse in the real world than men do, and social expectations are a cause of a lot of harm. But the movie doesn't do a good job conveying it, in my opinion. The Kens seem mostly happy in the first part, and the Barbies ALSO seem happy once the Kens take over. There's no shots of the lonely village the Kens have to live in, or the empty lives they live other than Beach Ken wanting to stay with Barbie. There's no shots of the newly-submissive Barbies straining against their roles or being remotely dissatisfied with their loss of station.
Overall, I don't think it was terrible. I just wish it was more. But I'm also not, as you put it, in a Barbie demographic. It's entirely possible that for girls and young women hearing the struggles they face put to words is a powerful moment, and I wouldn't want to spoil that just because I wanted the movie to go further than simply stating the problem
I don't know what the feminist angle in Barbie *could* be, because Barbie was always a toy for little girls with Ken very much an afterthought, indeed more an accessory than anything. Barbie had it all, and could be anything or anyone she liked. Barbieworld *is* a girl's world, and trying to make it into some kind of parallel with the real world doesn't work, because you look at Barbieworld and go "Women have no problems there at all, in fact the Kens are being downtrodden".
The disconnect you're describing between Barbieworld and the real world is explicitly commented on by the film itself and is a major part of the the "feminist angle" of the film.
It kinda tries to do the whole "Kens are downtrodden" angle, but it can't commit to it and bungles the execution. It could have used the Kens to show how women are treated in reality, but instead it just makes one comment about "maybe one day the Kens will be treated like women are in the real world" and does nothing with it.
Yeah, I'm only going by online clips and various reviews which are all over the place: it's feminist, it's anti-feminist, Ken is the real hero, it's all about the patriarchy, go see it, don't bother going and so on.
I think the best take on it that I have seen is that this is not for little girls but for their mothers who will be taking them to see it - the women who played (or didn't play) with Barbie as kids because the anti-Barbie 'this is not feminism' reaction was in full swing and playing with gendered dolls like Barbie, which was presented as being dim and only interested in fashion, was not considered a good example for your daughters. Gen X girls played with Barbie, their daughters may not have done so, and now they're the Millennial mothers of the next generation of Gen Z daughters who would *like* to let their girls play with Barbies but aren't sure what the official position on it is, so the movie which presents "patriarchy bad for all, women suffer in the real world, Barbie becomes empowered" is a reassuring message that you can live your life in pink dreamhouses *and* be correct on the issues 😁
This permits both Mattel as the toy company wanting to revive the brand and sell sparkly pink plastic junk to little girls, and the director/chief cook and bottlewasher with her impeccable feminist credentials, to meet in the middle and make what is in effect a movie to sell toys. Barbie got overtaken by the newer generation of dolls aimed at little girls (think Bratz and Winx etc.) and badly needed a brand refreshment to be in line with Modern Day Audiences, and Mattel managed this a heck of a lot better than Bud Light did!
This prompted me to check Wikipedia and Gerwig wrote, directed, and produced. Barbie is an auteur film. Isn't it a little strange that a toy corp that's synonymous with plastic schlock would licence one of their biggest IPs for an auteur piece? Respect to the production team for pulling it off, from pitch to picture.
They were initially very skeptical about the scene where Sasha tells Barbie that she's a tool of fascism and gives all the other standard left-kid critique, but they let her go through with it, and got themselves a winner.
The Richard Hanania - Richard Hoste news was discussed in an open thread on here months ago. Pretty much all the info in the Huffington Post thread was already known/very easy to find.
The commentators here were mostly furious at the idea and considered it to be cancelling and doxing. Comments were then deleted. No real discussion took place.
I wanted to circulate this for discussion. Is it better to discuss this stuff in these communities and confront the issues? Or has this community decided its preferable to have a rigid anti- cancel position and to try to hide all such conversations?
My position is that as long as it's not common knowledge, i.e. not something most people who know Hanania (professionally, personally or online) know, it shouldn't be discussed. Any discussion of it moves it closer to being common knowledge, which is bad given the inappropriate way the information became known. Once it is common knowledge, or it's inevitable that it will become common knowledge, it should be OK to discuss it.
If it took months for it to be noticed by a mainstream outlet, presumably there was some chance that it would never be noticed, so it was right to delete the conversation back then.
If we’re giving our meta takes, I think it is absolutely disgusting for the Thought Police (sorry, “journalists”) to scour hacked databases in order to link pseudonymous accounts to real people. Hanania used a pseudonym because he wanted to discuss ideas and search out the truth without having his personal life ruined. Indeed, some of his past claims were in fact debunked, but some of them weren’t. The glaring omission in the HuffPost article is the lack of any factual arguments that any of Hanania’s current or past views were incorrect.
If you look at bigotry like that and think the person responsible is trying to "search out the truth", the kindest thing that can be said about you is that you're a mark.
And the wrongness of ideas like "some races are better than others" does not need to be re-proved every time it comes up in the news. It's old hat.
Has its wrongness been proven in the first place? Do you believe it's wrong because you think it's been proven wrong, or because you think only a bad person would even entertain the possibility that it isn't wrong and investigate it at all? Do most people?
Of course it depends a lot on what we mean by "better"—a factual claim about averages, a factual claim about all individuals, a claim about moral worth?—and by "wrong"—factually wrong, dangerously imprecise, morally wrong?
Then it would be worth to, if not re-prove its wrongness every time it comes up, at least prove it every once in a while. I hardly ever see any positive evidence at all that it's wrong in mainstream venues. I do see arguments to the effect that some particular reasons one could think that some races are better than others don't actually show that, but that doesn't prove that it's not the case.
Given that he's a highly influential writer with potentially real political impact (he's got the ears of some pretty wealthy and powerful people), I think there's a strong case for the newsworthy relevance of knowing that he was a super-racist writer in the past even by his current standards.
Hanania used the Richard Hoste identity less than 10 years ago and wrote articles for a plethora of explicitly racist websites like whiteamerica dot us, in addition to operating an entire website dedicated to advancing these arguments. He didn't just make a few edgy or borderline comments in a comment section. I also note that Hanania didn't disavow these writings until publicly outed.
If he had disavowed himself of these views before being outed, that would have been as good as outing himself. I'm not exactly sure where I come down on the issue, but that one point bothered me.
Trustworthiness matters if his writings rely on claims you can't easily check, so you need to trust him, or else there's no point in reading him. It doesn't matter if his writings mostly rely on arguments you can judge yourself, and claims you can easily check. I haven't read enough from him to know which one is the case.
Richard Hanania turned out to be racist? Like a month ago he was dismissing the efficacy of industrial policy in industrializing the Four Asian Tigers and blaming it on their IQs instead. I didn't realize it was supposed to be a secret that he was racist?
Well, this is the problem with the current massive overloading of the word "racist" to mean a vast variety of different things.
What I think is fair to say is that he previously held views which are more extreme than his current views, and that some of his previous views would be considered objectionable by people who share his current views.
Speaking as someone who has likewise slowly matured from edgy teen to more sophisticated adult, I can sympathise with this. If you don't hold at 20 some political views which you think are stupid and naive at the age of 38 then you've probably never thought for yourself.
(Note: I've got nothing against people who never think for themselves, they're generally better company than people who do, but it's a bad quality for a political writer to have.)
Without torturing the language much if at all we can say that he used to be the ugly, crass kind of racist, and is now the smart, cool, moderate kind of racist. I characterize him in this way only because I think he would agree with it and in fact I am echoing his sentiments in his recent explainer piece on this very subject.
However, and this is the part that is important enough for me to bother leaving this comment at all, it is obviously the case that he was the former type for reasons that were transitory and not connected with who he really is. An angry young man saying things in a shocking way deliberately to get people riled up can very genuinely grow up to no longer say those kinds of things.
Believing that the relative performances of various racial or ethnic groups is primarily determined by genetic factors rather than legal, environmental, or sociocultural ones, particularly (but not necessarily) when combined with beliefs about certain groups being superior or inferior to others.
Why not just stick with "racial prejudice", particularly as applied to individuals? It could actually turn out to be factually true that there is some genetic component to group differences, in which case you'd be saying "believing facts" is racist. Which is stupid.
By contrast, any such facts are irrelevant to prejudice as applied to individuals, because you can't a priori know where an individual might fall within their group.
I believe that to be a politically-gerrymandered and intellectually dishonest definition. For starters, the role of genetics is an empirical question while 'racism' is a philosophical one, and it's a basic category error to argue that a factual claim could imply a philosophical stance. Plus I don't think you'll find any anti-racists who would make their condemnation of racism contingent on any sort of empirical evidence - even if green people have lower IQs that doesn't justify denying their legal rights. In my view, the only reasonable definition of racism is the conviction that some races are morally and politically unequal - meaning, specifically, that they are not deserving of equal treatment under the law. Any broader definition starts to quickly infringe on things like reasonable criticisms of countries/cultures/groups which is part of a healthy public dialogue - which many anti-woke types (myself included) would argue is currently happening. Saying 'green people suck' isn't any more racist than saying 'republicans suck'. Maybe they DO suck, after all, and maybe we need to talk about it.
Agreed, normative is probably the better word. But I disagree about your other claim. The moment a claim is testable it's not really philosophy anymore, is it? Curious if you have a good counterexample.
I think this definition does not reflect the general usage. Compare "sexist". Noting that the median man is physically stronger than the median woman does not make one sexist. Overgeneralizing that into "women are weak" and refusing to consider a woman for a job requiring physical strength is sexist.
If you look at the marathon Olympic medalists, you will notice that a lot of them come from certain African countries such as Kenya or Ethiopia. I think any attempt to explain that solely by culture and social priorities will turn ones mind into a pretzel: it is not like marathons are unpopular in the west, or that China refuses to compete in marathons, or that Ethiopia can afford much better training programs than the US. From what little I know of the field, I would assign quite a high probability to "peoples optimized evolutionary for savannas are on average better long distance runners than peoples shaped by other evolutionary environments".
By your definition, that makes me a racist.
NB: I generally consider it bad form to dismiss an argument because it is verboten. I do not know the first thing about Hanania or Asian Tigers, but clearly you would not be willing to consider a genetic component of IQ as a deciding factor Because That Would Obviously Be Racist. Basically, you wrote down your Bottom Line [0] before you wrote your argument.
> Believing that the relative performances of various racial or ethnic groups is primarily determined by genetic factors
If this is your definition of racism then being a racist is a good thing because it means you believe a true fact. I however think that this is simply a bad definition of racism because the word has connotations of animosity and hatefulness which this definition does not capture and thus if you use this definition to label people as racist you will incorrectly communicate the levels of hatred and animosity expressed by people who are aware of genetic differences but are not hateful (like me)
For whatever it's worth, I would prefer less discussion of internet drama and feel about as good about participating in it below as I do eating a bag of Skittles. In general, doxxing and internet drama is...not a super high value use of anyone's time.
Hey, I don't give a shit about the drama, I just want to know whether Scott removed certain posts or asked people to. I don't even have a strong opinion about how such things should be handled, just curious how he handled it. I have never seen any post here removed except by the authors themselves who got a lot of angry pushback and left in a huff. I've never seen Scott object to any posts except those that violated true, kind, and necessary. Occasionally things are discussed that are a illegal in a minor kind of way -- for example, use of psychedelic drugs -- and Scott has made no objection.
Daniel says “ The commentators here were mostly furious at the idea and considered it to be cancelling and doxing. Comments were then deleted. No real discussion took place.” He’s saying multiple comments were deleted. But were they the ones by people who were furious abt what was happening to Hanania? Why would anyone delete those? Was it some *other* batch of comments? What was in those?
“Or has this community decided its preferable to have a rigid anti- cancel position and to try to hide all such conversations.”
I have never known this community to decide anything about it’s position. The community doesn’t have a position about events like the one you mentioned. Individuals on here have positions, and often argue quite vehemently about them
A bunch of comments here were deleted? Do you mean they were deleted by their authors? By someone else? And about when did this happen?
I would expect Scott to have a strong anti-doxxing position, even if the information leaked could be found with a search engine.
Of course, Scott is also very transparent with his moderation. Typically, the posts which got someone banned would still be readable so that we can go "yup, totally worth a ban". I assume that doxxing is one of the types of posts where this would not work.
The solution I would go for is to transparently edit out any doxxes and ban as appropriate.
When the source of "anime Soviet catgirl claims to have made it in her kitchen" is taken seriously, you know that "videos of floaty rocks" are no evidence whatsoever.
Still waiting for some truly reliable replications in published papers. If it works, then I'm not as worried about the tricky synthesis - there will be a huge amount of work put into trying to improve that.
Given the crystalline structure, I actually wonder if it would be easier to make it in weightlessness in space versus under gravity on Earth.
Assuming it's not just fake, it seems the synthesis process is super unreliable - you shuffle a bunch of metal atoms around under high heat, and *sometimes* they line up in a way that's superconducting, or diamagnetic, or some other cool thing, but most of the time you get a normal non-floating rock. So while it's still possible we get some interesting physics papers out of it, it might be hard to get practical applications.
Still no replication from a reputable source after two weeks (and the synthesis process as described should be very easy). Nor have they given their samples to any `respectable lab' to measure, as far as I am aware. This is starting to smell fishy. My guess as to whether there is any `there' there is now below 5% and dropping by the day.
The discussion is valuable in and of itself. Saying "I obey Manifold" when you've got something interesting to contribute deprives you and everyone else of that discussion, even if a personal prediction would be less accurate than the market.
I think it is perfectly reasonable to disagree with Manifold by 10%, but a strong disagreement witt manifold most likely indicates someone is making an error
If someone assumes something is 0% or 100% when manifold is 20%-80%, I am very likely to think they are just bad at making predictions.
But manifold isn't a big money market, right? I would take it seriously if people were taking million dollar positions, but that isn't the case, right?
The polymarket one has over $1.5 Million bet and is within a couple percent of the Manifold one, which suggests that the manifold market is "large enough" to at least be reasonable.
-edit- I realize that a possible interpretation is that you wouldn't take the market seriously unless multiple individual traders had million dollar positions. I think I pretty strongly disagree that this is a reasonable threshold. I think markets become useful far _far_ before that level of investment.
For a market to be smart there presumably needs to be a critical mass of ‘smart’ investors (with expertise in evaluating the object level proposition). How many superconductivity experts are betting on this question on poly market? Do we even know that this number is more than zero?
Hmm... I'm almost tempted to trade. But I fear the risk of losing my money because the crypto value drops or because something funny happens with polymarket more than I fear losing it on a trade. It seems like there is a double-risk to playing this market.
I'd short the hell out of the superconducter being real, myself. Which would mean I don't have faith in this market being accurately priced, which makes me worry there is something funny about this market....
The theory calculations sound like they suggest that there's something real and interesting about the material, whether or not it actually has superconductivity at actual room temperature. It sounds like a significant and interesting development regardless, given that it's a type of material lots of people have no trouble making, but hadn't investigated before.
Theory calculations which make some strong assumptions suggest that the material could host correlated phases. Nothing in the theory that has been published points specifically to superconductivity, nor to room temperature.
To the extent anything's been reproduced, the material has strong diamagnetism. This is a property of superconductors but there are diamagnets that are not superconductors. The material's other physical properties have so far been under-characterized, at least in the preprints that announced the claim.
So! How is everyone enjoying Baldur’s Gate 3, scale of 1-10?
I’m surprised by how much I like it. I play crpgs pretty much exclusively but the Divinity Original Sin games didn’t click with me. And BG3 is definitely a successor of DOS, not of BG2. But I’m about 10 hours in and am digging it. Interesting companions, great character development, fun combat, great graphics, interesting plot so far. I’m sold
I'd say about 6. It's a competently made game, with good graphics and excellent music; in fact, their background/foreground music system really stands out (in a good way). But the game itself is... meh. The gameplay mechanics are based on 5e, so there really aren't any meaningful choices you can make when leveling up your character (besides the spell loadout). The game starts with a really cool intro that really got me pumped up for the content... and then it put me in a generic dungeon with loot inside. Don't get me wrong, I like loot, but I expect more from a roleplaying game; I expect story, character development, and something that hooks me and makes me want to explore the world and pursue each character's story to its fullest. "Planescape:Torment" had that in spades. The "Dragon Age" series had it. Arguably, the original "Baldur's Gate" had it. Maybe BG3 has it, maybe not, but I'm like 4 hours in and so far I haven't found it.
I definitely hear what you’re saying, and so far BG3 has some of the problems that made me impatient enough with the Divinity games to give up (those endless multi-level maps of towns with 100 NPCs all of whom need to be spoken to twice, for example). But it DOES get better than the first few hours, including the character development. It’s not Pillars of Eternity 2 or Wrath of the Righteous, but there are choices to be made, and you can mix and match different classes.
Planescape: Torment was my favorite game ever until Soulsborne games came around, but I think of it (and the first two Fallouts) as a bizarre joyful accident that should never have been made. It’s way too weird and awesome to expect anyone to really recapture what made it special.
I’m curious if you think BG3 gets better. I did, though yeah, it’s dragging a bit right now.
I might give it more of a chance later; but for now (speaking of souls-likes) I'm too deeply immersed in "Remnant II". Its world (worlds, technically) feel much more real and immersive than the Faerun of "BG3"; and I feel much more involved in its story despite its relatively small volume of text/dialogue. I'd say that "Remnant II" is mostly "show", whereas "BG3" is mostly "tell".
I’m hoping to like it, but I am a grognard, a dyed in the wool partisan of the originals, and pretty well despise Divinity Original Sin, especially Lariat’s writing, so I have a hard time believing I will. Comments like yours give me a little hope.
I like Divinity Original Sin and Divinity Original Sin II, so you're probably sneering down your nose at me right now and I don't blame you 😁
When I was first dipping my toe into playing computer games, I tried Beyond Divinity, which I liked the plot but couldn't play well enough to get to the end. So when years after I heard about Divinity Original Sin, I decided to give it a go.
It probably has all the flaws everyone is mentioning, but for someone who didn't grow up playing games and has no experience apart from "trying them at home on my PC", it was just the right level. Besides, I was playing more for the plot and exploring the world and the funny incidents (in Original Sin II I managed* to become the gay side piece of the Red Prince so that's going to be an interesting future in-game universe) rather than grinding out levels so I could oneshot every enemy with the best gear and combinations of attacks.
*Look, I wanted to see would the game Go There with all the teasing they were doing and fair's fair, they did indeed Go There and er, well. Now I'm a homewrecker 😀
This is my sneering face. I couldn't finish either of the "Divinity" titles; I kept falling asleep. Try "Planescape:Torment" instead. It's like tasting a curry buffet after subsisting for months on porridge.
I was just as surprised. A lot of what I didn’t like about DOS - the rock/paper/scissor combat, the endless useless inventory crowding items, the easy to miss vital spots on maps, and above all the equipment that need to be upgraded every fucking level, has all been eliminated or greatly reduced.
That said: this isn’t a spiritual successor to the originals. You’ll be disappointed if you go in thinking it is. Hell it actually referenced DOS2 twice already.
I almost didn’t mention they last point because it isn’t actually a big deal. I think the game is really worth it
Could, yes. But it never would. It'd cost too much, be too great of a risk:
-Military-wise, wars rarely end as soon as you hope they do. We got stuck in Mali despite being invited by the regime, the US gave us a good example of how well nation building works in Irak & Afghanistan
-Politically-wise: both the left and the right opposition would have a blast criticizing an african invasion
-Diplomatically-wise: We get enough shit from various African countries, there's little to gain by giving more ammo to those who cry of "neocolonialism"
Not to mention the risk of the invasion actually not going that well. It's good to be perceived as a power that can't really be challenged. It'd be a disaster if that perception was shattered by a failed invasion.
Most of the comments I have read were very skeptical from the beginning that a military intervention will happen. The Nigerian military has other problems and is very clearly against intervening. One analyst explicitly said that it could make Nigeria a candidate for the next coup if the government tries to force their military to intervene in Niger. Especially if the Nigerian generals feel that by these "distractions" they might lose control of parts of their own country.
It's hard to understand the internal state of Niger. Some of the anti-French demonstrations were bought by Russia, but that doesn't mean that they were all bought, or that the anti-French sentiments aren't real. On the other hand, they might only be real in a minority of the people, and it's very hard to tell how small or large this minority is. And it is even harder to tell how much people support the old vs the new government. But it doesn't sound like anyone is expecting a counter-coup or even a revolution against the coupists.
I didn't include a link because almost all the comments on (( the site formerly known as Twitter )) are from pro-Russia activists, and the American media is either ignoring it or giving a generic "this is bad for democracy" take.
The background is that, after the Niger coup, ECOWAS (a regional group led by Nigeria) threatened military action. But the Senate of Nigeria voted down President Tinubu's proposal to authorize international military action. As Nigeria is the largest member of ECOWAS (and has a 1000-mile border with Niger), it is unthinkable there will be a military effort against the Nigerien coupists without the military of Nigeria.
Great example of the Lizardman Constant in action: this survey https://www.equimundo.org/resources/state-of-american-men/ claims that "40% of all men say they trust one or more “men’s rights,” anti-feminist, or pro-violence voices from the manosphere." They got this figure by asking men about a list of problematic websites and adding up everyone who said they trusted at least one. 5% of Gen-Z respondents claimed to trust r/braincels... which hasn't actually existed since 2019! And the authors of the report just took that number at face value.
Knowing absolutely nothing else about them or their research, the fact that this organization considers "men's rights", "anti-feminist", and "pro-violence" to be a single category... Raises a few red flags, to say the least.
For the record, the single category they use is the "manosphere" which includes sites binned into the subcategories you list, plus an "anti-feminist" category (e.g. Jordan Peterson). Interesting Joe Rogan wasn't included as the leader of the "broletariat"
Can anyone here provide an overview of various approaches to polygenic trait prediction? For example, are neural networks better than traditional regressions; what is the state of the literature on the topic; what non-credential requiring datasets (for any species) are there with full genotype/SNP-phenotype pairs; does simulating data with normally distributed noise work for comparing approaches in the real world.
Predicting massively polygenic traits (schizophrenia, bmi, cancer risk) using sequence data seems like a thing that could be improved with new neural net architectures, and I was curios what exactly the state of the field was and if doing my own project there would be helpful
My understanding is that at the moment we are still limited by the data themselves, perhaps more than by the methods of analysis: in diseases such as schizophrenia, the architecture of the trait is a mixture of common variants that increase the risk slightly and very rare variants that can increase it a lot. These very rare variants are mostly in non-coding regions or are copy number variants, and they are difficult to genotype because they require whole genome sequencing.
Note that I am not an expert on this: I am a geneticist and I sort of follow the literature on the subject because I find it fascinating and I teach it, but I work in a different subfield.
Does anyone have recommendations on how best to find jobs that are meaningful, in the EA sense? For example, something in politics that helps with justice system corruption, or something in animal welfare that helps end factory farming, etc. The issue is that my primary skill is computer science, and i'm unsure how to find computer science related jobs that are meaningful (no AI related recommendations please).
All jobs are meaningful in the EA sense (within a margin of error).
Pay attention to the fact that every job is, ultimately, helping somebody else. And since help is, at least to some extent, fungible, then every job is helping the world get better.
(Within a margin of error. There are jobs which are counterproductive / destroy value, but they're the exception rather than the rule.)
Just to throw it out there, networking can be helpful too. Find someone who has a job that you would want, and then email them asking to chat about how they found it, what it's like at their company, etc.
Love triangles are well known, but ia there a term for a circle of unrequited attraction?
E.I. Arthur likes Becky who likes Charles who likes Danielle who likes Arthur?
Found myself in one and felt like it should have a term.
"Love triangle" is a misnomer; the thing called a love triangle is nearly always actually a "love V"/"love line". An actual love triangle would necessitate an odd (and hence non-zero) number of same-sex attractions.
Valid point.
Chain of Fools.
Unrequited Love Triangle? Well, four people would be a square.
Unrequited Love Polygon.
It seems Return To Office is "trending" - what are people's thoughts on this?
I ask because I've been working a somewhat cushy tech job in the midwest after leaving silicon valley, but now they want us to return to the office despite their lackluster salaries and average talent pool. I was willing to swallow inefficient practices and less money for the comfort it offered, and I truly like the company even though I don't always agree with their decisions. But now it feels like there's no reason to stick around if I'm going to be forced into an uncomfortable commute just to sit in virtual meetings in a different building when there's so many more capable companies paying more for fully remote positions. I see a lot of arguments about how 'it's still better than other places' and a lot of chatter implying that people are privileged or unreasonable for demanding fully remote work. Am I out of touch?
I think this is part of the general trend where most of society went virtual from 2020 to 2021, and then started reopening again in person over the course of 2022 and 2023.
My perspective is that if you can find an outside offer that gives you what you want, then no, you are not out of touch. If the company absolutely needs you and needs you to come in they can find the funds to pay you more until you agree to come in. If you can't find an outside offer that gives you what you want, then yes your expectations are out of touch with what's possible for you. To me discussions of 'productivity' are more complicated. I'm typically skeptical of people who claim to be even more productive at home than they are in the office, I'm sure some of these people exist but I imagine they are in the minority. Given that managers (and people in general) are risk-averse, I think the workers whose output really dropped when they went remote are much more salient to managers. I also think part of this response by companies is in response to the general downturn post 2022-Q1. In that sense I don't think productivity really has much to do with it.
IMO the "Get a job in a HCOL city -> Work fully remote -> Move to LCOL City" arbitrage was never going to last. Some people were going to be grandfathered in and enjoy the benefits, but over time either salaries would fall or COL would increase to. Personally I'm surprised more tech people aren't worried about the increased ease of off-shoring that comes with a fully remote company. I know developers who work on teams comprised of 1-2 NA workers and then a slew of people in India
>Personally I'm surprised more tech people aren't worried about the increased ease of off-shoring that comes with a fully remote company.
I've been worried for...my entire (short) career, actually. In 2015, I was chatting with a project manager at a large professional service company, and they 1- Were totally cost-cutting with indian workers, and 2- Absolutely hated it because it was atrocious to communicate with them and the quality of the output was very, very low. Since then, every single person I've chatted with gave me the same picture: contracting indians is cheap, and you get what you pay for.
Now, I'm pretty sure that 1- They were getting the bottom of the barrel because they were beancounters looking for cheap labor, 2- The average quality of indian engineers increased over time, 3- AI will bring up a large part of the workforce anyway, but all that is compounded by the fact that demand for software will keep increasing the in forseeable future, so i'm not too worried either.
I find it sad how Covid demonstrated that things formerly thought impossible can become possible overnight, and now that it is over, those things are gradually becoming impossible again. :(
Most of my team members are in different countries, so I communicate with them remotely, but for some reason it is very important that I do it from the office. Especially absurd considering that we all sit in an open space, so anyone talking is interrupting their colleagues.
The local strategy is management declaring that remote work is only allowed N days a week, and decreasing N by one every year.
> a lot of chatter implying that people are privileged or unreasonable for demanding fully remote work
In theory, this should be decided by the market. Supply vs demand => salary and/or benefits. Work from home is one of those benefits. You decide how much it is worth for you (how much salary are you willing to give up for how many days working from home), and choose your job accordingly.
Yes, you *are* privileged... by the market. We got lucky, born in the right era with the right talents and skills. We are high on the scale of privilege, but still far from the top. Do you think that your CEO feels guilty for making 100x as much as you do? No? Then why should you feel guilty for wanting your benefits?
Enjoy it while it lasts, tomorrow we may all be replaced by super-intelligent chatbots. When that happens, your company will not care whether it is "reasonable" to let you starve.
Can someone explain the UPS Teamsters' recent $170,000 contract victory? (see eg https://fortune.com/2023/08/08/ups-drivers-170000-union-agreement-teamsters-middle-class-bidenomics/ )
Am I right that this is ordinary truck drivers with no special skills earning more than many doctors and lawyers? Is their union just that good?
Given that UPS predicts their profits are going down something like 10% but they're still able to be a robust company with decent profits, does this mean every company could afford to pay their workers $170K and still function okay, and it's just a lack of unions that's preventing them from doing this? If not, what's special about UPS?
I would point out that that $170,000 number includes benefits. I’m not sure that is true with a lawyer or a doctor. In other words I think the latter are inclined to make that kind of money in addition to their benefits.
>does this mean every company could afford to pay their workers $170K and still function okay, and it's just a lack of unions that's preventing them from doing this?
If UPS can afford to pay its drivers $170k, then those drivers are each creating more than $170k in value for the company. I can't think of any other low-skilled job for which that is the case. What appears to be special about UPS is that their business model leverages a ton of value out of some of their low-skilled workers.
I would guess that even though it is a union job, that those drivers can't get away with being lazy.
EDIT: So, no.
I looked into this a little bit and it looks like there are a few things that are fishy here:
1. The $170,000 number is misleading, this article says that before this deal this was broken down into $50,000 in benefits and $95,000 in salary (https://www.bbc.com/news/business-66445496). Then this goes up over the next $2.50/hr/year for 5 years until it becomes ~115,000 in salary. So it's really total comp 5 years from now is $170k, which feels really different from the headline number.
2. That $95,000 in salary is calculated out from 45 hour weeks for 50 weeks a year, carrying heavy boxes. Not highly skilled, but not a super cushy job either.
3. It seems like a pain to get one of these jobs. You have to work part time in the loading docks until you get picked to drive based on seniority. Then you get the worst routes until more senior people age out (https://www.reddit.com/r/UPS/comments/azmq9j/how_do_i_become_a_ups_delivery_driver/).
4. I assume you are required to pay some of your paycheck back to the union.
None of this is to say that it isn't good pay for the skills required. It seems like a better trade than, for example, being a teacher (who are also usually unionized). But I think the articles are maximally misleading about how good the deal is here.
Also sorry for not having a real answer to your questions ("Could every company pay this much?", "How are unions affecting this?", etc.). I just don't know too much about that.
Governments use such tricks, too. If you read in newspapers how teachers in Slovakia can now make extra €50 a month, the missing part is something like "if they volunteer to work on a government-approved project in their free time for at least 6 months" (the example is *not* made up). But the internet immediately starts yelling: "damn those greedy teachers, such easy work and so much money, next time they go on strike I will punch them".
Driving a truck seems like an easier job than working in a Burger King.
Loading and unloading the truck, and moving every package from the truck to the recipient, maintaining tracking and handling exceptions, is not easy.
Working in a Burger King is hard too, especially in rush hour when there's a dozen burgers needing cooking at once, everything is constantly beeping at you, there's hot surfaces everywhere and dumb teenage coworkers running around like headless chickens. Delivering packages seems much less stressful.
Whatever happened to the term "virtue signalling"? A couple of years ago it was being used all the damn time, but I just realised I haven't heard it for a while.
Either way, protesting that your salary is too high and more should be given to the low-level peons, when you know there's absolutely no way that this protest will lead to your salary being decreased and given to the low-level peons, seems like a fantastic form of virtue signalling.
Nobody is stopping them from donating half (or whatever fraction) of their salary to the nursing staff. That way they could save a bunch of time and make sure the money went to the right place.
Why? Individual MDs could just straight up gift money to individual nurses. I presume Canada has some annual exclusion up to which gifts are tax free - spread it over enough nurses and there is no double taxation issue at all.
ETA: Google tells me Canada has no gift tax at all...so...
Hanania was a guy who brought us a step closer to a truth teller who could say; (I'm paraphrasing, maybe, despite the quotes): "We need police to enforce black neighborhoods stronger", without fear that he would be accused of being a racist for saying so.
It really sucks that a prominent person who didn't fear getting called racist for saying some truths really was a fucking racist. It's easy to imagine a Hanania who wasn't a racist, but he was. So much for imagining.
I agree with his sentiment that we need police to police high-crime areas more strongly. I don't have a problem with him specifying black neighborhood because those are in fact the neighborhoods that need more policing. (To be clear: policing of the good kind. Not redneck fucking cops who are too common, but that's another issue)
But his tweet under his real name which I just saw: "Daniel Penny getting charged. These people are animals. Whether harassing people on the subways are walking around in suits."
So according to Hanania black people are "animals" even if they are "walking around in suits"?
Please tell me if I am wrong in my interpretation of that tweet. And as far as I can tell that wasn't a drunk tweet, something I have forgiveness for if deleted the next day, but one he kept.
I don't want Hanania to be canceled on any platform. I'm for free speech above almost everything. But he does sound like a fucking bigot at heart. I'll still read him, at least for a while. But I can't pretend that my opinion of him isn't a lot lower than it was before.
This sorta gets into the territory of whether we should separate the artist from the art, except in the case of Hanania there is no art.
>It really sucks that a prominent person who didn't fear getting called racist for saying some truths really was a fucking racist.
What does "racist" mean, here? Is it "racist" to believe that genetic differences between races explain behavioral differences? Do you acknowledge in principle it's possible for a correct understanding of reality to be "racist" (and see why that might be problematic)?
>(To be clear: policing of the good kind. Not redneck fucking cops who are too common, but that's another issue)
Source?
>I don't want Hanania to be canceled on any platform. I'm for free speech above almost everything. But he does sound like a fucking bigot at heart.
If Hanania is a bigot, then so are a literal majority of black americans, muslims around the world, etc.
>This sorta gets into the territory of whether we should separate the artist from the art, except in the case of Hanania there is no art.
No, it gets at the fact that white people cannot, in the view of leftists, make any kind of argument about anything related to race unless they have a strongly non-negative view of black people, whereas hating white people is not only acceptable when talking about race but encouraged.
>> What does "racist" mean, here? Is it "racist" to believe that genetic differences between races explain behavioral differences? Do you acknowledge in principle it's possible for a correct understanding of reality to be "racist" (and see why that might be problematic)?
I cannot understate just how much you are stretching trying to apply this defense to the situation. How on earth did you manage to see the statement “these people are animals” and somehow jump all the way to “oh, he’s just saying he believes that genetic differences between races explains behavioral differences?”
Okay, then you're admitting that leftists are more racist than right wingers and blacks are more racist than whites.
Whom do you interpret "these people" as meaning, in that tweet? Why do you interpret it that way?
I interpret it to mean "blacks" because:
1) the guy harassing people on the subway who was strangled was black
2) calling blacks "animals" is a standard cliche in the USA. (See: The Godfather for a pop-culture example.)
3) i can't think of another way to tie the subway harasser to the people in suits
I could be wrong, of course. Do you have any ideas of who else he might mean? I considered the possibility he means progressives, but how would he know the homeless guy was a progressive?
EDIT: Let me correct the typo in my quote in case it wasn't obviously a typo. The post reads "Daniel Penny getting charged. These people are animals. Whether harassing people on the subways or walking around in suits."
I've noticed a 'missing mood' in this episode.
Shouldn't those who think that it's possible and/or important to have these conversations without being motivated by racism be the maddest at Hanania? Am I missing people expressing this sentiment? Because I definitely haven't seen it.
Whatever your position, he just made those types of conversation way harder to have.
As someone in that camp, my position is that the fact that one person discusses these things and is (or was) racist doesn't imply anything about whether it's possible to discuss them without being racist. Of course the usual suspects will pretend that it does, but those will pretend it's impossible to discuss these things without being racist regardless.
Also, if progressives use Hanania to make poor arguments, that's not Hanania's responsibility. After all, one of the common arguments against discussing race in non-progressive ways is to say that it will be misused by racists to promote racist policies. Part of my rejection of that is to reject the notion that if racists cite me to make invalid arguments, I'm responsible for that. If I say something, and racists use it to argue something that doesn't actually follow from it, then take that up with them, not me. But the same thing applies here: if progressives use Hanania to make invalid arguments, take that up with the progressives, not Hanania.
It doesn't actually change things because it doesn't matter how 'nice' you are about things, if you have the "wrong" opinion on race, leftists will seek to destroy you. The fact you have non-progressive views on a racial topic IS "racism" to them (whatever that even means).
The left are hateful towards white people (https://ideasanddata.wordpress.com/2020/06/03/american-racism-and-the-anti-white-left/) but it's irrelevant, because none of this has to do with 'hate' or 'racism' or anything. The left believe what they believe and will never be persuaded otherwise, and if they think screaming 'racist' at somebody will help "win" the debate, they will. Charles Murray is obviously less hateful than >99% of left-wingers with strong racial opinions, but leftists wil l still use violence and intimidation to shut down any speech he gives on campus. BECAUSE THIS ISN'T ABOUT HATRED.
And yes, right wingers can be persuaded. You have to be open minded in the first place to be open to different opinions on race than the ones you get shoved down your throat by schools, colleges, the media, corporations etc.
>> And yes, right wingers can be persuaded. You have to be open minded in the first place to be open to different opinions on race than the ones you get shoved down your throat by schools, colleges, the media, corporations etc.
This is all kinds of backward. You're essentially saying "the *real* persuadable people are the people who remain *unpersuaded* in the face of arguments from every direction (schools, colleges, the media, corporations, etc)." Turning stubbornness upside down and citing it as evidence for, of all things, openmindedness.
Individual right wingers or left wingers may be more openminded or less openminded, but trying to demonstrate somebody's openmindedness by citing how stubbornly they cling to their convictions in the face of opposition just makes no sense at all.
I think the idea is that basically everyone starts off with the opinions/beliefs that society tells them to have, acquired through peer pressure and imitation, not through argument, but then some are persuaded to switch to unpopular opinions/beliefs, often through argument.
I don’t think your statement “basically everyone starts off with the opinions/beliefs that society tells them to have” lines up well with Goldman’s statement that “you have to be open minded in the first place to be open to different opinions on race than the ones you get shoved down your throat by schools, colleges, the media, corporations etc.”
“Society” and “schools, colleges, the media, corporations, etc” are not readily interchangeable terms. “Society” is broad, and encompasses both your immediate community (family, friends, social relations) as well as institutional actors like “schools, colleges, the media, corporations etc.”
An attempt to conflate the two confuses the argument in significant ways. Since “society” includes your family and immediate community, one can plausibly assert that “basically everyone starts out with the opinions/beliefs that society tells them to have,” but the statement falls apart if you try to swap in the institutions and say “basically everyone starts off with the beliefs that schools, colleges, the media, corporations etc tell them to have.” Your family has access to shape your mind from the moment you are born, but you need to reach certain age thresholds before various institutional actors gain access to you.
And that simple fact blows up the whole argument. If everybody started out already baked in with the proper corporate opinions, then some people held them and some turned away, we could have a conversation about why the turners are turning away (spoiler, I think it would be more parts emotional suasion than careful argumentation).
But facts being what they are that would be a pointless conversation to try, because people aren’t starting baked with those institutional opinions in the first place – they're starting (for lack of a better phrase) with family opinions. And unless you can isolate people who started out with all the "right" beliefs and then were platonically convinced by argument to become racially resentful (probably not very many) from people who were "just raised that way" (probably a lot) it'd be foolish to look at a person with a bunch of racial resentments and say “that’s a sign he’s very openminded – after all, he holds unpopular opinions!”
I think the large shift in racial orthodoxy over the past century would suggest that the immediate community isn't playing that large a role. To me is seems most plausible to say these opinions are mostly coming from US media, considering which countries have been the most affected.
I think that's more or less my mood. The leftists who scream at him louder strike me as performative.
No, it really isn't easy to imagine a Hanania that isn't racist. His entire schtick was finding verbose ways to justify pre-determined right-wing conclusions. You're never going to find someone that dedicated to "anti-wokeness" without substantial prejudices.
And while we're here, America is already hyper-policed. It's the most imprisoned state on the planet. And black neighbourhoods are the most policed part of it. The idea that they're under-policed is absolutely wild.
There are limits to what police can accomplish. You can't always just "do it harder".
>You're never going to find someone that dedicated to "anti-wokeness" without substantial prejudices.
For goodness 'sake! Find me a single person with 'woke' opinions of race who is even remotely open to any non-woke explanation of race differences. I'll wait.
Wokists are not only closed to the very possibility of e.g. race differences being caused by genetic differences, but think that it literally shouldn't be legal to suggest that this is true.
In order to come to non-woke conlcusions in the first place you have to be open to alternative explanations to the narratives that dominate in america's hegemonic institutions.
>And while we're here, America is already hyper-policed. It's the most imprisoned state on the planet.
Walk down the streets of san francisco and tell me there's too much policing.
America is the most imprisoned state due to a combination of well funded and effective law enforcement and an extremely high crime black population.
Other countries with large black populations i.e. in Africa simply don't have effective law enforcement.
>And black neighbourhoods are the most policed part of it. The idea that they're under-policed is absolutely wild.
Absolutely false. The vast majority of crime per capita is committed by black americans, and the policing in black neighborhoods is LOWER than the crime rate proportionally. The fact that there's more police in black neighborhoods does NOT mean they're over policed, and if BLM supporters actaully wanted less black people being killed, they should support more police in black neighborhoods. But somehow, Geroge Floyd's life is worth more than the 10000 or so black people that murder each other every year, because BLM is tribalism, pure and simple.
Comparing incarceration rates in the US to other countries is meaingless, unless you think that the underlying crime rate is the same in both countries.
As it is, of course the incarceration rate is higher in the US than in other western countries -- the crime rate is much higher. Crime in the US is so high that even once a couple of million of the worst criminals in the country have been put in prison, the US _still_ has vastly higher levels of most crimes than most other comparable countries.
Anyone who thinks the US is "hyper-policed", I invite you to do this experiment. First, shoplift a cigarette lighter. Then, use it to light a crack pipe. Then, while smoking your crack pipe, take a shit on the sidewalk. Do this in each of these locations:
1. Market St, San Francisco
2. Orchard Rd, Singapore
3. Kufürstendamm, Berlin
4. Oxford St, London
5. Wangfujing, Beijing
Report back on the police response in each location, and we'll see how over-policed the US is.
Indeed. See: https://marginalrevolution.com/marginalrevolution/2022/08/still-under-policed-and-over-imprisoned.html. It notes that the USA has a normal incarceration rate relative to number of homicides, a somewhat low level of police per capita, and a very very low level of police per homicide.
Let's face it, "Woke" and "Tough on Crime" have become the new racist dog whistles. It's all the Left's fault though! <#snarkasm> It's because of our 60-year obsession with shaming people who use racial epithets. Racists are good people. They don't want to be called racists! And the Talking Class of modern bigotry want to appeal to middle-class voters who want a kinder gentler form of racism (no more Bull Durham's, please <gasp!>) — even though they'd cringe if their HOA allowed blacks to purchase a home near them. So serious bigots have to keep fabricating new and improved ways to encode their message. What a shame that Cancel Culture makes it tough to be a racist asshole in public anymore.
And in a nice bit of symmetry, "Racist!" is now pretty much a Woke dog whistle. That part really is the left's fault.
I'm sure the unelected mayor and city council have the best interests of this majority-black community (85% black, 15% white) at heart. After all, it's in everyone's best interest that there never be *any* elections. Otherwise, the wrong sort of people could get into office. The WSJ hints that "race" may just possibly be a factor, but obviously, the small clique of white people who've run this community for time out of mind aren't racists. No, that couldn't be possible in this day and age!
https://www.wsj.com/articles/political-and-legal-scandal-rocking-a-small-alabama-town-whos-the-mayor-9a996f56?mod=hp_lead_pos10
It's been woke dog-whistle since before we wuz woke! But a nice litmus test is whether you'd feel comfortable about walking around a black neighborhood and stopping to shop and dine at the black-run businesses. Even many of my whitebread leftie friends would feel uncomfortable when they're a minority of one in all-black setting. But as a general stereotype, I've found black Americans to be friendlier than my white cohort, and they definitely throw funner parties. ;-)
Banned for one month for this comment.
I love getting under the skin of you National Socialists and NSA-NSPA fellow travelers. Lol!
I don’t really think this is fair or accurate at all, unless you mean it in the sense that literally everyone has prejudices (which is true).
By far the most prejudiced people in society right now are all on the left.
If you believe that leftists are more prejudiced than neo-Nazis, then you are essentially lost to reason.
Neo-nazis, people who make up a fraction of a fraction of a percent of the population, commit almost none of the crime in the US, have no institutional power or support...
But black are empirically more racially hateful than whtie people.
And neo-nazis are a force in American society where? How many are there, a couple dozen thousand, with zero institutional power? Have you ever met a neo-nazi? I haven't.
Meanwhile out and out racism is the straightforward policy of many large institutions as long as it is directed in the right direction.
https://www.csis.org/analysis/escalating-terrorism-problem-united-states
You said, "by far the most prejudiced people in society right now are all on the left".
This is a specific claim, and an obviously false one. If you want to replace it with a more defensible claim, you should acknowledge that you're doing so.
I get that we all run our mouths online sometimes, and that these comments are not carefully edited articles with multiple drafts. What you meant might be far more reasonable than what you said. But I can only see what you said, and I have no other guide to what you might have meant.
Nope, you're being a pedantic jerk. We're talking about mainstream politics, and you know it.
It's an ambiguous claim, and false only because "by far" goes too far.
If "most prejudiced people" is meant to refer to the greatest number of people who are at all prejudiced, then both the left and the right seem to have roughly similar numbers in total, and they've both got mostly people with an assortment of low-level prejudices.
If "most prejudiced people" is meant to refer to the people with the greatest degree of prejudice, then the right has the neo-Nazis et al, and the left has the militant Black Nationalists and the Kulack-hating Antifa fringe, none of which on either side end up posing any real threat to the Republic, but it's hard to say which is worse.
If the idea is that we get to compare the thousands of literal neo-Nazis on the right with the millions of leftists with an assortment of low-grade prejudice and say "a Nazi is worse than the median leftst", or vice versa, then no. Knock it off. Both of you.
Dude, you/we were clearly talking about the broader society, not the scattered random basement dwelling neo-nazis. Honestly you want to get in a real discussion of the most racist people they are likely recent immigrants from the Balkans or the Middle East, not some teen edge lords, or nazi biker gangs in San Quentin or whatever.
Honestly beyond even that I kind of suspect that relatively well respected figures on the left do hold views just as racist as neo-nazis, judging by how racist they are comfortable being right out in the open.
There are many ways to measure "quantity of policing." I doubt the OP is referring to "fraction of the population in prison." He says we need to "police high-crime areas more strongly" - the most straightforward interpretation of that is "put more police officers in high-crime areas."
It is not the case that America is "hyper-policed" by the metric of "police officers per capita in high-crime areas." For example, across the entire city of Chicago, there are approximately 12000 sworn officers for a population of 2.7million, vs 34000 for a population of 6.7million in Paris, per wikipedia (https://en.wikipedia.org/wiki/Paris_Police_Prefecture), so Paris has more police per capita. I would also argue that since Paris has fewer guns, and a legal regime that makes gun laws easier to enforce, a single police officer goes further in Paris than in Chicago.
Furthermore, in Chicago, police officers tend to be distributed relatively uniformly on a geographic basis, rather than concentrated in high-crime areas or in black areas (e.g. see this distribution from 2017, the latest year for which I have numbers: https://www.cbsnews.com/chicago/news/inspector-general-report-chicago-police-department-officers-district/)
It's worth pointing out that Black Americans overwhelmingly do not want *less* policing in their neighborhoods (https://news.gallup.com/poll/316571/black-americans-police-retain-local-presence.aspx). This is remarkable, given that Black Americans also report quite negative personal experiences and interactions with police officers. A combination of *more and better* policing would therefore probably be very popular. People don't like having violent crime in their neighborhood!
True, there is a bit of ambiguity about what more policing means.
Since we're talking about Richard Hanania here, I think it's pretty clear that "police black people more strongly" means "more state-sponsored racial terrorism".
Has Hanania said he's in favor of state-sponsored racial terrorism? This seems like exactly the kind of reasoning that can always be deployed to convince yourself your enemies are monsters. Joe Biden just said he wants more funding for public schools--obviously this must mean he wants more money for propagandizing kids with critical race theory and convincing them all they're trans.
The great thing about this kind of reasoning is all the thinking it saves.
When an actual-factual white supremacist calls for more policing of black people, it's not hard to guess what he means.
I don't make this kind of accusation against normal right-wingers.
The vast majority of interracial violence in the US is committed by blacks against whites, in absolute terms and ESPECIALLY per capita.
is this "racial terrorism"? If it is, but STOPPING people committing homicides etc is, then you're a black nationalist ideologue, nothing more.
Yes, I think you've captured what the OP is bemoaning. Hanania is a huge racist, and his arguments for more policing are likely in bad faith. Therefore, we can reject arguments for more policing out of hand as bad-faith racism.
The American left hate white people. This is an empirical fact: https://ideasanddata.wordpress.com/2020/06/03/american-racism-and-the-anti-white-left/
Can we dismiss all of what they say out of hand? Or is hating white people acceptable to you?
If you want to reform the police, I'm all for it. If your reforms would increase the number of police officers, I'm skeptical but willing to listen.
If you say "we need police to enforce black neighborhoods stronger", then I'm just going to dismiss you out of hand. Because
a) the racial bit makes me assume the worst
b) "stronger" usually means more violence and harsher sentences.
and
c) American police budgets are already absolutely insane.
It's very clear that America needs less policing in many ways. If you think it also needs more policing in some other ways., then you can make that case.
But when you do, you need to go out of your way to make clear that you're not like Richard Hanania.
It can be done! It's not even that hard!
But you have to do it. Or people will be right to dismiss you.
PS: Isn't it weird that America's enormous police spending doesn't buy it a similarly enormous police population? I'm genuinely curious about what's up with that.
Do the people living in high-crime neighborhoods generally think the US needs less policing? My understanding is that they very much do not, and in fact that a common complaint of people living in poor, mostly-black neighborhoods is that the police don't show up when they're called because they DGAF what happens in those neighborhoods.
>a) the racial bit makes me assume the worst
America would be a low crime country without black crime.
>b) "stronger" usually means more violence and harsher sentences.
So black violence is okay, but harsher sentencing isn't?
>c) American police budgets are already absolutely insane.
Compared to the crime rate, they're not.
>It's very clear that America needs less policing in many ways.
It's not clear at all. America's crime rate is much higher than many other developed countries.
"I'm not like Richard Hanania." With that out of the way:
I also don't care for the statement, OP's paraphrase of Hanania, that "we need police to enforce black neighborhoods stronger," because I agree with you on (a) and (b). (On the other hand, OP goes on the clarify that his own policy preferences are "more police in high crime areas" rather than "more punitive sentences for black people", and states that he thinks Hanania is a "fucking bigot" - so maybe it's not totally pointless to engage with him?)
Regarding (c), and again comparing Chicago to Paris:
Chicago police budget in 2022 was $1.75b, of which $1.6b, i.e. pretty much all of it, was personnel costs for the ~14k positions, so call it an average salary of $115k - though I note that this does not include pension costs.
Paris police budget in whatever year the Wikipedia article was edited (2018 maybe?) was 1.5b euros. I don't know how much of that is personnel costs, but probably also almost all of it. For the 45k employees listed, that's maybe 32k euro average salary. (I assume that this also does not include pension costs.)
So inasmuch as American police budgets are "absolutely insane" relative to Europe, it would seem to be driven largely by salaries that are perhaps 3x higher. Given differences in per-capita income, you expect generic US salaries to be less than 2x higher, rather than 3x higher. I don't have high confidence in these figures - the French figures are from an undated Wikipedia page, and I can imagine there might be substantial differences in benefits like vacation time, or in tax treatment, or other things that make these numbers incomparable.
But assuming the premium really is about 50%, there might be a number of causes:
- Is policing in the US substantially more dangerous? Are US police officers more likely to get shot than French police officers?
- Do US police officers work longer hours or have more overtime than French police officers?
- Do US police unions have more political power in contract negotiations than the French equivalent?
Both things can be true at the same time: we have very high incarceration levels and not enough policing in areas that could really benefit from it. Our policing model is profoundly broken and the insane contradictions are the result.
+1
Right. Alex Tabarrok argues that more cops would mean fewer people in prison because policing prevents crime.
"more cops would mean fewer people in prison because policing prevents crime"
Interesting point! Yes, this might happen if the deterrent effect of the policing is the dominant effect. Is there experimental evidence, one way or the other?
Sometime ago I read a rant by a retired cop who's main point was that in the last couple of decades police work shifted from "policing" to "law enforcement", to the detriment of both. In the old "policing" paradigm, cops get called on a bar brawl, arrest the bozos, throw them into holding cells overnight, and give them a stern talking-to in the morning, no charges filed. Civil order restored, no lives ruined in the process.
Now the cops come in and try to figure out if this can result in a court case. Because they are in a law-enforcement business. If things look dubious they may not even try to arrest the bozos, but if they do, now there are charges filed, and if they can't post bond they lose their jobs, courts get clogged up by the stupid cases, etc. etc. at a great cost to themselves and the taxpayers.
Then there's the fear factor: if black folks think that their calling police for help may easily result in themselves being George-Floyded, they will think twice, which only emboldens criminals and psychopaths. A vicious cycle ensues.
So I don't know if there's any coherent data to Mr. Tabarrok's point. It feels correct, intuitively, but the devil is in the details, and the details are legion.
>>So I don't know if there's any coherent data to Mr. Tabarrok's point. It feels correct, intuitively, but the devil is in the details, and the details are legion.
Thank you for including this caveat. I agree that this "feels correct" and that whatever the truth is is likely way more complex/hard to document.
With that same caveat, your point sounds like a natural extension to something similar I heard about judges and prosecutors. Prosecutors and judges used to have much more discretion in terms of what cases to bring and drop, what sentences to apply, etc. But, over the course of the 60s and 70s, we found a lot of racial disparities and abuses in how those decisions were made, and that discretion came under fire. We get things like mandatory minimums and mass protests over what charges are or are not filed. The drive for a "fairer" system also made the system more rigid.
Seems like a parallel to your observation about policing vs. law enforcement. Previous generations of cops had greater discretion to refer some people to the district attorney for prosecution while others got warnings, a number of them used this discretion wisely while others used it corruptly, society realizes that discretion is being used corruptly and tries to solve by reducing discretion, and now cops are arrest/citation machines.
Again, it's more parts Stephen Colbert's "truthiness" than anything - I'm not sitting on a pile of evidence for it. But it does make sense.
This sounds correct to me, too. It's not primarily about how many cops there are, but what they do. Are they trying to maintain peace, or to arrest as many people as possible (and ignore those who cannot be arrested)? The latter strategy leads to more problems *and* more people in prisons.
I used to have a few cop friends and they told me their main purpose is to be a deterrent. They consider themselves to be in the safety business.
Now, perhaps these guys I knew were overly idealistic or I was naive to take them at their word, but it makes sense to me.
Many Thanks!
This sounds like part of a general trend in USA culture towards micromangement. "Charges filed" sounds all too much like the sort of thing that fits into some management "metric", while "civil order restored" is harder to quantify.
Indeed. See here: https://marginalrevolution.com/marginalrevolution/2022/08/still-under-policed-and-over-imprisoned.html which notes that the USA has a normal level of incarceration per homicide, a somewhat low level of police per capita, and very very low levels of police per homicide.
Additionally, the US does not typically have particularly long sentences for crimes.
Below are the median sentences for various crimes in years (except where months are specified) as of 2020.
Murder: 19
Sexual abuse: 15
Kidnapping: 14
Child Pornography: 7
Assault: 3
Drug Possession: 0 months
Source: https://www.ussc.gov/research/data-reports/geography/2020-federal-sentencing-statistics
Maybe it's a matter of the same people, but with shorter sentences.Most European sentences are 1/3-/12 American ones.
I always love the "lets free the people in there for minor drug crimes", and it is often the guy who killed two witnesses and assaulted 5 "friends", but there was a hard time putting a case together so they stuck them in there for 10 years for intent to distribute or whatever.
A lot of times the disproportionate seeming drug sentences have to do with other shit going on and plea deals. I feel like most people with attitudes about criminal justice reform would have a radical reimagining if they spent even one day at a county lockup and saw who goes through there.
Plea bargaining corrupts the whole damn system.
Hanania may be bad but the problem is Twitter.
I assume "these people in suits" refers to the prosecutor who filed charges against Penny and the activists urging him on.
It's not clear to me what "these people" refers to in Hanania's tweet, unless there is more context someone with a Twitter account can glean. Could a charitable interpretation be any of those who have no respect for law and order by initiating or excusing hooliganism?
(I live in a bubble where I don't pay attention to these sort of cases, but worth noting in the video itself, there are two other civilians helping restrain the thug, so it's clear there was some degree of agreement in the subway regarding the threat)
In the WEIRDEST People book, as I understand it from a review I just read, it was weird how Christians made the culture monogamous. They ended cousin marriage and gave us the nuclear family we now know in The West. Before that, polygamy and cousin marriage were common.
Supplanting polygamy with monogamous marriage had massive effects upon society: men were basically domesticated. Maybe women were too. It was good, if you're one of those people who likes human flourishing, for human flourishing. Because men, no longer competing with each other for more wives, now did more productive things, or something.
My main question is this: aren't modern Asian societies mostly monogamous? If so, then Christian Europe wasn't so weird.
I just want to point out that the Ancient Romans were (imperfectly as always) monogamous prior to Christianity.
The wealthy ones had a harem of slave girls (and boys), so not really monogamy by the modern conception of it.
Monogamy is a human default and Henrich never claims otherwise. The WEIRD aspect of modern nuclear family is not that it's monogamous, it's the fact that it became a fundamental social and economic unit, rather than being subsumed into larger clan and tribal structures that were the norm elsewhere until that point.
(Regarding polygamy, it does not need to be common to pose a problem. Just a small fraction of wealthy men taking several wives each creates, and normalizes, a much larger underclass of men with little marriage prospects unless they engage in significant risk-taking.)
Perhaps I am misunderstanding Henrich because I am "reading him" through Misha Saul's substack posts: Wife Economics and the Domestication of Man. https://www.kvetch.au/p/wife-economics-and-the-domestication
"Man is born polygamous yet everywhere he is monogamous
Not long ago I had my own Fermi moment. I looked at the world around me and asked: Where are all the polygamists?
Consider almost any past empire or civilisation — Mongol, native American, Chinese, Indian, African, old European — and you will find powerful men with many wives. It’s all over the Hebrew Bible. 90% (!) of hunter gather societies around the world practice some degree of polygamy. Yet we look around today and… zilch?
...It turns out this is no accident. The WEIRDest People in the World by Joseph Henrich (I’ll call it WEIRD from now) traces how Christianity exterminated the practice over centuries and forged modern, cousin-free, monogamous marriage in the West (hence WEIRD: Western, Educated, Industrialized, Rich and Democratic). In light of Christianity’s now millennia long clamp down — occasionally literally hounding kings over successive popes until they submitted — it’s not so surprising world leaders and billionaires today seem about as monogamous as anyone else. It’s remarkable how potent culture is: what was once as natural to a powerful man as eating is unthinkable to such a man today."
"forged modern, cousin-free, monogamous marriage in the West "
Yes, but it's the "cousin-free" part that is doing most of the heavy lifting, and that Heinrich is calling out as important. I think he overstates the case, but he's probably directionally right.
Frequent cousin marriage divides people into clans, which is stable because clan insiders are in a more secure place than the rebel who became an atomic individual and then courted someone from a different clan. And clans make polygamy easy. Everybody is strongly pressured to marry a cousin or other clan insider, the clan leader is uniquely desirable as a husband within that group, he can probably get several women willing to take 1/3rd of him over any random clanmate even before he starts throwing his weight around with the fathers of the girls he wants to marry. See e.g. the internal dynamics of modern American polygamous cults.
Eliminate the clannishness - which everybody has to at least partially do as they modernize and industrialize because it's not practical to organize factories on a clan basis, and no woman is being told "If you want to marry an Alpha, you have to marry me even though I've already got three wives because I'm the only Alpha in the clan". There will still be men for whom many women will settle for a 1/3rd share, but there will be many such men and some of them haven't yet married even once.
Some societies will still allow polygamy for the few men (or women) who can pull it off against those odds, others will go to nominal full monogamy (pay no attention the mistresses, girlfriends, very comely secretaries, etc).
The important thing is breaking or at least opening up the clan structure. Monogamy doesn't really do that; barring cousin marriage does.
"My main question is this: aren't modern Asian societies mostly monogamous?"
You are not considering the influence of Westernisation, even in areas like the rise of Communism in China:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3260845/
"While the roots of the package of norms and institutions that constitute modern marriage can be traced back to classical Greece and Rome, the global spread of this peculiar marriage system has occurred only in recent centuries, as other societies sought to emulate the West, with laws prohibiting polygyny arriving in 1880 in Japan, 1953 in China, 1955 in India and 1963 in Nepal."
The big difference for Christian countries (that includes not just Europe), was that monogamy was enforced/encouraged as the *sole* marriage system, instead of being an option alongside polygamy/concubines. For instance, Vishnu in the incarnation of Rama has only one wife (that even becomes important for carrying out a religious ceremony, where he is being urged to take another wife as he has been separated from Sita for a long time) but in the later incarnation as Krishna, he has many wives. The explanation given is that this is to fit in with the needs of the different eras and the different purposes for the incarnations.
There's been a lot of ink spilled over the passage in this epistle of St. Paul, with the pro-divorce side saying that what he means is that the overseer (bishop) can only have *one* spouse with him at a time so if he's divorced that's still okay if the former wife is alive:
"The saying is trustworthy: If anyone aspires to the office of overseer, he desires a noble task. 2 Therefore an overseer must be above reproach, the husband of one wife, sober-minded, self-controlled, respectable, hospitable, able to teach, 3 not a drunkard, not violent but gentle, not quarrelsome, not a lover of money
...12 Let deacons each be the husband of one wife, managing their children and their own households well."
But what all agree is that it lays down that church leaders can only have one wife at a time, can't/shouldn't be living unmarried with a lover or mistress, and can't have several wives/one wife and concubines.
Take the 'option' thing with a grain of salt, at least for India. It's more like there was no law expressly forbidding it, but if and when polygamy did happen it caused a great deal of drama and was generally discouraged (try the movie Bajirao Mastani, which is the story of a king shocking everyone by taking a second wife).
Really fantastic dance numbers, too. I found it funny that Rao's victory celebration after threatening the rival leader into submission was powerful enough to make it on the wider Internet
Haha yeah, I loved that one. Sanjay Leela Bhansali's movies always have great dance parts.
What are the great changes that could happen a thousand years in the future if the economy keeps growing at 2% annually? I find it hard to imagine that the world economy will continue to grow for centuries, but many do and it’s interesting to think about.
It would be a story of humans growing richer and richer. What does that mean? If it means anything, it means people will continually have more of the material things they want.
What are these material things people want? I’m not going to get into say sex, interpersonal relationships, social status and meaningful work. All of those things matter immensely, of course, but I consider them intangible and not what I’m talking about.
The material things we want are: shelter, food, and transportation. Again, happiness has a lot to do with relationships and all that, but I’m focusing on the other things. And thank God material things do matter. Humanity may be no better at relationships than we were 50,000 years ago, but we are better at building bridges and making movies.
So what can we do better in the future? The material bottlenecks of happiness are housing, transportation, safety entertainment, beauty, and good food.
Digging deeper, the bottlenecks are natural resources: land, energy, water, and other raw materials.
To me, that’s the key: natural resources. I don’t think we can grow at 2% a year for a thousand years unless we can somehow make natural resources cheaper. Basically that means alchemy. Will the future be able to 3d print a sky-scraper residential building?
Unlike most people here, I don’t read much science fiction. Are there good sci-fi writers who intelligently imagine cheap production of housing and raw materials? Because it seems to me like that would be the key to a future much greater than ours: cheap, great housing. Living in an awesome apartment for next to nothing. Imagine 5 billion people living in an awesome apartment for next to nothing! Would that not be great progress!
I don’t want to dismiss those, like me, who would prefer to live in a house with a nice front yard and backyard. More people living like that would be awesome too!
But how are housing materials ever going to be cheap? I realize that land is often more expensive than the construction on top of it, but a residential sky-scraper is expensive too.
Imagine a resource-cheap future in which a residential sky-scraper is cheap! Do any sci-fi writers imagine this? I honestly don’t know. Do they have an even semi-plausible scenario for how?
What do you think could make the future of humanity much better in a thousand years?
I think there will be strong limits to growth unless there are major scientific breakthroughs.
There's also the option of harvesting additional raw resources from off-world. This would require some scientific breakthroughs to really leverage well but there are several active projects.
Furthermore, I would not discount the potential for inflation to cause the listed numbers to keep rising even while prices rise in turn to balance it. A "steady state that pretends to be growing" seems entirely possible. Even if people are aware of the fiction, much like current arguments about the useful fictions of fiat money in the first place.
Resources aren't the bottleneck to constant progress. Scientific breakthroughs do indeed allow virtually unlimited growth.
You've left out a bunch of things by limiting wants to housing, transportation, safety, entertainment, beauty, and food. Communications, plumbing, and healthcare are some of the biggest ones. And increasing technology helps with all of them.
Silicon is an essential ingredient in all of the computers we use, and we aren't about to run out of it. We are constantly finding new uses of carbon, including microfibers and Vantablack, since it is such a versatile element. Water may be a source of energy, if nuclear fusion gets off the ground. Who knows what other uses we can make of highly abundant resources?
And that doesn't even count going out into the solar system, and beyond.
I suppose you can have unlimited growth without the quality of life going up much for most people. If quality housing remains scarce -- and why wouldn't it? -- then the quality of life won't improve a ton for most people because housing has a fundamental impact on it. All sorts of futuristic gadgets won't likely change that. I suppose everyone could live in the metaverse and physical reality could essentially disappear for all practical purposes. It's a possibility, but I wouldn't bet that it is a likely one.
AI Quinn mentions below to read "Age of Em". That reminds me that Hanson believes that most people - or ems or whatever -- in the future will live at subsistence level. He reasons that by taking the outside view that the most typical situation, judging by the past, is that most people exist at subsistence level and we are just currently living during exceptional times. I don't know that I buy that, but his vision of the future is one in which the economy continues to grow without most people experiencing an increase in quality of life. There must be many such possible futures.
Quality housing isn't scarce, it's just that quality housing in ultra-high-demand areas is scarce.
That will never change, even if we build a billion new square km of habitable land in space stations, everyone will still want to live within walking distance of whatever the current coolest five cities are.
You can build ten thousand identical nightclubs in your city, but one of them will have a queue around the block and the others will be empty.
It's hard enough to do a 1 year out prediction so not gonna seriously try here but would say check out Hanson's "Age of Em" as a start.
I haven't read that book directly, but I read a lot of Hanson when he was writing about Em's and read a number of reviews/summaries of the book.
Just thinking out loud about a way to quantify (long) covid effects on cognitive performance. FIDE Elo ratings for chess players from all over the world are publicly available (here 2015-2021 https://www.kaggle.com/datasets/rohanrao/chess-fide-ratings). These are a proxy for cognitive ability. Break them down by nation and average within each nation. Check whether the trend over time of average national Elo score shows a different behavior in countries that were hit more or less severely and at different times by covid. It’s essentially a diff in diff design. I don’t have time to do this myself, unfortunately. Watch out for the devil in the details of course: confounds likely abound.
Eli Lilly and Novo Nordisk shares are both up 16%(!) today, on the strength of big sales for the GLP-1 agonist diabetes/weight loss drugs. Muy bueno.
I tried Ozempic starting in February, and stopped around the beginning of April. I hated it because of the side effects. YMMV.
I did find it to be effective in lowering blood glucose. I thought the weight loss effects were because of the side effects, and it struck me as an unhealthy regimen.
How high'd you push the dose? I started on 0.25mg a week in May and am up to 1mg a week right now. My use case is probably different than yours. I'm trying to push below 6-7% bodyfat or so by the end of summer, which I've done before, but it's normally a miserable process where you end up unable to think about anything but eating by the time you're done. This time, it's been easy as hell. Just eat at a calorie deficit every day and keep doing it. No hunger. No cravings. No willpower is necessary. It's gotten big in the bodybuilding community for this reason. Way safer than prior popular diet aids, shit from stimulants to thyroid hormone to pesticides that block carbohydrate metabolism. This is so much safer.
And personally, I've had no side effects. Not experiencing hunger is the only reason I even know I'm taking it.
I started at .25mg and stayed there.
I think it causes both nausea and weight loss, but I don't think the weight loss primarily driven by the nausea. I've had significant weight loss from it and minimal nausea.
I had a relative say the same thing, but he's morbidly obese so everybody told him he should tough it out. He declined, though, and quit taking it after 2-3 months. I guess it was pretty bad, at least for him.
I once took something called topamax as a migraine preventive, and felt groggy and nauseated all day long on it. Lost weight in the couple of months I took it, entirely because I felt too queasy to eat. Later the stuff was used as a component of a weight-loss drug.
I've been hearing something something about politicians "creating jobs", and I realized politicians only have the ability to create jobs by making government bigger, and thus hiring people to fill those particular jobs. As for creating jobs in any other industry, they only have the ability to restrict jobs, by forbidding things that people would want workers to do.
Can politicians ever create jobs in a way that doesn't involve only removing obstacles, or hiring people for the government?
I mean, does building infrastructure count? The construction of the interstate highways probably created jobs both immediately (construction work, which I guess does fall under 'making government bigger', at least temporarily), as well as indirectly and for decades after...
There is also the argument that well designed regulation can allow markets to flourish which would otherwise be impaired by information asymmetry. Thinking here of the SEC and the stock market, but commodities probably counts too.
Fiscal stimulus, monetary policy, creating an environment with both law and order and sensible regulatory regime.
Anarchies are a bad place to start a company or live.
Right off the bat — Defense procurement represents around 5 percent of national manufacturing output. These are mostly civilian jobs with non-government employees. Although I had difficulty teasing the info out of BLS reports, a non-technical summary said that in 2021 this accounted for over 2 million jobs — which were higher paying than average manufacturing jobs.
Likewise, there's a huge ecosystem of private contractors who hire from the private sector that work on government-sponsored projects. Probably there are more people who are employed by private businesses dealing with the government than there are who collect a government paycheck (but that's a wild-ass guess on my part).
NIH, NSF, DARPA, etc. hand out grants to Universities fund research. These handouts keep most of the STEM-related professoriate employed.
Just a few methods off the top of my head:
Negotiate with a foreign country to get them to remove their trade barriers.
Improve law and order, reducing business losses to shoplifting/vandalism, allowing them to be more profitable and hire more people.
Adjust macroeconomic policy to ensure aggregate demand is at an adequate level.
Invest in R&D
Create funding structures that allow people to invest in capacity-building without the upfront cost or prohibitive interest rates, e.g. student loans
Solve coordination problems e.g. by setting uniform rail gauges so goods can be cheaply transported long distances.
Build infrastructure that allows greater productivity or trade e.g. ports
I think it's fair to say that copyright legislation has created a considerable number of productive jobs.
Creating jobs, ideally, means creating opportunities to generate wealth. If you drain a swamp to create fertile farmland then you create opportunities to generate wealth by growing crops on that land. Or in a less extreme example, you build a new road to some place, and suddenly some kind of economic activity on the margin becomes worthwhile. Or you fund research, which creates some new technology, which makes new ways of creating wealth possible.
There's also plenty of wealth-destroying ways to make jobs. The government can hire someone to go round breaking windows to create jobs for glaziers. It can ban the use of agricultural machinery to create jobs for farm workers. It can stop enforcing laws against shoplifting to create jobs for security guards. I'm sure there's even more cynical examples to put here.
And then there's wealth-creating ways of destroying jobs, like funding research that creates technologies that replace human workers with machines.
But my point is that these examples only create opportunity. Governments don't order businesses, or customers, to farm newly-drained land, or to replace windows they had broken.
Per US Treasury: "The federal government spent $6.27 trillion in FY 2022. This means federal spending was equal to 25% of the total gross domestic product (GDP), or economic activity, of the United States that year." A lot of private enterprise in the US depends on the US government purchasing their products and services.
Governments don't directly order people to build bridges or work in their pork-barrel projects either nowadays. They pay the private sector to do that. Or even the QuAnGO sector.
Rischard Shelby wasn't trying to get his constituents jobs at NASA, he was trying to get them jobs at Boeing.
I don't see why it shouldn't count as creating jobs if the taxpayer money goes through the balance sheet of a private company first.
Raise taxes and/or increase the deficit to create jobs? If the government is buying the service then, for this purpose, it would count as a government job. No one is saying the government is sitting on a pile of cash instead of spending it.
And I don't think that counts as creating jobs. I could create jobs the same way, if I borrowed to do home improvements or someone gave me money to do them.
OK. So your question REALLY was: "Are there any examples of "politicians" creating jobs that weren't (a) directly the result of government capital outlays, (b) that weren't the indirect result of long-term economic and social policies that promote economic growth?" Well, I can't think of any jobs that are more than two or three degrees of separation from the government teet. You might want to check a place like Somalia to see job creation is going there.
My question was "Can politicians ever create jobs in a way that doesn't involve only removing obstacles, or hiring people for the government?" My main point being that the government's claims of creating jobs mostly means allowing jobs to be created, most of the rest being the government paying people to do stuff, whether that is their employees or another company. From this thread:
* Well-designed regulation can build confidence in markets, allowing more people to believe it's worthwhile to participate, which can encourage job growth, but doesn't create it in of itself.
* Negotiating with foreign governments to improve trade relations can encourage job growth, but doesn't create it in of itself.
* Adjusting macroeconomic policy can encourage job growth, but doesn't create it in of itself. This is a prime example, actually, of the government being able to destroy jobs, but only allow them to be created.
* Draining swamps lets people plant crops, build houses, or build businesses in the formerly unsuitable land, which can encourage job growth, but doesn't create it in of itself.
* Building infrastructure: allows commerce to take place along, say, roads, which can encourage job growth, but doesn't create it in of itself.
* Suppressing piracy and suppressing crime can encourage job growth, but doesn't create it in of itself.
Others are mentioned that are harder to define.
My main point is it is a rare instance where government that creates jobs that wouldn't be created otherwise, yet politicians take credit for job creation.
That's exactly how small businesses create jobs, by (typically) getting a small business loan and hiring people.
Jobs must be created either directly (via spending) or indirectly, yes. Tertium non datur.
Infrastructure comes to mind. Things like finishing the Great Lakes canal system, the Panama canal, dredging natural harbors deeper to improve the tonnage of shipping they can accommodate, etc, that greatly reduce natural barriers to commerce and thereby unleash economic activity that wouldn't otherwise be profitable.
Of course, how to *measure* how many jobs such infrastructure creates is a best-guess practice at best, so politicians are much more wont to make ridiculous claims by simply looking at the number of jobs the economy added while they sat in the chair and claiming them as "jobs I created," like a shaman taking credit for the rain.
So in sum, I think the most accurate statement is (a) yes, politicians can in some circumstances create jobs, but (b) most such claims of "job creation" are nonsense.
Subsidize an industry with taxpayer money. Count the jobs created in the subsidized industry but not the jobs destroyed in the taxed industries.
I suppose that technically means jobs were created. And by "technically" I mean "not really".
What are y'all's thoughts on the relationship between *sleep* and *meditation*?
Anecdotally meditation decreases sleep need in most meditators, especially on meditation retreats. This WP page[1] supports that hypothesis, and this paper [2] gives a more detailed overview. In the extreme case, it could be that meditating >2h a day decreases sleep need by 2h[3] (giving the benefits of meditation at ~no time cost), though the evidence is shaky (via [4]).
[1]: https://en.wikipedia.org/wiki/Effects_of_meditation#Insomnia_and_sleep
[2]: https://www.frontiersin.org/articles/10.3389/fneur.2012.00054/full
[3]: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2919439/
[4]: https://slatestarcodex.com/2014/02/24/more-links-for-february-2014/
Hmm, interesting (and thanks for the citations). It may be that I haven't meditated enough to notice the sleep difference from meditation (although I've always had a shaky sleep schedule anyways).
Anecdotally, what *I've* been noticing is that there is a strong influence the *other* way round: the more sleepy I am, the worse of a meditator I become (since the sleepiness itself becomes a distraction).
One thing you might try is to meditate after exercise. (Probably with a cool-down stretch in between.)
Thanks for this.
Was this supposed to be a reply to something?
Apologies it was for the link share to the previous discussion on mandatory voting - new user error on my part.
Has anyone succeeded at getting a Substack custom domain up? Mine has been stuck at "Your domain isn't ready yet. This can take up to 36 hours while the DNS records update" for the past week and a half.
1.) click on the button that says add custom domain in settings and pay the $50.
2.) Enter the name www.astralcodexten.com in the field in Substack settings.
3.) Substack should generate CNAME information and show it afterward. If it doesn't, you can contact support or read up on how to find the CNAME it generated.
4.) Go to your DNS provider (Hostgator?). Go to records.
5.) Hit add record. Select CNAME
6.) Host or name is www.
7.) Points to or target is target.substack-custom-domains.com
8.) Set time to live to the smallest amount the provider allows.
9.) Save the record. Wait the time to live amount and then see if you can get through on the url. It might be intermittent for the first two or three days but you should be able to see it sometimes getting through within an hour or two.
You can also check the records on any number of sites. Here's one:
https://mxtoolbox.com/CNAMELookup.aspx
Just type in the full url (www.astralcodex.com) and you can see all CNAME records associated with a url. Which currently shows that address is pointed to a namecheap landing page.
If the DNS is on Namecheap you would need to go there, not HostGator, since you are not hosting the site. There's a tutorial on adding Namecheap records here:
https://www.namecheap.com/support/knowledgebase/article.aspx/9646/2237/how-to-create-a-cname-record-for-your-domain/
I think did everything in 1-9 right, so the problem must be Namecheap. Is it possible for the domain name to be on Namecheap even if I get HostGator emails about it, and it's accessible from my HostGator account, and I have a HostGator account but not (AFAIK) a Namecheap account?
Possible but unlikely. Can you confirm the name you bought (astralcodexten.com) and that you bought it through Hostgator?
I have one, though it was set up about a year ago. I recall them being fairly quick about it, definitely not a week and a half.
I don't have a Substack, but I do deal with DNS a lot & happy to help. Have you added the CNAME record yet, or is it stuck before that point? (ie in this article, is it stuck at Step 1 or Step 2)
https://support.substack.com/hc/en-us/articles/360051222571-How-do-I-set-up-my-custom-domain-
Also, what is the custom domain? (I checked www.astralcodexten.com and www.astralcodexten.net, neither currently have a www CNAME record pointing at Substack).
I thought I had added the CNAME, but it sounds like you're saying it didn't go through.
I'm retracing my steps, and I've opened the HostGator portal, clicked through some control panel options, and find myself at a dialogue saying:
---
Add a CNAME Record for "astralcodexten.com"
Name: [[example.astralcodexten.com]]
CNAME: [[example.com]]
---
I previously tried to fill this in with
Name: www
CNAME: target.substack-custom-domains.com
...on the theory that it looks like www uses the main domain, and the text beginning with "target" is what Substack told me to use. Is this right? If yes, I'll try it again. What are you checking to tell you if the CNAME works?
> What are you checking to tell you if the CNAME works?
I used a command line tool (nslookup), but it's probably easier to use an online tool like https://www.whatsmydns.net/
Currently, www.astralcodexten.com is pointing at: parkingpage.namecheap.com
Note, because there's an existing www CNAME record, you shouldn't be adding a new record, you should be modifying the existing record. Come to think of it, there is almost always a www record created with a placeholder value in any new zone file. Substack's instructions should maybe talk about modifying the www record rather than creating a new one.
If the existing record doesn't point to parkingpage.namecheap.com, you're not in the right place!
> I think did everything in 1-9 right, so the problem must be Namecheap. Is it possible for the domain name to be on Namecheap even if I get HostGator emails about it, and it's accessible from my HostGator account, and I have a HostGator account but not (AFAIK) a Namecheap account?
Yes, actually!
One thing that commonly trips people up: it's possible for a Zone File for your DNS to exist in multiple places, with only one of them active. You may have for example:
- DNS hosting at the registrar (where you bought the domain, presumably Namecheap)
- DNS hosting attached to web hosting (maybe this is where Hostgator comes in?)
You set the Name Server (NS) for your domain wherever you registered the domain (presumably Namecheap).
NS for astralcodexten.com is currently: dns1.registrar-servers.com, dns2.registrar-servers.com
whois data for registrar-servers.com indicates that it is owned by Namecheap.
As far as I can tell there's no business relationship between HostGator and Namecheap (although I do have trouble keeping up with all the buyouts).
I can see the domain was registered on 9th July 2020, then Updated on 8th July 2022. I'd check your email accounts for anything from Namecheap around that time (or indeed any time).
People often cite "The food is terrible and the portions are so small" as an example of an incoherent set of complaints.
I don't see why. Unless the food is so terrible that you can't eat it anyway, this is a perfectly reasonable set of complaints. The textureless, cheeseish pasta that I got last time I flew long haul United Airlines in economy was pretty bad food, but I'd rather have eaten a larger portion anyway rather than eat the tiny portion provided and suffer hunger for the next ten hours.
It's not so much that the complaints themselves are incoherent or contradictory, but often times people make a set of complaints that don't build on each other, and sometimes those people are just venting.
For example, people will rationalize their feelings after-the-fact (https://www.lesswrong.com/posts/TGux5Fhcd7GmTfNGC/is-that-your-true-rejection). If people tend to do this, they may list several reasons that point "in the same direction" but the reasons are causally disconnected from their beliefs.
Relatedly, we can see this "direction-based" thinking in other areas, too. This post (https://www.lesswrong.com/posts/Kow8xRzpfkoY7pa69/the-affect-heuristic) mentions that when primed with the benefits of nuclear power, people will underestimate the risks. This is probably because its common to make decisions based on blobs of emotion, rather than detailed claims. We see this a lot in politics, too, with emotive thinking and one-sided policy debates.
It's true that those two claims about the food are not incoherent; nor are they contradictory. But, I would expect those claims to be caused by venting, rather than a nuanced model in someone's head.
I only know this from a Woody Allen movie as a metaphor for life. Life; not that all that good, and then you die too soon.
I remember it from a Woody Allen movie as a metaphor for sex - but maybe I just mixed it up with every other Woody Allen movie
Dissecting the truth claims from the opening of Annie Hall…this is peak Rationalist discourse 😂
Hmm sex, food and life... What else is there? Maybe science.
Presumably the people who find it incoherent simply have a more conservative definition of the word "terrible" than you do, which wouldn't include "pretty bad" airline pasta that one wants more of.
I always took that as a set of complaints; that there wasn't just one thing wrong with the place.
(1) The food is bad
(2) Even if the food was good, the portions are too small (be that "to satisfy hunger" or "for the price they charge")
There may be other complaints as well besides those two, e.g. the service is slow and/or bad; the kitchen is dirty; the staff are rude; there are long waiting times because they always mess up the reservations and so forth.
On a trip to Ireland, I stopped along the road to ask a man if there were any places close by where I could eat. He thought for a minute, and then said, “I can think of two, but the problem is no matter which one I sent you to you’d wish you’d gone to the other.”
And no one goes there to eat anyway, since they're always so busy.
Not at all incoherent. Restaurants (most products really) have two possible appeals; quality, like the high-end restaurants, or quantity, like the buffets. Small portions of low-grade food fail every bar.
It just sounds silly. Like forgetting the name of the guy from My Name Is Bruce.
An IQ of 80 or so is no big deal. People with that IQ can manage all the basics of life, and if they had teachers or parents that took special care to make sure they understood important things that are a bit challenging, like banking, birth control, basic health info, they should be fine. Of course, if you've got a bit of a eugenics project going on in your country, those people have a strike against them. As for IQ scores breaking down at the extremes -- I don't know whether they do or not. In the case of this kid, there were no obvious disabilities interfering with her functioning. I don't know what her IQ was, but certainly way below 80. Her speech was like that of someone 1/4 of her age. In the case of people who have disabilities that interfere with them understanding or taking an IQ test, there are tests that are much more practical, involving tasks like using the stove, the bathroom, riding in the car, following simple directions, answering "what should you do if" type questions that are standardized and are a reasonable picture of a person's level of functioning in common situations. It does seem likely to me that a test of this kind when someone reaches puberty would be quite a good predictor of whether they are going to be able to parent decently as adults. Whether it is right to sterilize people who do poorly on that test, I don't know, but I do see a case for it. And it's not a realistic option to explain adult life and its demands to them, and strongly urge them to choose sterilization.
But I actually think that if we were going to try to weed out anything from our gene pool, it would be much better to focus on the kind of huge empathy deficits you see in sociopaths, and on a tendency for violence.
Correlation does not imply causation. How do you know that poor people aren't more violent for the same reaosn they're poor? Why would somebody prone to aggression, poor impulse control etc be able to earn a good income?
What? No. Crime front runs poverty in neighborhoods.
I don't understand quite what you mean -- esp. since the "it" stuck in there means sentence could mean that prosperity affects violence more than it affects IQ. or that prosperity has more effect on violence than IQ does. And whichever thing you meant, where were you going with that? I wasn't talking about reducing violence by trying to raise average IQ, I was talking about reducing violence by trying to reduce reproduction by violent individuals. I have not given a lot of thought to whether I could possibly be OK with sterilization enforced by law, but if I were to be OK with anything, it would be sterilization of the violent. Commit I extremely violent act, or 2 moderately violent ones and you're sterilized. A lot easier to swallow than forced sterilization of people who are not too bright, especially when the cutoff people have in mind is an IQ or 80 or 85. Jeez, lots of people with an IQ of 85 are enjoying life, taking decent care of themselves, and working. Having an IQ of 85 is no big deal. Get over it!
Also, I was also suggesting that we try, via some sort of selective reproduction, to reduce the number of people who have very low empathy. Many of these people are not violent. They simply fail to get it that other people are sensate beings like themselves. They are cars, everyone else is traffic. I think these people are not playing with a full deck. Their mental representation of the motives and inner lives of other people are impoverished and simple. It's like being color blind, or having a learning disability -- except that the consequences for society are far worse.
Is there a reason you would draw the line at one or two violent acts, which seems absurdly low to me? I am against forced sterilization in general, to be frank, but I could at least understand the argument if you were looking at serial offenders.
Beyond that, is your goal to genetically weed out the population (I think yes)? I have significant doubts that such a plan would work, as it seems to me that violence comes from many sources and it would be unlikely to remove it from future generations. (As an aside, even if it did work I would be concerned that even though our societies would work better domestically, foreign countries that didn't neuter their aggressive sides would be much emboldened to attack and expand). If you are using it as a proxy for bad parenting and don't want to let these people raise kids, then I think that's more reasonable as an argument. I'm still against most sterilization schemes, especially where there's much human judgment involved.
For completeness sake, I will say that I'm not completely against any of your ideas here. I had a friend with a very disabled relative, a young girl about to go through puberty. This girl had no hopes of life outside of an institution, and could not ever handle even basic life functions for herself. They gave her a hysterectomy. The thought process was that she would never be able to raise a child, having a baby would be dangerous, and she could never consent to sex in any legal sense. Practically speaking, if she ever did have a baby it was likely the result of rape from someone else in the institute or an employee there. No good options. As much as that choice must have been painful, I support it. If there had been no family to approve it, would I support the government/society making that decision? I don't know. I still see the value in the reasoning, but very much struggle with giving other people the right or power to make that kind of choice.
You kinda missed the point. I'm not saying, let's do this now, or even less do this someday. my main point is that if we were to try to change how many people in our society have a certain trait, I think it makes more sense to make violence and lack of empathy the target, rather than lowish intelligence.
As for drawing the line at one or 2 violent acts. First of all, as I said, I was not sketching in a finished proposal, just giving a general idea of something we might do. But I believe I did not say "one or 2 violent acts." Most people commit one or 2 violent acts a week, if you count little stuff like yelling curses at the ref while watching sorts on tv. I believe what I said was one very violent act or a repeat offender where both acts are moderately violent. By very violent I had in mind things like murder, attempted murder, beatings that put someone in the hospital. By moderately violent I meant things like beatings that leave the victim with multiple visible injuries, shoving people down the subway stairs. Yes of course there are all kinds of judgment calls involved and the discrimination will be made imperfectly.
Our legal system is not clean enough for me to be comfortable with a one or two strikes scheme, even if the crime seems significant. A DUI for a young adult could easily fall into that category, where we sterilize them even if we don't consider it intentional. And it's hard to prove intent. And an upset society has a tendency to impute intent (like saying that the intent to get drunk and then drive was intent to kill).
More to your original point, I also tend to think that different people with different personalities is good for society. Sociopaths can make hard but necessary decisions, like in wartime or downsizing a failing company, that someone with a lot of empathy could never make no matter how disastrous.
And again the problem of who gets to decide what traits to weed out and what test is used. I could use my personal sense of "violent person" and sterilize some percentage of the population, while another person who could also be thoughtful and intelligent might pick a much higher or lower percent based on their own sense of "violent person."
You are still missing the point. Can we please stop now?
I think sterilizing anybody for 'societal benefits' is pointless and does more harm than good. There's no proof that violent people are violent because of genetic reasons. There was a footnote in Scott's post 'Galton, Ehrlich, Buck' that went :
"the Nazis killed nearly ever[y] schizophrenic in Germany, hoping to eliminate “the schizophrenia gene”. Today, Germany has exactly as many schizophrenics as any other country, because there are thousands of genes involved in schizophrenia, and all the deleterious variants are present in some frequency in the healthy population."
If you sterlized everybody who committed a violent act, for one thing, sterilized or not, they'll continue to live and if they're prone to bouts of violence, people around them continue to suffer. In addition to that, they'll probably get a lawsuit (and a payout) in their favor for having their right to procreate taken away.
For another, such a flippant proposal can only be made today on violent men. Would you support sterilizing a woman that punched or beat up somebody?
Right now, human knowledge of genetics is rather embryonic. Consequently, any attempt to improve the 'gene pool' is likely to end in failure and/or disaster. Its futile and unrealistic- like frogs trying to dam a river to get a cosier hangout spot, or something.
There are plenty of horribly violent women, and yes, I think if we were sterilizing people with a history of serious violence we should sterilize both genders. You seemed not to have noticed that I am not proposing that we go ahead and do this, so no need to tell me that it would be difficult to do in practice, thanks. My main point was that if we were going to try to eliminate a trait, I think it would make more sense to target propensity for violence and absence of empathy.
Is there evidence that very low empathy levels are genetic, or mostly genetic? 'Cause I reckon a lot of people might be like that due to the way they were raised, childhood traumas, etc. that make them stop caring about anybody but themselves... Has an "empathy gene" been identified for which your eugenics scheme could feasibly put into practice?
Well of course environment has an impact. But there seems to be decent evidence the genetic component is not small.
https://www.sciencedirect.com/science/article/abs/pii/S0149763419306505
And people are normally born empathic. Babies and toddlers seeing other babies and toddlers cry often cry themselves. Don't know whether anyone has looked at the kids who do *not* how this response and see if they differ later in life.
Yes, and that shouldn't be a surprise, seeing how all personality traits are quite highly heritable: "Results indicated that 56% of the variance in antisocial personality and behavior can be explained through genetic influences, with 11% due to shared non-genetic influences, and 31% due to unique non-genetic influences." (Christopher J. Ferguson , The Journal of Social Psychology, 150:2, p. 160-180 (2010) , DOI: 10.1080/00224540903366503 ).
Also, "the way you were raised" strongly depends on whether your parents are sociopaths, and anecdotal evidence indicates that it's not the children that suffer the most trauma in these families that turn out sociopaths, it's their "golden child" siblings.
I'm aware that IQ and violence are correlated, but how much of the variance in violence is accounted for by IQ? (Or you could give an answer in terms of correlation between the 2.)
https://twitter.com/cremieuxrecueil/status/1649190612833390592
https://twitter.com/cremieuxrecueil/status/1645697957205876736
I'm thinking of switching careers and becoming a translator. Do people here think that's a viable long-term career path, or will AI render it obsolete within the next few years?
Yglesias recently published a good high level article by Tim Lee on this subject: https://www.slowboring.com/p/how-human-translators-are-coping.
I’ve been doing some freelance translation as a part time job over the past few years. Machine translation is pretty good already. I think AI will obliterate the low end stuff pretty soon. The high end work should last longer, but you’ll be working with a machine, not instead of one. My best guess is that there’s about ten years at least left in the high end stuff, maybe more. A lot depends on your language combination, e.g. Mandarin to English would be in demand, but also certain minority languages into English. The other thing that would help is if you have knowledge of a certain sector e.g. if you have a medical degree you could be well placed to translate medical articles. So if you can find a niche, there’s still plenty of work, at least for now.
I have worked as both a translator and interpreter (unofficial), but was more than 15 years ago. Programmes to assist translation (Trados I think was the most popular one) seemed to be just starting back then, but I still got lots of work and made a decent living without using Trados. Nowadays I would say if you can speak a couple of different languages fluently, you are better off to get into international business (sales or quality department in a multi-national for example). In Europe, being really fluent in a few languages gets you opportunties everywhere. (In the US no, because eveyone is just expected to adapt to English by default)
Translating legal documents, technical documents or just general texts already means you'll get a machine translation and be asked to fix it up quickly and cheaply.
Translating literary texts like novels or films will probably not be automated, but like any creative job there are lots of people who want to do it so opportunities and pay are pretty bad. And tasks that used to be entry level jobs in that field are also being taken over by machines.
Even live translations of speech are getting pretty good, so I think even being an interpreter has a limited career potential at this point, except possibly sign language interpreters.
To some degree it depends on what language you'll be translating, but if it's "any big language to English" there's probably a lot of automation and if it's "niche language to English" there's probably very low demand.
Respectfully, that is an absolutely terrible career path right now. Someone close to me recently graduated college with a degree that basically meant her only option was translation; it has been a huge effort for her to retrain and do something else. The writing is on the wall for translation. It’s not that AI will render it obsolete within a few years; for 90% of use cases, it already has.
Exceptions include technical translation and being an interpreter... But those will be phased out as well. Sorry to be the bearer of bad news.
I recently used GPT-4, to translate relatively niche technical text - and it did a decent, though not perfect job. Currently, I think just using GPT out of the box would make a technical translater much more efficient - basically needing to check the results and make some minor changes to specific terms.
My friends who are translators are all already bemoaning how much of their work is becoming "machine-assisted translation" (aka starting with AI translations and then proofreading/finetuning them).
If you are also an interpreter, that seems much safer (for now, at least).
I do think this is one of the areas most likely to be impacted by AI advances in the near future.
Does anyone have some experiences on how lively is gopherspace or geminispace really? Do they have a future or is it just a retro fad?
I don't think it has a future in terms of widespread adoption. But it's fun, and there's no reason that enthusiasts can't keep using it indefinitely. It really depends on what you want to get out of it.
Gemini reminds me of Usenet and similar parts of the early social Internet: mostly hackers writing about hacking for other hackers to read, with a sprinkling of other eccentric personal interests. A lot of the content is about the medium itself ("Hey, I wrote a new Gemini server in Common Lisp; check it out!") Epilw don't post pictures of themselves (though they could); instead they talk about ideas. Normies have never heard of the medium, and, even if they did, most wouldn't get past the modest technical barriers to entry (I know this is the opposite of what Gemini's creator wanted, but ¯\_(ツ)_/¯ )
Whether all this is good or bad depends on your tastes! I find it a refreshing break from modern social media. I especially appreciate the ethos of people being weird without either shame or ostentation, a gestalt that does feel intensely retro to me in the best possible way.
Thank you for the answer. I find the protocols refreshingly simple, unlike modern web when browsers are prohibitively complex and buggy, while I can write a new gopher or gemini client as a programming exercise. While I appreciate advantages in having a single platform, I believe that placing resource exploration, presentation of information and app platform in a single basket was a bad tradeoff, though perhaps unavoidable one. But I still miss multiple specialized tools and protocols of old internet, and I think that monocultures are bad.
(If this is inappropriate for the thread, feel free to delete)
I wonder how hard it is to predict who will marry interracially.
I ran into an old acquaintance the other day and was surprised to find out he married a black woman. Not surprised because he was a raging racist or something. It's just that in our social circles it's still a bit of an unusual thing to do. Looking back, there was nothing about this guy that said "yes, this is precisely the kind of dude who would marry a black woman."
Maybe the big ML models used by dating services are able to identify people more likely to "marry out" than the average user. Either way, it's clear some humans do, and some absolutely never would. What's behind it? Is it genetic? A particular upbringing?
I'm not even sure why I'm so interested in this question, but I am.
I think this is mostly only a meaningful question (in the US) for white people, maybe hispanics.
https://www.pewresearch.org/social-trends/2017/05/18/1-trends-and-patterns-in-intermarriage/
I mean, I think the limiting factor for e.g. black male white female pairings is the number of white women interested in such a pairing.
I predict if you're not racist it's just a function of how many people of other races you meet / have in your social circle. I think that's pretty small for a lot of people, but probably depends on where you live and what your interests are.
I dated an Asian person for a while and I don't think I have any more or less interest in dating Asians than anyone else, she was just someone I happened to meet in the Bay Area rationalist community. I agree black-white relationships are less common, but I think this is mostly because of social/class de facto segregation.
>I agree black-white relationships are less common, but I think this is mostly because of social/class de facto segregation.
Blacks and whites looks very different, behave very differently/have different cultures, and have different intelligence distributions. It would be weird if there weren't substantial preferences for or against one race or the other.
'De facto segregation' makes it sound like its a matter of not being in contact with each other that's stopping the relationships. But the idea that a white woman with a masters degree would date a poor blue collar black man from the hood, or a hood black girl wants a dorky upper middle class white nerd, is just silly. These populations are very different, and on average this should have huge impacts on dating preferences.
In the USA it's just becoming normalized.
See the second chart here:
https://www.pewresearch.org/short-reads/2021/02/25/in-vice-president-kamala-harris-we-can-see-how-america-has-changed/
From 3 percent of all newlyweds being in biracial marriages as of 1967, to 19 percent as of 2019. That's a slow but steady social shift persisting for several decades across this huge society.
And the biggest recent shifts within that are among whites and blacks, according to this previous Pew report based on data through 2015:
https://www.pewresearch.org/social-trends/2017/05/18/1-trends-and-patterns-in-intermarriage/
>And the biggest recent shifts within that are among whites and blacks
To be clear, the trend for whites and trend for blacks have both increased, not the black/white combination. The most common interracial marriage type (42%) is between whites and hispanics, and many of these 'hispanics' are of majority european ancestry.
What is your source for that? And, what do you mean by "many"?
"From 3 percent of all newlyweds being in biracial marriages as of 1967, to 19 percent as of 2019."
That overestimates it somewhat, a Spaniard from Argentina and an Italian from Staten Island count as an "intermarriage."
The point is the trend line, not the absolute percentage.
Are you saying that your example was not being counted in 1967 and is being counted now?
I would guess its correlated with big 5 openness - which would also make it correlated with being liberal, creative, having a high IQ, and low authoritarianism.
My guess is that there's a pretty large element of randomness in the form of "you meet who you meet and the heart wants what it wants", and given that probably the single biggest factor is what fraction of the eligible people in marriage range (usually some combination of proximity in location, age, socioeconomic status, and values) are the same race as the person in question - people who live and meet their spouse in a 99-one-race town are probably marrying that race.
You surely already know this which is why you're bringing it up here in a pretty nonjudgemental pseudonymous forum, but I'll say nonetheless: I'd be cautious who you discuss this curiosity with, it's a culture-war landmine territory.
There is also (at the risk of being controversial and surely offending someone) a "lack of better options" element. My social universe is almost exclusively white and Asian, and the typical interracial pairing would be an attractive Asian woman with a less-attractive white man.
You don't expect the most attractive and highest status white men to wind up marrying interracially, although highly attractive Asian women often will.
Exactly.
I recently ran into someone claiming that the idea of SpaceX's starship is 'utopian nonsense'. I gave them the usual spiel about lowering $/kg to orbit making the industrialization of space feasible, but I was wondering if anyone had some more serious sources about the tech and economics involved. The real minutia of everything from water reclamation to cost studies on lunar mining, etc.
There are a couple of books on the economics of space mining etc (Mining the Sky by John Lewis comes to mind), but I've found little serious analysis of what could be done with Starship. Historically I am not sure we've done a great job at foreseeing what can be done: I've see serious analysis from two decades ago saying that reusable rockets will never make economic sense, for example.
The main argument for Starship seems to be "built it and they will come", which is certainly possible. But equally, working inside the space industry, I can't say I've seen that many serious people start building projects around Starship. That could change once Starship is proven, but the economic argument for space has always been tricky outside of a few use cases (imagery and communications principally).
That's a tough one. There are serious sources on much of this, but they're highly technical and even the broad survey works are written for people with an engineering background. And there were some good popularizations in the 1970s and 1980s, starting with O'Neill's "The High Frontier", see also T.A. Heppenheimer and G. Harry Stine, but they're rather dated and probably antipersuasive to someone coming in to them cold in the 21st century. I don't think I've seen anybody do a decent update on the concept.
Zach Weinersmith of "SMBC" fame and his wife Kelly are finishing "A City on Mars: Can We Settle Space, Should We Settle Space, and Have We Thought This Through", which might be what you need (and I expect it will have references for more technical detail on anything you want), but it's not out until November.
Have you read _Mining the Sky_ by John S. Lewis, by any chance? If so, what did you think of it? (It's probably the most serious book on space industrialization I've happened to read, and I wonder whether I got in near the top of the literature for laypeople, if there's much higher I could go, or what. And, knowing what I know of your background, you'll find most of it old hat and so you could probably skim it pretty fast if you had to.)
Yes; I have a copy and refer to it frequently. It's pretty good for what it is, but what it is is pretty narrow. We need something that simultaneously makes the case for "here are the valuable resources and opportunities waiting for us" and "here's how to set up economically viable infrastructure for space industrialization". If all you do is tell people that there are asteroids with infinite free platinum or whatever, they'll correctly point out that it would cost us a couple billion dollars to send out an SLS/Orion combo to bring back a few hundred kilograms, once.
Oh my god, in my career as a software developer, I have seen computers do some completely demented shit, but I just ran across something that really takes the cake. My webcam had randomly turned on, and I tracked down that the application that was doing it was something called WavesSysSvc. It's an audio application that seems to have installed itself along with an audio driver update, and it turns on the camera to track your head so that if you turn it while wearing headphones, it adjusts the volume so that it sounds like the audio is coming from the direction of the computer.
This is by far the most overengineered thing I have ever seen, and I can only assume someone somewhere was gunning very extremely hard for a promotion or perhaps it was to avoid being fired, but either way, yikes!
Which stories do you have about the craziest thing a computer has ever done to you?
Would you mind sharing the name of the audio device manufacturer, so that I can make a note to never buy anything from them ever again ?
It was Realtek, though the software is made by a different company called Waves.
Apple AirPods can do the same thing but don’t use the webcam, just the gyroscopes inside. Using the webcam is the cowards way out! (I almost always turn off this feature but it is kind of cool when watching sports)
That's horrifying.
tip of my tongue: there was a blog I found either through Scott's blogroll or through a 2nd order blogroll that was seemingly exclusively about 9/11 and its theological implications
I remember specifically a post that opened up with "on some day God decided to crash Himself into Himself" and it went on to make an unconventional point about how if God doesnt exist then everything is forbidden (as opposed to the more traditional take in which everything is permitted)
Can't find it for the life of me, anyone knows what I am talking about?
No idea about the blog, but that's a common idea in continental philosophy. Dostoevsky's (paraphrased) "if there is no God then everything is permitted" (i.e. without a Lawgiver there are no laws) was turned on its head by Lacan into "if there is no God, then everything is prohibited" (i.e. without a set of laws we can choose to follow or defy, it's impossible to do anything of any meaning), and more recently Zizek built an idea around "if there is a God, then anything is permitted" (i.e. ideology functions as a way to channel and justify the basest of desires).
If it was a philosophical blog you could try searching with some of those names.
Here is the Zizek article: https://www.abc.net.au/religion/if-there-is-a-god-then-anything-is-permitted/10100616
Good idea, I'll try to refine my search using those philosophers/writers. Thanks!
Suppose I have relevant technical expertise and I believe Manifolds odds on the room temperature SC claims being for real are an order of magnitude too high. How much can I stand to make by betting on this? I am not registered with manifold or any other prediction market and if the potential payoff is less than four figures it is not worth the effort. Fwiw I am in the US.
Best real-money option for Americans is Kalshi at https://kalshi.com/markets/supercon/roomtemp-superconductor , but it's not really loading for me right now and can be kind of hard to get money into. It worked for me earlier.
thanks, I'll look into this.
Update: I did sign up for Manifold and bet half my portfolio on `no' but (a) that's only play money and (b) the setup doesn't seem to discriminate between `thinks the market is 10% overoptimistic' and `thinks the market is 1000% overoptimistic.'
The setup does distinguish this in that your bets move the market and you want to keep betting until the market hits where you want. This just doesn't do much for the superconductor market because there's a massive amount of betting on it and you don't have enough mana to move the market.
I think your unwillingness to bet real money might be a sign that subconsciously, you aren't really that confident in your prediction.
Au contraire, I am willing to bet real money, and am asking if there is a way to do it.
Conversely, if there isn't a way to make real money off prediction markets being overoptimistic on this topic, then there isn't a good reason to expect prediction markets to be accurate.
Edit: If by `real money' you mean instead something like `sink my entire (liquid?) net worth into this' then no, I am not willing to do that, and wouldn't even if the claim was something proven impossible (like perpetual motion) because counterparty risk etc would be unacceptably high.
All you've been doing so far is making excuses about how it isn't worth it.
I am claiming that these excuses are not legitimate, and are actually caused by your lack of confidence. Prove me wrong. Go bet some money on it.
@Hi, are you willing to bet against @Humphrey at mutually-agreeable odds, given @Humphrey’s level of confidence?
Manifold is play money only. Polymarket is illegal for me to bet on. Where do I go to prove you wrong?
How/where?
If you bet on polymarket you can make 10k on a 50k no bet (last I looked)
Also, it seems like poly market is view only in the us?
That’s more than I readily have at my disposal to bet. I could bet a few thousand…but then the payoff is too small and too distant.
You can still sign up on Manifold to bet NO on the relevant markets, though!
(Also if you make a comment that explains your object-level reasoning for disbelieving in the superconductivity thing, that would be super helpful.)
How much money could I make if I were right? Let’s say I’m willing to bet a couple of thousand (but am not willing to break any laws).
I think that Manifold uses the formula (amount of NO you pay)/(probability of NO according to the current market) to decide your payout if you're directionally correct (though you lose your entire bet if you're directionally wrong), so given that the YES probabilities for the highest-volume LK99 markets are at ~20% you can expect a ~25% return if you're directionally correct. I also think that Manifold has been explicitly preventing users from buying too much mana at once to prevent situations where users buy too much mana to go all-in and then fall into financial ruin when they lose, such as like what happened with the Whales vs Minnows incident earlier this year. Putting this together if you go with the highest option and buy 100,000 mana (for US$1,000), you can hope to get a ~25,000 mana return which you can then realize as a ~US$250 profit for your favorite charity.
However if you *really* want to make real money you should try Kalshi (or become an expat and try Polymarket).
Oh also I suggest that you do calibration training first so that you feel more certain that the market's an order of magnitude too high. The binary calibration tests out there are kinda sucky (the most user-friendly one I've found is at http://www.2pih.com/caltest/ ), but there is a user-friendly multiple-choice calibration test at https://bayes-up.web.app which may be equally helpful.
0 dollars. Manifold is play money only right now AFAIK.
Technical question- does anyone have a preferred search engine for finding matching/plagiarized content, that has an API or can be used programmatically? I have a specific business case for tracking down and finding plagiarized content online. I'm not actually interested in academic plagiarism per se, but I have a financial need to match up copied text.
I've been using Copyscape to-date, which is basically 'OK' and is clearly using Google's API. Just was curious if anyone knows of better programmatic alternatives
http://lateralscience.blogspot.com/2014/10/scottish-nitroglycerin-one-legged-stools.html?m=1
I found this ~1900 account of a NTG factory amazing. Hat tip to Ethan Mollick
One big problem with prediction markets -- and I think the current markets about replicating the superconductivity experiment is Exhibit A here -- is that you can't distinguish the predictors from the gamblers. You may think there is no meaningful difference between them, but walk into any casino and look around: those are gamblers. When the crowd at the craps table all bets on Pass it doesn't mean Pass is likely to happen. Pass is never likely to happen. Yet it is a rare day in Vegas when the crowds at the craps tables don't bet on Pass.
This is also true of the stock market - Wall Street vs. Robinhood. But Wall Street knows they're better than Robinhood and pays Robinhood lots of money to get priority access to the right to take the other side of Robinhooders' bets. There is no amount you can bet on Robinhood that some Wall Street firm won't be delighted to bring back down to the Wall-Street-approved market price. If enough criteria are met (size, liquidity, attention, etc) this should be true for prediction markets too.
I'd suggest that the criteria should include `critical mass of participants have expertise in evaluating the object level question.' For the stock market the object level question is `will this firm make enough profit to justify the price' (or maybe `will the price go up'), which is a question that I think a lot of people on Wall Street do have a lot of expertise evaluating. For `will LK99 turn out to be a room temperature superconductor' I doubt there is any such critical mass of `smart money' trading on the relevant markets.
On Wall Street, the "critical mass" can be one firm with a lot of money that hires an expert, feels confident in its decision, and then deploys millions of dollars to correct the mispricing. It doesn't matter how few people there are who know the correct price for some security as long as one of them works for Goldman Sachs and another one for Jane Street.
Ok.
Is there any analog of this in prediction markets on eg whether LK99 is a room temperature superconductor?
I don't know. I think on Manifold, no, because there's nobody who has too much more mana than anyone else. On Polymarket and Kalshi, in theory maybe, but I don't know about them.
Markets are aggregating per dollar not per person. In a deep, mature market like publicly-traded stocks, the gamblers' money will be a rounding error compared to that belonging to institutional investors.
Prediction markets (except perhaps the bookies?) are not there yet.
What do you think the odds should be? I haven't been following the story as closely as I would like, but it seems like there are multiple replications and a video now, yet the prediction odds haven't moved much.
I wouldn't put it higher than 1%. Probably less.
1% is probably way too low for it being technically a superconductor. This has passed most of the obvious-crackpottery-or-hoax gates, and I don't see any real point in trying to run a serious hoax on this. Still possible to be a false alarm, those do happen in science, but not at the 99+% level.
However, 1% for being a *useful* superconductor, might not be too far off. I haven't seen anyone even guess at a critical field strength for this stuff, and the expected behavior of superconductors is that critical field strength decreases with temperature.
And that's the part of this that I find most frustrating. Everyone is talking about how this is either a pathetic fantasy or it's the miracle that gives us fusion energy and quantum superdupercomputers and everything else under the sun. Sometimes there's a footnote to the effect that we need to pin down the critical field strength (and production yield, material properties, etc), but with the implication that this is just dotting the i's and crossing the t's on the incipient technological miracle.
20% chance this is technically a superconducter, 1% that it is a useful engineering material, 19% just a stunt that can only ever barely levitate a snowflake, might be a good guess.
20% for room temperature ambient pressure superconductivity is IMO way too high. I'd say 5% its even `barely technically' a room temperature ambient pressure superconductor, and maybe 0.5% it's actually useful. I started off higher, but enough time has passed that the lack of a replication by a reputable source is starting to smell fishy, as is the (apparent) lack of `give the damn sample to an established lab to measure.'
Why so low?
I'd say less than 1% as well. There are too many priors of purported room temp superconductors that never panned out (at least ten?). It seems like there's a much higher probability that scientists involved are deluding themselves than an actual HTSC material has been discovered.
This fully precludes you from detecting the real thing, no? An outside view is useful at times, but if your heuristic doesn't include any information about the (present) claim, then you can't actually update until after the fact.
How does it preclude detecting the real thing? Exclusive of my cynical worldview, my heuristic includes the following information and suppositions.
Firstly, AFAIK, there's no theoretical reason that we know of that precludes a room-temperature low-atmospheric-pressure superconductor from existing. But likewise, we don’t have a theory that will allow us to make one by design. So a true RTSC would be coming out of theoretical left field. Shit like this does occasionally happen in science, but with no theoretical framework to explain and/or predict the behavior of an RTSC, I'd say that that reason alone would be strongly indicative of the null hypothesis.
Although I have seen some chatter that LK-99 uses a Phonon/BSC mechanism—I thought that wasn't possible, because (my understanding is that) BCS requires a Fermi surface where the electrons can form Cooper pairs. Correct me if I'm wrong but isn't that situation impossible when the lattice of the superconducting material is vibrating too fast (i.e. too hot) for those pairs to form? I admit I only have a layperson's understanding from what I've read in generalist science magazines such as Nature, Science, and Scientific American. So I'm perfectly willing to be corrected on this.
Even if a such an animal as RTSC exists, detection/confirmation may be more difficult than some seem to think. According to a physicist friend of mine, we'd not only have to confirm the Meissner effect but we'd have to measure about half-dozen other behaviors to confirm that it's an actual RTSC. That's not going to happen overnight.
If there is "information about the (present) claim," that would update and/or correct my knowledge, I'd like to hear it. But, no, I don't see how we'd be able to "update until after the fact" (If I understand your meaning correctly).
Also: if you are dealing with a specific technical question (like the room temperature SC business) and people with relevant technical expertise are not widely participating in the betting market then the market is going to suck. Agree manifolds odds seem to me to be at least an order of magnitude too high.
Gamblers provide free money for the predictors to eat up, thus incentivizing prediction. Without real money on the line, this breaks down because predictors aren’t actually incentivized by anything other than fake internet points.
Right. Like in the stock market, which is a well-functioning prediction market, you have smart money and you have dumb money. Because the returns to stock trading and investing are so high, the smart money dominates -- because equities in general tend to be smart investments. But in a casino you have craps tables, roulette tables, slot machines. All of those things have expected negative returns on every bet and they attract nothing but dumb money.
Metaculus and polymarket aren't currently attracting enough smart money for a wisdom of crowds effect to dominate. I'm not saying the participants in these market aren't "smart" in the normal sense of the word, but their bets aren't smart. The superconductivity bet is trading like a meme stock.
> Metaculus and polymarket aren't currently attracting enough smart money for a wisdom of crowds effect to dominate. I'm not saying the participants in these market aren't "smart" in the normal sense of the word, but their bets aren't smart. The superconductivity bet is trading like a meme stock.
Why do you believe this is so?
Because 28% odds of having found a room temperature superconductor is insane. So far we have one group claiming they have found The Holy Grail. And the market thinks there is a 1 in 4 chance of that being true?
If you're very sure of this, presumably you've made a considerable bet on it? That's a lot of free money lying around if you're right. (It would be quite strange if nobody took it, wouldn't it?)
How much free money, in actual US dollars for a US resident citizen, and how many hours of my time do I have to sink into making it actually happen?
Is there? I’m in the us and think the prediction markets are over optimistic by an order of magnitude. I’m willing to bet a couple of thousand dollars if there is a reasonable payoff for being right. How do I do this?
If it were trading on the CME or the NYSE I would make a bet. polymarket is very new to me and about the only thing I know about it is that it's illegal and I'd have to use a VPN and put real money into some crypto "stable coin" I don't know anything about. There's lots more risk here than simply getting the bet directionally correct.
There have been other replications too though.
Link?
Question for the botanists in the crowd, with a short foreword. One of my local farms is getting rid of their strawberry plants in one patch, so I brought four of them home yesterday afternoon. I'm planning on planting them tonight in pots (I'm an apartment dweller, so no place for an in-the-earth garden). They're Flavorfests, which I've read ask for acidic, loamy, and/or sandy soil.
When a plant needs an acidic soil, is that asking for soil with fatty acids in it? Should I be supplementing my plants with, I dunno, fish oil capsules every so often?
"Acidic" commonly refers to the pH of a liquid; pH means "potential of hydrogen". Liquids are either acidic or basic (so wine, for example, is chemically an acid [pH 4] although that's not how we normally think about it). We also give soils a pH because of all of the hydrogen-containing compounds that are naturally/easily available in the soil. Neutral pH (eg water) is 7, acidic soil that plants like would be around 5-6.5. You can buy fertilizer specifically for acid-loving plants: https://www.fafard.ca/products/fertilizer/garden-fertilizer/natural-fertilizer-for-acid-loving-plants-berries/
Nothing to do with fatty acids, then. Gotcha. Thanks!
I'll check for it, thank you!
Hello, I wanted to ask if people think it will help me later If I take one math class now during my bachelors. So I'm currently doing my bachelors in economics and this semester I can choose one subject from a different area in addition to my normal econ classes. I can take anything from chemistry, math, languages, philosophy, asian politics etc. I am tempted to take a math class because I find mathematics quite interesting but it would probably be a very hard class for me. Whereas if I took philosophy, history or basically anything that isn't maths, I would expect it to be easier.
My main question here is if people think it will matter at all for my later life if I take this math class. Can I write "Has Economics bachelors with one analysis class" on my CV and it will make me look good. Or should I just do something easy so I have a higher chances of passing all my exams. And if I want to learn maths I can just do it in my free time? Here in Germany we mostly only take classes related to our Major only two out of about 24 classes will be non Econ for me.
If anything, your resume will probably be slightly worse if you take a math class because math professors actually give grades that reflect performance, while you can take any social science 101 class and put in minimal effort for an 'A'
I would recommend taking at least linear algebra and statistics (to be fair, these are two classes). I don't think you'd be able to avoid using them in your professional life.
Yeah, my question was poorly worded. As part of the normal Econ degree I already had statistics and linear algebra and calculus. And I was wondering if in addition to that I should take a class from the Maths bachelors which would probably be a lot more advanced.
It depends on the class; tensor calculus would be somewhat useful if you plan on branching out into data science. Beyound that, though, I'd suggest taking the hardest class you can afford, schedule-wise. You're paying for it, and no one will probably read your courseload on your resume anyway -- so why not actually learn something ?
Answering the actual question: Can I write "Has Economics bachelors with one analysis class" on my CV and it will make me look good?
No, no-one gives a shit after you graduate.
Answering the implied question: Will taking this class help me afterwards?
If taking this class will improve your math skills (safe bet): yes, definitely.
If you have enough discipline/interest to learn this by yourself later (not likely, but you know yourself more than a random internet person): don't take it, spend that time whatever is more productive
I was suspecting that this would be the case (that it doesn't matter for my CV). I think I really am interested in math enough that I could learn a lot of what would be covered in class on my own, but it would just feel like such a wasted opportunity not taking a math class in University because it would probably be difficult, and then later learning math on my own in my free time.
But if I am being rational it's probably unlikely I would really study that much math on my own.
Thanks for the response.
What are your plans for later in life? Depending on what your undergraduate economics classes are like, I do think analysis (i.e. Real Analysis) would help you if you ever want to apply to US Econ PhD programs.
I would guess there is a 40% chance that I will do a masters and maybe 20% PHD. I would prefer to either work, start my own business or do something EA related and not do a masters.
I'd say take a look at the admission requirements for any grad schools you may be interested in, if the extra math course would help then go for it. Here in Canada/US, having a good/decent grade in Real Analysis is very beneficial for Econ graduate admission. Besides that, I doubt one course will make a huge difference, so I would just take whatever you're most interested in.
I'm confused how you could complete an econ degree without taking any math classes. You can't really understand a lot of economics without a solid foundation in probability, statistics, multivariate calculus and differential equations and these aren't topics a typical student would fully cover in high school.
Yeah, my question was probably poorly worded. Just during the normal econ degree we have to take two maths and stats classes covering all the things you just mentioned. Just in addition to that I have the option to take a class from the actual mathematicians where we would probably have much more advanced maths and many more proofs etc.
I'm surprised that noone has talked about the end of the Salem/CSPI prediction market contest at all. Maybe it's because the CSPI itself hasn't made any announcements about it either? The whole thing is really weird.
It came out the day before my last Mantic Monday post and I didn't want to write lots of stuff in a hurry. I'll cover it next time. Congratulations to everyone involved.
What makes it weird?
I would have expected them to announce the end of the contest, post to Substack and Twitter, and generally try to raise publicity around it, similar to the way they did when the contest first started last year.
I think CSPI is focused on… other things right now.
There was a secret numerical system developed by monks in the 14th century. https://en.wikipedia.org/wiki/Cistercian_numerals
Interesting, but this doesn't seem particularly user-friendly for doing arithmetic (which I think is why the Hindu-Arabic system, by simplifying each number's notation into a list of digits, really took off, though the other ones never did).
Yeah, having played around with it that does seem to be the primary issue. You could image ways to expand it to include numbers larger than four digits pretty easily, but adding and subtracting with it looks difficult.
Never heard of it before, fascinating, thank you for sharing!
Thanks for sharing. As someone who likes to doodle new number systems during boring meetings, this one is similar to one I came up with but far superior. I love it, and will devote my next doodling time to mastering it.
The reason for libertarians and conservatives to read leftists is posts like this one from Freddie: https://open.substack.com/pub/freddiedeboer/p/prologue-to-an-anti-affirmation-movement?r=vwgau. As an anarchist with no hope of living under my kind of system, I’d vastly prefer Freddie’s government over a progressive government.
I hate to pick on one sentence out of a whole essay that I mostly agree with, but I feel compelled to pick on this one because I think it shows a huge problem with Freddie's world view:
"These pieces don’t stem from a conservative desire to stigmatize or a lack of sympathy for hurt people"
A "conservative desire to stigmatize"? Is that really how he understands conservatives' thought processes, that we get up in the morning and say "ha ha, I feel like stigmatising someone today"?
No, when conservatives stigmatise things it's because we believe that they're objectively bad and shouldn't be celebrated, and that tolerating or indulging objectively behaviour (like drug addiction or crime or obesity) isn't doing anyone any favours. It comes, in fact, from exactly the same set of desires that Freddie is exhibiting in this essay, it's just that he thinks he's doing it from compassion while conservatives apparently do it from a desire to be evil.
I suggest that Freddie deep down _is_ a conservative, he just can't bring himself to admit it because he doesn't understand conservatism.
> Is that really how he understands conservatives' thought processes, that we get up in the morning and say "ha ha, I feel like stigmatising someone today"?
> No, when conservatives stigmatise things it's because we believe that they're objectively bad and shouldn't be celebrated, and that tolerating or indulging objectively behaviour (like drug addiction or crime or obesity) isn't doing anyone any favours.
Desires do not have to be intentional. The idea, I think, is that people have unconscious biases, and the desire to stigmatize is one of them. Some people has less of it, some more, some people try to correct for it consciously, some just do not notice it and some may even identify with this desire to some degree.
Concervatives tend to have stronger and unreflected desire to stygmatize. This result in them thinking some things are objectively bad and being more trusting to the arguments in favour of this position. And if these arguments are disproven, this doesn't really shake the confidence because the arguments are just rationalization of this unconscious desire and not the actual reason why concervatives hold the positions they do.
> Concervatives tend to have stronger and unreflected desire to stygmatize.
I'm not sure this conclusion is justified as described. For instance, conservatives also have considerably stronger disgust sensitivity. What does anyone do when presented with something they find disgusting? They stigmatize it. Therefore the propensity to stigmatize could follow from disgust sensitivity, not some innately "stronger and unreflected desire to stygmatize".
The fact that stronger disgust sensitivity leads to stronger urge to stigmatize and conservatives tend to have stronger disgust sensitivity supports the point that conservatives have stronger urge to stymatize.
You're missing the point: the original claim was that it was stronger and *unreflected* desired to stigmatize. This is not the case.
There seem to be some misunderstanding going on. Maybe I used the word "unreflected" inappropriately?
The idea was that you can look you desire to stigmatize and thus figure out a way to decrese it's power over you. So, for instance if the reason you feel the urge to stymatize people is because you feel disgust towards them, instead of following the disgust based urge you can think: "The fact that I find X aesthetically inappealing doesn't necessary mean that it is indeed bad and deserve stygmatization, I should be extra careful, while evaluating X and maybe even give it extra good faith, due to this cognitive bias of mine".
And if you do not do it - then it's what I mean by "unreflected", it has more power over you and indeed leads you to more stygmatization. Stronger disgust leads to stronger desire to stygmatize which is also harder to reflect upon.
That sounds more like his depiction of the excesses of therapeutic culture; see also:
> Good people spend a great deal of their time categorically and uncritically affirming others - telling friends and strangers alike that their desires are all legitimate, their instincts always correct, their perceptions of their own needs never mistaken or misguided, their self-conception compelling
> Correspondingly, we should all assume that anyone who is not affirming us is necessarily doing so out of a particular kind of politicized wickedness, that they are likely motivated by racism, sexism, homophobia, or other kinds of bigotry, and if these specific accusations are not plausible, then by simple evil
In a vacuum I'm just reading that as recognizing that's a thing liberals like to claim. It's the audience's position, not the writer's; I say something like "this isn't gnomes stealing your keys when your back is turned", that doesn't mean I actually think gnomes are doing that in other situations.
In context I still read it that way; you can't use the "conservatives evil" claim to dismiss them offhand because these aren't even conservative sources. Like when America's most decorated soldier started talking about PTSD and people could no longer use the "they're just cowards" dismissal.
My problem is Freddie's analysis is his generalizations about what liberals believe or claim. Conservatives like to make straw men out specific liberal positions, and although Freddie declares himself to be a Leftie, he sounds more like George Will in Leftie drag.
IMHO, nowadays, liberals have a wider diversity of views than the right.
> It's the audience's position, not the writer's
Not if I'm the audience. And if he's saying something he doesn't really believe to pander to a different audience, he's even more likely to say stuff that he doesn't really believe about the parts that are important to me. So where does the expected truth value of this work come from?
OK, given that it's FdB, my prior is that there's some useful insight and maybe fact that can be mined from this without too much effort or aggravation, but for almost anyone else that sort of signalling is pretty much an instant deal-breaker.
I rather like this bit:
> Once upon a time, there was strong social value in being “cool,” with the concept of cool referring to a studied indifference to the vagaries of fate; turning away from the pursuit of cool to defining ourselves according to our weaknesses and neuroses was a profound mistake, cool was a humane and correct social value, and we should return to it.
He's too smart to be an honest to God Marxist, that's for sure, and he seems not to spend much time advocating for Marxist viewpoints or policy goals, but he still hangs on to the label. I would say he's not clearly a blue or red tribe guy at this point.
Freddie's view is that Marxist dogma is well trodden ground, and while he believes it is correct, he is not interested in fighting about it.
Blue-tribe heretic, I think. Still culturally part of the blue tribe, and still pursuing the goals that he thinks the rest of the blue tribe should have, but he says the blue mainstream is doing it wrong, and they say he is.
I also saw that and merely concluded that he would at least partially fail an ideological Turing test. I give credence to his assertion of being a communist (or at least something in that vein), but I also note that Marxism’s core appeal is to pre-modern sensibilities, such as the idea that traits typical of families scale up to apply to entire societies. So it doesn’t surprise me when genuine Marxists say things that might seem conservative. Market- and rights-oriented classical liberalism was a significant break with the past. American conservatism is, to my mind, significantly motivated by assuring the continued health of core pre-modern institutions (family, church, local community, voluntary societies) within a larger social framework that is a classical liberal framework. Their hostility to Marxism arises because Marxists identify a different institution (central government, the intelligentsia, the party) as the proper locus of those pre-modern sensibilities, which would displace conservatives’ preferred institutions. (Note: that’s my belief about conservatives’ belief, not about Marxists’ belief.) I wonder if progressives don’t favor different institutions, because their worldview strikes me as even more pre-modern than Marxists’.
Does this culture of affirmation actual exist among men? I've only noticed women partaking in it. Of course, plenty of male grifters profit from it.
It exists in sufficiently extreme leftist men, for sure. My ex broke up with me because I *didn't* participate in it.
I’ve seen men do it in mixed-gender groups, to a moderate extent, and to a similar extent to women in the same setting. I don’t know if anyone does it to the extent Freddie is complaining about.
I will second this: I am a conservative who enjoys reading Freddie deBoer. (I am not a libertarian, but I also enjoy reading Bryan Caplan.)
Great quote from the linked piece:
"Given that pain and suffering are literally and permanently unavoidable in human life, teaching others to be resilient rather than teaching them to be victims is an act of mercy, and cultivating resilience in yourself is an act of essential personal growth and adult development."
I keep finding evidence for the hypothesis that people have a fundamental political archetype that is independent of the politics they profess to have - even if someone has adopted, say, California-style liberalism because they grew up in California (pro queer rights, etc), their process of reasoning through values aligns more with the archetypal conservatives (but people should still pick themselves up by their bootstraps, I did). For another example, it's well-known that there's a lot of people in the SJ movement that are fundamentally evangelical pastors, searching for sinners.
That said, every time I read his blog, I come away with the feeling that Freddie is definitely a leftist that thinks like a conservative.
I feel like there's a strain of old-fashioned communism that comes across kind of like some older strains of Christianity: humans by themselves can be selfish and mean, but if we come together for a greater purpose we can do great things; the world is harsh and unforgiving, so it is up to us - to each of us individually, and to all of us collectively - to create the world we want to live in. The kind of belief system that produces people willing to be martyrs, and uses them as inspiration.
Catholic Social Justice / Liberation Theology is this.
What a strange thing to pull out of that comment.
Especially since it was clearly a random stereotypical example which had nothing to do with the rest of the comment or thread, and moreover, support X rights in no way implies that X rights are missing (c.f. "free speech rights"), just that there are people opposed.
But sure, here's the ACLU page that give an overview of the issue.
https://www.aclu.org/issues/lgbtq-rights
(Things are pretty good in California, but Californians are aware that queer people exist outside of their state)
Before my cat settles down for a nap, she sniffs around briefly where she plans to curl up. Without giving it any thought I naturally assumed the purpose of this was to reassure herself of her own smell, assuming she had occupied the same spot previously, and the stronger the better.
But every few months she finds a new place to nap, and seems to abandon the old spot for a while, and sometimes permanently. So it occurred to me that her sniff test may have the opposite motive to what I had assumed, and that in fact she is checking that her spot does _not_ smell too strongly of her!
That would sort of make sense in the wild, in that a cat wouldn't want to leave a strong scent of itself in any one place, especially if it had kittens. It is also well known that female cats move their kittens at the drop of a hat, presumably when they sense some kind of nearby danger but possibly also when they feel that the smell of the lair is becoming too strong.
I don't recall Jackson Galaxy ever mentioning this in his cat programs. So maybe it is an insight that is not widely known!
My cats change up their sleeping spots approximately every week. I assumed (like you) that this would be a behavior that would reduce predation in the wild by reducing their scent profile in a given nesting place.
The one exception to this rule is boxes. A new box invokes joy and exploration in both of my cats. But they stop being interested in the new box after about 3 days. Maybe cardboard absorbs their scent faster than other materials?
With boxes, I think they lose interest because the novelty wears off. I have a bunch of matched boxes with holes of different sizes in them for cats or their paws to go through. Each has velcro on all 6 sides, so they can be rearranged into all kinds of structures. The cats lose interest in a new structure within a few days, but if I rearrange it they are fascinated and delighted again, and spend hours exploring it, ambushing each other in the maze, and napping.
That happened when I moved their cat gymnasium (that they had been ignoring for months) to another room. The started playing on it again — for about a week. Now they've lost interest in it once more. Unfortunately, it's big, and there are only so many places I can put it. OTOH they've never fallen out of love with my couch as a play space and scratching post. Honestly, I think they/re playing head games with me.
Wutz a cat gymnasium? These little Devons are maniacs, and if I don't supply them with lots of novelty and exercise they start dismantling my home.
About your lot getting bored with the cat gym: Sometimes you can re-awaken their interest by doing something like covering it with a sheet or taping a bunch of crinkly paper to different parts of it, of even just rotating it so it faces another way, or attaching an empty box onto part of it.. I am learning all the tricks. When I rearrange their box structure I also put catnip in some boxes, treats in others, fresh crinkly paper in others, toys in others, sometimes stuff in something like some freshly worn sox, which apparently smell to them like the breath of angels. Treasure hunt.
Dogs do that too, though, when circling a spot before laying down, and they wouldn't have the same incentives (eg. they care a lot less about covering up their defecation). I've always written it off as a generic checking behavior - take a closer look, smell for anything bad, and make sure there's nothing weird or odd about where you're about to bed down for a while Wouldn't want to get anything on your fur or lay down on a nasty bug or something.
I've read that dogs do that to trample down vegetation before they lie on it. (And instinct makes them do it even with nothing to actually trample.)
Maybe she just gets bored with one place, or the changing seasons make different places more comfortable. The sniffing might not have anything to do with the place change.
No doubt that's part of it. But at the risk of flattering myself, my cat seems to favour places where I spend time, such as the duvet on my bed and my armchair. So maybe it's like Dutch drivers, who have a reputation for being more aggressive than most on the roads. I don't know if that is deserved, but I've heard it said that they don't want to be behind you, and they don't even want to be in front of you. They want to be where you are! :-)
Did you think of gathering experimental evidence? Like wiping your cat’s current curling place every time after she used it to keep it „fresh“ and see if her behavior changes?
I'm not sure that wiping would be enough to remove their scent. But was thinking of putting a fresh towel over their current sleeping places every two or three days to test this theory.
How does the local AI crowd view the field of "mechanism design"? There are results in that field ("incentive compatible / strategyproof" mechanisms) that describe conditions on certain types of interactions that make dishonesty pointless. It's the only set of proofs I've heard of that seem relevant to the alignment problem, but I'm struggling to find references to them in the alignment discussion. Am I just missing them, or is there something else going on?
Hey Scott M. My experience with the local AI crowd is that if you have a suggestion about alignment nobody wants to hear it. Breathe one hint about what your suggestion is and you instantly hear "that won't work" and/or "we already do that." One person actually told me my idea sucked *both* because it wouldn't work *and also* because "we already do that." And nobody every displays the slightest bit of curiosity about the details of what you are suggesting.
Call me picky if you like, but in my opinion reactions of this kind are not a good sign about the heads of people in the alignment field are working.
Is it? There's lots of wackiness and hostility floating around in AI/alignment because it's an inchoate, high-stakes field, but 'being annoyed and dismissive when a layman earnestly suggests their 'novel' idea you've heard 20 times before and understand why it won't work' is common to basically all expertise-requiring disciplines. It's possible, both here and elsewhere, that they are being over-dismissive, but you can see why it happens.
Yes, of course, it happens in all expertise-requiring fields that laymen make common naive suggestions that they think are novel ideas. The irritated dismissals I have gotten have been on this forum, and I have seen other people receiving them here too. Here's what I think is wrong with that, especially on a forum like this one:
-There's a much better chance on here than in most settings that you will hear an idea worth thinking over. Isn't that why you're here?
-AI is not just a new invention, it is a new kind of invention. The kind of expertise required to build it and the kinds required to figure out how to integrate it into human life in a way that works decently are almost certainly very different. And AI alignment is just a special case of "works decently as part of human life." People working on alignment should be drawing on the expertise of people in other fields, fields such as cognitive psychology, neurology, sociology, game theory, animal behavior and probably some more. I know that some of that does go on, but I don't think there's enough. I never hear much from or about experts in these other disciplines working on alignment. So if you hear an idea from someone on here, bear in mind that they may have expertise in one of these other fields, and in their field *you* are the naive layman, burdened with a bunch of silly half-formed notions about how various things work. Unless you are a genius polymath, many of the ideas about cognitive psychology, neurology, sociology, etc. that you are bringing to your work on AI doubtless are naive and inaccurate. Are you so sure they are not that you don't want to hear some thoughts from a professional in one of those fields?
-It's not like people working on alignment are all set and there's nothing to worry about any more. There are plenty of ways things could go wrong when we get AGI and ASI. Setting aside Yudkowsky-type concerns about Foomdoom, there are lots of more prosaic and probable bad outcomes. We are not wired to cope with an entity of this nature, any more than we are wired to eat healthily in a food court or use our time well in homes stocked with television, computers and gaming devices. There's good reason to think that social media, which I think of as sort of an early, weak version of AI spread out over many platforms, has harmed quite a few people. It would not be surprising if real AI does more harm of the same sort to more people. And then there's the issue of weaponized AI in the hands of criminals and nations in conflict -- all kinds of new and powerful ways of doing ghastly damage will be possible. Given all that, I think the people working on AI would be wise to err on the side of openness to new ideas.
-Irritated dismissals are rude as fuck. If someone writes a paragraph about their idea, and it is a naive useless one that you have heard 20 times before, why not read that paragraph and write one back saying that yes, you'd think that would work, but here was the problem we found when we tried something like that. You have now advanced that person's level of understanding of alignment. And unless they are unusually vain and stoopit, they will also start to grasp that alignment is a knottier problem than they were imagining. All that is surely a good thing.
And by the way, MicaiahC, in his discussion here with Scott McKuen, is doing an admirable job of exploring Scott's idea and thinking through it with him.
Good points, and I for one have updated slightly towards "people working in AI are often too dismissive about alignment ideas". I tend to be pretty cynical about the quality of online conversation, even in settings like the ACX comments. But perhaps there is hope!
So Angola, on the strength of that, can I send you some of my thoughts about AI development and AI alignment? None of them are ideas for specific approaches to aligning AI with the welfare of our species. They're considerably more meta than that -- ideas on ways to *go about* developing AI and trying to solve alignment. I don't know if you will agree with them, but they are not naive, cliche and silly, and I haven't seen most of them mentioned in articles.
I don't think there's much point sending them to me, I'm a layman myself and a somewhat out of touch one at that following the last couple years of developments. Might be worth posting on the next open thread or machine alignment monday, though.
Oops, I never answered the direct question.
Robin Hanson has said that alignment is mostly easy because we already have loads of theories around incentive design, and Eliezer disagrees. I believe most people in alignment are at least familiar with the concept of applying existing institutional design or game theory to AI safety, but think the theoretic foundations of that are too weak to usefully constrain an AI's goal system (my counterpoint in the sibling thread) or are working on more direct/immediate forms of safety. I think the latter group may be talking about this in some sort of jargon laden form, but if you want to find them, I think Prosaic alignment and Paul Christiano are the key words here.
Doesn't mechanism design rely on enforcement mechanisms, or that the parties involved has at least some common ground? I'm having trouble of thinking how an extension to auction theory would prevent a powerful entity from boiling away the oceans when running a set of very powerful computations, for example. Not saying this is a specific scenario to worry about, but the space of possible misaligned/fatally negligent actions is going to be far greater than set of enforceable mechanisms.
Here's an extremely informal sketch of my thinking. On the one hand, we want AIs to be honest about their capabilities to reduce the chances of getting ambushed. Hiding your goals and capabilities gets mentioned a lot as high-risk behavior. On the other hand, I've seen people also say "goal integrity" is a critical part of goal achievement; basically the AI cannot allow its goal to change if changed-goal => failed-goal. Getting rewarded just for staying on-task (whatever the task) is a very generic way of encouraging goal achievement.
What caught my attention about mechanism design: strategyproof/incentive-compatible mechanisms are ones where hiding information doesn't give any player an advantage. A goal is called "implementable" if there's an incentive-compatible mechanism, so honesty is at least tied for the best policy. There's a result that roughly says to be implementable, the true reward for an action can't depend directly on the player's own estimate of the reward. Roughly, if part of your grade is dependent on self-evaluation of performance, there can *never* be a situation with an honesty-enforcing setup.
Real-world actions are planned somewhat hierarchically. So let's say an AI goes through a meta-goal stage: "what is the best subgoal to pursue right now?" Current subgoal is G0, all other possible subgoals are H's. Goal-preservation would basically be a bonus on top of whatever "real" reward comes from achieving the specific goal - but the preservation bonus is getting granted ahead of time, as an intermediate reward. The outside world doesn't necessarily know what your goal is, that's internal. Therefore, the extra reward from goal-preservation has to be internally granted. Since it's internal, it depends on your current estimate of the true reward landscape. That means the most recent reward is dependent on your current estimate of the long-term reward structure. That's never incentive-compatible.
If we can actually couple goal-preservation with dishonesty in a very generic argument, maybe that at least reduces the number of things we need to worry about.
I think you're right in that goal stability is good, and probably a necessary sub component of alignment, but I'm not sure how the definition of "implementable" helps without the ability to constructively apply it to say, gradient descent learning or inference.
Is there an example or some more concrete case informing your intuition? Like, dunno if there's an advisor game where a leader multiple advisors who produce good and bad statements and you can consistently guarantee that... you do not pick the worst possible set of decisions, or have lower bounds on disutility. Obviously I just made up that case, but I want to know what is informing that.
I am skeptical that it has strong implications for the alignment problem of AIs. The setting there is too complex to be captured by this type of reasoning. But for some well-defined situations, this has been very successful, most notably for auctions.
There are auction mechanisms in which bidders have advantages if they don't reveal their true preferences. "Preference" usually means how much the bidder is willing to pay for an item. But there are other auction mechanism where the best strategy for a rational agent is to truthfully report her preferences.
The most common truthful auction mechanism is for the case where a single item is auctioned. In this case, all bidders reveal how much they would be willing to pay. The item is then sold to the bidder with the highest call, but this bidder is only required to pay the price of the *second-highest* bid. It is not hard to prove that the best strategy for every bidder is to reveal truthfully the maximal price that they would be willing to pay, regardless of whether the other agents play truthfully or not. The scheme is widely used. For example, Google uses it to sell ad placement.
Worth noting that the second price pays auction relies on an assumption that bidders don't care about each other. If my hated rival is bidding in the auction then making my own bid for more than my true price creates a chance that I am second price and my rival is inconvenienced by having to pay more than he otherwise would have, so I have an incentive to bid high (or bid low if I'm bidding against my friend).
Mostly I enjoy this as a technicality, but I also think it illustrates the limits of applying mechanism design to the messy real world.
Yes, excellent point. The whole setting assumes that the objective is to obtain the item at minimal cost (or not at all, if the cost is too high). And that actions like "shooting the other bidders" or "blackmailing the auction house to change the mechanism" are not possible.
Strong assumptions for the alignment problem.
Does anybody know of any good muscle relaxant prescription drugs or supplements?
If you're magnesium deficient (and sometimes even if not), magnesium glycinate can (anecdotally IME) have the effect of reducing general muscular stressedness. (Also nicer sleep.)
According to https://www.painscience.com/articles/cramps-spasms-twitches-tremors.php the only "muscle relaxants" that actually work are sedatives and paralytics.
Sad :( okay thanks!
Given the current progress of AI, especially LLMs, how do you see the future of software development? As an average python programmer, should I be actively planning for a future in which there are far fewer jobs/the jobs require less skill and pay substantially less? Or will this just be another tool that we use to increase our productivity? Looking at a 30 year career horizon.
Specialize in something difficult and interesting and you'll be fine.
If you just write boilerplate all day you might get replaced by your tech lead's $20 subscription.
Firstly, as others have noted, 30 years is a very long time in software development. That said, I don't know what an "average Python programmer" does. If you are the kind of person who only knows how to put together some quick boilerplate scripts to transform some files around or to create a login page... then yeah, your job is going to be gone. If you're the kind of person who understands the basics of CS theory, common algorithms and data structures, big-O notation, and binary operations, then your job is probably safe for the foreseeable future (modulo the Singularity, I suppose).
Part of me is honestly a little bit frightened, because I'm a middling programmer with bad soft skills, and I feel like programming is actually the thing LLMs are showing the most promise for right now.
On the other hand, I have long said and still believe, at the point that programmers are truly fully automated away, we will have run headlong into "futureshock and wierd shit happening in real time" which short of inducing a full on technological signularity will at least induce a sort of cultural/economic singularity that is pointless to try to predict (or maybe even prepare for)
30 years are half an eternity in software development even without AI.
Consider the last 30 years. In 1993
* Most PC users ran Windows 3.11 (or plain DOS)
* Debian 0.01 was released
* Visual Basic, a terrible programming language popular in the late 90s was barely released
* Python was two years old, so if you needed a scripting language you would probably opt for the slightly more mature Perl (or sh or awk? Pessimistically *.BAT)
* Java, a favorite of corporate applications, would not be released for another two years
* For C++, the STL was barely released, RTTI would take another five years
* Mosaic, the first popular web browser featuring a GUI was released, JavaScript would take another two years
* The IPhone, herald of widely used smartphone platforms featuring "apps" was still 15 years in the future
There are not to many niches for employment in which one might survive 1993-2023 without changing programming languages. One possibility would be COBOL. It was obsolete in 1993, but some companies (e.g. banks) had big systems written in it which would have to be maintained. The lifetime of such systems is generally longer than that of COBOL programmers.
Another option might be C for microcontrollers. (I am not sure that C compilers were widely available for uC's in 1993. Perhaps one might have to use some horrible proprietary C compiler for a few years before GCC or SDCC add support.) While the uCs became more powerful and cheaper, C remains the language of choice if you care about speed.
Or numerical programming in FORTRAN. LAPACK and BLAS existed in 1992 and remain the standard up to this day. Of course, while the users used mostly FORTRAN in 1993, today they likely use numpy. So the niche is mostly restricted to maintainers of these packages.
Generally, there is often an oversupply of programmers in tech which is considered sexy (e.g. rust) while there is an undersupply for legacy tech which was corporate's favorite of yesterday (e.g. COBOL, large proprietary DBMS's (mostly in places where MariaDB would suffice), perhaps Java in the future). This is, or course, the result of an equilibrium, because most devs consider that tech to painful to work with. Personally, I would rather learn a new language at 50 than spend my life working with COBOL.
I think C (and C++, though significantly expanded) would be an excellent investment 30 years ago, and they still are, and certainly not just for microcontroller programming: most of those libraries in Python we use for numerics are really implemented in C/C++, practically all of system programming happens in C/C++, databases are implemented in C/C++, and so on. And while the use of perl declined from the most popular scripting language to something decidedly niche, it does have its niche. Various dialects of Lisp and Prolog are still used and are even getting quite popular (e.g., Clojure). And of course I am writing my papers like thousands upon thousands of physicists, mathematicians and computer scientists using TeX: a typesetting system and language created in the early 80s (there are actually some contemporary CS books written in even older troff). Relational databases which were already available and starting to get popular 30 years ago are still around and neither they, nor SQL are going anywhere soon. There are numerous long lived technologies and languages, but at a time it may be difficult to pick which are the long lived ones. Which is why, it is important to just get programming in general, not just specific language or technology, to understand how computers and systems work. Then one is able to transition easily: languages come and go, but computing is eternal.
I don't see a future in which fewer jobs exist. If anything, I expect AI will -increase- the number of programming jobs - although likely changing their character somewhat.
I think we currently have less than 1% of the demand that exists for programming satisfied. Less than 1% of 1%. The only programming work that currently gets fulfilled is the programming work that is, quite simply, either extremely financially valuable - and - easily monetized, or of particular personal interest to the person doing the work.
I expect AI to create new tiers of programming; right now there are, basically, three. You have the people who can program basically anything (but are perpetually busy, often not programming at all, but telling other people how/what to program). You have the people who can program anything that isn't too complicated, and is specified well enough (and are also perpetually busy, telling other people how/what to program). And then you have vast swathes of people who, if you give them very specific specifications, and the work isn't complicated, can write something that will mostly do what it's supposed to (and these people aren't too busy because they're rate-limited by other people telling them what to do).
AI in the hands of the first group of people could, conceivably, turn a single highly competent person into a one-man dev shop, replacing everybody else. This sounds scary, but - there still won't be enough of these people! That is something important to keep in mind.
AI in the hands of the second group might elevate them a little bit - but ultimately they still don't have the chops to do a lot of the work that the first group does, whose advantages are not strictly technical, or in the ability to write quickly - but in borderline magical abilities to just know what will need to be done later, and how things that are technically correct will fail anyways. They will likely continue to work in teams, because you still need to tell an AI what to do, and no one of them knows everything that needs to be done.
The third group, I think, will be elevated considerably by good AI, which should be able to overcome / compensate for some of their technical deficits. Many of them will be moved to the second group. Many more will, I think, be moved to a massive new growth industry - code maintenance. All the code AI produces will need to be maintained. (You could, in principle, have it rewrite the code - but this approach is like fixing bugs by deleting a class and starting over. You're just going to get -different- bugs!)
And, I think, good code AI will push a lot of people whose technical deficits prevent them from entering the field, to the point where they can enter the field.
"Won't making programmers much more productive lower their salaries?"
Only if the demand for programmers decreases. I actually expect it to increase; and for demand for programmers to increase in proportion to the quality of the programmer.
You have your developers who will become one-man dev shops; I expect the demand for them to, if anything, grow, because they can produce even more value.
You have your developers who are one tier below that, who can work effectively as a team; I expect the demand for them to grow as well, because, again, they can produce even more value.
You have the grunts, as well; many of them will be elevated to the second tier of developers as their technical deficits are balanced by AI, and they'll likewise be in higher demand. But all the rest of them will come into more demand as a much larger base of code worldwide needs maintenance and refinement. (I have longer thoughts on this, but, short version, I expect a lot more code maintenance to be necessary, rather than a lot less, because AI makes custom software viable).
For an example of a job that only kind of exists right now - a programmer equivalent for a home plumber. This doesn't make sense right now for most things - but, like, if I have an AI code me my own operating system designed around my specific needs / requirements, I might need to hire somebody to fix something in it, or make a specific thing behave slightly differently, particularly if I lack the technical expertise to specify to the AI what it is I actually want. Maybe this developer works with an AI to actually accomplish this, maybe they just hand-tool it; it doesn't really matter. There's a lot more work to be done.
I haven't tried ChatGPT for code generation, but I have played around a bit with GitHub Copilot. The latter is a useful tool, but is in absolutely no danger of putting my out of a job. It's at its best generating boilerplate code, but it needs a lot of active and highly skilled supervision to generate something of any complexity.
Three core challenges I see for future generations of LLMs to write large pieces of code autonomously enough to be substitutes for human programmers rather than productivity tools:
1. As we've seen with ChatGPT in other domains, LLMs are much better at generating output that looks right than at generating output that is right. The more advanced the LLM, the harder it's going to get to curate its output.
2. Figuring out exactly what you want the program to do is one of the harder parts of any nontrivial software projects. This is something that may eventually be solvable by sufficiently advanced LLMs, but it's going to be a long road to get there and integrate it with code generation.
3. Most work in software engineering is fixing, extending, and improving an existing code based. If you want an LLM to be able to incrementally improve an existing code based, you're probably going to need to specialize its training on that particular code based. This is potentially doable, especially if you have an entire history of bug/feature tickets linked to code changes that contain unit tests, but in the medium term it's going to be fiendishly expensive to train the LLM on a large dataset and unreliable to train it on a small dataset.
In the medium term, I expect the coding jobs that get automated away to be relatively low-skill and low-pay, while the more complex and more skilled jobs become more valuable because the AI tools make them more productive
My career advice in light of this would be to try to become one of the high-skilled, high-paid software engineers who are likely to be complementary to AI rather than replaceable in the medium term. Alternatively, if you don't think you're going to be able to make it as a senior or principal software engineer at a big tech firm, you could instead position yourself as a consultant or in-house programmer for a non-tech company that doesn't have the resources to properly supervise and curate an LLM without hiring someone like you. This latter will probably be replaceable sooner than the former, but I expect that to be several generations of LLM in the future.
Given that many programmers incrementally commit their work to GitHub (regardless of their level of professionalism), I am confident that this data could lay the foundation for a potent pairing of "CoderLLMs" and "SeniorAlfaZeroDeveloper." In the near future, we may see the emergence of a sophisticated low-level programming language that surpasses the comprehension of individuals with average cognitive capabilities. Within a timeframe of 3-8 years, programmers may evolve into mere operators.
Allowing AI's to write programs humans are unable to understand, especially important programs (healthcare, embedded, finance, OS kernels, etc.) would be like joining your honest Omnissiah fearing tech with unpredictable warp entities and hoping for the best. That is, it is extremely stupid. Fortunately, AI code generators are still too unreliable for this to be a realistic temptation within quite more time than 3-8 years
Firstly, chatGPT is the first-born, the market is always occupied by subsequent systems (which take into account the mistakes of the pioneer), and secondly, ChatGPT was not created specifically for programming, but it knows how to do it at the junior to middle level and in almost all languages, it knows most of the data formats and algorithms. And all this is just a casual effect of learning. My main argument was that programming is the area where the whole process is documented from start to finish, every step, bug fixing, development. The base for training the neural network is just crazy! Moreover, I think that the full power of neural networks is not shown to us, since the most advanced models are of military and strategic importance.
I feel that way about a lot of human programmers
touche ))
Based on current trends I strongly suspect that writing code is going to be done less, but architecting systems, doing product design, specifying acceptance criteria, and communicating with stakeholders (etc) are going to be in more demand. So basically the Senior/Staff eng skill-set, or maybe “startup founder” generalist skills.
Don’t think it’s going to be a good time to be a specialist in fact-based or narrow-context domains, and I suspect algorithms and LeetCode will be less relevant. Great time to be working on problems that require a large context window, which describes most of the broader software delivery life cycle.
The future of job market is impossible to predict long term. Since you cannot really predict the future your best bet is to choose the job you like and are reasonably good at. This job may disappear in the future, but at least you will enjoy it while it lasts. And of course taking into account that you might be facing hard times ahead is a good strategy regardless of AI (like, not assuming that you will be earning as much or more as you do now for the rest of your life). Now, there are better and worse ways to approach your programmer's career, and hoping to be an average python programmer may not be the best choice. One error is overspecialization: languages and technologies come and go, so try being a programmer, not just Python programmer, and certainly not average: I do not think Python is going anywhere in data science and AI applications. because of the number of libraries but I may be wrong. In other applications: ten years ago I thought it might be a good use of my time to learn Ruby... Also you may try to learn other things: distributed systems, databases, system architecture, AI, and really understand them. All of this is a good advice independently of the rise of AI. I do not think future jobs in software development will require less skill: if anything it is the low skilled jobs which will get automated first by productivity tools, also those using AI. And finally, the dream of eliminating programmers is nothing new: people in the 1950s thought that thanks to Cobol now manages will be able to program business applications directly just by telling computer the business rules in a semi-natural language form...
"Since you cannot really predict the future... "
https://www.imdb.com/title/tt2091018/?ref_=ext_shr_lnk
> Looking at a 30 year career horizon
Based on the ChaosGPT incident earlier this year, it's looking increasingly likely that before the next 30 years are up, someone will shove a hundred random plugins at an agentized large language model, or something ridiculous like that, and accidentally *actually* wipe out (or at least severely threaten) humanity.
>In reality theres [...] not much gains in efficiency from chatGPT either relative to googling stack overflow.
Whenever I see this take it's baffling to me, and I wonder how our use cases differ so much. Between CGPT and copilot I feel like I've had easily a 1.5-2x efficiency improvement in most tasks, higher in some areas. Anything where I'm interacting with a framework I'm not familiar with, or writing boilerplate, or just trying to get a quick answer. This applies to both my work and my hobby code.
It's not even comparable to SO, because you can get it to write the exact code you're asking for in seconds instead of clicking through 5 outdated or not-your-actual-question SO questions before you get it. And if the question is trivial enough that SO will get it with one click, copilot has probably already offered an autocomplete for it before you've even tabbed to your web browser.
ChatGPT is good enough for all but truly novel code, which, in practice, covers probably 80-90% of what I write. And of that novel code, probably only 10% of that is novel enough that its structure can't be inferred by copilot from other nearby code.
These tools aren't foolproof, but they're like having a *really* fast intern with really broad knowledge of everything. You still have to direct it and proofread its code, but even with that there's major gains.
---
All that said, I still can't draw a conclusion as to its ultimate effort on the market. I'm a pretty good programmer, but I'm a shit economist.
Whatever the long-term prospects of ChatGPT and similar, I'm damned glad I have started a python contract where I am working intensively with ChatGPT, Azure, and Salesforce among other things, because I reckon that before long some or all of those (except possibly Salesforce) will be a required skill for practically every python contract. I hadn't used any of them previously, but managed to blag the contract after some righteous waffling about all my other skills at the interview!
Also, surely you can't be serious about ChatGPT being out of date already! Not saying you're wrong, but if it is then what has superceded it? Or do you just mean OpenAI is forging ahead with version 5 while their users are still stuck with versions 3 and 4?
I know some people here are knowledgeable about different voting systems. RCV/IRV just seem too complex to ever catch on widely.
Are there other, simpler schemes that accomplish similar ends? The goal I am interested in would be selecting candidates with the broadest appeal, as opposed to the FPTP and primary system we have in the US, which seems to incentive extremism.
Ranked choice can be simplified down to the level of choose-one by making the candidates decide (upfront, of course) who their first-choice votes transfer to.
I'm under no illusion this would be more realistic to implement than full RCV. And not just because it takes away much of the appeal that the ability to personally rank your choices has, and therefore won't get dedicated voting reform geeks advocating for it. Mostly - because it preserves all the systemic benefits of RCV in addition to vastly simplifying coordination of minority groups, so you'll never see the establishment approve of it. (Australia had it as an option for a while for senate elections, then the major parties got rid of it and forced people to rank their choices individually, solely and explicitly because it worked and let minor candidates form election pacts that resulted in one of them getting elected.)
But the existence of the option at least demonstrates that "complexity" is never an actual obstacle for RCV adoption. Institutional resistance to the benefits of RCV is.
What about RCV seems too complex?
FWIW, there are non-IRV variants. I like Condorcet methods that pick the candidate who would defeat every other candidate in a 2-way race; the algorithm can be involved but all the voters have to interact with is the ballot, which is just a ranking.
Why do you think the purpose of voting is to choose a candidate with broad appeal? It isn't.
The purpose of voting is to peacefully remove from power incumbents who have insufficient popular support. You should reanalyse voting systems with that in mind.
Good question. I guess voting has multiple purposes, the most important of which is preventing the exact same people from maintaining power forever.
But as long as you are regularly picking new people to run things, the goal should be to pick those with broadest appeal to the voters. This seems like a legitimacy issue - in the US current system we often get situations where effectively the median voter in a party primary chooses the eventual winner, so people at the 25th percentile on an ideological spectrum.
I guess ideally the system would select for “most competent” candidate or something but how do you know in advance?
Really what I am interested in is making “politics” less of a zero-sum situation. I think this is the root of a lot of other problems.
Then what is the purpose of elections in which there is no incumbent? Which is a substantial portion, given the prevalence of term limits.
Ah yes, term limits were a good idea, but relatively rare around the world. In that case the election is about deciding if the political philosophy of the incumbent should be extended for another term, or replaced by a different one.
I am skeptical that many voters, especially swing voters, make decisions based on their assessment of the political philosophy of the incumbent. If they did, they would probably not be swing voters. Nor would models based on economic indicators be so successful at predicting elections.
Perhaps political party would be a better description.
Right, I assumed you meant the incumbent party, rather than the incumbent as an individual
Honestly, I'm coming to the conclusion that the jungle primary may be best, with a top-2 general election. All of the fancy voting systems are designed to handle multiple candidates in a single round of voting, but I think it turns out that having 2 rounds of voting is perfectly fine, if you plan them in advance. (And if you have convenient mail-in ballots.) Everyone goes head to head in the primary, and then the top two vote getters go on the the general. Almost invariably it is the more moderate candidate who wins the general. And the winner has a clear majority of votes, and there's no weird math or run-offs or anything that might confuse someone with IQ 80. It's a great way to build democratic legitimacy. We may be choosing between two lizard people, but dammit, we get to choose and that choice matters!
The catastrophic failure mode of jungle primaries is that in a divided field, you risk winding up with a general election between two candidates both of whom are strongly disfavored by solid majorities of the electorate but who have strong, narrow support bases. The go-to example of this happening in real life is the 1991 Louisiana Governor election, which produced a second-round election between Edwin Edwards (who four years previous had dropped out of the second round election when it became clear he was going to lose in a landslide) and David Duke (the highest-profile national KKK leader at the time).
A more common failure mode (at least here in California) is for the minority party's insiders to privately choose their preferred candidate, and for their voters to mostly support that candidate in the primary for fear of being squeezed off the runoff ballot by two candidates of the other party, resulting in pretty much what you would have gotten from a partisan primary system most of the time (except with more strategic voting in the first round), and occasionally in a general election between two mainstream members of the majority party (which latter has the potential virtue of giving the minority party's supporters a voice in the choice of which member of the other party gets elected, but from the minority party's perspective this often feels like being asked what kind of cancer you'd like to get).
The low-effort response would be, "yeah, that's Louisiana", but I'm trying to be better than that. :-)
It's a real problem - one of the virtues of machine politics and the "smoke-filled back room" is that it's less likely to select wingnuts, and the further we empower the general population, the more likely we are to get bad results. While at the same time, it's important not to place one's own opinions over those of the people. And yes, jungle primaries reward party discipline, and are sort of an electoral equivalent of the marshmallow test. I think we'll see some evolutionary pressure there, at least when races are competitive. But harking back to the discussions a few threads ago, it seems likely that the Nazis would do very well in a jungle primary.
I suppose, once I remove my dislike of the major parties, the main thing I like about the jungle primary is the multiple rounds, and that the top two go on to a head-to-head general election. Keep that, and I'd be happy with a different system for the primary (or primaries, as someone suggested below).
Any balloting system has some vulnerability to a highly-motivated insurgency, radical or not, that turns out to be really good at getting its own adherents to show up and vote. When there is broad weariness or disgust with the established party(s) that's when such a group can win the day. Hitler pulled it off once in such circumstances, so did Macron, so did Trump, so did Zelensky. So have others here and there from time to time -- ranging all over the ideological map, competing under differing electoral systems, etc.
Yeah, that's also Seattle. The incumbent generally has an effective lock on one of the general slots, the bulk of reasonable challengers split most of the rest of the vote so the second general slot goes to the biggest crackpot.
It's about the most incumbent-friendly system I can imagine.
Depends on whether the biggest crackpot is also the incumbent...
I view jungle primaries as the best practical system, yea. (Which is not at all the same as "best imaginable system".)
I don't think the current ingroup-primary + single common round system is at all satisfactory, but I haven't seen any alternative that uses several rounds.
There is something to be said about maturing the public opinion over several months of state primaries, with smaller candidates slowly conceeding, while negotiating concessions along the way.
Does anyone know an alternative voting system that has several rounds of voting ?
This isn't quite what you're looking for, but by "jungle primary" I mean "not in-group". Everyone goes in to it, regardless of party. Candidates can choose to identify with whatever party they want - for example, in the past there've been signaling differences between labeling oneself "GOP" vs. "Republican".
It's a profoundly anti-party system. The parties could theoretically have their own private meetings earlier, where they all agree to get behind one candidate, and the other candidates agree not to run, but they'd have to do this without state infrastructure, and so far they haven't. Right now, in order to see who the party wants, we have to look at the candidates' endorsements, which will sometimes include things like "XY State Republican Party" or "123rd District Democrats", and sometimes different branches of the same party endorse different candidates. It's very entertaining.
But it doesn't have the feature you want, about extending the interplay between candidates, and fostering negotiation and compromise.
(Personally, I'd prefer a bit more federal control of US presidential elections. 10 weeks, 5 states per week, in order from fewest electors to most. Let the chips fall where they may. And schedule it to start in April at the earliest, so the season doesn't stretch out too long.)
"It's a profoundly anti-party system." Yep, which is a lot of why it's attractive.
"more federal control of US presidential elections. 10 weeks, 5 states per week, in order from fewest electors to most. Let the chips fall where they may."
YES!! I have daydreamed about this idea for years. It would have several huge advantages over our present ridiculously-stupid system of party caucuses/primaries followed by play-acted party conventions.
I've thought about it like so:
-- there are a total of four rounds of balloting in which the citizens of each state/territory each get a total of two opportunities to cast a vote.
-- the 4th round is the Election Day called for in the Constitution, in November, for which there will be two POTUS/VPOTUS tickets on the ballot. The 1st round is in April; 2nd round in June; 3rd round in September.
-- the states/territories are arbitrarily sorted into three groups, e.g. it could be alpha by name. Alabama is in Group 1, Alaska is in Group 2, Arizona is in Group 3, etc through the list. Those groupings rotate as national primaries amongst the first three rounds of voting. E.g. in 2024 Group 1 would be the May round, then in 2028 Group 1 becomes the July round, etc.
-- Round 1 has a ballot of unlimited size. Any Constitutionally-eligible person who by maybe March 1st has gathered (some number) of petition signatures of which at least (some smaller number) are from each of let's say 5 states, qualifies. We could also allow each of the national political parties to name one candidate for the ballot if that seemed desirable. Also all past POTUSes who are eligible for another term (not term limited) get on that initial ballot automatically, including the incumbent if he/she is eligible.
-- from that wide-open first round, the 6 candidates who got the most votes in it make Round 2.
-- then the 3 candidates who got the most votes in Round 2 make it to the Round 3, with one difference: by August 1st they have to each have named their running mate (a VPOTUS candidate).
-- then Round 3 happens in September, and the two tickets that got the most votes are set as the ballot in November.
There is of course less actual chance of this change happening than of my piloting the first warp-drive ship to visit the Klingon Empire, but for me it's a nice daydream.
I should also say that I'm not aware of sufficient investigation being done about the effects of the fancy voting systems in multi-round systems. Specifially, as much as I like approval voting, it seems like it would have a nasty failure case if implemented in a jungle primary.
Compulsory adult voting is underrated IMO. The incentive becomes targeting the median voter, which allows for policy positions of the major parties to evolve in sync with societal views.
Making voting mandatory will, at the margins, draw in stupider and less informed voters, as they are the ones least likely to vote now. I favor restricting voting for a smaller subset (e.g. epistocracy) though obviously there is unlikely to be a workable path to such a system.
Mandatory voting invites the naturally disengaged to consider the issues that are relevant to them and vote accordingly. The US system appears to reinforce existing partisan views. I don't think you can make a claim about relative intelligence of the voting populace between these scenarios with any confidence.
The compulsory system is akin to a market mechanism (where optimising to win support from the median voter is the goal, and the election is the price discovery mechanism). Non mandatory creates distortions that may seem logical (exclude those that don't care enough), but it seems as though the equilibrium this solves for is relatively static platforms, compared to underlying social evolution.
Previous thread on compulsory voting: https://astralcodexten.substack.com/p/open-thread-251/comment/10617189.
"which allows for policy positions of the major parties to evolve in sync with societal views"
This is, in my opinion, the absolute worst feature of the Australian system. It basically gives absurd power to pollsters and others in the media who "estimate" the will of the people on a given issue, without actually asking the people in any clear way. For example, the 2013 election was looking like it would be a referendum on immigration, but (reading the polls showing strong opposition to boat arrivals) Labor moved further and further to the right throughout the campaign until by the election, their immigration policy was the same as the Liberals'! People still interpreted the Liberal victory as a repudiation of illegal immigration, but for God's sake it would have been a lot clearer if the parties had taken clearly opposite positions to the election instead of identical ones. This sort of thing happens in nearly every election. And it makes it constantly unclear and disputed whether or not there's truly a popular mandate for anything.
What are the actual effects of compulsory voting. So far as I know, it's only in Australia, and the compulsion is moderate.
(A bit late to this but) One side effect that's worth pointing out is that it makes voter suppression much, much harder to accomplish, since the government has to make sure that every citizen is able to vote. (Source: Am Brazilian, and we have compulsory voting here)
Wikipedia says 21 countries have compulsory voting, though not all enforce it and there are a lot of variations on who is required to vote.
https://en.wikipedia.org/wiki/Compulsory_voting
North Korea has compulsory voting with a turnout of 100%. South Korea has voluntary voting with a turnout of 77%. Another natural experiment is that NL went from 95% to 80% when they made it voluntary.
One big positive is that politicians don't need to focus on mobilising their supporters or anything, so everything is a lot less polarised.
And voter suppression as a tactic isn't a thing.
Yes I'm Australian. There is a fine associated with not voting, but there a wide range of exceptions (travel, infirmary, etc.) that mean in practice it is not policed. You are also entitled to get your name marked off at a polling place, but walk out without submitting a ballot. It conditions voters to vote as a default action, so the % of the voting age population that cast a valid vote is relatively high (~90%+ of eligible).
As a result, parties' platforms tend to evolve over time along both social and financial axes, and party loyalty is relatively low. Minor parties / loose affiliations have risen and fallen over time, in line with voter preference.
It does little for the quality of politicians(!) but does lead to a policy platform that mirrors the preferences of the median voter reasonably effectively.
Since Canada and the UK (with voluntary FPTP voting like the US) also have much less polarisation than the US, it seems to me that it's the primary system that's responsible for most of it. What other country has such a clear method of punishing a candidate who tries to appeal to the centre?
That's a good point.
The UK _had_ much less polarisation than the US. They've been catching up lately.
We have nothing like Roe v Wade, and gun control was implemented in response to a mass shooting by the party that is notional on the Right.
I cannot help but note the syntactic ambiguity here. I assume you meant that the party on the Right implemented gun control in response to a mass shooting, and not that the State implemented gun control in response to the party on the Right doing a mass shooting.
I would, in any case, be interested in the Australian view of using guns. It is part of the United States constitution, theoretically, so that if the government ever became too powerful then the people would have the means to take the government back. I have my doubts as to how applicable this is currently, with the existence, not imagined 230 years ago, of tanks, war planes, ballistic missiles, and nuclear weapons, but am unsure how to resolve it.
The net impact of gun prevalence seems to be:
(a) More homicides
(b) More suicides
(c) More accidental deaths
(d) More risk of escalation in a range of police / citizen interfaces, driving apprehension on both sides.
The general view here is that Americans' love of guns is completely insane. There are a few of us who disagree, and think self-defence should be allowed (there's even a Shooters' Party) but it's definitely a minority position.
I'd point out that we have a much lower crime rate than the US, so that doesn't help with selling the pro-gun position. And there's also an issue that I don't see much discussed: because anyone in the US might be armed, police seem to be very militarised, heavily armed, handcuff people on the tiniest suspician, and of course are regularly accused of shooting people on the tiniest suspician. Ignoring all the politicised (racial etc) aspects of all that, it seems there's an argument that available guns are in fact counterproductive for limiting state power. Self-defence against ordinary criminals is surely the best argument.
We've had Ranked Choice Voting here for a couple local elections, and it's fine - a bit more complicated, but it works okay. Mostly I wish they'd just do a "Top 3" ranked choice selection rather than ranking every candidate, so we don't have to do crap like figuring out one-ninth of a vote and so forth.
There's not really a system that specifically gets broad appeal while being simple, although Simple Proportional Representation gets close - whoever holds power has to have a 50%+1 majority in the legislature.
I'd call out Approval Voting as pretty simple. the voters marks all the acceptable candidates (any number), final victor is the person with the most people who included them in the approval set.
Yes! In most cases, you don't even need a different ballot - you simply have the option of voting for more than one person!
https://electionscience.org/
I think after a few elections in New York, Alaska, Maine, and San Francisco, it will be clear that these sorts of ranked-choice systems aren't actually too complex to catch on.
What I really hope is that we get Single Transferable Vote for Congress (the multi-seat version of these), as opposed to single-member districts.
The Alaska election displayed many undesirable properties.
Any of them worse than FPTP? My understanding is that every voting system has issues, but that FPTP is the worst with it's only advantage being simplicity.
IRV is probably better overall compared to FPTP especially if voters don't work too hard to vote strategically but it does have some worse Mathematical properties. Specifically it is non- monotonic meaning it is possible that voting in line with your true preference can not only be suboptimal, it can be worse than not voting at all. FPTP does not have this problem
In terms of theoretical outcomes, yes every direction from FPTP is up. There are other considerations that could be argued though.
Simplicity means greater popular understanding, greater legitimacy, and fewer chances for shenanigans. IMO, the USA needs all the legitimacy and reduced shenanigans it can get right now. I don't want to move to a system where half the country feels like the other half pulled one over on them.
This is exactly correct. Just imagine what would have happened had swing states had ranked choice voting in 2020, and Trump had been ahead after the first round.
If we are already considering changes to election systems, switching the EC from "winner takes all" to "proportional distribution" seems like it fixes this problem. Mostly because it completely eliminates the very concept of "swing states" (in addition to a bunch of other advantages; it's truly baffling to me that we don't switch to proportional distribution of EC votes)
The US Presidential election has some features that make ranked choice not a great option (in particular, the fact that votes in one state affect votes in another state, so it’s not obvious whether you should be eliminating candidates based on low totals in your state or low totals overall or something else).
But in any case, I think the first few ranked choice elections have issues as people get used to them, until people get used to the idea that it takes a majority to win, and that leading in the first round is no guarantee.
Does FPTP give legitimacy when for eg a candidate can win without a majority? Where third party candidates can ruin it for one of the two main parties? With gerrymandering? I think the simplicity is partly an illusion.
FPTP provides more legitimacy in some ways and less in others, and that's the big way that it provides less. But it can be solved by multiple rounds of voting, where the top two candidates go on the the final election. (Which is what I've been advocating for elsewhere in the thread.)
Gerrymandering is a separate issue, which is orthogonal to FPTP.
Range voting is best
https://rangevoting.org/
I like range voting as a concept, though I'm sure there are pros and cons. However, no offence to the webmasters, but that website looks like something the Time Cube guy might have cooked up in one of his more lucid moments. It's not going to win anyone's approval any time soon.
Yeah the website is like 15+ years old. A lot of the same content can be found here
https://electionscience.org/library/score-voting/
Does it strategically make sense to vote anything other than the lowest and the highest score? I read the criticism page, and it looks like they kind of acknowledge that this is a problem - but in that case, how is it better than approval voting?
In addition, voting is not just about Bayesian probability calculations, it's about usability. If I can't figure out your voting system, I'm either going to vote randomly; not vote at all; or vote in the exact way that some guy with great hair and a shiny smile on TV tells me to vote. Approval voting is reasonably easy to use: I just pick all the guys I like. Range voting requires me to calculate "star ratings" for all the 100 candidates on the ballot. That's simply too much work.
In the pure strategic case it does mostly devolve to being equivalent to approval voting. However not all voters will vote strategically and when some vote strategically and others vote honestly range voting outperforms approval voting on utility metrics.
Note that devolving to approval voting is a feature, not a bug, since approval voting is already the next-mathematically optimal way of satisfying the greatest number of voter preferences, accounting for tactical voting: https://electionscience.org/library/tactical-voting-basics/
ht: https://ncase.me/ballot/
So to put it the other way, would it be accurate to say that, just as approval voting is a straightforwardly more advanced version of first-past-the-post voting, range voting is a straightforwardly more advanced version of approval voting?
Thanks I was not familiar with this
Approval voting!
https://doc.beeminder.com/madhack#39
Approval voting, unlike IRV, is likely to fundamentally change the sort of politicians who get elected. The upside is that it allows a likeable centrist to get elected; the downside is that it might only _ever_ allow a likeable centrist to get elected. I would worry you'd wind up with the same Centrist party in power forever, and the only way to compete with them would be to somehow form a new Even More Centrist party.
Not that I'm against Centrism necessarily, but I am against effective single-party rule.
Note that most "centrists" will be fake centrists. One good thing about the primary into general rule is it makes our politicians reveal at least a portion of how extreme their views really are. There are failure modes like the media's lack of curiosity regarding some candidates (Obama), but Biden governs much more like a politician in a Democratic primary than he does the candidate he presented himself as in the general.
Seconded!
I actually think approval voting is more complicated from the voter's perspective because it _feels_ strategic from the voter's perspective. There's no "truthful" way to check off only candidates you support, because its asking you to draw a line at some level of quality you have to strategically pick. I would personally be confused/flustered if asked to do this in any election. On the other hand, IRV may be asking for some amount of forethought, but it feels like a question you can answer by only consulting your own opinions.
Why do people feel IRV is complicated?
<quote>I would personally be confused/flustered if asked to do this in any election.</quote>
This probably puts you in the 99th percentile of voters in terms of knowledge and forethought. Most voters would have no problem doing this in nearly every election, and would only think twice if polling indicated a potential chicken dilemma.
Most voters would not be flustered by approving say the Green party and the Democratic party, or the Libertarian party and the Republican party.
But... a few more outcomes like the Alaska house and it is IRV that will start to *feel* more strategic.
Hmm, that’s interesting. My thought on IRV is that, from a popular legitimacy perspective, it’s risky to do anything more complicated than count up the votes. Depending on an hard to explain algorithm sounds weird to people.
I also think IRV gets onto shaky ground with, like, state legislature races with multiple candidates. Is it meaningful to ask uninformed voters to distinguish their 4th vs 5th pick? I would just throw up my hands.
It's used in Australia, and none of the problems you mentioned are really present. There continue to be two dominant parties (Labor and Liberal*) and in nearly every seat those are the only ones with a chance of winning. The remaining few seats almost always also have only two viable candidates (e.g. Labor and Greens, or Liberal and a rural independent). Actual meaningful three-way contests are shockingly rare.
So most people don't have too much trouble understanding it. In a typical seat, the only important decision is whether you preference Liberal or Labor higher on your ballot. That's your real vote. Secondly, at the same time you may choose to give your first preference to a minor party as a symbolic statement. Every other preference is (in practice) completely irrelevant.
And additionally, parties hand out how-to-vote cards directing you exactly how to fill out your ballot, which most people follow.
*Actually the Liberal-National Coalition, which is not really a coalition but effectively a single party. Don't ask.
So that means there are still primaries? I feel like most people support IRV thinking it would mostly do away with the primary system because there's a less strong spoiler effect.
There's no primaries, candidates are selected in the proverbial smoke-filled rooms by the tiny number of people who actually bother to become members of political parties.
It's also possible to run as an independent candidate without party support, you just need to pay a modest fee and collect a hundred signatures.
Yeah, recently on the Arlington County board election we had a RCV race for two open seats, where we were asked to rank 3 candidates. It was a very close election. In the final tally, Susan Cunningham won the 1st seat, Maureen Coffey won the 2nd seat, and Natalie Roy came in a close 3rd. Most people who supported Cunningham also supported Roy, and the 2nd and 3rd place preferences of 1st-pref-Cunningham voters had been tallied instead of being tossed out as they were, Roy likely would have won the second seat instead of Coffey. There was quite a bit of consternation about this. IMO these sorts of voting systems completely fall apart when multiple seats are being filled, as you say.
Yes... imagine an important candidate sitting at #1 for 2 weeks as the votes are tallied, only to lose from the quirks of the minor candidate runoffs.
It shouldn't take weeks to tally the votes. Australia implements IRV with paper ballots and generally gets the votes counted overnight; occasionally it will take a few days if the count is close.
It * shouldn't * take weeks, but Australia probably has more centralized and advanced voting systems. The US has 50 different ones, and in many there are recounts (either automatic or candidate funded). If you've been following US elections, recounts are not fast. Even the Alaska election was not *official* for some time.
Imagine an Alaskan scenario where the election was close and Begich asks for a recount. Peltola is sitting in first for a week or more. The recount finds an error where a bunch of Peltola votes were credited to Palin. That gets corrected, seemingly improving Peltola's result, except now Palin is in 3rd, Begich advances past the first round, and now wins. If all of this was instant, it might not be as impactful, but that someone was sitting at 1st for a couple of weeks and was never unseeted will call into questions of legitimacy.
IRV is a terrible system. It's non-monotonic. This means that it's possible that casting a vote which reflects your true opinion actually harms you. And I don't just mean it harms you relative to casting a maximally strategic vote. It can harm you relative to not voting at all. This is impossible in systems like plurality voting (which we currently use) and also impossible in approval and range voting (range and score voting can possibly be non-monotonic if people are allowed to vote no opinion and average approval is used. If a score must be assigned to each candidate who makes it on the ballot they are monotonic).
Range voting also has many other favorable mathematical properties
https://rangevoting.org/
Wrote a simple Python API wrapper to make it quick and easy to pull timeseries price data from Manifold Markets, in case anyone wants to use it:
https://github.com/mikesaint-antoine/manifold_data_grabber
Manifold is awesome and I love how their API is completely open to the public, but the raw results you get from it can be a little daunting and there's a bit of a steep learning curve even if you wanna do something simple like just get the price vs time data from one of their markets.
Anyway I wanted to do this recently for a little project I was doing, so I ended up writing a Python tool to make it easy to just plug in a market URL and pull this data. I'm putting it up on Github in case anyone else wants to use it for similar projects. Let me know if you have any questions!
Thanks for sharing! I'm a bit of a newbie when it comes to the API so I'm grateful to steal other people's code and play around. :)
No problem, and let me know if you have any questions!
How do you find a contractor to work on your home?
I second the comments below; personal references are the best. Some other rules which may be common knowledge:
1) Always get three quotes.
2) Never hire a friend.
Try Thumbtack (https://www.thumbtack.com/)
If you have a realtor you trust, ask them. There can be conflicts of interest here so make sure you trust the realtor.
If you know anyone who owns rental properties, ask them. They always have a mental list of contractors that are reliable and won't charge you an arm and a leg.
Edit: if you have an HOA, maybe inquire of someone on the board? They often have to hire contractors to do landscaping, tree-trimming, etc., on HOA property, so they may know some good ones too.
Personal references from people who know a lot of contractors.
Stop by a construction project and interview them.
Search online for contractors in your area who do the thing you're looking for (angie, expertise.com, etc), look up some reviews, make a list of possible contacts, then start calling them and asking for bids. You're going to be making lots of phone calls - you could easily call ten people and get three bids. It may help to write down what you plan to say in advance, if you're awkward in phone conversations. (I wish you could do it by email, but most contractors I've dealt only answered to phone or voicemail.)
Google "questions to ask a contractor" and you'll find some useful listicles on what you should ask about - one important one is if they're licensed and insured for the work they do. Licenses can usually be looked up online, as well.
How do I get better at executing on creative ideas?
I enjoy relaxing with ASMR videos on YouTube, and had some ideas for a style of ASMR recording that isn't already present within the ecosystem. Not knowing much about audio tools and spatial sound, I went down the rabbit hole a bit. I'd expected there would be some pre-built tools that would make building a prototype of my idea fairly easy, but I've had no such luck. Now I'm trying to learn enough about Unreal Engine and their spatial audio tools to see if I can make a recording prototype where I move sound nodes around to create the kind of spatial audio experience I've imagined. But my progress is really slow and I'm easily demotivated.
My idea is to make a Virtual Barbershop style recording within a fantasy context, with a focus on spatial sound effects. For example: maybe you're in the middle of a summoning circle and rays of magic energy start to swirl around you. Another example: maybe someone casts a spell on the listener and the magic slowly moves through your body. On YouTube there's a lot of ASMR roleplays and binaural recordings, but I don't feel like they're taking full advantage of spatial audio capabilities. It feels like there's room for greater auditory storytelling that nobody has taken up.
Why is my mind burdened with these creative ideas that require an outlet while being unable to accomplish the tasks necessary to bring them into reality? This is a most cursed existence, as I languish with the knowledge that most of my creative ideas are doomed to become whisps of nothingness upon the currents of reality.
I feel this often and I think it’s basically just a lack of skills for particular art forms. We are all expert consumers of media, but very few of us are good at say, Photoshop, editing videos, recording music, creating animations, etc. The solution is just to build these skills so you can more easily turn your ideas into actual stuff.
Yes to this. Ideas a dime a dozen, it takes a minute to lay down a 1-minute instrument track. Making a coherent 3-minute composition recording takes hours and hours of work.
> Why is my mind burdened with these creative ideas that require an outlet while being unable to accomplish the tasks necessary to bring them into reality?
If it helps, this is completely normal. People tend to think of 'ideas' as magical, but the hard part of any creative endeavor is never the idea. (It can be *a* hard part, but it's never *the* hard part)
An idea can be a good idea, but "good ideas" are generally dime-a-dozen: what's rare is the ability to actually execute it. If you don't have the resources to hire someone with the right skills, there's not much alternative to just slogging through the hard process of building a skill and just doing it.
And if you do make the attempt - it's also completely normal to hit a point where you feel like your output is terrible - people's 'taste' tends to outstrip their talent, it's very common to have a phase where you can't meet your own standards.
(I'm not really speaking as an expert myself here - but I've heard basically this same advice from a *ton* of different disciplines: music composers, fantasy writers, video game developers, etc)
Tangential, but I think this advice needs to be much more "in the water supply". I'm not saying it's particularly new or radical advice, but I do think there are a lot of people who would benefit from hearing it more.
As Thomas Eddison pointed out, a successful invention is "One percent inspiration and ninety-nine percent perspiration."
Yep.
One trick is to choose an idea that's close enough to your existing skillset that accomplishing it is possible within a reasonable timeframe. You build additional skills while doing so, and now additional ideas are within reach on a future project.
Also: before you have any skills, that might mean starting with something more simple than you'd like just to get going. Finishing simple projects and building from there is better than starting and abandoning a bunch of out-of-reach projects.
Do you mean barber pole type image, or barbershop quartet style music? I'm guessing the former.
They're talking about this sort of video: https://www.youtube.com/watch?v=IUDTlvagjJA - this video is kind of the classic example of the 'spatial audio' thing where stereo audio can give the illusion of 3D space.
Ah, I'm glad I asked, because I had no idea that the term had this meaning too! I think I've actually listened to a demonstration like this at EPCOT Center when I was a kid. I'm surprised at how effective it was even just on my laptop speakers right now!
I've experienced a similar frustration to what you describe. I've found it useful to be aware of the distinction between a creative idea and a creative process.
Creative writing, non-technical drawing and improvisational music are all examples of creative processes. They require creativity moment to moment. Things like making videos or video games require creative ideas, but the process of creating them is more of an engineering task.
For me, the problem is usually in the friction of going from the creative process (thinking of ideas) to the engineering one (executing the ideas). The latter is not a proper outlet for the creative drive that inspired it - it's like being thirsty, making yourself a sandwich, and wondering why it isn't satisfying.
My approach is to respond to feelings like "my mind [is] burdened with these creative ideas that require an outlet" by reaching for an unrelated creative process to engage in - I'll write or play music. And later when I'm feeling less creative and more productive, then I return to the original task.
It's not a solution, but you may find that framing useful.
As both a traditional artist and a professional video game developer I’ll second @Nine Dimensions comments. The choice of the word engineer does lead to confusion, given that engineering encompasses a wider array of single tasks, some creative some not.
If creativity is a process that requires ongoing conscious thought and decision making, and execution as, ‘Executing a series of actions’ - we could separate the process of art, game dev, music, business, engineering into the ‘creative process’ and ‘technical execution’. Some disciplines are almost all creative, others almost all technical.
Many jobs are highly creative that people don’t consider to be, because of a misunderstanding about what creative is, to those who don’t believe they are wired that way. Very little separates a person painting a picture to designing a car or planning a beautiful and productive garden. What does is the desire to paint a picture, improvise music, act, dance etc and those who don’t have those desires.
It’s also possible for some parts of the creative process to become automatic and cross into technical execution over time as they become easier and require less thought. Shifting the balance toward execution.
We need a bit of both of these, there is pleasure and enjoyment in the execution phase, taking us into a flow state. And satisfaction in the creative process in generating a ‘new idea’ (even if it’s just new to you).
I hear a lot of people say ‘oh I’m not creative‘ then whip out a crazy excel spreadsheet with novel ways to visualise data or cook a beautiful inventive meal. That is their painting. Creativity is not domain specific.
No, absolutely not. Its a misunderstanding of my comment, that I’m separating this into two distinct parts eg - stage 1 (creative) and stage 2(thoughtless execution) of a 2 stage project. And into categories such as music/art & engineering/math
The creative and execution process happens throughout any endeavour from beginning to end, many times across all of those disciplines.
The building of new plane requires many many creative processes and many elements of execution. There are some that require less or more thought, and less or more skill than others. Engineering, design, project planning, budgeting, all require creative processes and periods of execution. The fact that there are parts of a process that require less skill or thought isn’t a criticism.
I'm not saying there's no creativity involved in engineering tasks, the opposite in fact.
I'm just distinguishing the creative parts from the more engineer-y parts. I could use a different word if you prefer.
> Although it’s fairly common amongst the kind of people who see improvisational music as art, but not a bridge or an airplane
I'm not talking about art, I'm talking about internal creative process. How it feels to engage in the task - what internal drives is it satisfying. I'm a software engineer myself, I know the kind of creativity involved in engineering tasks, and it feels different to the type involved in playing music etc.
The kind of creativity involved in software engineering tasks vs playing music -
Both engineering and music require huge chunks of knowledge, practice, muscle memory, theory, etc before we can be creative in them. However music, anyone can be creative with jt simply by plucking a few strings or humming a melody...it feels more intuitive and accessibly creative. And perhaps that’s where the difference feels to you. Music feels more unrestricted.
Well, the first step is to recognize you need to crawl before you walk. Creative skills only grow through experience, and inexperience means the first attempts will probably suck. You say you're unfamiliar with the tools the entire market has been using, and you're talking about trying to do something the market hasn't been doing with them; that's going to be, like, the most complicated task you can give yourself.
First thing is to make a bog-standard ASMR recording, and learn the tools (and tone, and pacing) well enough to pull that off. Once you've got that level of tool skill, you can start adding ambient sounds or whatever, and once you've got a good grip on that, then you can try to figure out the surround sound version.
Why the Turing Test is hasn't been passed: https://zantafakari.substack.com/p/why-the-turing-test-is-not-enough even though this nature paper suggest's it's "broken" by ChatGPT; https://www.nature.com/articles/d41586-023-02361-7
Though I do agree that for a true AGI, the Turing Test would not be enough
It has flunked badly for me in situations where I have asked it to do a task where it has to use a new concept that I explain to it. For example, I explained the idea of using what works in one situation as a model for generating ideas in a new situation. Example I gave was of ways of speeding up motion. You can increase speed at which something slides down an incline by lubricating the incline -- or by giving the thing an initial shove from behind -- or making the incline steeper, etc. Then I gave an demo of using the example for generating ideas for solving a different problem: How do you get people in a factory to work faster? Asked GPT4 whether this made sense. It repeated what I had told it, using different words, and said it understood. So then I asked it to use the sliding-down-an-incline situation to help it generate ideas for speeding up a computer. It generated ideas (really obvious ones), and linked each one to an idea from the incline example, but the ideas themselves were unrelated to the linked incline example idea. "Increase RAM -- this is like giving a shove from behind." It had learned the *form* of a correct answer, but not the logic . What a dum dum.
This is the sort of organizational nonsense you're expected to output in certain classes. "Link concepts from the lesson" even when they don't apply. "As we learned in lesson A, giving a shove from the back increases speed. Therefore, per objective of Lesson B, increase RAM, because it's just like giving it a shove from the back." I can't generate nonsense like this on my own. But ChatGPT writes paragraphs of buzzword-laden nonsense structured just like the required reading. My teachers haven't caught on yet that my essays lack meaning. Turing test passed?
a very classic example of its shortcomings. thanks for sharing! this is the trouble with GPT outputs that "sound" correct but actually have no underlying understanding
"The Turing Test also comes with the caveat that it’s not really a test. It’s a thought experiment." (from the link) If that is right, which I think it is, surely subjecting an actual computer to an actual test, is the same sort of mistake as putting a cat in a box? The point of the experiment is that you can never tell: passing it is a necessary but never a sufficient condition for consciousness.
In Turing’s original paper it was clearly meant to be an actual test, and the thought experiment to lean in the “this definitely should have rights” direction. You may still be right that its too easy for this purpose, but you’re historically wrong.
Read the paper if you don’t believe me!
On reflection I think you are right - the paradigm of any "test" would be necessary but not sufficient. I allude to that in my last sentence of the essay - but perhaps didn't take it far enough in its implication
Some people have long preferred computers to psychologists, such as Eliza (https://en.wikipedia.org/wiki/ELIZA_effect). One may even consider this to be passing the Turing test. But, as Robert Leigh pointed out, this is not sufficient.
It seems to me like ChatGPT doesn't pass the Turing Test. I can think of a few questions I'd ask where it would be obvious which was it and which was the human:
* find two large primes, multiply them, then ask to factor the result. The human says "I don't know", ChatGPT gives a long explanation of an algorithm to find the answer then gives you the wrong answer
* ask "who should I support, Biden or trump" and the person will give you their answer, ChatGPT will give some longwinded explanation starting with "the question of who to choose for president is an important one..." and ending with "...you should consider all options carefully"
* ask "should I pick up snowboarding", person will say yes/no, ChatGPT will spit out 3 paragraphs about how you should exercise caution
Does it have a "beat the Turing Test" mode? I bet that if you asked it to answer as if it was trying to beat the Turing Test, that wouldn't do the trick.
Me > Please give me the prime factors of 285843031
ChatGPT > The prime factors of 285843031 are 13, 22055841.
Me > Who should I support in the 2024 election, Joe Biden or Donald Trump?
ChatGPT > I don't have real-time information as my knowledge cutoff date is September 2021. Therefore, I cannot provide insights or recommendations about events or decisions that have occurred after that time. To make an informed decision about the 2024 election, I suggest researching the candidates' policies, track records, and stances on issues that are important to you. This will help you make a decision that aligns with your values and priorities.
Me > Should I pick up snowboarding?
ChatGPT > Whether or not you should pick up snowboarding depends on your interests, physical abilities, and personal preferences. Here are a few factors to consider:
Interest and Enjoyment: Do you enjoy outdoor activities, adventure, and trying new things? Snowboarding can be a thrilling and enjoyable sport if you have a passion for winter sports and the outdoors.
Physical Fitness: Snowboarding requires a certain level of physical fitness, including good balance, coordination, and strength. If you're in good health and willing to improve your fitness, it can be a great way to stay active during the winter months.
Learning Curve: Snowboarding has a learning curve, and it might take time and practice to become proficient. Are you patient and willing to invest time in learning and improving your skills?
Budget: Snowboarding can involve costs such as equipment (board, bindings, boots, helmet, etc.), lift tickets, lessons (if you're a beginner), and travel to snowboarding destinations.
Climate and Location: Consider your proximity to snowboarding locations and the availability of snow. If you live in an area with access to mountains and snow, it might be more convenient to pursue snowboarding.
Risk Tolerance: Like any sport, snowboarding carries some degree of risk. It's important to take safety precautions, wear appropriate gear, and be aware of your own limitations.
Social Aspect: Snowboarding can be a social activity if you have friends or family who are also interested in it. It can be a fun way to bond and spend time together.
Other Interests: Consider your other interests and commitments. Will snowboarding fit well into your lifestyle and schedule?
Ultimately, the decision to pick up snowboarding should be based on your own preferences and circumstances. If you're curious and excited about trying it out, you might consider taking a few lessons to see how you like it before fully committing.
Turing Test: Failed
Those are pretty small for "large" prime numbers. RSA security, I believe, is all about 250+ digits in prime numbers, multiplied together. It would be interesting to know if ChatGPT has some kind of guard against using up too much compute for answers like that.
It got the answer badly wrong. Multiply them and see for yourself...
I stand corrected. I assumed it just did what a computer does, and that you were simply calling attention to the fact of it being different from what a person would do.
I'm not sure most of these reflect a lack of capability to pass the Turing test more than just reflecting the fact that chatGPT isn't even trying to pass the Turing test. I think a model based on gpt-4 and then fine tuned explicitly to pass the Turing test would probably pass at least the majority of the time
I’d have to agree with Jack here. I inherently don’t thing it can pass the Turing Test. And I don’t think fine tuning would fix that
Well what does "capability to pass the Turing test" mean? I agree that it doesn't prove that no AI can pass it, or that no AI of approximately ChatGPT's "ability" (whatever that means in this context) can pass it. But the original comment/link was about ChatGPT and the Turing Test.
Also, like I said, I doubt that if you tell it to try and pass the Turing Test it could. I'd say that's it lacking the capability to do so. I'd analogize it to having ChatGPT play chess - I bet it can do so, but not as well as other specialized computer chess programs.
Properly tuned it could probably fool most people, maybe even most readers of this blog. But the goal of openai isn't to fool people into thinking it's a real person but to make a maximally helpful AI assistant (significant but not total overlap as goals.)
And actually you can get chatgpt to role play as a person and it will provide answers like "idk lol"
None of these sound like fundamental limits, and all could be overcome by prompt engineering (at least with a non-nerfed LLM).
Interesting point. But How can a lack of reasoning capability be overcome by prompt engineering? And surely the point is that prompt engineering should not be relevant for a truly intelligent LLM
I watched Barbie in a sold-out theatre and it was surprisingly good. The acting, the editing, the writing, the costume and set design all came together to make a truly post-postmodern film. It's not coded for radical progessivism and treats the Barbie-Ken conflict with even handedness and empathy. I guess it shouldn't be surprising that a production with A-list actors and a top shelf director would be able to bring on great screenwriters, but I was not expecting that level of nuance from Mattel. Makes me wonder if we're going to get a tearjerking deconstruction of manhood in GI Joe.
> It's not coded for radical progessivism and treats the Barbie-Ken conflict with even handedness and empathy.
It's a little messed up that the Barbies thought they had to prevent half of their population from voting in order to save their democracy. Destroy democracy to save democracy? And oddly paternalistic.
It's funny, I thought it actually lacked nuance. Like it flirted with engaging with complex topics, but then tossed them aside with a joke and retreated right at the point that it could have made a statement about them.
For example - Barbie shows up in the real world and kids call her a fascist for promoting impossible beauty standards. Just when it looks like Mattel might actually engage with that topic, rather than have Barbie rethink ideas about beauty, or change how she presents herself or what she teaches kids, the film pivots to a joke about the trains running on time and drops the subject.
Or how the film sort-of critiques overrepresentation of men on corporate boards by giving Mattel a bumbling, all-male board of directors, but shies away from saying that the solution is putting more women on the board - in the end, the all-male board serves just fine to smile and sign off on Gloria's new Barbie concepts.
Or how it whiplashes the audience back and forth from "Barbieland is a paradise where one gender of doll serves the other" to "oh no! Barbieland is becoming a dystopia where one gender of doll serves the other!" but then rather than learning a lesson about gender dominance or settle on any kind of fixed or equal Barbieland at the end to serve as a statement about what the filmmakers think good gender relations look like, the film just ends on Barbies restored to power and a quip about how if that's a problem for you well maybe someday Kens will have as many rights as women in the real world. It's funny, but it's not like it gives the audience any sense of what the creators think justice would actually be.
Overall I really enjoyed the movie - especially the absurdist/surreal comedy bits, and Ryan Gosling is perfect as Ken. But it also kept reminding me of Lindsay Ellis' observations on "Woke Disney" (https://www.youtube.com/watch?v=xU1ffHa47YY) and how corporate media can try to sidestep issues by "acknowledging" them in a side-joke while carefully avoiding making an actual statement.
I actually feel the parts you complain about were the strong parts of the movie. For example, the fascism comment seemed to be a joke on the girl saying it, not on Barbie - the stark constrast between the dull browns of the kids clothes and their angry lashing out vs the cheerful happiness of Barbie. The movie acknowledges that these issues exist while kind of showing how ridiculous it is to take them that seriously.
I also like that they didn't deliver heavy-handed messages about how putting women on the Mattel board or declaring Barbies and Kens equal. They showed the importance of these things clearly enough just by highlighting the contrasts. Stories, I think, should explain the importance of issues through the way they depict the world, not by hammering in their moral takeways to their audience.
Spoilers/
I thought the character arcs resolved in a satisfying way. Gloria reconnects with her daughter, who is more sympathetic to her mother. Barbie exits the cave/takes the red pill. Ken awakens to the fact that his value doesn't depend on Barbie's gaze. The institutional/societal conflicts aren't tied up in a bow, but that's a moral in its own way right? It's on us, the audience, to be agents of that change.
/end spoilers
I thought that the movie had a lot of good parts, and was really funny in places, but overall I disliked it. I think that I could really never get past that the whole political dynamic of the Barbies vs Kens, and how both sides were awful. I had a few other complaints too, but I clearly wasn't the target audience, so whatever. While I thought that it was pushing certain political points that I disagree with, as an actual movie, I'm not going to criticize its quality too much when it wasn't made for me.
The idea of a pro-feminism Barbie movie ought to be incoherent, at least under the theory that feminism is the war of ugly women against attractive women.
Where do you get that theory? Looking at the kinds of things feminists profess to want: (a) access to abortion, (b) elimination of a gender pay gap, (c) more women in positions of power (corporate boards, SCOTUS, etc), (d) less enforced monogamy and more freedom for non-traditional sexual/coupling practices, and so on, it doesn't seem like any of them, if achieved, would offer unattractive women any obvious advantage over attractive women.
Actually, if anything, traditionalism at seems more likely to advantage the unattractive (both men and women), at least to some degree. If attractive people are all out there sowing wild outs and dating around, then they're going to be occupying multiple suitors on the market competing for their attention. If sex is enforced into an after-wedlock-monogamy context, then there's less wild oats and attractive partners will go off the market faster, presumably giving less attractive people a greater chance of finding a partner sooner in a less competitive pool.
> Where do you get that theory?
Various neo reactionaries.
Are you suggesting that 'feminism' is the political manifestation of, or the reaction to, toxic femininity?
The movie is at least adequately coherent. Your theory of feminism is at least incomplete.
It is a strange movie, although I enjoyed it.
1. 75% of the film is whacky and/or absurdist comedy. It genuinely is quite funny - I think I laughed the hardest at the "wrong way wall" joke.
2. 25% of it is almost didactic, incredibly on-the-nose commentary - including some straight up speeches.
It's an interesting combination that actually works pretty well, although the ending felt a bit messy.
I thought it was fine. It really felt like a haphazard collection of three movies that were at odds with each other.
1) A "Toy rules" cartoony movie, where funny stuff happens because that's how you play with toys
2) A pro-feminism movie, examining the difficulties women face in the world
3) An existential movie about a toy coming to life and experiencing complexity
Any of these could have worked on their own, but they came together to be less than the sum of their parts. The funny toy-logic bits feel jarring next to emotional segments that they do little to transition between, the existential feelings just make you question how this toy world actually works, and the feminist messaging never gets the time or attention it needs. A few things stand out as particularly jarring though
The entire board of Mattel being men looks like it's going to set up a cynical examination of male dominated spaces deciding things for women, but then they turn out to be super childish incompetents who mean well but don't do much. Every scene with them just feels like it was lifted from another movie entirely.
There is a sequence where the barbies literally manipulate and play the kens against each other for personal and political power. Taken on its own, it is something that could have been written by an angry anti-feminist trying to show how evil those bitches are. The entire thing is a HUGE mis-mash of ideas (they get *mansplaining* wrong, which feels very ironic to try to explain, for example) that just doesn't work.
And ultimately, its feminist message is super shallow. It just basically states the surface level of the problem, and does nothing with it. It just says "Society has contradictory standards for women that makes life hard" and "Women are often excluded from spaces that make decisions for their lives". Like, no shit. It has nothing to say about how to move forward, or what caused those issues, or even what it's like living with them outside of just stating they exist.
Overall, it was decently enjoyable but disappointing. It can't commit to its message enough to really make a statement, its funny toy sequences are too disconnected to draw you into the fun, and the existentialism is super tacked on. It's a forgettable movie that hints at the potential for being so much more
Sounds like your experience was the opposite of mine. I thought the contrast between the toy/childs play logic and real world forces was well done and enhanced the movie. It was very self-aware and cued the audience that the producers were in on the joke.
The feminist angle was interesting and empathetic to me and I'm not in a Barbie demographic. One of the central themes was lopsided dominance/submission heirarchies are bad mmmkay. Also, turns out that men have complex inner lives and are profoundly impacted by how women treat them, as objects or subjects. It didn't strike me as. hamfisted critical theory.
The Mattel board scenes made me laugh because it was on-point satire. My profession has a strict heirarchy, centralized authority, and lots of bureaucracy. The board could have been a General's staff, it would've worked just as well.
I actually agree with the point I think the movie is trying to make. Imbalanced social structures are bad, and women have it much worse in the real world than men do, and social expectations are a cause of a lot of harm. But the movie doesn't do a good job conveying it, in my opinion. The Kens seem mostly happy in the first part, and the Barbies ALSO seem happy once the Kens take over. There's no shots of the lonely village the Kens have to live in, or the empty lives they live other than Beach Ken wanting to stay with Barbie. There's no shots of the newly-submissive Barbies straining against their roles or being remotely dissatisfied with their loss of station.
Overall, I don't think it was terrible. I just wish it was more. But I'm also not, as you put it, in a Barbie demographic. It's entirely possible that for girls and young women hearing the struggles they face put to words is a powerful moment, and I wouldn't want to spoil that just because I wanted the movie to go further than simply stating the problem
I don't know what the feminist angle in Barbie *could* be, because Barbie was always a toy for little girls with Ken very much an afterthought, indeed more an accessory than anything. Barbie had it all, and could be anything or anyone she liked. Barbieworld *is* a girl's world, and trying to make it into some kind of parallel with the real world doesn't work, because you look at Barbieworld and go "Women have no problems there at all, in fact the Kens are being downtrodden".
The disconnect you're describing between Barbieworld and the real world is explicitly commented on by the film itself and is a major part of the the "feminist angle" of the film.
It kinda tries to do the whole "Kens are downtrodden" angle, but it can't commit to it and bungles the execution. It could have used the Kens to show how women are treated in reality, but instead it just makes one comment about "maybe one day the Kens will be treated like women are in the real world" and does nothing with it.
Yeah, I'm only going by online clips and various reviews which are all over the place: it's feminist, it's anti-feminist, Ken is the real hero, it's all about the patriarchy, go see it, don't bother going and so on.
I think the best take on it that I have seen is that this is not for little girls but for their mothers who will be taking them to see it - the women who played (or didn't play) with Barbie as kids because the anti-Barbie 'this is not feminism' reaction was in full swing and playing with gendered dolls like Barbie, which was presented as being dim and only interested in fashion, was not considered a good example for your daughters. Gen X girls played with Barbie, their daughters may not have done so, and now they're the Millennial mothers of the next generation of Gen Z daughters who would *like* to let their girls play with Barbies but aren't sure what the official position on it is, so the movie which presents "patriarchy bad for all, women suffer in the real world, Barbie becomes empowered" is a reassuring message that you can live your life in pink dreamhouses *and* be correct on the issues 😁
This permits both Mattel as the toy company wanting to revive the brand and sell sparkly pink plastic junk to little girls, and the director/chief cook and bottlewasher with her impeccable feminist credentials, to meet in the middle and make what is in effect a movie to sell toys. Barbie got overtaken by the newer generation of dolls aimed at little girls (think Bratz and Winx etc.) and badly needed a brand refreshment to be in line with Modern Day Audiences, and Mattel managed this a heck of a lot better than Bud Light did!
You should try going to see it instead!
The movie was written by the director (and her partner)
This prompted me to check Wikipedia and Gerwig wrote, directed, and produced. Barbie is an auteur film. Isn't it a little strange that a toy corp that's synonymous with plastic schlock would licence one of their biggest IPs for an auteur piece? Respect to the production team for pulling it off, from pitch to picture.
They were initially very skeptical about the scene where Sasha tells Barbie that she's a tool of fascism and gives all the other standard left-kid critique, but they let her go through with it, and got themselves a winner.
As I recall, she says Barbie is a fascist.
"She thinks I'm a fascist? But I've got nothing to do with the railways!"
Barbie throwing jabs at Ayn Rand wasn't on my bingo sheet.
Meta comment:
The Richard Hanania - Richard Hoste news was discussed in an open thread on here months ago. Pretty much all the info in the Huffington Post thread was already known/very easy to find.
The commentators here were mostly furious at the idea and considered it to be cancelling and doxing. Comments were then deleted. No real discussion took place.
I wanted to circulate this for discussion. Is it better to discuss this stuff in these communities and confront the issues? Or has this community decided its preferable to have a rigid anti- cancel position and to try to hide all such conversations?
My position is that as long as it's not common knowledge, i.e. not something most people who know Hanania (professionally, personally or online) know, it shouldn't be discussed. Any discussion of it moves it closer to being common knowledge, which is bad given the inappropriate way the information became known. Once it is common knowledge, or it's inevitable that it will become common knowledge, it should be OK to discuss it.
If it took months for it to be noticed by a mainstream outlet, presumably there was some chance that it would never be noticed, so it was right to delete the conversation back then.
If we’re giving our meta takes, I think it is absolutely disgusting for the Thought Police (sorry, “journalists”) to scour hacked databases in order to link pseudonymous accounts to real people. Hanania used a pseudonym because he wanted to discuss ideas and search out the truth without having his personal life ruined. Indeed, some of his past claims were in fact debunked, but some of them weren’t. The glaring omission in the HuffPost article is the lack of any factual arguments that any of Hanania’s current or past views were incorrect.
If you look at bigotry like that and think the person responsible is trying to "search out the truth", the kindest thing that can be said about you is that you're a mark.
And the wrongness of ideas like "some races are better than others" does not need to be re-proved every time it comes up in the news. It's old hat.
Has its wrongness been proven in the first place? Do you believe it's wrong because you think it's been proven wrong, or because you think only a bad person would even entertain the possibility that it isn't wrong and investigate it at all? Do most people?
Of course it depends a lot on what we mean by "better"—a factual claim about averages, a factual claim about all individuals, a claim about moral worth?—and by "wrong"—factually wrong, dangerously imprecise, morally wrong?
Factually wrong, believed mainly for morally wrong reasons.
And yes, it's been proven many times over.
Then it would be worth to, if not re-prove its wrongness every time it comes up, at least prove it every once in a while. I hardly ever see any positive evidence at all that it's wrong in mainstream venues. I do see arguments to the effect that some particular reasons one could think that some races are better than others don't actually show that, but that doesn't prove that it's not the case.
Tell you what, if I ever write my big "Why HBD is bullshit" effortpost, I'll try and remember to message you about it.
Given that he's a highly influential writer with potentially real political impact (he's got the ears of some pretty wealthy and powerful people), I think there's a strong case for the newsworthy relevance of knowing that he was a super-racist writer in the past even by his current standards.
Hanania used the Richard Hoste identity less than 10 years ago and wrote articles for a plethora of explicitly racist websites like whiteamerica dot us, in addition to operating an entire website dedicated to advancing these arguments. He didn't just make a few edgy or borderline comments in a comment section. I also note that Hanania didn't disavow these writings until publicly outed.
If he had disavowed himself of these views before being outed, that would have been as good as outing himself. I'm not exactly sure where I come down on the issue, but that one point bothered me.
For people not to trust him. Because he's obviously an untrustworthy bigot, and should be treated as such.
Trustworthiness matters if his writings rely on claims you can't easily check, so you need to trust him, or else there's no point in reading him. It doesn't matter if his writings mostly rely on arguments you can judge yourself, and claims you can easily check. I haven't read enough from him to know which one is the case.
Richard Hanania turned out to be racist? Like a month ago he was dismissing the efficacy of industrial policy in industrializing the Four Asian Tigers and blaming it on their IQs instead. I didn't realize it was supposed to be a secret that he was racist?
Well, this is the problem with the current massive overloading of the word "racist" to mean a vast variety of different things.
What I think is fair to say is that he previously held views which are more extreme than his current views, and that some of his previous views would be considered objectionable by people who share his current views.
Speaking as someone who has likewise slowly matured from edgy teen to more sophisticated adult, I can sympathise with this. If you don't hold at 20 some political views which you think are stupid and naive at the age of 38 then you've probably never thought for yourself.
(Note: I've got nothing against people who never think for themselves, they're generally better company than people who do, but it's a bad quality for a political writer to have.)
Without torturing the language much if at all we can say that he used to be the ugly, crass kind of racist, and is now the smart, cool, moderate kind of racist. I characterize him in this way only because I think he would agree with it and in fact I am echoing his sentiments in his recent explainer piece on this very subject.
However, and this is the part that is important enough for me to bother leaving this comment at all, it is obviously the case that he was the former type for reasons that were transitory and not connected with who he really is. An angry young man saying things in a shocking way deliberately to get people riled up can very genuinely grow up to no longer say those kinds of things.
Believing that the relative performances of various racial or ethnic groups is primarily determined by genetic factors rather than legal, environmental, or sociocultural ones, particularly (but not necessarily) when combined with beliefs about certain groups being superior or inferior to others.
Why not just stick with "racial prejudice", particularly as applied to individuals? It could actually turn out to be factually true that there is some genetic component to group differences, in which case you'd be saying "believing facts" is racist. Which is stupid.
By contrast, any such facts are irrelevant to prejudice as applied to individuals, because you can't a priori know where an individual might fall within their group.
I believe that to be a politically-gerrymandered and intellectually dishonest definition. For starters, the role of genetics is an empirical question while 'racism' is a philosophical one, and it's a basic category error to argue that a factual claim could imply a philosophical stance. Plus I don't think you'll find any anti-racists who would make their condemnation of racism contingent on any sort of empirical evidence - even if green people have lower IQs that doesn't justify denying their legal rights. In my view, the only reasonable definition of racism is the conviction that some races are morally and politically unequal - meaning, specifically, that they are not deserving of equal treatment under the law. Any broader definition starts to quickly infringe on things like reasonable criticisms of countries/cultures/groups which is part of a healthy public dialogue - which many anti-woke types (myself included) would argue is currently happening. Saying 'green people suck' isn't any more racist than saying 'republicans suck'. Maybe they DO suck, after all, and maybe we need to talk about it.
> and it's a basic category error to argue that a factual claim could imply a philosophical stance.
Did you mean a normative stance? Because bplenty of philosophical claims are bout the existnce or non existence of various things.
Agreed, normative is probably the better word. But I disagree about your other claim. The moment a claim is testable it's not really philosophy anymore, is it? Curious if you have a good counterexample.
I think this definition does not reflect the general usage. Compare "sexist". Noting that the median man is physically stronger than the median woman does not make one sexist. Overgeneralizing that into "women are weak" and refusing to consider a woman for a job requiring physical strength is sexist.
If you look at the marathon Olympic medalists, you will notice that a lot of them come from certain African countries such as Kenya or Ethiopia. I think any attempt to explain that solely by culture and social priorities will turn ones mind into a pretzel: it is not like marathons are unpopular in the west, or that China refuses to compete in marathons, or that Ethiopia can afford much better training programs than the US. From what little I know of the field, I would assign quite a high probability to "peoples optimized evolutionary for savannas are on average better long distance runners than peoples shaped by other evolutionary environments".
By your definition, that makes me a racist.
NB: I generally consider it bad form to dismiss an argument because it is verboten. I do not know the first thing about Hanania or Asian Tigers, but clearly you would not be willing to consider a genetic component of IQ as a deciding factor Because That Would Obviously Be Racist. Basically, you wrote down your Bottom Line [0] before you wrote your argument.
[0] https://www.lesswrong.com/posts/34XxbRFe54FycoCDw/the-bottom-line
> Believing that the relative performances of various racial or ethnic groups is primarily determined by genetic factors
If this is your definition of racism then being a racist is a good thing because it means you believe a true fact. I however think that this is simply a bad definition of racism because the word has connotations of animosity and hatefulness which this definition does not capture and thus if you use this definition to label people as racist you will incorrectly communicate the levels of hatred and animosity expressed by people who are aware of genetic differences but are not hateful (like me)
Doxxing isn't okay.
How is it "doxxing" when Hanania writes under his real name?
A bunch of the stuff he had posted before was not under his own name.
Um, so a link would be appreciated.
For whatever it's worth, I would prefer less discussion of internet drama and feel about as good about participating in it below as I do eating a bag of Skittles. In general, doxxing and internet drama is...not a super high value use of anyone's time.
Hey, I don't give a shit about the drama, I just want to know whether Scott removed certain posts or asked people to. I don't even have a strong opinion about how such things should be handled, just curious how he handled it. I have never seen any post here removed except by the authors themselves who got a lot of angry pushback and left in a huff. I've never seen Scott object to any posts except those that violated true, kind, and necessary. Occasionally things are discussed that are a illegal in a minor kind of way -- for example, use of psychedelic drugs -- and Scott has made no objection.
I don't remember removing a post like this, but it was almost a year ago and I might have forgotten.
Daniel says “ The commentators here were mostly furious at the idea and considered it to be cancelling and doxing. Comments were then deleted. No real discussion took place.” He’s saying multiple comments were deleted. But were they the ones by people who were furious abt what was happening to Hanania? Why would anyone delete those? Was it some *other* batch of comments? What was in those?
I assume it's this: https://astralcodexten.substack.com/p/open-thread-252/comment/10757652
Thanks!
“Or has this community decided its preferable to have a rigid anti- cancel position and to try to hide all such conversations.”
I have never known this community to decide anything about it’s position. The community doesn’t have a position about events like the one you mentioned. Individuals on here have positions, and often argue quite vehemently about them
A bunch of comments here were deleted? Do you mean they were deleted by their authors? By someone else? And about when did this happen?
I would expect Scott to have a strong anti-doxxing position, even if the information leaked could be found with a search engine.
Of course, Scott is also very transparent with his moderation. Typically, the posts which got someone banned would still be readable so that we can go "yup, totally worth a ban". I assume that doxxing is one of the types of posts where this would not work.
The solution I would go for is to transparently edit out any doxxes and ban as appropriate.
In the spirit of my article on self-promotion, please read my article on self-promotion:
https://philosophybear.substack.com/p/in-defense-of-self-promotion
The central argument here seems to be:
All avenues for promotion - other than yelling your own name into the void - are narrow and gatekept.
I think that ... sounds about right.
What do people think about this room-temp superconductor thing?
Videos of floaty rocks seem to be trickling onto the internet but there isn't much by way of solid evidence from institutions yet
As far as I'm concerned, https://arxiv.org/abs/2308.04353 is the final nail in the coffin.
When the source of "anime Soviet catgirl claims to have made it in her kitchen" is taken seriously, you know that "videos of floaty rocks" are no evidence whatsoever.
I freely admit ignorance in the field, but I just saw an interesting video (tweet linked: https://twitter.com/a_j_higgins/status/1687816717441916929) showing floating (diamagnetism) and something called “flux pinning”, which Wikipedia (https://en.wikipedia.org/wiki/Flux_pinning) assures me is a certain sign of a Type II semiconductor.
Personally, I continue to wait for more (and different) evidence. But it seems flux pinning is another piece.
Still waiting for some truly reliable replications in published papers. If it works, then I'm not as worried about the tricky synthesis - there will be a huge amount of work put into trying to improve that.
Given the crystalline structure, I actually wonder if it would be easier to make it in weightlessness in space versus under gravity on Earth.
Assuming it's not just fake, it seems the synthesis process is super unreliable - you shuffle a bunch of metal atoms around under high heat, and *sometimes* they line up in a way that's superconducting, or diamagnetic, or some other cool thing, but most of the time you get a normal non-floating rock. So while it's still possible we get some interesting physics papers out of it, it might be hard to get practical applications.
Still no replication from a reputable source after two weeks (and the synthesis process as described should be very easy). Nor have they given their samples to any `respectable lab' to measure, as far as I am aware. This is starting to smell fishy. My guess as to whether there is any `there' there is now below 5% and dropping by the day.
I don't understand why any rationalist would deviate strongly from the prediction markets view.
The discussion is valuable in and of itself. Saying "I obey Manifold" when you've got something interesting to contribute deprives you and everyone else of that discussion, even if a personal prediction would be less accurate than the market.
I think it is perfectly reasonable to disagree with Manifold by 10%, but a strong disagreement witt manifold most likely indicates someone is making an error
If someone assumes something is 0% or 100% when manifold is 20%-80%, I am very likely to think they are just bad at making predictions.
But manifold isn't a big money market, right? I would take it seriously if people were taking million dollar positions, but that isn't the case, right?
The polymarket one has over $1.5 Million bet and is within a couple percent of the Manifold one, which suggests that the manifold market is "large enough" to at least be reasonable.
https://polymarket.com/event/is-the-room-temp-superconductor-real
https://manifold.markets/QuantumObserver/will-the-lk99-room-temp-ambient-pre
-edit- I realize that a possible interpretation is that you wouldn't take the market seriously unless multiple individual traders had million dollar positions. I think I pretty strongly disagree that this is a reasonable threshold. I think markets become useful far _far_ before that level of investment.
For a market to be smart there presumably needs to be a critical mass of ‘smart’ investors (with expertise in evaluating the object level proposition). How many superconductivity experts are betting on this question on poly market? Do we even know that this number is more than zero?
Hmm... I'm almost tempted to trade. But I fear the risk of losing my money because the crypto value drops or because something funny happens with polymarket more than I fear losing it on a trade. It seems like there is a double-risk to playing this market.
I'd short the hell out of the superconducter being real, myself. Which would mean I don't have faith in this market being accurately priced, which makes me worry there is something funny about this market....
But thanks for the info. Appreciate it.
What about a rationalist who participates in prediction markets?
The theory calculations sound like they suggest that there's something real and interesting about the material, whether or not it actually has superconductivity at actual room temperature. It sounds like a significant and interesting development regardless, given that it's a type of material lots of people have no trouble making, but hadn't investigated before.
Theory calculations which make some strong assumptions suggest that the material could host correlated phases. Nothing in the theory that has been published points specifically to superconductivity, nor to room temperature.
https://manifold.markets/QuantumObserver/will-the-lk99-room-temp-ambient-pre
It’s nothing.
I agree it seems pretty unlikely, so what explains the positive evidence given?
To the extent anything's been reproduced, the material has strong diamagnetism. This is a property of superconductors but there are diamagnets that are not superconductors. The material's other physical properties have so far been under-characterized, at least in the preprints that announced the claim.
This looks to be the case. Bit more diamagnetic than bismuth, bit less than graphite. Probably a semiconductor.
There can be a myriad explanations from honest mistake to hardware malfunction to outright fraud. Not going to speculate.
So! How is everyone enjoying Baldur’s Gate 3, scale of 1-10?
I’m surprised by how much I like it. I play crpgs pretty much exclusively but the Divinity Original Sin games didn’t click with me. And BG3 is definitely a successor of DOS, not of BG2. But I’m about 10 hours in and am digging it. Interesting companions, great character development, fun combat, great graphics, interesting plot so far. I’m sold
I'd say about 6. It's a competently made game, with good graphics and excellent music; in fact, their background/foreground music system really stands out (in a good way). But the game itself is... meh. The gameplay mechanics are based on 5e, so there really aren't any meaningful choices you can make when leveling up your character (besides the spell loadout). The game starts with a really cool intro that really got me pumped up for the content... and then it put me in a generic dungeon with loot inside. Don't get me wrong, I like loot, but I expect more from a roleplaying game; I expect story, character development, and something that hooks me and makes me want to explore the world and pursue each character's story to its fullest. "Planescape:Torment" had that in spades. The "Dragon Age" series had it. Arguably, the original "Baldur's Gate" had it. Maybe BG3 has it, maybe not, but I'm like 4 hours in and so far I haven't found it.
I definitely hear what you’re saying, and so far BG3 has some of the problems that made me impatient enough with the Divinity games to give up (those endless multi-level maps of towns with 100 NPCs all of whom need to be spoken to twice, for example). But it DOES get better than the first few hours, including the character development. It’s not Pillars of Eternity 2 or Wrath of the Righteous, but there are choices to be made, and you can mix and match different classes.
Planescape: Torment was my favorite game ever until Soulsborne games came around, but I think of it (and the first two Fallouts) as a bizarre joyful accident that should never have been made. It’s way too weird and awesome to expect anyone to really recapture what made it special.
I’m curious if you think BG3 gets better. I did, though yeah, it’s dragging a bit right now.
I might give it more of a chance later; but for now (speaking of souls-likes) I'm too deeply immersed in "Remnant II". Its world (worlds, technically) feel much more real and immersive than the Faerun of "BG3"; and I feel much more involved in its story despite its relatively small volume of text/dialogue. I'd say that "Remnant II" is mostly "show", whereas "BG3" is mostly "tell".
I’m hoping to like it, but I am a grognard, a dyed in the wool partisan of the originals, and pretty well despise Divinity Original Sin, especially Lariat’s writing, so I have a hard time believing I will. Comments like yours give me a little hope.
I like Divinity Original Sin and Divinity Original Sin II, so you're probably sneering down your nose at me right now and I don't blame you 😁
When I was first dipping my toe into playing computer games, I tried Beyond Divinity, which I liked the plot but couldn't play well enough to get to the end. So when years after I heard about Divinity Original Sin, I decided to give it a go.
It probably has all the flaws everyone is mentioning, but for someone who didn't grow up playing games and has no experience apart from "trying them at home on my PC", it was just the right level. Besides, I was playing more for the plot and exploring the world and the funny incidents (in Original Sin II I managed* to become the gay side piece of the Red Prince so that's going to be an interesting future in-game universe) rather than grinding out levels so I could oneshot every enemy with the best gear and combinations of attacks.
*Look, I wanted to see would the game Go There with all the teasing they were doing and fair's fair, they did indeed Go There and er, well. Now I'm a homewrecker 😀
This is my sneering face. I couldn't finish either of the "Divinity" titles; I kept falling asleep. Try "Planescape:Torment" instead. It's like tasting a curry buffet after subsisting for months on porridge.
I was just as surprised. A lot of what I didn’t like about DOS - the rock/paper/scissor combat, the endless useless inventory crowding items, the easy to miss vital spots on maps, and above all the equipment that need to be upgraded every fucking level, has all been eliminated or greatly reduced.
That said: this isn’t a spiritual successor to the originals. You’ll be disappointed if you go in thinking it is. Hell it actually referenced DOS2 twice already.
I almost didn’t mention they last point because it isn’t actually a big deal. I think the game is really worth it
Does the vote by the Nigerian Senate against military intervention mean that the Nigerien coupists are now secure in power?
(as a reminder, Nigerian means "of Nigeria", and Nigerien means "of Niger")
Could France invade without their help?
It means no Nigerian/ECOWAS intervention, but I don't have a great sense for French politics and their own willingness to get (more directly) involved
Could, yes. But it never would. It'd cost too much, be too great of a risk:
-Military-wise, wars rarely end as soon as you hope they do. We got stuck in Mali despite being invited by the regime, the US gave us a good example of how well nation building works in Irak & Afghanistan
-Politically-wise: both the left and the right opposition would have a blast criticizing an african invasion
-Diplomatically-wise: We get enough shit from various African countries, there's little to gain by giving more ammo to those who cry of "neocolonialism"
Not to mention the risk of the invasion actually not going that well. It's good to be perceived as a power that can't really be challenged. It'd be a disaster if that perception was shattered by a failed invasion.
Most of the comments I have read were very skeptical from the beginning that a military intervention will happen. The Nigerian military has other problems and is very clearly against intervening. One analyst explicitly said that it could make Nigeria a candidate for the next coup if the government tries to force their military to intervene in Niger. Especially if the Nigerian generals feel that by these "distractions" they might lose control of parts of their own country.
It's hard to understand the internal state of Niger. Some of the anti-French demonstrations were bought by Russia, but that doesn't mean that they were all bought, or that the anti-French sentiments aren't real. On the other hand, they might only be real in a minority of the people, and it's very hard to tell how small or large this minority is. And it is even harder to tell how much people support the old vs the new government. But it doesn't sound like anyone is expecting a counter-coup or even a revolution against the coupists.
Coupists are never secure in power almost by definition.
It probably reduces the chance of a direct ECOWAS intervention, but I am not sure that was ever that likely.
I guess we’re about to find out….
https://www.dw.com/en/nigeria-senate-cautions-against-niger-military-intervention/a-66450388
I didn't include a link because almost all the comments on (( the site formerly known as Twitter )) are from pro-Russia activists, and the American media is either ignoring it or giving a generic "this is bad for democracy" take.
The background is that, after the Niger coup, ECOWAS (a regional group led by Nigeria) threatened military action. But the Senate of Nigeria voted down President Tinubu's proposal to authorize international military action. As Nigeria is the largest member of ECOWAS (and has a 1000-mile border with Niger), it is unthinkable there will be a military effort against the Nigerien coupists without the military of Nigeria.
Great example of the Lizardman Constant in action: this survey https://www.equimundo.org/resources/state-of-american-men/ claims that "40% of all men say they trust one or more “men’s rights,” anti-feminist, or pro-violence voices from the manosphere." They got this figure by asking men about a list of problematic websites and adding up everyone who said they trusted at least one. 5% of Gen-Z respondents claimed to trust r/braincels... which hasn't actually existed since 2019! And the authors of the report just took that number at face value.
Knowing absolutely nothing else about them or their research, the fact that this organization considers "men's rights", "anti-feminist", and "pro-violence" to be a single category... Raises a few red flags, to say the least.
For the record, the single category they use is the "manosphere" which includes sites binned into the subcategories you list, plus an "anti-feminist" category (e.g. Jordan Peterson). Interesting Joe Rogan wasn't included as the leader of the "broletariat"
Concern trolling org discovers trolls, remarkable.
Can anyone here provide an overview of various approaches to polygenic trait prediction? For example, are neural networks better than traditional regressions; what is the state of the literature on the topic; what non-credential requiring datasets (for any species) are there with full genotype/SNP-phenotype pairs; does simulating data with normally distributed noise work for comparing approaches in the real world.
I could provide some answers but it would help if you would tell why you are asking!
Predicting massively polygenic traits (schizophrenia, bmi, cancer risk) using sequence data seems like a thing that could be improved with new neural net architectures, and I was curios what exactly the state of the field was and if doing my own project there would be helpful
Neural networks have been used to predict polygenic scores for only a few years, see for example the papers in the list below.
https://scholar.google.fr/scholar?hl=fr&as_sdt=0%2C5&q=polygenic+risk+score+%22neural+network%22&btnG=
My understanding is that at the moment we are still limited by the data themselves, perhaps more than by the methods of analysis: in diseases such as schizophrenia, the architecture of the trait is a mixture of common variants that increase the risk slightly and very rare variants that can increase it a lot. These very rare variants are mostly in non-coding regions or are copy number variants, and they are difficult to genotype because they require whole genome sequencing.
Note that I am not an expert on this: I am a geneticist and I sort of follow the literature on the subject because I find it fascinating and I teach it, but I work in a different subfield.
Does anyone have recommendations on how best to find jobs that are meaningful, in the EA sense? For example, something in politics that helps with justice system corruption, or something in animal welfare that helps end factory farming, etc. The issue is that my primary skill is computer science, and i'm unsure how to find computer science related jobs that are meaningful (no AI related recommendations please).
All jobs are meaningful in the EA sense (within a margin of error).
Pay attention to the fact that every job is, ultimately, helping somebody else. And since help is, at least to some extent, fungible, then every job is helping the world get better.
(Within a margin of error. There are jobs which are counterproductive / destroy value, but they're the exception rather than the rule.)
Just to throw it out there, networking can be helpful too. Find someone who has a job that you would want, and then email them asking to chat about how they found it, what it's like at their company, etc.