Man, the question of whether schools should have a prohibition on professors and grad students dating was the topic of probably the single nastiest extended argument on the internet I've ever seen...
I followed Gwern at some point and have been puzzling for over a year now to try and figure out who he is (not like, his identity, I know he’s pseudonymous). What does Gwern do? What is his thing? Who is Gwern??
Regarding 35., there are some interesting stories regarding Stalin and litterature. A famous one is the relationship between Stalin and Mikhail Bulgakov (who wrote the falmous The Master and Margarita). It seems Stalin and him exchange letters and Stalin protected him despite Bulgakov being quite critical of his regime.
So, some people use artist names in AI art prompts, which is lazy even if you completely dismiss all ethical issues. Stable Diffusion (the FOSS one people use for porn) really incentivizes this in its prompting meta, because SD's prompting meta is generally a mess. One thing not that many people on all sides of this ideological hypercube notice is that *this is noise* -- for all but the most famous artists, the association between their names and their styles is pretty weak. It's less that SD demands you do that specifically and more that it's wound so tightly it wants you to write noisy twenty-billion-token prompts (though there is not *no* degree to which it wants you to do it). Glaze is built under the assumption it isn't noise, which means it doesn't actually do 'the thing it should do' all that much, but I'm no fan of prompting with artist names and would rather see fewer cringe prompts, and if all else were equal I'd be cautiously optimistic about something that could bring a ceasefire across various parts of the hypercube, even if it's placebo.
Here's where things get *really* funny, considering the copyright infringement accusations that fly around all AI art: Glaze infringed the hell out of its licensing. Not in, like, an ambiguous fight-it-out in the courts way, in a "massive chunks of plagiarised code in violation of multiple incompatible licenses" way:
This is fixed now, which is nice. Nonetheless, it's kind of incredible that people who take the hard "all generative art is plagiarism" stance did...unambiguous plagiarism.
According to this link, the increase in partisanship *is* universal and simultaneous across countries in a way that’s hard to explain without invoking the changes in media structure and funding.
I think it took a long time to invent brazilian jiu-jitsu because unarmed one on one combat between people matched by weight is a pretty niche activity not subject to much optimization pressure.
Also short stories probably got replaced by tv episodes and comic books
Tulsa article is very interesting, trying to think of a good pretext for forwarding it to some people I know.
One thing that the linked article didn't mention is that Tulsa Benevolent Dictator for Life George Kaiser is 79. I'd be worried about what happens when he passes; will his heirs just rug pull the entire city and move to a coast?
If those dense cities don't have any farms, then they have to import their food. In order to import food, they have to produce goods and services that farmers will trade for, or goods and services that are desired by other people who make goods and services that farmers want.
This is basic international trade economics, but with borders drawn around arbitrarily small regions instead of large nation-states.
So if these 15 minute cities don't have light-industrial zoning, then the economy will be propped up by finance types, or professional bloggers, or journalists, or any type of profession that generates revenue from outside the city. Those professions are not necessarily stable; what happens if the internet goes down for long periods of time?
These facts appears to be the reason that American cities have fluctuated between dynamic, energetic utopias (a la Thoroughly Modern Millie) and crime-ridden dystopias (a la Taxi Driver) over the past decades. Making cities even *more* insular might compound the problem.
Although I will concede that having grocery stores on the first floor of an apartment building is more efficient than zoning commercial buildings far away from residential building.
RE: Stop AI Development for 6 months, for those interested, Lex Fridman has talked with Max Tegmark who is the father of the letter (if I understood correctly; he is at least the President of FLI) and he elaborates on his reasons for the letter and what he hopes to achieve (https://www.youtube.com/watch?v=VcVfceTsD0A, time stamped).
Interesting tidbit - he (obviously) feels like all the companies are bound together in a race to the bottom and hopes that the letter will break this system... and the main analogy he uses for this (and for most of the talk) is Scott's very own Meditation's on Moloch article (start at about minute 40 for full context, reference at 41:30). So, @Scott, how do you feel about this? It appears you contributed to the letter indirectly as well :P
29. I still want to see this data broken down by places that instituted the Counter-Reformation voluntarily and places where it was enforced after being conquered by Catholics. There would be no significant reverse effect because the Protestants mostly lost those wars in terms of net territory. The conquest of Prague alone could skew data significantly. Meanwhile places like Italy, Vienna, Paris, etc all continued on. They did eventually suffer their own declines. But mostly due to things completely unrelated to the Counter-Reformation. Italy had an economic crisis. Paris had Louis XIV being Louis XIV and insisting that "absolute monarch" really means absolute. Including over the universities. Poland had The Deluge. Etc.
Diets: my understanding is that due to rebound weight gain outpacing weight loss, dieters tend to gain weight over time faster than comparable non-dieters. This is deceptive, because the first diet tends to yield better results than subsequent attempts , giving the impression that better adherence or finding the 'right' diet that's easy to adhere to will change things. They don't, and the the cycle perpetuates.
Creating more diets to try will only contribute to accelerating obesity long-term through the cycle of dieting.
30. I tend to think the media as cause narrative is a bit off. But most countries have become more polarized and are facing fairly severe crises over this and populists. The exact severity varies significantly. But this is a common international phenomenon among the democracies.
While I haven't actually ran the numbers eyeballing it the non-polarized democracies mostly seem to be those which are growing fairly rapidly economically (in real per capita terms) and have high trust in institutions. (And they still have polarized media.)
Specifically, this doesn't decrease feelings of difference. But it decreases the feeling that these differences can't be overcome or pessimism about the future of the country or the feeling compromise is bad. For example, a lot of Indians feel Indian society is very deeply divided but mostly believe these differences can be overcome. (There's also countries where the population feels polarized but politicians have managed to keep a lid on it. Often strong party systems where the ruling party defeated a populist revolt.)
As for how this happened the pattern in trust is easy. An institution does something and loses trust. It sometimes slowly grows (or can grow rapidly if something happens). But often it just stays low. For example, for the US government it was Vietnam. Trust in the US government at the start of American major troop presence (1961) was 77%. When the US pulled out (1975) it was 34%. And ever since then the average has been in the 30s-40s. For the banks/Wall Street it was the slowdown then crash of 2008 (from 53% in 2004 to 18% in 2010) and it now stays in the 20s. This has happened in succession such that the only trusted institutions left in the US are the military, first responders (police, firefighters, etc), and small business.
Reasons for economic growth are more complex. But the average real per capita growth since 2000 has been 1.2%. It was double that, 2.4%, from 1945 to 2000. The statistic has been consistently below the trailing century average since 2000.
Yeah, so anyway, grow the economy and get higher quality institutions. Easy to say, hard to do.
> Why do East Asians have so many famous numbered lists (“Four Noble Truths”, “Thirteen Classics”, etc)?
As Gwern himself mentioned as well, monosyllabic digits play a role (I'm not familiar with India), but I think it's mostly a matter of style. Certain ways of writing can just feel more natural for no particular reason. Reminds me of how English speakers love creating acronyms that spell a word, even though English is not really a language well suited for pronounceable acronyms, with only 5 vowels.
35. Georgian was somewhere between heavily discouraged and banned so Stalin wasn't exactly facing a lot of competition. But no, it wasn't anything to do with his later dictatorship. He actually went out of his way to hide them. As a child Stalin was an excellent student and a gifted poet and singer. He was also fairly religious and had a strong interest in art and music. (The interest would stick with him for the rest of his life.) He got into a good school and had a bright future ahead of him.
Then he read Marx and Lenin and a bunch of socialist literature and became a socialist revolutionary. He walked away from his educational life to be a full time, professional revolutionary (and/or criminal). He claimed he was kicked out of school for revolutionary activities but in reality he started skipping class and eventually just didn't show up. Instead he spent his time being a bank robber, extortionist, and general criminal in an attempt to disrupt the Tsarist system and raise funds for the revolution.
30. <i>My only concern about this story is that some other countries haven’t become any more extreme/partisan since the advent of the Internet - is this because their media didn’t undergo this process? </I>. I‘d argue it is because media in other countries was already very partisan before the internet. This was certainly the case in countries I know well - Italy and Austria. There were (and still are) newspapers across the political spectrum. In Italy if you wanted sports news you bought Corriere dello Sport. Sports coverage in the „quality“ newspapers was meager. Same was true of classifieds. There were specialized publications for that. Italy even has a specialized publication for crosswords - La Settimana Enigmistica. In Austria Socialist voters read Der Standard, centrist conservatives read Die Presse. If you cared about soccer you bought Kicker. What held everything together in these countries was linguistic/ethnic identity. Now that we are also losing that as social glue thanks to mass immigration it’s not clear how the center holds.
I know the answer to Gwern's "Whatever happened to short stories" question.
Short stories were basically found in magazines, which were hugely popular in the interwar period, into the early 1950s. All the famous writers of the period wrote for them - Hemmingway, Fitzgerald, Steinbeck, even Dahl and Vonnegut got their start, at the very tail end of the period, writing in magazines. Gwern is focused on sci-fi, perhaps because that's his interest, but just about every genre was in just about every mainstream magazine, from Cosmopolitan to Colliers to the Woman's Home Companion. Importantly, because the magazines were big business, short stories were where the money was - you wrote short stories for money, and novels for literary kudos. And that money is why everyone was writing short stories.
Then the bottom dropped out of the magazine market due to television, subscriptions and advertisments both collapsing. Those magazines that survived had to pivot away from competing with television, and be complementary to it. One thing that TV can do very well is tell a short story - that's what an episode is, but TV can bring it to life much better than a magazine can, so magazines dropped short stories, and those few that still carried them were no longer paying much.
A lot of people still liked short stories as they were used to them, so for a while the genre limped on through anthologies and collections. But there was no longer an economic way to release a standalone short story, and TV wasn't just a competitor to the magazines, it was a competitor to short stories in total, so the short story died as a popular form, with only a few weird literary hold-outs remaining.
I had thought the Internet would see a revival of short stories, but it hasn't happened to any great extent.
In Gwern's list of questions, he asks if there are any drugs that work as "anti-psychedelics": i.e. are there any drugs that strengthen your sense of identity, rather than dissolving it. Gwern asks "Why [isn't there a drug that makes me feel] with the same absolute certainty that I feel ‘all is one’, that I am a unique snowflake utterly unlike any other human being or organism?"
Gwern thinks that stimulants don't qualify on the grounds that they don't induce psychedelic effects. I disagree. During coke highs, people often report that they believe themselves to be possessed of superhuman qualities and their demeanor changes to become more arrogant and narcissistic.
This sure sounds like it's approaching "unique snowflake" territory to me. It's almost as though stimulants are turning up the amplifier knob on somebody's sense of identity, magnifying their sense of themselves to mythic proportions.
How reasonable does this sound to other people? I don't have much expertise on drugs, and I'd appreciate somebody who knows more about habitual stimulant use who could explain how frequent this egotistical high really is.
(Tantalizingly, this entry on "ego inflation" on the Psychonauts wiki mentions that habitual stimulant use can lead to increased narcissism and egocentricity, even when the addict is sober. Can anyone, maybe a clinician, shed some light on whether or not this is a real feature of stimulant addiction? https://psychonautwiki.org/wiki/Ego_inflation)
Re approval voting in North Dakota: I'm confused, because this more recent article states that a more recent vote on Wednesday failed to override the governor's veto of the ban. Meaning the ban has not been put in place, and approval voting is still ok? Has my brain been overwhelmed by the double negatives? https://www.valleynewslive.com/2023/04/20/bill-eliminate-approval-voting-fails-override-veto/
> surely it would be even better if you spotted someone wearing the ring and then you could use your smartphone to call up their dating profile.
What? If you've already spotted them in-person, you can just go up and talk to them. Why would you want to dance around with dating profiles and instant messaging to set up a time and place to meet in person, when you can just walk up and say hi?
25. My big question with the El Salvador crackdown is whether it's financially sustainable. That's 2% of the population, and a higher percentage of young men - can the government afford to keep them locked up for a long period of time?
27. The opposition from the North Dakota state government is bizarre. They're getting hung up on approval voting when there's a ton of different ways that local governments have experimented with government structure? It's a pity.
36. I was wondering why all the 5-over-1s all had the same block design (the three-story walkups seem to have more variety in design).
21. I speculatively attribute this to the aggressive campaign of misinformation about the burden of student loan debt. In reality, for about 70% of four-year graduates, it's less than or equal to about one year of the college wage premium ($30,000).
Granted, there are taxes, and the median college graduate differs from the median high school graduate in terms of cognitive ability and/or conscientiousness, so college doesn't literally pay for itself in one year, but in a large majority of cases it pays for student loan payments.
34. I can't figure out what Loeb is doing here. He's a serious scientist who has made good discoveries. But the whole thing with Oumuamua being an alien light sail, despite little evidence of that, and this new thing with the "interstellar" alien technology at the bottom of the ocean is weird. Is he using it for publicity to attract funding for research? Astronomy could certainly use more money, but I'm not sure "aliens!" is the way to go. I mean there's a small possibility he's right, but for Oumuamua we really have no way of telling and the chances of finding a small rock that might have come from interstellar space at the bottom of the ocean are basically zero...
7. I'm having trouble understanding the 15 minute city idea. The PDF someone linked to said there was a priority on protecting local centres, which... I guess means strangling out their competition? I don't know how else you're protecting them.
The first link makes it sound like they're trying to turn Oxford into Midgar. That's a good way to get yourself stabbed in the back, and also blown up by Godzilla and smushed by rocks.
7. It seems worth clarifying that this was an Oxfordshire County Council plan. While Oxfordshire County Council is a government entity within the UK, it's not "the British government"!
Good prank if the pear ring thing becomes slightly successful but not to the point where most people know about it: give someone else such a ring, and see hijinks ensue.
28. "Trevor’s understandably morally anxious about benefiting from his Jewish ethnicity. Should he be? Is this any worse than benefiting from white privilege or male privilege or whatever other forms of privilege they’ve invented since I last checked?"
I imagine Scott is just trying to provoke us into saying the following (but I think the following is correct):
This is a false analogy. To the extent "white privilege" and "male privilege" are still a thing, they are *statistical* realities, or even just decoys for the real thing (whiteness correlates with class, which in turn determines much of what gets called "white privilege"). The right analogues here would be "in the US, "Jewishness" [defined as straight maternal descend from Orthodox Jews living in 1800, say, or whatever else you want] correlates with class" (very likely true) or "being brought up in Judaism makes you less likely to be molested by a priest" (also true).
The right analogy would be with Old-Style white privilege - being entitled to certain things if and only if you were white (or white enough for whoever is testing). A thing in Apartheid South Africa, certainly. The G.I. Bill also worked out this way in the U.S. South, though perhaps in a slightly less openly declared way (though my understanding was that it was completely deliberate, and other people saw through it). Would you sign up for a generously funded benefit program open to whites and only to whites? Or Asians, people of Latin American descent, etc.?
"Is it any worse than benefiting from being an American, and so having access to social support and benefit programs that Sudanese and Bangladeshis can only dream of?"
That's a thornier issue. Your average citizen of a rich country probably does not think very often of what an enormous privilege it is to be born with that status - and they would be baffled to hear that that privilege is questionable. (Of course, that doesn't make it *not* questionable.) I've even heard it said that, if it weren't for those privileges, citizenship would be meaningless - that is saddening and unintentionally funny, as that is exactly what you would associate with Roman decadence, say: citizens thinking of citizenship _merely_ as a set of privileges (which are great, because they have them).
Of course that may also why most Israelis wouldn't wink twice at the proposition: unlike most American Jews, they are brought up to think that "Jewishness" is a nationality - and most Israelis are completely comfortable with having preferences for the dominant nationality in their country (and in fact, according to a recent poll, most of them openly believe citizens of that nationality should get preferences over fellow citizens not of that nationality).
Yglesias' point about crackdowns - that the first requirement was an honest and effective force capable of doing the cracking down, and that once you have such a thing, the exact method by which you use it to deal with the gangs is really irrelevant - seemed to me the really useful insight.
If your police (or military or paramilitary) force is capable of being bribed by the gangs, then it doesn't matter what orders you give it, it's not going to crack down in an way that is effective.
30. 15-20 years ago, I remember seeing someone point out the contrast between the WSJ, which at the time had made its editorial page free and kept its news behind a paywall, with the NYT, which at the time had done the opposite. It might have been Tyler Cowen, but I can't find the post now.
23: My hunch is that people partially support those policies because they support / like Trump (i.e. reverse causality). So once he is elected his supporters no longer really care cause goal accomplished.
>My only concern about this story is that some other countries haven’t become any more extreme/partisan since the advent of the Internet - is this because their media didn’t undergo this process? Why not?
The American "city newspaper" is not a typical model for most countries. I live in Britain; we have city newspapers, but they are the second-tier publications, like the Manchester Evening News (my city paper). The first-tier newspapers are national newspapers. All of them are based in London, though one (the Guardian) used to be based in Manchester until the 1960s. Because there have usually been around ten national newspapers (there are seven that have been continuously in print from before WWII; there are currently three others in print), they have had to compete with each other, and have often done so on the basis of their political positioning.
The city newspapers were usually exactly the sort of inoffensive entities stuffed full of ads that was described, but the national newspapers have always been viciously partisan. A famous 1992 headline, on election day, ran "If Kinnock wins today, will the last person to leave Britain please turn out the lights" https://pressgazette.co.uk/wp-content/uploads/sites/7/2016/04/The-Sun-If-Kinnock-Wins-Today-600x755.png This was in the then highest-circulation newspaper in the country.
The reasons for this are pretty simple: you can get an overnight train from London to just about anywhere in the country, so the first copies of the paper off the presses at about 10pm were loaded onto trains for the most distant parts of the country, with later printings progressively headed for closer regions. The upshot is that the entirety of Great Britain could, by the 1860s, read the national press. Americans, accustomed to newspapers being associated with a city, tend to refer to The Times as "The Times of London", to distinguish from the New York Times; this isn't incorrect (its editorial offices are, indeed, in London), but The Times of England or of Britain would be more informative.
It wasn't possible until the arrival of air freight to physically distribute a newspaper around the entire United States overnight, so a late-evening printing of the New York Times could be delivered in San Francisco the next morning, and air freight would make newspapers very expensive. The only national American newspaper is USA Today, which operates by having 37 printing presses spread around the country and electronically transmitting page layouts to each, so identical papers are printed in a variety of places. The technology to do this didn't exist until the late 1970s, and USA Today was founded in 1982, by which time the American "city newspaper" model was established.
The other upshot is that the UK distribution model (until it broke down recently with the internet) was that there were large numbers of independent newsagents who hired deliverers who would deliver the paper of choice to each household, so the newsagent would have a few hundred houses on their books, and each would order the paper they chose, and then the deliverer would deliver that paper to that house, then a different paper to the next house, and so on. The US distribution model was for the newspaper itself to organise delivery; this made it much harder for new newspapers to break into the market in a city. USA Today was usually only available at newsstands.
Just as an example of how different it was; in the 1990s and 2000s, if you stayed in a hotel in the UK, they would ask you which newspaper you wanted in the morning and would then deliver the requested paper to each room. In the US, you generally just got USA Today, or the local city newspaper, and they didn't ask.
> They also hoped that sometime in the future there would be a “fire alarm” - something would happen to get people and policy-makers’ attention - and then the political route would open up. [...] I personally am pretty surprised it was just “GPT-4 got released and was very good”. [...] I’m starting to worry that the problem won’t be building a political coalition against unsafe AI, the problem will be not overshooting and banning all AI forever.
It's a bit early to worry about *that*.
AFAIK there no mainstream politician has publicly come out in support of slowing down AI research. The idea of AGI risk is seen with contempt and derision by your average tech-savvy internet nerd, let alone your average 40-something blue-collar worker.
Sure, big names are coming around, but as far as the public goes? There's about as much support for AGI safety as there would be for a carbon tax in the 1980s.
I really love everything about HustleGPT, especially because it turned out to be absolutely useless and the guy is desperately trying to convert his 15 minutes into money, but it looks like that's not going to well either.
2. "Why are furries so artistically and economically influential compared to other fetishes?"
I don't know, 50 Shades of Grey sold a lot of copies, people are using ideas of consent drawn from kink in vanilla sex (sometimes to unreasonable lengths in my view), role-playing games are using the stoplight system...
5. Extremely un-PC hypothesis: a lot of the academic rules on dating are invented by lesbians trying to compete for bisexual women by driving out straight (or bi) men. I don't know where else to put this that will actually get reasonably smart people to come up with critiques but won't dismiss it out of hand or simply believe it as part of the 'gay agenda'.
6. So more money won't make chronically unhappy people happy, but will make everyone else happier? Yeah, I'd believe that. Didn't someone show here it was actually logarithmic, such that going from about $100K to $200K is like going from $50K to $100K, but it was hard to recruit enough rich people to prove it? Tracks my personal experience but that proves nothing.
15. I suspect the cat's out of the bag on that one, but worth trying.
21. With the college thing I think the debt problem has grown to the point it's reaching parity--i.e., jumping ahead in the job queue with a college degree no longer justifies the now-huge debt you acquire. That's assuming you don't get into Harvard, of course, and what point in the prestige hierarchy breakeven is reached is probably questionable and depends on a bunch of other things.
22. Seems like a nice idea, silly though it is. A lot of people don't want to be flirted with, so they made it socially dangerous to do so (at least in liberal areas)...so this solves that problem. I kind of doubt it'll catch on. If there's one thing we know we nerds suck at it's romance.
23. The obvious answer is their policies don't work. ;) You could say the same about communists, and I wonder if there's a certain symmetry there. My guess is right-wing populists drive off the professional class necessary to implement policy successfully, which was certainly a problem with Trump--he had to rely on people more interested in tax cuts to staff his administration. Honestly he wasn't doing so bad in some ways until COVID hit--the economy was growing without inflation and he didn't start any wars. A less narcissistic person (which is to say, anybody) might have pulled out a close reelection.
28. Go for it, dude. The thing I've realized after reaching middle age is, everyone takes advantage of their privileges, and then lies about it to make well-meaning, honest people feel guilty so they don't take advantage of theirs. This one seems pretty good for everyone--Middle America seems underinvested right now. I don't know about the culture clash between the 'blue' Jewish arrivals and the 'red' locals, but as David Brooks has pointed out, blue cities in red states are doing pretty well. Maybe the challenge is to keep either side from getting all the power.
35. Well, that other genocidal dictator was a painter.
Regarding research about the Catholic Church from Martin-Luther-University: I'm working there (science department) and can assure everyone that we're pretty much the most secular institution in the least religious part of the world (former East Germany). ;-)
By the way, regarding nuclear power, it's worth noting how utterly different the attitudes to nuclear power are in comparable countries. At the same time as Germany has proudly switched off its last power plants, Finland - smaller, but similar to Germany in many ways, though also different in many ways - has basically all parties in the new parliament supporting more nuclear power, and in a recent poll 68% of people see nuclear as "positive" or "very positive" and only 6% as "negative" or "very negative". https://twitter.com/Kaikenhuippu/status/1649011146362433539
2. The twins thing has an intuitive explanation for me:
There’s probably so few “extremely successful” identical twins because the marginal difference between “highly successful” and “extremely successful” is often razor thin. I’m going to use sports as an example because it’s very easy to evaluate objectively, but I believe the talent distribution we see so clearly in sports exists in other domains as well.
The difference between a guy who, for example, ends up playing 12 seasons in the NBA and a guy who never makes a roster and plays overseas is extremely tiny, especially from an outside perspective. If you have 2 twins, one is 6’8” with a 7’ wingspan, and one is 6’6.5” with a 6’9.5” wingspan, that alone is enough to make one person an NBA prospect and the other just a great D1 basketball player. Even though both guys are functionally the same, in todays world of max efficiency those small differences determine whether someone is “extremely successful” or just regular old successful.
I assume the same principle applies to the difference between, say, a world renowned physicist at MIT vs his twin brother in the physics department at Ohio State, or whatever (example is purely hypothetical). Both brilliant and successful, and on a scale of all humanity pretty much equally smart.
I think this, combined with just how rare identical twins are pretty much explains it fully.
Pear ring seems like a great idea, but I wish that someone would invent and publicize a gay version so I can advertise my singleness to the correct gender. Rainbow? Purple? I pass as straight, and I definitely don't want a piece of jewelry to cause me to be constantly approached by men.
Re #28, why would these billionaires think this is a good idea? They do realize this exactly plays into (and in this case confirms) the anti-Jewish trope of Jews being more successful because of nepotism, where other Jews in high places preferentially grant jobs and benefits to their "tribesmen" over the "goyim," right? I fail to see how they think creating the exact classic conditions for resentment against a local Jewish community would be a good idea for their local Jewish community. Maybe a few more talented Jewish people move in because of this program, at the cost of all the locals who find out about this program viewing them (and by extension the Jews already there) as privileged interlopers and not wanting anything to do with them. Seems like a pretty obvious net loss to me.
(Edit because I realized I confused the words "interloper" and "interlocutor")
With regards to Gwern's question on writing vs. coding in the morning. My best coding time is usually around 9-11am. 8-9am I'm still waking up, 11am I'm starting to get hungry for lunch. I usually eat at noon, then I'm tired from eating lunch from roughly 1-2:00. 2:00-3:30 can sometimes be ok if I'm well rested enough and I'm usually getting tired again as the afternoon wears on. The later I code, the more easy mistakes I make, which I usually catch the next morning while wondering what on earth I was thinking.
I've also noticed that most of my best ideas come in the morning, oftentimes I'll wake up with them. To me, there is very clearly some mental background process that continues to work on whatever problem I'm dealing with at the time, and whose results generally filter into my consciousness in the morning after waking. If we assume this process is similar, this would explain why writers feel more effective in the morning: they are getting the creative output of their sleep processes when they wake up.
As to why coders stereo-typically like to stay up late, while I might be biased, I think coding is probably more rewarding that writing is and this leads people to get wrapped up in it in a way writers typically don't. You can write a bit of code that adds a bit of functionality, then you can test it and figure out all the little bugs that inevitably creep in, fixing each one is a little reward like solving a puzzle. And when you fix all of them and the code works and you can see the new feature, that's a bigger reward and so, so satisfying. Fixing the bugs can also task you, because you are so close to having it working and you just need to iron out those kinks so 5 more minutes to try 1 more thing.... can become 1 hour, then 2 (it's pretty easy to get into a flow state when coding). That said, as I mentioned before, my ability to code degrades the more tired I am, especially as the evening wears on and irrespective of how focused I am on the task, so I place being well rested at a premium to maintain optimum coding ability. Everyone else seems to just rely on caffeine (I don't use it so I'm pretty sure I'm an outlier, for whatever that's worth).
> My only concern about this story is that some other countries haven’t become any more extreme/partisan since the advent of the Internet - is this because their media didn’t undergo this process?
UK newspapers were always partisan. UK broadcast media aren't allowed to be. If you want to understand US polarisation, the repeal of the fairness doctrine should be your first port of call.
A big reason behind #21, the drop in people saying college is worth the cost, is that the opportunity costs went way up: youth unemployment went from 16.3% in 2013 to 7.5% now.
"The alignment community always figured their concerns sounded too weird for normal people to care about, that politics was a lost cause, and that our best hope lay in technical research. They also hoped that sometime in the future there would be a “fire alarm” - something would happen to get people and policy-makers’ attention - and then the political route would open up. I personally am pretty surprised it was just “GPT-4 got released and was very good”."
And this was always a bizarre belief. Rationalists did not invent AI fear, or even AI doomerism. How much of science fiction is about a mad scientist creating a being that turns on him? That's what Frankenstein is about, and that was written in 1818! Isaac Asimov's "I, Robot" series, written in the 1940s and 1950s, is about the problems of AI alignment. One of the most iconic evil AIs, HAL 9000, is from the 1960s. So are Daleks, who want to exterminate the whole universe. More recently, we have movies like Matrix or Terminator, TV series like Battlestar Galactica, and video games like Horizon Zero Dawn.
Storytellers have been taping into the public's fear of AI since before AI existed. Whether they like to admit it or not, rationalists are embedded in this culture and take their inspiration from it. It's no surprise whatsoever that improved AI capabilities would heighten AI fears and potentially lead to the government snuffing out the most revolutionary technology since electricity.
I've been saying for several years that remote work is not sustainable for businesses, maybe any large business. That we're seeing significant drops makes perfect sense to me, and I don't think we'll ever get to the point where a majority of jobs are even partially remote. Honestly, it seems that any discussion of significant remote work is a symptom of the types of people who have the freedom to write about such things being one of the few professions where it might work (journalism).
The core problem is that most employees do not have the knowledge and drive to start a career remotely. Even if experienced employees can complete their tasks from home, new employees need direction and training (this includes experienced employees starting at a new business in most cases).
If you allow your long term employees to work remotely, there's no one around to enculturate and train the new employees. You've also got potentially significant problems making sure that work is being completed and that employees are using their time effectively. Many information jobs that could work remotely involve long term projects from multiple people, which permits some people to ride on the accomplishments of others or pretend to get their work done before some deadline looms and stuff crashes. This exists in person, but proximity allows managers to have more opportunities to review and notice problems.
Compounding this is the fact that the vast majority of jobs require a physical output at a particular location. You can't be a remote nurse, mechanic, farmer, dock worker, etc. That he thinks only 40% of workers will be fully in person is pretty much insane, if you understand the types of work that exist. We're not going to have WFH for the service industry, manufacturing, construction, retail sales (distribution network), let alone the more obvious like mining and agriculture. These fields make up about 80% of the private sector workforce. That a few people could work remotely for companies in these fields is probably true, but we're talking well under 1%, probably far less.
Even speaking theoretically, it's probably the case that fewer than 10% of jobs can even be done partially from home. The chart pictured above is probably only a subsection of jobs - likely not polling too many restaurants or nursing homes to determine the percent of people working from home.
"A few years ago a weird object called Oumuamua entered the solar system and a few astronomers speculated in might have been an alien spacecraft (it’s since left). "
I was taught that the Catholic Church becoming anti-science for the Counter-Reformation is what allowed Protestantism to seize the pro-science position.
Copernicus, after all, did his fateful research in the employ of the Church trying to develop a better calendar that correctly dated holidays with respect to the life of Jesus. His book didn't really get much traction even in the scientific community until it got an endorsement from Tycho Brahe, on account of some testably wrong predictions Copernicus made based on a much too small solar system. (The endorsement being posthumous and probably actually attributable to Brahe's student Kepler is another story.)
Initially, the Protestants opposed heliocentrism as an "innovation". They complained of a pattern of behavior by the Church of interpreting the Bible to the point of, in the Protestants' eyes, changing the meaning. Heliocentrism is one of the flimsier ones. First, the orthodoxy here was more Aristotelean than Biblical. But even the Biblical textual argument is weak; as I understand it, the argument comes down to two passages: Bereshit 1:1-3 (God split Heaven and Earth before creating light), and a single line from Joshua (*something* happened to the sun by Joshua's command; the Protestants translated it as "stopped", modern Jewish scholarship has it as "disappeared", perhaps like an eclipse). Both were sort of flimsy arguments! But in the Protestants' eyes they fit into a larger pattern, and taking cheap shots at an opponent's controversial theory is a common way people score points.
But as part of the Counter-Reformation, the Church chose to go fully hard-ass on heliocentrism. Possibly *because* the Protestants were so ineffective at stamping science out, the Church might have felt they could make the case that they were holier and more originalist than the Protestants. They had the institutional advantage to run those inquisitions, after all. But by this point the Scientific Revolution was clearly in full swing, and while Galileo was a jerk his trial was clearly a bit funky, and so the Protestants had the opportunity to pivot to the pro-science position.
That equilibrium lasted a few hundred years. For my money, it seems less true now; these days Evangelical Protestants seem at least as skeptical of science as the Church.
6. From your summary, if they were both right, then neither was right. My understanding is that they concluded "some people get happier with more money, whereas some people don't."
31. I never agreed with Guzey's conclusion from that first debunking article (the only one of his that I read) -- IMO plenty of sleep is obviously important. But his meta-point seemed more important: that most evidence is pretty muddled and popularizers disingenuously present it to the public. But reading this link, maybe I was just projecting this meta-point onto him if he actually tried to sleep 4 hours a day.
One of Gwern's questions: "Nicotine is one of the best stimulants on the market: legal, cheap, effective, relatively safe, with a half-life less than 6 hours. It also affects one of the most important and well-studied receptors. Why are there no attempts to develop analogues or replacements for nicotine which improve on it eg. by making it somewhat longer-lasting or less blood-pressure-raising, when there are so many variants on other stimulants like amphetamines or modafinil or caffeine?"
Good question. It seems a shame that the stigma that rightly arose against cigarette smoking has probably prevented progress in finding and exploiting the many beneficial uses of nicotine. If you're not familiar, google it... May help for depression, Alzheimer's, ADHD, etc.
Seems like perception vs reality are reversed for alcohol and nicotine. Nicotine per se isn't actually that harmful or dangerous, it's just the common routes of administration. With alcohol, otoh, it is the substance itself that is inherently harmful, especially if you take it in the amounts required for the main effect which AFAICT is the only reason anyone drinks the stuff. If you're drunk, you're not drinking in moderation. There is no safe, healthy way to get drunk. It really sucks, I know. But people latch on to studies saying it's potentially healthier in moderation (researchers have a weird idea of what that means...I'd think 3-5ish, as long as I'm not hammered, but ONE or TWO 12 oz beers, why not just have a single shot glass of wine cooler? What's the point?), and say, "See, it's healthy! Stop being a prudish scold; everyone does it (as it used to be with smoking), real men drink, always have." And continue with their drinking habits that are almost certainly unhealthy.
Anyway...
I occasionally use nicotine in patch form, which has never been addictive. I rarely continue doing it for more than a week or two because I just don't feel like anymore. But it does help me get work done; it seems at least as good as any of the other stimulant meds for ADHD.
Speaking of modafinil and variants... Recently tried armodafinil. Many apparently say they like it better, but I definitely don't. It's weird... It seems very different to me. Feels less like a stimulant, I feel less sharp, more fuzzy, which is kind of good, but also bad. Seems less effective at helping me get work done... I seem to waste more time on BS like TikTok (or writing pointless rants for strangers to read) when I do it. OTOH, it seems easier to sleep on it. Experimenting with a little bit of both... Seems like they kind of balance each other out a bit.
The answer to why Brazilian jiu-jitsu took so long to invent is simple: for all it likes to pretend, MMA is not like actual combat. If you really want to win a fight, you bring a friend or use a weapon. Both of those give you huge advantages, and it's what everyone did: norms of "help your friend in any fight and ask questions later" and "always carry at least a knife" are extremely widespread across cultures. Extended, one-on-one, completely unarmed combat just didn't happen often at all. No one got good at MMA until people started playing by MMA rules.
"Also that AI wiping out humanity is “not inconceivable . . . that’s all I’ll say”
Completely unsurprising. We have two centuries of literature conceiving of AI being bad in one way or another (I count Frankenstein's Monster as AI), and as soon as "wiping out humanity" became a thing, AI was fingered as a possible villain the next day.
And has anyone checked up on the babies he used as living experiments? Are they okay? Did it all work out? If we're going to decide that cut-n-paste humans are the way forward, at least let us find out if it worked okay or not the first time.
“I can’t tell whether they’re claiming Stalin was actually very good - or that he pretended his poems were anonymous, and everyone else praised them lavishly while pretending not to know Stalin wrote them.”
Stephen Kotkin’s biography suggests his talents were recognized even when he was an unknown revolutionary under the Tsar:
“As a teenager, he had abandoned a successful trajectory, with high marks in school, to fight tsarist oppression, and published first-rate poems in a Georgian newspaper, which he recited in front of others. (‘To this day his beautiful, sonorous lyrics echo in my ears,’ one person would recall.)”
> 27. [...] One of the few rare victories for these voting systems was Fargo, North Dakota...
Approval voting has struggled to gain momentum. But ranked-choice voting has actually been having a run of successes in the past few years: the mayoral primaries in NYC, the mayoral election in Burlington, VT, statewide elections in Alaska and Maine...
I absolutely agree that I'm doing a "special crackpot that works mostly for me" diet but hoping to find what the magic button is ;) It would be great if we could interpolate from all those one-off crackpot diets that work and find one common ingredient. Maybe not, but maybe.
On 5, British Universities are different to American ones. In particular, 18-year-olds are viewed entirely as adults, and the university is a business they pay to provide them instruction in a specific subject. There’s no expectation that they’ll provide any pastoral care, and you only rent a room from them in your first year (where they’re basically just your landlord). They’re not in any sense responsible for you; if any of them tried to impose an honour code, that would be as weird as McDonald’s imposing one on its customers.
Dating your lecturer (who may or may not be a professor) doesn’t seem any stranger to me than dating your dentist, with the caveat that they shouldn’t mark your exams to stop them being biased in your favour.
(Exceptions to this are that there’s been a slow pressure towards Americanising for the last few years driven by the general infantilisation of society, and Oxbridge works completely differently from everywhere else).
The Gathering Place in Tulsa (funded by oil millionaire turned banking billionaire, but who's counting) is really quite amazing. Both Tulsa and Oklahoma City are doing impressive things in terms of family-oriented public infrastructure, even if they are still essentially giant sprawl exurb metropolitan areas. Purple cities in red states can be pretty nice places to live.
>New weird either-genius-or-crackpot anthropology blog (origin of pronouns, snake cult theory of everything). My heuristic is that as soon as someone uses the phrase “Basque-[Anything]” they’re beyond salvation, but this person is trying very hard!
Would point out it's only the second most radical take about snake mythology. The anthropologist Jeremy Narby was studying a tribe in the amazon, and he participated in their Ayahuasca ceremonies. Like them, snakes appeared to him. Then he notices that creation myths link snakes to knowledge across the world. Egypt, Jewish, Australian, Amazonian, Indian...why are there always snakes? He suggestion is that western science is wrong, and that there are mystical ways of knowing, somehow connected to snakes.
My take is that snake venom is itself a hallucinogen, maybe it was used as part of a self-domestication package when recursion / self-awareness was first emerging.
There really was a bear in Georgia that OD'd on cocaine, but the bear corpse in KY is not that bear, and their story of what happened to the real bear is a complete fabrication. The outfit displaying the supposed cocaine bear finally admitted as much.
I'm the anon commenter quoted in that article,, and I'd looked into the corpse-journey story a couple of years ago and found it was clearly BS (some parts didn't make sense, and I tracked down multiple people who could have verified parts of the story, and got "definitely didn't happen", "definitely not the bear", etc.)
Re the Tulsa thing, I am not sure how much he's correct in crediting Tulsa's transformation to Kaiser et al. Obvious disclaimer, I live in the other city in Oklahoma and am not super-familiar with Tulsa. But a lot of what he talks about (at least the non-Jewish parts) I'm seeing here in OKC, too, despite a lack of billionaires funding it (at least as far as I know). Which makes me suspect that structural factors (land is cheap, the government is pretty hands-off) and general demographic shifts are driving a lot of it. If you have people who want Culture, Culture will appear, and both OKC and Tulsa are big enough to have demand.
I'll definitely second the message that Oklahoma is a pretty nice place to be. My mortgage is about the same as my rent was in deep suburban LA for a place a third the size (and I had a really good deal there) and the government mostly leaves you alone. Gas is cheap, and it's safe. The only downside is the lack of battleships and that the local ACX community isn't that strong, but if you move, we can change that.
> Why are furries so artistically and economically influential compared to other fetishes?
I have a few theories:
First, unlike many other fetishes, the furry fetish can ONLY be satisfied by art. Into bondage? There's plenty of videos on PornHub. Into feet? The industrious editors of wikiFeet have you covered. But for furries, there is no photography, only art that has to be drawn, painted, or created digitally. Naturally, this results in the creation of a lot of art. A minority of furries are interested in photographs and videos of actual people having actual sex in fursuits, but they're definitely a small percentage (and even then, someone has to buy a fursuit from an artist who created it).
Second, the furry fandom is identity based. It's commonplace to have a "fursona" to represent yourself online (which is where furries spend all their time anyway). This creates a lot of demand for art, since people will want a visual for their furry avatar. In terms of porn, this makes it rather unique in that you can pay to have a representation of yourself in the pornography. I think it's pretty easy to see why that would appeal to people and lead to increased investment in furry fetish artwork relative to other fetishes.
Third, it's an internet-based culture. This leads to certain demographics associated with economic success. There's an overrepresentation of people in tech/IT. It's mostly male. When folks from other countries participate in English-speaking furry communities, you'll only be interacting with the people in those other countries who speak perfect English (a correlate of intelligence and economic success). And for what it's worth, the community is extremely autistic, something like 4x as likely to be autistic as the rest of the population (though I'm not sure what impact this has on their income).
Although...
I'm not actually sure if furries are that economically influential. It certainly seems that way based on the quantity of furry art, but is there actually data to support that idea? Sure, they appear to have a lot of money to burn on art, but furry art isn't always expensive. A lot of furry artists live in third world countries and make a living off doing commissions for people in first-world countries. Even if those commissions are only $10 USD, that can go a long way in a developing economy.
Plus, appearances could lead to a biased sample. I think a good analogy would be related to somebody hating gay people because they think they're all flamboyant and annoying. The issue here is that they don't notice the normal behavior of non-flamboyant gays because they don't realize non-flamboyant gays are gay at all. For furry art, we only see commissions from the wealthiest and most successful furries who can afford it. For every one of them, there might be 10 furries living in poverty.
Seems like a good question for someone like aella to answer via survey data.
> The Roman Emperors thought Christians not performing the relevant sacrifices to the Gods were an externality, too.
Yeah, that's some real high-level discussion there. They've definitely got a grip on what externalities are and have good ideas on how to handle them. This definitely isn't just throwing out the laziest possible argument to avoid having to consider the possibility that you're having a negative effect on other people. I wonder if the response is the same to the externalities imposed by homeless people on the sidewalk or drug markets in public places.
Re 5 on professors dating students - I expect that most people are actually much more fine with that than most people affiliated with universities would be. I suspect the American poll results would be pretty much in line with the British ones. You can find plenty of examples of it happening in TV shows (here's a list that includes a lot of *high school* students dating teachers, as well as college students, not too many of which are portrayed very negatively: https://www.refinery29.com/en-us/2017/08/168506/teacher-student-sexual-affairs-on-tv-shows - and it doesn't include cases like Jeff Winger dating his math professor in the first season of Community).
Meanwhile, within universities you have a lot of sentiment that is too far in the opposite direction. The Texas A&M University System (i.e., the thing controlled by the legislature that has authority over the College Station campus, as well as the others, so not the "woke" part, more the Republican part) a few years ago dropped a rule banning *all employees* from having "romantic, sexual, or amatory relationships" with undergraduate students, with only 24 hours notice to the campuses before the rule went into effect. Whoever wrote the rule realized they needed an exception for undergraduate student employees, but didn't understand that basically all graduate students count as "employees", and that the university is large enough that incoming grad students who are single are going to accidentally hook up with people from other parts of the university if they try to meet anyone in a forum other than carefully controlled online dating.
Re: the African birth rate decline: Scott’s gloss of it makes it sound like it’s consensus that Africa was terrifying everyone and pushing us towards overpop catastrophe.
This is the OPPOSITE narrative from what I’ve been hearing for the past several years in my circles: that overpopulation is mostly a eugenicist myth, that huge numbers of countries of developing countries are headed for UNDERpopulation crises, etc. So what gives?
Regarding #37, article is paywalled but here is the nut of it:
"....few have noticed a wealth of new data that suggest that Africa’s birth rate is falling far more quickly than expected. Though plenty of growth is still baked in, this could have a huge impact on Africa’s total population by 2100. It could also provide a big boost to the continent’s economic development. “We have been underestimating what is happening in terms of fertility change in Africa,” says Jose Rimon II of Johns Hopkins University. “Africa will probably undergo the same kind of rapid changes as east Asia did.”
The UN’s population projections are widely seen as the most authoritative. Its latest report, published last year, contained considerably lower estimates for sub-Saharan Africa than those of a decade ago....
Yet even the UN’s latest projections may not be keeping pace with the rapid decline in fertility rates (the average number of children that women are expected to have) that some striking recent studies show. Most remarkable is Nigeria, where a UN-backed survey in 2021 found the fertility rate had fallen to 4.6 from 5.8 just five years earlier. This figure seems to be broadly confirmed by another survey, this time backed by USAID, America’s aid agency, which found a fertility rate of 4.8 in 2021, down from 6.1 in 2010. “Something is happening,” muses Argentina Matavel of the UN Population Fund.
If these findings are correct they would suggest that birth rates are falling at a similar pace to those in some parts of Asia, when that region saw its own population growth rates slow sharply in a process often known as a demographic transition.
A similar trend seems to be emerging in parts of the Sahel, which still has some of Africa’s highest fertility rates, and coastal west Africa. In Mali, for instance, the fertility rate fell from 6.3 to a still high 5.7 in six years. Senegal’s, at 3.9 in 2021, equates to one fewer baby per woman than little over a decade ago. So too in the Gambia, where the rate plunged from 5.6 in 2013 to 4.4 in 2020, and Ghana, where it fell from 4.2 to 3.8 in just three years....
Demographers are divided over how much to read into these recent surveys, particularly since the data they produce can be noisy. “When you see a precipitous decline in fertility, your starting-point is that something is wrong with the data,” says Tom Moultrie of the University of Cape Town....Yet even comparing only within iterations of the same survey (as The Economist has done with the figures above), the trend is evident. Comparing across them in the case of Niger, which has the world’s highest fertility rate but few surveys, shows a decline from 7.6 in 2012 to 6.2 in 2021.
Others are also reducing their projections. In 1972 the Club of Rome, a think-tank, published an influential book, “The Limits to Growth”, warning that consumption and population growth would lead to economic collapse. Now it says the population bomb may never go off: it reckons sub-Saharan Africa’s population may peak as soon as 2060, which is 40 years earlier than the UN’s projections.
Even so, fertility rates are not dipping uniformly. Some countries, including Angola, Cameroon and Congo, are seemingly stuck at relatively high rates. And there are often big regional differences within countries such as Kenya. Almost everywhere in Africa, fertility rates are much lower for urban women, who typically have 30-40% fewer children than those in the countryside....
Family planning, especially when promoted by outsiders, has often caught the ire of religious leaders. Yet in some places that may be changing. Clerics talk more often about family planning these days, notes Amina Mohammed, a devout mother on the outskirts of Kano. “There is no verse in the Holy Koran where Muslims are forbidden from controlling, planning or restricting the number of children they have,” says Shuaib Mukhtar Shuaib, one such cleric. The Prophet Muhammad tacitly approved of the withdrawal method, he continues. These days Idris Sulaiman Abubakar, a gynaecologist in Kano’s biggest public hospital, is more worried about the impact of Nigeria’s film industry on contraception than that of religion. “They’ll bring a story-line that the woman’s reproductive system was damaged because she uses pills,” he explains.
Girls’ education also makes a big difference to fertility rates. In Angola, for instance, women without any schooling have 7.8 children, whereas those with tertiary education have 2.3. Educated women have a better chance of a job, so the opportunity cost of staying at home to look after children is higher and they are more likely to win arguments with their husbands over how many kids to have.
Research by Endale Kebede, Anne Goujon and Wolfgang Lutz of the Wittgenstein Centre for Demography and Global Human Capital suggests that a stall in Africa’s demographic transition in the 2000s may have stemmed from the delayed effect of cuts in spending on education in the 1980s, when many African economies were in crisis. The rapid falls in fertility rates that now seem to be taking place could be because of the huge push to improve girls’ schooling in the past few decades....
On the topic of right-wing policies becoming less popular once people are elected - "Reflections on the Guillotine" comes to mind, I think a mind-boggling percentage of people do not have the faintest understanding of what it means for someone to undergo any number of terrible things, from literally execution to homelessness to hunger. A policy that so many people agree with in theory that involves holding someone "accountable" or people "pulling themselves up by their bootstraps" or "giving criminals what they deserve" ends up horrifying a lot of people who were previously sheltered by their policies being mere ideas. It's easier to love "justice" when you're not watching someone's blood pouring out of their neck, just like it's easier to love "austerity" when local government functions so well you hardly know it exists.
Man that North Dakota one is frustrating. It's just maddening to have local political change rugpulled on you. Like when the Supreme Court ruled states weren't allowed to put term limits on their federal representatives.
Tulsa has a long* history of oil billionaires using their wealth to indulge their particular whims and create cool stuff (see: anything beginning with "Phil," The History of Science collection (now in Norman), etc.)
"In the 80s and 90s, Tulsa experienced a decline. It got gross and dangerous. "
As someone who lived there (and was a stage electrician), this is completely wrong. You could run around downtown after midnight. It had a symphony orchestra, an amazingly good youth orchestra, a ballet company, and at least three theater companies.
If you want to know what percent of Americans think college is worth the cost, you should just look at what percent of Americans go to college. Any delta between this and your poll results is cheap talk.
What most interested me about #6 was actually the magnitude of income's effect on happiness, rather than the non-linearities in the most + least happy groups:
> KD reported that the effect of an approximately fourfold difference in income is about equal to the effect of being a caregiver, twice as large as the effect of being married, about equal to the effect of a weekend, and less than a third as large as the effect of a headache.
Interesting to put this into perspective - I wish this had been emphasized! It was buried in a single paragraph in the middle of the paper.
#20: "In fact, I’ve updated so far that I’m starting to worry that the problem won’t be building a political coalition against unsafe AI, the problem will be not overshooting and banning all AI forever. " Wow, that seems really implausible to me. Here's why: (1) Government has done very little so far to regulate the big tech companies, even when they are doing things that are harmful in a ways that are much more immediate and easy to understand than Foom -- for example, providing powerful vehicles for spreading misinformation. Why would AI change that? (2) There are enormous amounts of money to be made from AI. That will motivate the makers, and also the many many industries that can benefit from it, to support its continued development. (3) The main danger is hard to understand. I am smart, well-educated and have spent a lot of time reading about and discussing AI risks, and I still feel unable to judge how great a risk it poses to humanity. (4) My impression is that most of the public's not a bit concerned. Even most of my professional friends sort of chuckle when the subject comes up, as though it's an urban myth like that killer bees are about to invade.
I can think of 2 things that might nudge the US in the directing of reining in AI development: One would be some well-publicized bad events resulting from widespread use of GPT4: Cybercrime greatly increases; or a bunch of business go under because they give AI too much responsibility and it makes a big mistake; or some AI use in the financial industry causes alarming events in the stock market. The other is the reaction of groups of professionals whose jobs are shinking because AI can do so much of them. Recently saw a study of accuracy of AI and radiologists at making a yes/no judgment from a medical imaging result relating to a certain pulmonary problem. It was a problem where after a few days you knew for sure whether correct answer was yes or no. Radiologists alone were about 80% accurate, AI alone was about 84%. If the images AI was unsure of were passed to the radiologist, combined accuracy of AI plus radiologist was about 86%. Seems likely to me that training AI a bit more on the difficult-to-judge cases will result in AI alone being the most accurate way to handle ALL the images. Of course there will still be things for the radiologist to do -- but a LOT fewer of them. Seems like GPT4 & similar, unlike other advances, is going to be taking away the jobs of some upper middle class professionals.
But still I doubt that even the 2 things above that might make the public more afraid of AI and more resentful of it are enough to overcome the very powerful forces that nudge events in the direction of continued AI development.
(8) has been a fun story to watch unfold from the inside - for a lot of physicists in condensed matter room temperature superconductivity is the ultimate holy grail and absolutely nothing gets them wound up more than a good old fashioned USO (Unidentified Superconducting Object). These happen occasionally but the fact that this one actually got published in Nature makes it totally unique - they are going to get a huge black eye at the end of this debacle.
This has basically devolved into the physics version of a slapfight between Dias and Hirsch, conducted in the traditional way via weekly combative back-and-forth rants very thinly disguised as arxiv papers (see: https://arxiv.org/abs/2304.00190).
Referencing an earlier comment from the hidden thread but I knew someone had to be doing this! Don’t understand how this doesn’t solve almost the whole thing.
#6 matches something that I've been thinking for a while. Money doesn't buy happiness if you're a baseline miserable person, but it does if you're a baseline-happy person.
As a baseline-happy person, I tend to be content unless I've got some particular reason to be miserable. As I've increased my wealth, I've managed to use money to remove many of those reasons to be miserable. I used to be too hot in the summer, now I have air conditioning. I used to sleep less comfortably on a cheap mattress, now I sleep very comfortably on a fancy mattress. I used to be miserable doing housework (or miserable living in a filthy house), now I pay someone to do it for me. There are still some other miseries that I'm not rich enough to solve, like flying economy class or needing to go to work, but with more money I could solve those too and be even happier.
Regarding right wing populists: It's because they usually advocate some degree of "law and order." Everyone wants to have law and order, no one wants to do it. High effort, high resistance.
It's like people voting to clean their room, and then support dropping when it's time to actually clean their room.
>keep it a good place to live while the rest of the Midwest hollows out
I beg your pardon? First of all, Oklahoma isn't centrally Midwestern. Neither is rust-belt Ohio or barren Wyoming. The states in between have plenty of nice healthy towns and cities (sure the rural population is declining, but that happens everywhere) that are great places to live if you can handle a little snow. We like our rich farmlands and tight-knit communities; California can't have them.
(Yes, I'm still salty about the Wall Drug bit in Unsong literalizing "flyover states"; why do you ask?)
#2 : I have a hypothesis that numbered lists is just a good propaganda/educational tool from a purely memetic point of view. I don't have evidence, it's just a plausible just-so story that came to me when I thought about the question deeply.
As a starting point, I note that modern clickbait journalism and buzzfeed-style content mills have always loved the numbered list, "7 REASONS TO <insert random thing>, NUMBER <insert number between 1 and 7> WILL SHOCK YOU". Communism was also famous for its 5 year plans and 4 things that need to abolished and things like that. Judaism has the Ten Commandments, Islam has (off the top of my head, and translated non-professionally from Arabic) the Five Pillars of Islam, the Ten Companions of The Prophet Foretold To Enter Heaven, the Seven People Who Will Enjoy Allah's Shade On Judgment Day, etc.... Martin Luther upended Europe and Christianity with a 95-point list.
What are the common factors of clickbait journalism, Judaism, Islam, Communism, and East Asian philosophy ? I can find nothing except that they were all ideologies or families of ideologies that needed to indoctrinate, to preach, to capture attention and hold it. Indoctrination and propaganda are not necessarily bad things, some forms of education are and should be indoctrination, and lots of good Art is propaganda. Indoctrination and propaganda are just general purpose tools to capture attention and convince people, they are understandably favorite tools for ideologies like Islam, but they are neutral tools nonetheless.
Anyway, can we find reasons why numbered lists are a good indoctrination and propaganda tool ? Here are the 7 reasons why numbered lists are a good linguistic tool :
(1) The number serves as a simple checksum against data corruption or forgetfulness. You might forget how many The Pillars of Islam were or mis-remember the number to be 3 or 4 or 7, but you can never (or very rarely) make that mistake with The Five Pillars of Islam.
(2) The number serves as an attention grabber that builds tension and invites a listener or a reader to continue paying attention to get to the final thing. A classic template for a joke is "3 Xs are Y", where X is any general category and Y is any activity or action. Stories has evolved numbered chapters, television shows has numbered episodes nested inside numbered seasons, and movie series has numbers on every movie. Media players has a timer countdown to tell you how many hours and minutes and seconds remain. Etc..., Numbers are a "Progress Bar" service that tells you how much of the thing remains, and this is valuable from an attention regulation point of view. It's a simpe musical rhythm that starts at 1 and gradually builds up to climax.
(3) Lists are a good info organization structure in general, striking a sweet spot between free-form text on the one hand and elaborate diagrams or other overly-structured complex data structures on the other. Lists are how lots of writers outline essays, speeches, stories, screenplays, etc...[1]. When Dan Bricklin, the inventor of spreadsheet software, was thinking about a way to reference variables, he striked upon the grid system (i.e. a numbered, lettered, 2D list) as a way of organizing and storing the values in a spreadsheet[2]. Arguments in Logic and Analytical philosophy are written as numbered lists of premises and conclusions. High-level overviews of algorithms and code in Computer Science and Software Engineering are often written as ordered lists of steps. Laws and Constitutions are written as lists. The oldest surviving writings are lists of financial transactions. TODO lists.
(4) Numbered lists invite sorting along some natural or important axis. For example, the Pillars of Islam is sorted according to difficulty. First you simply have to profess that there is no God but Allah, just say a single phrase with your mouth. Then you have to pray 5 times a day, slightly harder but still within the realm of easy. Next you have to donate 2.5% of your money yearly, now we're talking money, but still a relatively small amount. Next you have to fast in Ramadan, spending anywhere betwen 12 and 16 hours without food or water or sex, for 30 days. It's hard, especially in the summer, especially in the near-equator regions that Islam is the main religion in. Finally, you have to go to mecca to perform the Haij, an ancient Arabic religious practice that Islam kept. The Haij is so difficult that it's optional, you can be a Muslim whose health or finances don't allow you to perform the Haij.
Other lists may sort according to importance or historical recency or any other topic-specific notion, numbered lists allow and suggest that to you by their very nature.
(5) The number can invite comparisons, metaphors, imagery and analogies, whether justified or not. Humans have built a lot of culture and mythology around numbers, and plenty of important things in our biology and in nature in general has distinct numbers. Two eyes, Two hands, Five fingers per Hand, Ten fingers in total per pair of limbs. 7 as a holy number, 13 as a lucky (or unlucky) number. Addition to mean conjunction or unity (e.g. "A man and a woman, 2 people unite in love to be 1"). Etc... If your numbered list can invoke a number with already a rich cultural imagery or history, it can be more effective and memorable. One such cultural imagery is the concept of the numbered list itself : you can piggyback off a more popular and established numbered list to get your numbered list to the audience quicker and more effectively. For example, if I name my list The Three Laws of X, Newton's and Asimov's (both Issacs, coincidence ?) immediately come to mind, boosting the attention my audience give me, even if its mocking or hostile (e.g. "Who this guy thinks himself to be ?").
(6) The list serves as a FAQ that followers of the ideology can always refer new comers or hostile opponents to in order to know the ideology better or to correct misconceptions or reiterate the ideology's core values.
(7) Lists are flexible, each point has a variable scope so you can always shorten or lengthen the list as you please by selectively splitting or fine-graining points and conslidating them. This point itself is such an example, I thought of it purely to get the list's length to 7.
tldr; numbered lists are a natural tool for storytelling and oration, they have plenty of educational and memetic advantages and disparate ideologies and propagandists have gravitated to them independently.
"I was listening to Stalin on, uh, Spotify, on the way over here...
And I'm talking about I listen to early Stalin. Not the later shit when he sold out. Early Stalin, when he was paranoid and he had the military behind him. Not when he started talking about wheat!"
29. The paper starts the clock on counting scientists in 1450. But what about before then? Before Francis Bacon was Roger Bacon. During the Renaissance there were lots of scientists in Italy, and then they went away. But were there many before the Renaissance? I don't know. Maybe it was just a return to normal.
Re: 5 I think what's going on is that the people most likely to talk about professor student dating are likely doing so either because:
1) They are in academia themselves in which case they have an incentive to signal that they totally aren't creepily into dating their students (and a lesser but still powerful signally concern for female academics).
2) Because it's an issue of concern for them. Few people who are just 'shrug it's not a big deal' are going to be very animated about it.
Overall this seems like an issue which suffers from a strong bias to overreport only one direction of concern. And that's bad because I think we are making things worse by imposing such strict rules for professor-student and, especially, TA-student dating.
Let's be honest, there is a great deal of temptation for this to happen. There are women who find that relationship sexy and men who are tempted to date younger women. So let's play out how this happens under a regime that says we fire professors who date their students and one that just requires reporting or class switches or mild punishment.
The student who dates under rule that says it's fine just give notice won't feel pressure not to ask to be transferred to a different class, advisor or even just to seek advice from someone else. OTOH, a relationship under the firing rule will necessarily start with the professor explaining that it's critical it be kept secret, that they'll lose their job if anyone ever finds out. That's going to make the student feel very reluctant to reach out (especially because they probably still have some feelings for the prof and just want get out of the situation not get them fired) for any help and absent that will have no way to extract themselves from the situation. And abusive relationships tend to make the victim doubt themselves. They, no matter how wrongly, may feel guilty for inducing the professor to put their career on the line.
Yes, you'll deter some professor student relationships from happening at all. However, that deterrence will be concentrated primarily among those who are best adjusted/most rule following and thus least likely to use psychological pressure, threats etc.. to manipulate a student. So the end result is you've probably denied some people a fun time and created a really awful trap for the most vulnerable students for the worst abusers of power to use against them.
The AI Risk phenomenon is a great example of why humanity will probably never implement most of the transformative technologies people have been theorizing about for centuries, if not millennia (i.e., nuclear energy - both fission and fusion, genetic engineering, gene therapy for common diseases, stem-cell therapy, space travel, etc.). Instead of working finding ways to properly integrate a new, transformative technology in ways that benefits our society, we consistently choose to regulate it either to death or near-death. Think about the number of problems that exist today because we have such low risk-tolerance and fail to organize enough cognitive capital to create ways to integrate these technologies in a sustainable way and make sure errors do not lead to a public outcry and subsequent bans.
As our world develops its communication technologies, any negative effect of a potential new technology will immediately be known to most of the world and will most likely be regulated to death, near-death, complete-uselessness, or near-uselessness.
Part of the problem is new tech is always on some pathway to worldwide use with time. Almost everyone will be using it given a reasonable amount of time. This propagates problems that otherwise would have been contained and effects which would be minimal. Part of the way to mitigate risk is to make sure the tech is only used in a constrained space where leakage is very low.
On BJJ and the related question of traditional martial arts vs. MMA, the explanation is simply that traditional martial arts have become stylized to fuck as it's become less acceptable *and* less needful to straight up murder a guy without using firearms. You can find video clips of traditional martial artists smashing someone's face in; they'll just look remarkably like e.g. a boxer or MMA fighter doing it. A fifty-year-old guy doing some feckless wavy-hands Tai Chi has just been wasting his time on ineffective martial arts because there was no pressure to be effective and he was probably too genteel to want to pay the cost of being effective in the first place.
23. Didn't Trumpism prevail on the anti-international trade issue? Biden signed that huge bipartisan bill with a lot of "America First" items on it. I doubt we would be here on this issue were it not for Trump making it hugely popular.
You could say that world events, not Trump, helped prod this attitude along. But I think it's hard to make the case that the USA is penalizing trade with England, Germany and France because of Russia, Covid or China.
In regards to El Salvador and "tough on crime policies" in general, I think a point that gets overlooked is that a lot of "tough on crime" rhetoric combines two policies that are often treated as a package deal, but in principle can be separated. These are:
1. Increasing the likelihood that a criminal will actually be arrested and punished.
2. Increasing the severity of their punishment.
"Tough on crime" policies like the one in El Salvador often do both at once. However, a lot of psychological research indicates that 1 is the policy that does most of the work. 2 isn't actually that effective. Criminals care much more about the odds of getting caught than what exactly happens if they are.
It seems to me that this might allow some kind of effective compromise between tough on crime activists and human rights activists. Give the human rights activists lighter sentences, nicer prisons, and more serious efforts at rehabilitation. Give tough on crime activists can get more cops on the street and more arrests.
The thesis-antithesis-synthesis process really does work a fair amount of the time.
I do it on Twitter a lot: I say X, somebody responds with "What about Y?", I'm stumped for awhile because Y is true, and then figure out Z would account for both X and Y being true some of the time.
I think the "do less AI" movement is pretty doomed. To me, it feels analogous to degrowth arguments related to climate change. The incentives are just overwhelming in the other direction, and trying to regulate against those incentives is just going to cause unintended consequences. You'll just end up with shadier AI researchers making more progress relative to those that care about safety, at the margin.
Again analogizing to climate change and how solar, batteries, nuclear and so on are apparently the best solutions to the problem, I think more positive technological progress in both AI alignment, in tracking safety risks like bioweapons, and in progress on rapidly rolling out vaccines and whatnot should be what we emphasize. I'm glad that more of these concrete proposals are coming out along with the ultimately not very useful ones like the 6 month pause and the EY Time article.
I also think we need a ground-up rework of the concept of AI alignment. I'm not even necessarily aligned with myself from 8 hours ago, so how are we going to align a nebulous concept of AGI with a nebulous concept of humanity? In my opinion, a 21st century version of Asimov's laws seems more promising: for instance, do no more than acceptable harm X (e.g., hurting someone's feelings by disagreeing with them) as a fundamental AI rule.
On "21 - College no longer worth the cost" - it's just... If you keep rising the cost and the real margin of benefit against something that you can learn yourself lessen in all possible metrics - what exactly were the expectations here? Real "surprised Pikachu face" moment.
Pretty sure if the question was about "is going to college beneficial" - the results would've been different. But is it beneficial ENOUGH to also go into the personal bankruptcy for years? No, it isn't.
Regarding the linguistics article and the mysterious prevalence of 1s pronouns based on the phoneme /n/, I don't have time or energy to properly read the full article or to give a very carefully thought-out response. But throughout a couple of decades of being an amateur comparitive linguistics enthusiast I've always been extremely skeptical of hypotheses proposing language proto-families, and, while I appreciate how well the article thinks through multiple sides of each question it treats, quite a few things in it subtly triggered my crackpottery radar even apart from its speculations about the dawn of self-awareness and snake mythology and so on.
For one thing, while I don't have time to actually check through the hypothetical roots for 1s pronouns among the ancestors dozens of language families, from the few I do know about I'm immediately suspicious that the author (and linguists he cites) might be cherry-picking certain roots and certain debated forms of roots. For instance, Austronesian (okay, mainly the Malayo-Polynesian branch) is one of the few language families where I am familiar with a few roots and knew that the 1s pronouns come from a pretty clear *aku root, no nasal consonants involved. I just checked this on Wiktionary, and not only does it back me up on this, I don't think it lists any 1s pronouns in the living daughter languages with /n/ in them. So why does this article list *’a(ŋ)kƏn as the root and count it among the evident roots based on /n/? It looks like this alternate proposed root has only possibly a nasal before the /k/ and that it ends in /n/, but (and this may be just an amateur assumption) the fact that a two-syllable root ends nasally seems like very weak evidence of it being related to another proto-language's root which begins with or is based mainly on /n/. It also goes to illustrate one of the main reasons (I'm sure) why linguists typically don't like to give much credence to proto-family hypotheses: the exact roots of the proto-language are already very hypothetical and up to much debate (in this case, apparently someone thinks that the 1s pronoun root in proto-Austronesian may have been *’a(ŋ)kƏn rather than *aku; that seems to reflect at best some disagreement among similarly qualified experts!).
Relatedly, a language family may have more than one root for a pronoun for a particular person/number combination. Indo-European is a good example of this: one of the roots is *me, and this is the one that the author lists under the proto-pronouns that don't involve /n/ (but then uses this list to point out that a bunch involve /m/ and that /m/ is not far from /n/). But the other root, from which Latin and Greek ego, Germanic ich/ik/ek/jeg/etc. and English I are derived, is something like *éǵh; this is highly contested, and in fact some common form (like *éǵhóm) evolving into both *éǵh and *me seems to also have been proposed, but I think we can agree that at first blush, *éǵh appears to have no similarity to *me. (And the author even points out how our pronoun I doesn't reflect a root involving /n/, now choosing to ignore our object form me which does have a nasal that he's pointed out is not so far from /n/!) So, how many of the 1s pronoun roots listed may have existed alongside other unrelated-looking 1s pronoun roots? Once you acknowledge that a proto-language may have distinct, unrelated-looking roots for a particular person/number pronoun (and why shouldn't they? many modern languages do!), the waters have already been muddied quite a lot.
I also don't know what to make of all the discussion in the article about time scales and how slowly over centuries and millennia pronoun roots may have evolved compared to other types of roots, and what this might mean in terms of evidence of anything. Are the reconstructed roots listed even proposed to have been in use during the same millennia? This may be a case of my not having read the article carefully enough.
I notice that the author mentioned in one place the near-universality of certain consonants in roots for some kinship terms most likely arising from the sounds babies make first and mentioned in another place how pronouns may arise from kinship terms. (In fact, this is a huge phenomenon visible in the present: in many East Asian languages there is a blurring between pronouns and kinship terms.) So why not propose as a possible explanation for the prevalence of /n/ (and other nasals) in 1s pronoun roots that in prehistoric times they may have often arisen from parental kinship terms which in turn almost certainly were influenced by consonants prevalent in baby sounds? Think of the obvious origins of "mama" and then think of mothers talking to their babies who refer to themselves as "Mama" in place of using a 1s pronoun. It doesn't seem like an unreasonable hypothesis to me.
I hope this parent comment might inspire some people who are more knowledgeable and have more time for research (or more research behind them) to chime in and do a little picking apart of the article.
re Newspapers. US has just two parties, and is huge. so you can earn easily from polarised columns. any country with a multi party system will have more camps, so harder to capitalize.
the UK has badly two parties. papers are partly having a political identity. I think lots of entertainment is involved. do has durability maybe with the internet. in the UK the political commentary is part comedy IMHO
I've meditated off and on for a bunch of years. I do think it makes you more calm, makes attention more controllable, etc, etc-that's all fine. My problem with meditation is something in my head I call the doormat problem: the more I meditated, the calmer I became, and there were a bunch of things I didn't solve which I would have otherwise solved if I had the adequate emotional impetus pushing me along. For instance, for anger/sadness, I'd just sort of meditate it away, which is not great.
In general, it felt like I was turning down the volume control knob on my emotions as a whole, which is not very pleasing to me, even if it gives me reduced despair/sadness/anxiety. YYMV, etc.
> 24: 200 Concrete Problems In AI Interpretability. “You can note which you're working on, and reach out to other people doing the same.”
Thanks for signal boosting my sequence! This is a link to the spreadsheet to track which problems you're working on, you can also see the actual sequence (with the problems, resources + tips, and a bunch more context) here https://www.alignmentforum.org/s/yivyHaCAmMJ3CqSyj/p/LbrPTJ4fmABEdEnLf
In answer to the question of whether it should feel okay for some position of privilege to be offered to "Jews only", I think the question you need to ask yourself is whether you'd be okay with someone else offering a similar set of benefits but advertising it as "No Jews".
If you're okay with this then I admire your consistency. If you're not then you should reconsider being okay with the first one.
29. Anything regarding Christian history is massively impacted by the belief status of the writer, particularly if they are Protestant, Catholic, or lately, secular. The writer’s institution is very relevant. The reason why you’ve heard so much from pro-Catholic side lately is because the internet has made it a lot easier to publicize and/or come across those opinions. English language liberal arts academia was an exclusively Protestant affair for a very long time.
Man, the question of whether schools should have a prohibition on professors and grad students dating was the topic of probably the single nastiest extended argument on the internet I've ever seen...
On the pear ring, what an utterly unappealing name.
Pears are some of the worst fruit, it's basically saying you're personified by a shitty fruit.
But this raises the question, if you had to describe yourself as a fruit, which would you be?
I followed Gwern at some point and have been puzzling for over a year now to try and figure out who he is (not like, his identity, I know he’s pseudonymous). What does Gwern do? What is his thing? Who is Gwern??
Regarding 35., there are some interesting stories regarding Stalin and litterature. A famous one is the relationship between Stalin and Mikhail Bulgakov (who wrote the falmous The Master and Margarita). It seems Stalin and him exchange letters and Stalin protected him despite Bulgakov being quite critical of his regime.
I fully recommend Vectors of Mind's blog. He's a smart ML engineer and this is my favorite post of his: https://vectors.substack.com/p/the-big-five-are-word-vectors
Oh, Glaze.
So, some people use artist names in AI art prompts, which is lazy even if you completely dismiss all ethical issues. Stable Diffusion (the FOSS one people use for porn) really incentivizes this in its prompting meta, because SD's prompting meta is generally a mess. One thing not that many people on all sides of this ideological hypercube notice is that *this is noise* -- for all but the most famous artists, the association between their names and their styles is pretty weak. It's less that SD demands you do that specifically and more that it's wound so tightly it wants you to write noisy twenty-billion-token prompts (though there is not *no* degree to which it wants you to do it). Glaze is built under the assumption it isn't noise, which means it doesn't actually do 'the thing it should do' all that much, but I'm no fan of prompting with artist names and would rather see fewer cringe prompts, and if all else were equal I'd be cautiously optimistic about something that could bring a ceasefire across various parts of the hypercube, even if it's placebo.
Here's where things get *really* funny, considering the copyright infringement accusations that fly around all AI art: Glaze infringed the hell out of its licensing. Not in, like, an ambiguous fight-it-out in the courts way, in a "massive chunks of plagiarised code in violation of multiple incompatible licenses" way:
https://jackson.sh/posts/2023-03-glaze/
https://twitter.com/essex7927/status/1636500422855630848
This is fixed now, which is nice. Nonetheless, it's kind of incredible that people who take the hard "all generative art is plagiarism" stance did...unambiguous plagiarism.
https://davidrozado.substack.com/p/gag
According to this link, the increase in partisanship *is* universal and simultaneous across countries in a way that’s hard to explain without invoking the changes in media structure and funding.
I think it took a long time to invent brazilian jiu-jitsu because unarmed one on one combat between people matched by weight is a pretty niche activity not subject to much optimization pressure.
Also short stories probably got replaced by tv episodes and comic books
Tulsa article is very interesting, trying to think of a good pretext for forwarding it to some people I know.
One thing that the linked article didn't mention is that Tulsa Benevolent Dictator for Life George Kaiser is 79. I'd be worried about what happens when he passes; will his heirs just rug pull the entire city and move to a coast?
Some speculation, re: 15 minute cities.
If those dense cities don't have any farms, then they have to import their food. In order to import food, they have to produce goods and services that farmers will trade for, or goods and services that are desired by other people who make goods and services that farmers want.
This is basic international trade economics, but with borders drawn around arbitrarily small regions instead of large nation-states.
So if these 15 minute cities don't have light-industrial zoning, then the economy will be propped up by finance types, or professional bloggers, or journalists, or any type of profession that generates revenue from outside the city. Those professions are not necessarily stable; what happens if the internet goes down for long periods of time?
These facts appears to be the reason that American cities have fluctuated between dynamic, energetic utopias (a la Thoroughly Modern Millie) and crime-ridden dystopias (a la Taxi Driver) over the past decades. Making cities even *more* insular might compound the problem.
Although I will concede that having grocery stores on the first floor of an apartment building is more efficient than zoning commercial buildings far away from residential building.
RE: Stop AI Development for 6 months, for those interested, Lex Fridman has talked with Max Tegmark who is the father of the letter (if I understood correctly; he is at least the President of FLI) and he elaborates on his reasons for the letter and what he hopes to achieve (https://www.youtube.com/watch?v=VcVfceTsD0A, time stamped).
Interesting tidbit - he (obviously) feels like all the companies are bound together in a race to the bottom and hopes that the letter will break this system... and the main analogy he uses for this (and for most of the talk) is Scott's very own Meditation's on Moloch article (start at about minute 40 for full context, reference at 41:30). So, @Scott, how do you feel about this? It appears you contributed to the letter indirectly as well :P
Twelve years of hanging around the rationality community and I've finally made it into an SSC post. I can finally rest now
29. I still want to see this data broken down by places that instituted the Counter-Reformation voluntarily and places where it was enforced after being conquered by Catholics. There would be no significant reverse effect because the Protestants mostly lost those wars in terms of net territory. The conquest of Prague alone could skew data significantly. Meanwhile places like Italy, Vienna, Paris, etc all continued on. They did eventually suffer their own declines. But mostly due to things completely unrelated to the Counter-Reformation. Italy had an economic crisis. Paris had Louis XIV being Louis XIV and insisting that "absolute monarch" really means absolute. Including over the universities. Poland had The Deluge. Etc.
Diets: my understanding is that due to rebound weight gain outpacing weight loss, dieters tend to gain weight over time faster than comparable non-dieters. This is deceptive, because the first diet tends to yield better results than subsequent attempts , giving the impression that better adherence or finding the 'right' diet that's easy to adhere to will change things. They don't, and the the cycle perpetuates.
Creating more diets to try will only contribute to accelerating obesity long-term through the cycle of dieting.
30. I tend to think the media as cause narrative is a bit off. But most countries have become more polarized and are facing fairly severe crises over this and populists. The exact severity varies significantly. But this is a common international phenomenon among the democracies.
While I haven't actually ran the numbers eyeballing it the non-polarized democracies mostly seem to be those which are growing fairly rapidly economically (in real per capita terms) and have high trust in institutions. (And they still have polarized media.)
Specifically, this doesn't decrease feelings of difference. But it decreases the feeling that these differences can't be overcome or pessimism about the future of the country or the feeling compromise is bad. For example, a lot of Indians feel Indian society is very deeply divided but mostly believe these differences can be overcome. (There's also countries where the population feels polarized but politicians have managed to keep a lid on it. Often strong party systems where the ruling party defeated a populist revolt.)
As for how this happened the pattern in trust is easy. An institution does something and loses trust. It sometimes slowly grows (or can grow rapidly if something happens). But often it just stays low. For example, for the US government it was Vietnam. Trust in the US government at the start of American major troop presence (1961) was 77%. When the US pulled out (1975) it was 34%. And ever since then the average has been in the 30s-40s. For the banks/Wall Street it was the slowdown then crash of 2008 (from 53% in 2004 to 18% in 2010) and it now stays in the 20s. This has happened in succession such that the only trusted institutions left in the US are the military, first responders (police, firefighters, etc), and small business.
Reasons for economic growth are more complex. But the average real per capita growth since 2000 has been 1.2%. It was double that, 2.4%, from 1945 to 2000. The statistic has been consistently below the trailing century average since 2000.
Yeah, so anyway, grow the economy and get higher quality institutions. Easy to say, hard to do.
> Why do East Asians have so many famous numbered lists (“Four Noble Truths”, “Thirteen Classics”, etc)?
As Gwern himself mentioned as well, monosyllabic digits play a role (I'm not familiar with India), but I think it's mostly a matter of style. Certain ways of writing can just feel more natural for no particular reason. Reminds me of how English speakers love creating acronyms that spell a word, even though English is not really a language well suited for pronounceable acronyms, with only 5 vowels.
35. Georgian was somewhere between heavily discouraged and banned so Stalin wasn't exactly facing a lot of competition. But no, it wasn't anything to do with his later dictatorship. He actually went out of his way to hide them. As a child Stalin was an excellent student and a gifted poet and singer. He was also fairly religious and had a strong interest in art and music. (The interest would stick with him for the rest of his life.) He got into a good school and had a bright future ahead of him.
Then he read Marx and Lenin and a bunch of socialist literature and became a socialist revolutionary. He walked away from his educational life to be a full time, professional revolutionary (and/or criminal). He claimed he was kicked out of school for revolutionary activities but in reality he started skipping class and eventually just didn't show up. Instead he spent his time being a bank robber, extortionist, and general criminal in an attempt to disrupt the Tsarist system and raise funds for the revolution.
Re: El Salvador: It's not even clear that the previous president's policies didn't work. The homicide rate had been on a downward trend for years by the time Bukele got elected. https://www.statista.com/statistics/696152/homicide-rate-in-el-salvador/
Non-American with zero geography knowledge here; can someone kill the joke by explaining #38 (the map) to me?
30. <i>My only concern about this story is that some other countries haven’t become any more extreme/partisan since the advent of the Internet - is this because their media didn’t undergo this process? </I>. I‘d argue it is because media in other countries was already very partisan before the internet. This was certainly the case in countries I know well - Italy and Austria. There were (and still are) newspapers across the political spectrum. In Italy if you wanted sports news you bought Corriere dello Sport. Sports coverage in the „quality“ newspapers was meager. Same was true of classifieds. There were specialized publications for that. Italy even has a specialized publication for crosswords - La Settimana Enigmistica. In Austria Socialist voters read Der Standard, centrist conservatives read Die Presse. If you cared about soccer you bought Kicker. What held everything together in these countries was linguistic/ethnic identity. Now that we are also losing that as social glue thanks to mass immigration it’s not clear how the center holds.
I know the answer to Gwern's "Whatever happened to short stories" question.
Short stories were basically found in magazines, which were hugely popular in the interwar period, into the early 1950s. All the famous writers of the period wrote for them - Hemmingway, Fitzgerald, Steinbeck, even Dahl and Vonnegut got their start, at the very tail end of the period, writing in magazines. Gwern is focused on sci-fi, perhaps because that's his interest, but just about every genre was in just about every mainstream magazine, from Cosmopolitan to Colliers to the Woman's Home Companion. Importantly, because the magazines were big business, short stories were where the money was - you wrote short stories for money, and novels for literary kudos. And that money is why everyone was writing short stories.
Then the bottom dropped out of the magazine market due to television, subscriptions and advertisments both collapsing. Those magazines that survived had to pivot away from competing with television, and be complementary to it. One thing that TV can do very well is tell a short story - that's what an episode is, but TV can bring it to life much better than a magazine can, so magazines dropped short stories, and those few that still carried them were no longer paying much.
A lot of people still liked short stories as they were used to them, so for a while the genre limped on through anthologies and collections. But there was no longer an economic way to release a standalone short story, and TV wasn't just a competitor to the magazines, it was a competitor to short stories in total, so the short story died as a popular form, with only a few weird literary hold-outs remaining.
I had thought the Internet would see a revival of short stories, but it hasn't happened to any great extent.
RE 2 "Why are there so few pairs of extremely successful identical twins?"
There are some, such as the Bryan brothers, who hold a bunch of tennis records in men's doubles.
https://en.wikipedia.org/wiki/Bryan_brothers
In Gwern's list of questions, he asks if there are any drugs that work as "anti-psychedelics": i.e. are there any drugs that strengthen your sense of identity, rather than dissolving it. Gwern asks "Why [isn't there a drug that makes me feel] with the same absolute certainty that I feel ‘all is one’, that I am a unique snowflake utterly unlike any other human being or organism?"
Gwern thinks that stimulants don't qualify on the grounds that they don't induce psychedelic effects. I disagree. During coke highs, people often report that they believe themselves to be possessed of superhuman qualities and their demeanor changes to become more arrogant and narcissistic.
This sure sounds like it's approaching "unique snowflake" territory to me. It's almost as though stimulants are turning up the amplifier knob on somebody's sense of identity, magnifying their sense of themselves to mythic proportions.
How reasonable does this sound to other people? I don't have much expertise on drugs, and I'd appreciate somebody who knows more about habitual stimulant use who could explain how frequent this egotistical high really is.
(Tantalizingly, this entry on "ego inflation" on the Psychonauts wiki mentions that habitual stimulant use can lead to increased narcissism and egocentricity, even when the addict is sober. Can anyone, maybe a clinician, shed some light on whether or not this is a real feature of stimulant addiction? https://psychonautwiki.org/wiki/Ego_inflation)
Re approval voting in North Dakota: I'm confused, because this more recent article states that a more recent vote on Wednesday failed to override the governor's veto of the ban. Meaning the ban has not been put in place, and approval voting is still ok? Has my brain been overwhelmed by the double negatives? https://www.valleynewslive.com/2023/04/20/bill-eliminate-approval-voting-fails-override-veto/
Typos:
In 32, the links are backwards.
In 33, the first link goes to the substack itself, not the pronoun article.
> surely it would be even better if you spotted someone wearing the ring and then you could use your smartphone to call up their dating profile.
What? If you've already spotted them in-person, you can just go up and talk to them. Why would you want to dance around with dating profiles and instant messaging to set up a time and place to meet in person, when you can just walk up and say hi?
Maybe I'm old but I don't get it.
25. My big question with the El Salvador crackdown is whether it's financially sustainable. That's 2% of the population, and a higher percentage of young men - can the government afford to keep them locked up for a long period of time?
27. The opposition from the North Dakota state government is bizarre. They're getting hung up on approval voting when there's a ton of different ways that local governments have experimented with government structure? It's a pity.
36. I was wondering why all the 5-over-1s all had the same block design (the three-story walkups seem to have more variety in design).
21. I speculatively attribute this to the aggressive campaign of misinformation about the burden of student loan debt. In reality, for about 70% of four-year graduates, it's less than or equal to about one year of the college wage premium ($30,000).
Granted, there are taxes, and the median college graduate differs from the median high school graduate in terms of cognitive ability and/or conscientiousness, so college doesn't literally pay for itself in one year, but in a large majority of cases it pays for student loan payments.
34. I can't figure out what Loeb is doing here. He's a serious scientist who has made good discoveries. But the whole thing with Oumuamua being an alien light sail, despite little evidence of that, and this new thing with the "interstellar" alien technology at the bottom of the ocean is weird. Is he using it for publicity to attract funding for research? Astronomy could certainly use more money, but I'm not sure "aliens!" is the way to go. I mean there's a small possibility he's right, but for Oumuamua we really have no way of telling and the chances of finding a small rock that might have come from interstellar space at the bottom of the ocean are basically zero...
7. I'm having trouble understanding the 15 minute city idea. The PDF someone linked to said there was a priority on protecting local centres, which... I guess means strangling out their competition? I don't know how else you're protecting them.
The first link makes it sound like they're trying to turn Oxford into Midgar. That's a good way to get yourself stabbed in the back, and also blown up by Godzilla and smushed by rocks.
7. It seems worth clarifying that this was an Oxfordshire County Council plan. While Oxfordshire County Council is a government entity within the UK, it's not "the British government"!
Wrote a new post on how gaming has shaped and informed Japanese and American cultures
Would love a read!
https://hiddenjapan.substack.com/p/pachinko-vs-pinball
I’m impressed, actually a bunch of interesting stuff here
22: "Comments on social media accounts have pointed out the possibilities of how women who wear the ring might be solicited more often."
...Not "might". "Would". That's the entire point.
Good prank if the pear ring thing becomes slightly successful but not to the point where most people know about it: give someone else such a ring, and see hijinks ensue.
28. "Trevor’s understandably morally anxious about benefiting from his Jewish ethnicity. Should he be? Is this any worse than benefiting from white privilege or male privilege or whatever other forms of privilege they’ve invented since I last checked?"
I imagine Scott is just trying to provoke us into saying the following (but I think the following is correct):
This is a false analogy. To the extent "white privilege" and "male privilege" are still a thing, they are *statistical* realities, or even just decoys for the real thing (whiteness correlates with class, which in turn determines much of what gets called "white privilege"). The right analogues here would be "in the US, "Jewishness" [defined as straight maternal descend from Orthodox Jews living in 1800, say, or whatever else you want] correlates with class" (very likely true) or "being brought up in Judaism makes you less likely to be molested by a priest" (also true).
The right analogy would be with Old-Style white privilege - being entitled to certain things if and only if you were white (or white enough for whoever is testing). A thing in Apartheid South Africa, certainly. The G.I. Bill also worked out this way in the U.S. South, though perhaps in a slightly less openly declared way (though my understanding was that it was completely deliberate, and other people saw through it). Would you sign up for a generously funded benefit program open to whites and only to whites? Or Asians, people of Latin American descent, etc.?
"Is it any worse than benefiting from being an American, and so having access to social support and benefit programs that Sudanese and Bangladeshis can only dream of?"
That's a thornier issue. Your average citizen of a rich country probably does not think very often of what an enormous privilege it is to be born with that status - and they would be baffled to hear that that privilege is questionable. (Of course, that doesn't make it *not* questionable.) I've even heard it said that, if it weren't for those privileges, citizenship would be meaningless - that is saddening and unintentionally funny, as that is exactly what you would associate with Roman decadence, say: citizens thinking of citizenship _merely_ as a set of privileges (which are great, because they have them).
Of course that may also why most Israelis wouldn't wink twice at the proposition: unlike most American Jews, they are brought up to think that "Jewishness" is a nationality - and most Israelis are completely comfortable with having preferences for the dominant nationality in their country (and in fact, according to a recent poll, most of them openly believe citizens of that nationality should get preferences over fellow citizens not of that nationality).
Yglesias' point about crackdowns - that the first requirement was an honest and effective force capable of doing the cracking down, and that once you have such a thing, the exact method by which you use it to deal with the gangs is really irrelevant - seemed to me the really useful insight.
If your police (or military or paramilitary) force is capable of being bribed by the gangs, then it doesn't matter what orders you give it, it's not going to crack down in an way that is effective.
30. 15-20 years ago, I remember seeing someone point out the contrast between the WSJ, which at the time had made its editorial page free and kept its news behind a paywall, with the NYT, which at the time had done the opposite. It might have been Tyler Cowen, but I can't find the post now.
That seems relevant here.
23: My hunch is that people partially support those policies because they support / like Trump (i.e. reverse causality). So once he is elected his supporters no longer really care cause goal accomplished.
>My only concern about this story is that some other countries haven’t become any more extreme/partisan since the advent of the Internet - is this because their media didn’t undergo this process? Why not?
The American "city newspaper" is not a typical model for most countries. I live in Britain; we have city newspapers, but they are the second-tier publications, like the Manchester Evening News (my city paper). The first-tier newspapers are national newspapers. All of them are based in London, though one (the Guardian) used to be based in Manchester until the 1960s. Because there have usually been around ten national newspapers (there are seven that have been continuously in print from before WWII; there are currently three others in print), they have had to compete with each other, and have often done so on the basis of their political positioning.
The city newspapers were usually exactly the sort of inoffensive entities stuffed full of ads that was described, but the national newspapers have always been viciously partisan. A famous 1992 headline, on election day, ran "If Kinnock wins today, will the last person to leave Britain please turn out the lights" https://pressgazette.co.uk/wp-content/uploads/sites/7/2016/04/The-Sun-If-Kinnock-Wins-Today-600x755.png This was in the then highest-circulation newspaper in the country.
The reasons for this are pretty simple: you can get an overnight train from London to just about anywhere in the country, so the first copies of the paper off the presses at about 10pm were loaded onto trains for the most distant parts of the country, with later printings progressively headed for closer regions. The upshot is that the entirety of Great Britain could, by the 1860s, read the national press. Americans, accustomed to newspapers being associated with a city, tend to refer to The Times as "The Times of London", to distinguish from the New York Times; this isn't incorrect (its editorial offices are, indeed, in London), but The Times of England or of Britain would be more informative.
It wasn't possible until the arrival of air freight to physically distribute a newspaper around the entire United States overnight, so a late-evening printing of the New York Times could be delivered in San Francisco the next morning, and air freight would make newspapers very expensive. The only national American newspaper is USA Today, which operates by having 37 printing presses spread around the country and electronically transmitting page layouts to each, so identical papers are printed in a variety of places. The technology to do this didn't exist until the late 1970s, and USA Today was founded in 1982, by which time the American "city newspaper" model was established.
The other upshot is that the UK distribution model (until it broke down recently with the internet) was that there were large numbers of independent newsagents who hired deliverers who would deliver the paper of choice to each household, so the newsagent would have a few hundred houses on their books, and each would order the paper they chose, and then the deliverer would deliver that paper to that house, then a different paper to the next house, and so on. The US distribution model was for the newspaper itself to organise delivery; this made it much harder for new newspapers to break into the market in a city. USA Today was usually only available at newsstands.
Just as an example of how different it was; in the 1990s and 2000s, if you stayed in a hotel in the UK, they would ask you which newspaper you wanted in the morning and would then deliver the requested paper to each room. In the US, you generally just got USA Today, or the local city newspaper, and they didn't ask.
> They also hoped that sometime in the future there would be a “fire alarm” - something would happen to get people and policy-makers’ attention - and then the political route would open up. [...] I personally am pretty surprised it was just “GPT-4 got released and was very good”. [...] I’m starting to worry that the problem won’t be building a political coalition against unsafe AI, the problem will be not overshooting and banning all AI forever.
It's a bit early to worry about *that*.
AFAIK there no mainstream politician has publicly come out in support of slowing down AI research. The idea of AGI risk is seen with contempt and derision by your average tech-savvy internet nerd, let alone your average 40-something blue-collar worker.
Sure, big names are coming around, but as far as the public goes? There's about as much support for AGI safety as there would be for a carbon tax in the 1980s.
The bill was authored by representative from WEST FARGO THEY WERE JEALOUS
I really love everything about HustleGPT, especially because it turned out to be absolutely useless and the guy is desperately trying to convert his 15 minutes into money, but it looks like that's not going to well either.
2. "Why are furries so artistically and economically influential compared to other fetishes?"
I don't know, 50 Shades of Grey sold a lot of copies, people are using ideas of consent drawn from kink in vanilla sex (sometimes to unreasonable lengths in my view), role-playing games are using the stoplight system...
5. Extremely un-PC hypothesis: a lot of the academic rules on dating are invented by lesbians trying to compete for bisexual women by driving out straight (or bi) men. I don't know where else to put this that will actually get reasonably smart people to come up with critiques but won't dismiss it out of hand or simply believe it as part of the 'gay agenda'.
6. So more money won't make chronically unhappy people happy, but will make everyone else happier? Yeah, I'd believe that. Didn't someone show here it was actually logarithmic, such that going from about $100K to $200K is like going from $50K to $100K, but it was hard to recruit enough rich people to prove it? Tracks my personal experience but that proves nothing.
15. I suspect the cat's out of the bag on that one, but worth trying.
21. With the college thing I think the debt problem has grown to the point it's reaching parity--i.e., jumping ahead in the job queue with a college degree no longer justifies the now-huge debt you acquire. That's assuming you don't get into Harvard, of course, and what point in the prestige hierarchy breakeven is reached is probably questionable and depends on a bunch of other things.
22. Seems like a nice idea, silly though it is. A lot of people don't want to be flirted with, so they made it socially dangerous to do so (at least in liberal areas)...so this solves that problem. I kind of doubt it'll catch on. If there's one thing we know we nerds suck at it's romance.
23. The obvious answer is their policies don't work. ;) You could say the same about communists, and I wonder if there's a certain symmetry there. My guess is right-wing populists drive off the professional class necessary to implement policy successfully, which was certainly a problem with Trump--he had to rely on people more interested in tax cuts to staff his administration. Honestly he wasn't doing so bad in some ways until COVID hit--the economy was growing without inflation and he didn't start any wars. A less narcissistic person (which is to say, anybody) might have pulled out a close reelection.
28. Go for it, dude. The thing I've realized after reaching middle age is, everyone takes advantage of their privileges, and then lies about it to make well-meaning, honest people feel guilty so they don't take advantage of theirs. This one seems pretty good for everyone--Middle America seems underinvested right now. I don't know about the culture clash between the 'blue' Jewish arrivals and the 'red' locals, but as David Brooks has pointed out, blue cities in red states are doing pretty well. Maybe the challenge is to keep either side from getting all the power.
35. Well, that other genocidal dictator was a painter.
Regarding research about the Catholic Church from Martin-Luther-University: I'm working there (science department) and can assure everyone that we're pretty much the most secular institution in the least religious part of the world (former East Germany). ;-)
By the way, regarding nuclear power, it's worth noting how utterly different the attitudes to nuclear power are in comparable countries. At the same time as Germany has proudly switched off its last power plants, Finland - smaller, but similar to Germany in many ways, though also different in many ways - has basically all parties in the new parliament supporting more nuclear power, and in a recent poll 68% of people see nuclear as "positive" or "very positive" and only 6% as "negative" or "very negative". https://twitter.com/Kaikenhuippu/status/1649011146362433539
Americans are also, in most polls, more pro- than anti-nuclear (see, for instance, https://www.pewresearch.org/fact-tank/2022/03/23/americans-continue-to-express-mixed-views-about-nuclear-power/), though not as positive as Finns.
Re why there are no more successful pairs of identical twins, until recently twin pregnancies were associated with many complications with sequelae such as lower IQ (-4 points or so on average). This is not seen in more recent cohorts, e.g. https://www.researchgate.net/publication/228853282_Is_there_still_a_cognitive_cost_of_being_a_twin_in_the_UK
2. The twins thing has an intuitive explanation for me:
There’s probably so few “extremely successful” identical twins because the marginal difference between “highly successful” and “extremely successful” is often razor thin. I’m going to use sports as an example because it’s very easy to evaluate objectively, but I believe the talent distribution we see so clearly in sports exists in other domains as well.
The difference between a guy who, for example, ends up playing 12 seasons in the NBA and a guy who never makes a roster and plays overseas is extremely tiny, especially from an outside perspective. If you have 2 twins, one is 6’8” with a 7’ wingspan, and one is 6’6.5” with a 6’9.5” wingspan, that alone is enough to make one person an NBA prospect and the other just a great D1 basketball player. Even though both guys are functionally the same, in todays world of max efficiency those small differences determine whether someone is “extremely successful” or just regular old successful.
I assume the same principle applies to the difference between, say, a world renowned physicist at MIT vs his twin brother in the physics department at Ohio State, or whatever (example is purely hypothetical). Both brilliant and successful, and on a scale of all humanity pretty much equally smart.
I think this, combined with just how rare identical twins are pretty much explains it fully.
Pear ring seems like a great idea, but I wish that someone would invent and publicize a gay version so I can advertise my singleness to the correct gender. Rainbow? Purple? I pass as straight, and I definitely don't want a piece of jewelry to cause me to be constantly approached by men.
Does anybody know of a transcript or commentary of the interview with Ryan Kupyn on forecasting AGI? Video doesn't work for me - esp. 1 hour.
Re #28, why would these billionaires think this is a good idea? They do realize this exactly plays into (and in this case confirms) the anti-Jewish trope of Jews being more successful because of nepotism, where other Jews in high places preferentially grant jobs and benefits to their "tribesmen" over the "goyim," right? I fail to see how they think creating the exact classic conditions for resentment against a local Jewish community would be a good idea for their local Jewish community. Maybe a few more talented Jewish people move in because of this program, at the cost of all the locals who find out about this program viewing them (and by extension the Jews already there) as privileged interlopers and not wanting anything to do with them. Seems like a pretty obvious net loss to me.
(Edit because I realized I confused the words "interloper" and "interlocutor")
With regards to Gwern's question on writing vs. coding in the morning. My best coding time is usually around 9-11am. 8-9am I'm still waking up, 11am I'm starting to get hungry for lunch. I usually eat at noon, then I'm tired from eating lunch from roughly 1-2:00. 2:00-3:30 can sometimes be ok if I'm well rested enough and I'm usually getting tired again as the afternoon wears on. The later I code, the more easy mistakes I make, which I usually catch the next morning while wondering what on earth I was thinking.
I've also noticed that most of my best ideas come in the morning, oftentimes I'll wake up with them. To me, there is very clearly some mental background process that continues to work on whatever problem I'm dealing with at the time, and whose results generally filter into my consciousness in the morning after waking. If we assume this process is similar, this would explain why writers feel more effective in the morning: they are getting the creative output of their sleep processes when they wake up.
As to why coders stereo-typically like to stay up late, while I might be biased, I think coding is probably more rewarding that writing is and this leads people to get wrapped up in it in a way writers typically don't. You can write a bit of code that adds a bit of functionality, then you can test it and figure out all the little bugs that inevitably creep in, fixing each one is a little reward like solving a puzzle. And when you fix all of them and the code works and you can see the new feature, that's a bigger reward and so, so satisfying. Fixing the bugs can also task you, because you are so close to having it working and you just need to iron out those kinks so 5 more minutes to try 1 more thing.... can become 1 hour, then 2 (it's pretty easy to get into a flow state when coding). That said, as I mentioned before, my ability to code degrades the more tired I am, especially as the evening wears on and irrespective of how focused I am on the task, so I place being well rested at a premium to maintain optimum coding ability. Everyone else seems to just rely on caffeine (I don't use it so I'm pretty sure I'm an outlier, for whatever that's worth).
> My only concern about this story is that some other countries haven’t become any more extreme/partisan since the advent of the Internet - is this because their media didn’t undergo this process?
UK newspapers were always partisan. UK broadcast media aren't allowed to be. If you want to understand US polarisation, the repeal of the fairness doctrine should be your first port of call.
A big reason behind #21, the drop in people saying college is worth the cost, is that the opportunity costs went way up: youth unemployment went from 16.3% in 2013 to 7.5% now.
https://www.bls.gov/news.release/archives/youth_08202013.pdf
https://www.statista.com/statistics/217448/seasonally-adjusted-monthly-youth-unemployment-rate-in-the-us/
"The alignment community always figured their concerns sounded too weird for normal people to care about, that politics was a lost cause, and that our best hope lay in technical research. They also hoped that sometime in the future there would be a “fire alarm” - something would happen to get people and policy-makers’ attention - and then the political route would open up. I personally am pretty surprised it was just “GPT-4 got released and was very good”."
And this was always a bizarre belief. Rationalists did not invent AI fear, or even AI doomerism. How much of science fiction is about a mad scientist creating a being that turns on him? That's what Frankenstein is about, and that was written in 1818! Isaac Asimov's "I, Robot" series, written in the 1940s and 1950s, is about the problems of AI alignment. One of the most iconic evil AIs, HAL 9000, is from the 1960s. So are Daleks, who want to exterminate the whole universe. More recently, we have movies like Matrix or Terminator, TV series like Battlestar Galactica, and video games like Horizon Zero Dawn.
Storytellers have been taping into the public's fear of AI since before AI existed. Whether they like to admit it or not, rationalists are embedded in this culture and take their inspiration from it. It's no surprise whatsoever that improved AI capabilities would heighten AI fears and potentially lead to the government snuffing out the most revolutionary technology since electricity.
Re #26 - remote work
I've been saying for several years that remote work is not sustainable for businesses, maybe any large business. That we're seeing significant drops makes perfect sense to me, and I don't think we'll ever get to the point where a majority of jobs are even partially remote. Honestly, it seems that any discussion of significant remote work is a symptom of the types of people who have the freedom to write about such things being one of the few professions where it might work (journalism).
The core problem is that most employees do not have the knowledge and drive to start a career remotely. Even if experienced employees can complete their tasks from home, new employees need direction and training (this includes experienced employees starting at a new business in most cases).
If you allow your long term employees to work remotely, there's no one around to enculturate and train the new employees. You've also got potentially significant problems making sure that work is being completed and that employees are using their time effectively. Many information jobs that could work remotely involve long term projects from multiple people, which permits some people to ride on the accomplishments of others or pretend to get their work done before some deadline looms and stuff crashes. This exists in person, but proximity allows managers to have more opportunities to review and notice problems.
Compounding this is the fact that the vast majority of jobs require a physical output at a particular location. You can't be a remote nurse, mechanic, farmer, dock worker, etc. That he thinks only 40% of workers will be fully in person is pretty much insane, if you understand the types of work that exist. We're not going to have WFH for the service industry, manufacturing, construction, retail sales (distribution network), let alone the more obvious like mining and agriculture. These fields make up about 80% of the private sector workforce. That a few people could work remotely for companies in these fields is probably true, but we're talking well under 1%, probably far less.
Even speaking theoretically, it's probably the case that fewer than 10% of jobs can even be done partially from home. The chart pictured above is probably only a subsection of jobs - likely not polling too many restaurants or nursing homes to determine the percent of people working from home.
"A few years ago a weird object called Oumuamua entered the solar system and a few astronomers speculated in might have been an alien spacecraft (it’s since left). "
It's basically just Avi Loeb and his students/postdocs. Loeb has been labelling everything in sight aliens, not just the interstellar comet, but also fast radio bursts (which have since been shown to originate from neutrons stars with strong magnetic fields): https://www.scientificamerican.com/article/an-audacious-explanation-for-fast-radio-bursts/
I was taught that the Catholic Church becoming anti-science for the Counter-Reformation is what allowed Protestantism to seize the pro-science position.
Copernicus, after all, did his fateful research in the employ of the Church trying to develop a better calendar that correctly dated holidays with respect to the life of Jesus. His book didn't really get much traction even in the scientific community until it got an endorsement from Tycho Brahe, on account of some testably wrong predictions Copernicus made based on a much too small solar system. (The endorsement being posthumous and probably actually attributable to Brahe's student Kepler is another story.)
Initially, the Protestants opposed heliocentrism as an "innovation". They complained of a pattern of behavior by the Church of interpreting the Bible to the point of, in the Protestants' eyes, changing the meaning. Heliocentrism is one of the flimsier ones. First, the orthodoxy here was more Aristotelean than Biblical. But even the Biblical textual argument is weak; as I understand it, the argument comes down to two passages: Bereshit 1:1-3 (God split Heaven and Earth before creating light), and a single line from Joshua (*something* happened to the sun by Joshua's command; the Protestants translated it as "stopped", modern Jewish scholarship has it as "disappeared", perhaps like an eclipse). Both were sort of flimsy arguments! But in the Protestants' eyes they fit into a larger pattern, and taking cheap shots at an opponent's controversial theory is a common way people score points.
But as part of the Counter-Reformation, the Church chose to go fully hard-ass on heliocentrism. Possibly *because* the Protestants were so ineffective at stamping science out, the Church might have felt they could make the case that they were holier and more originalist than the Protestants. They had the institutional advantage to run those inquisitions, after all. But by this point the Scientific Revolution was clearly in full swing, and while Galileo was a jerk his trial was clearly a bit funky, and so the Protestants had the opportunity to pivot to the pro-science position.
That equilibrium lasted a few hundred years. For my money, it seems less true now; these days Evangelical Protestants seem at least as skeptical of science as the Church.
6. From your summary, if they were both right, then neither was right. My understanding is that they concluded "some people get happier with more money, whereas some people don't."
31. I never agreed with Guzey's conclusion from that first debunking article (the only one of his that I read) -- IMO plenty of sleep is obviously important. But his meta-point seemed more important: that most evidence is pretty muddled and popularizers disingenuously present it to the public. But reading this link, maybe I was just projecting this meta-point onto him if he actually tried to sleep 4 hours a day.
One of Gwern's questions: "Nicotine is one of the best stimulants on the market: legal, cheap, effective, relatively safe, with a half-life less than 6 hours. It also affects one of the most important and well-studied receptors. Why are there no attempts to develop analogues or replacements for nicotine which improve on it eg. by making it somewhat longer-lasting or less blood-pressure-raising, when there are so many variants on other stimulants like amphetamines or modafinil or caffeine?"
Good question. It seems a shame that the stigma that rightly arose against cigarette smoking has probably prevented progress in finding and exploiting the many beneficial uses of nicotine. If you're not familiar, google it... May help for depression, Alzheimer's, ADHD, etc.
Seems like perception vs reality are reversed for alcohol and nicotine. Nicotine per se isn't actually that harmful or dangerous, it's just the common routes of administration. With alcohol, otoh, it is the substance itself that is inherently harmful, especially if you take it in the amounts required for the main effect which AFAICT is the only reason anyone drinks the stuff. If you're drunk, you're not drinking in moderation. There is no safe, healthy way to get drunk. It really sucks, I know. But people latch on to studies saying it's potentially healthier in moderation (researchers have a weird idea of what that means...I'd think 3-5ish, as long as I'm not hammered, but ONE or TWO 12 oz beers, why not just have a single shot glass of wine cooler? What's the point?), and say, "See, it's healthy! Stop being a prudish scold; everyone does it (as it used to be with smoking), real men drink, always have." And continue with their drinking habits that are almost certainly unhealthy.
Anyway...
I occasionally use nicotine in patch form, which has never been addictive. I rarely continue doing it for more than a week or two because I just don't feel like anymore. But it does help me get work done; it seems at least as good as any of the other stimulant meds for ADHD.
Speaking of modafinil and variants... Recently tried armodafinil. Many apparently say they like it better, but I definitely don't. It's weird... It seems very different to me. Feels less like a stimulant, I feel less sharp, more fuzzy, which is kind of good, but also bad. Seems less effective at helping me get work done... I seem to waste more time on BS like TikTok (or writing pointless rants for strangers to read) when I do it. OTOH, it seems easier to sleep on it. Experimenting with a little bit of both... Seems like they kind of balance each other out a bit.
The answer to why Brazilian jiu-jitsu took so long to invent is simple: for all it likes to pretend, MMA is not like actual combat. If you really want to win a fight, you bring a friend or use a weapon. Both of those give you huge advantages, and it's what everyone did: norms of "help your friend in any fight and ask questions later" and "always carry at least a knife" are extremely widespread across cultures. Extended, one-on-one, completely unarmed combat just didn't happen often at all. No one got good at MMA until people started playing by MMA rules.
First time I spent some time reading Gwern’s stuff. I really like the way this guy thinks. Special thanks for this link.
"Also that AI wiping out humanity is “not inconceivable . . . that’s all I’ll say”
Completely unsurprising. We have two centuries of literature conceiving of AI being bad in one way or another (I count Frankenstein's Monster as AI), and as soon as "wiping out humanity" became a thing, AI was fingered as a possible villain the next day.
2: off the cuff twins theory. What drives people to greatness is a fundamental insecurity and feeling of loneliness that is uncommon in twins.
Regarding Glaze, I saw a paper today that appears to defeat a similar sort of masking by just...JPEG-ing the image: https://arxiv.org/abs/2304.02234
(To be fair, their stated scope is to allow for AI editing of the image, not using it as training data. Still, there's bound to be SOME crossover)
Pic of example: https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F03f18c55-674f-4a1f-a7ca-13b8b43d93e5_1922x1614.png
"Jiankui He is back"
And has anyone checked up on the babies he used as living experiments? Are they okay? Did it all work out? If we're going to decide that cut-n-paste humans are the way forward, at least let us find out if it worked okay or not the first time.
I will definitely be reading the Lutheran paper 😁
It could be that Stalin was the best poet in the Soviet Union because he was the only poet in the Soviet Union who wasn't afraid of Stalin.
“I can’t tell whether they’re claiming Stalin was actually very good - or that he pretended his poems were anonymous, and everyone else praised them lavishly while pretending not to know Stalin wrote them.”
Stephen Kotkin’s biography suggests his talents were recognized even when he was an unknown revolutionary under the Tsar:
“As a teenager, he had abandoned a successful trajectory, with high marks in school, to fight tsarist oppression, and published first-rate poems in a Georgian newspaper, which he recited in front of others. (‘To this day his beautiful, sonorous lyrics echo in my ears,’ one person would recall.)”
> 27. [...] One of the few rare victories for these voting systems was Fargo, North Dakota...
Approval voting has struggled to gain momentum. But ranked-choice voting has actually been having a run of successes in the past few years: the mayoral primaries in NYC, the mayoral election in Burlington, VT, statewide elections in Alaska and Maine...
Oh hi! Thanks for the mention :)
I absolutely agree that I'm doing a "special crackpot that works mostly for me" diet but hoping to find what the magic button is ;) It would be great if we could interpolate from all those one-off crackpot diets that work and find one common ingredient. Maybe not, but maybe.
On 5, British Universities are different to American ones. In particular, 18-year-olds are viewed entirely as adults, and the university is a business they pay to provide them instruction in a specific subject. There’s no expectation that they’ll provide any pastoral care, and you only rent a room from them in your first year (where they’re basically just your landlord). They’re not in any sense responsible for you; if any of them tried to impose an honour code, that would be as weird as McDonald’s imposing one on its customers.
Dating your lecturer (who may or may not be a professor) doesn’t seem any stranger to me than dating your dentist, with the caveat that they shouldn’t mark your exams to stop them being biased in your favour.
(Exceptions to this are that there’s been a slow pressure towards Americanising for the last few years driven by the general infantilisation of society, and Oxbridge works completely differently from everywhere else).
The Gathering Place in Tulsa (funded by oil millionaire turned banking billionaire, but who's counting) is really quite amazing. Both Tulsa and Oklahoma City are doing impressive things in terms of family-oriented public infrastructure, even if they are still essentially giant sprawl exurb metropolitan areas. Purple cities in red states can be pretty nice places to live.
>New weird either-genius-or-crackpot anthropology blog (origin of pronouns, snake cult theory of everything). My heuristic is that as soon as someone uses the phrase “Basque-[Anything]” they’re beyond salvation, but this person is trying very hard!
Would point out it's only the second most radical take about snake mythology. The anthropologist Jeremy Narby was studying a tribe in the amazon, and he participated in their Ayahuasca ceremonies. Like them, snakes appeared to him. Then he notices that creation myths link snakes to knowledge across the world. Egypt, Jewish, Australian, Amazonian, Indian...why are there always snakes? He suggestion is that western science is wrong, and that there are mystical ways of knowing, somehow connected to snakes.
My take is that snake venom is itself a hallucinogen, maybe it was used as part of a self-domestication package when recursion / self-awareness was first emerging.
There really was a bear in Georgia that OD'd on cocaine, but the bear corpse in KY is not that bear, and their story of what happened to the real bear is a complete fabrication. The outfit displaying the supposed cocaine bear finally admitted as much.
https://www.savingcountrymusic.com/confirmed-cocaine-bear-at-ky-for-ky-origin-story-fictionalized/
I'm the anon commenter quoted in that article,, and I'd looked into the corpse-journey story a couple of years ago and found it was clearly BS (some parts didn't make sense, and I tracked down multiple people who could have verified parts of the story, and got "definitely didn't happen", "definitely not the bear", etc.)
Re the Tulsa thing, I am not sure how much he's correct in crediting Tulsa's transformation to Kaiser et al. Obvious disclaimer, I live in the other city in Oklahoma and am not super-familiar with Tulsa. But a lot of what he talks about (at least the non-Jewish parts) I'm seeing here in OKC, too, despite a lack of billionaires funding it (at least as far as I know). Which makes me suspect that structural factors (land is cheap, the government is pretty hands-off) and general demographic shifts are driving a lot of it. If you have people who want Culture, Culture will appear, and both OKC and Tulsa are big enough to have demand.
I'll definitely second the message that Oklahoma is a pretty nice place to be. My mortgage is about the same as my rent was in deep suburban LA for a place a third the size (and I had a really good deal there) and the government mostly leaves you alone. Gas is cheap, and it's safe. The only downside is the lack of battleships and that the local ACX community isn't that strong, but if you move, we can change that.
As a member of the Feline-American Community, the reason we like earwax so much is the proteinaceous taste.
> Why are furries so artistically and economically influential compared to other fetishes?
I have a few theories:
First, unlike many other fetishes, the furry fetish can ONLY be satisfied by art. Into bondage? There's plenty of videos on PornHub. Into feet? The industrious editors of wikiFeet have you covered. But for furries, there is no photography, only art that has to be drawn, painted, or created digitally. Naturally, this results in the creation of a lot of art. A minority of furries are interested in photographs and videos of actual people having actual sex in fursuits, but they're definitely a small percentage (and even then, someone has to buy a fursuit from an artist who created it).
Second, the furry fandom is identity based. It's commonplace to have a "fursona" to represent yourself online (which is where furries spend all their time anyway). This creates a lot of demand for art, since people will want a visual for their furry avatar. In terms of porn, this makes it rather unique in that you can pay to have a representation of yourself in the pornography. I think it's pretty easy to see why that would appeal to people and lead to increased investment in furry fetish artwork relative to other fetishes.
Third, it's an internet-based culture. This leads to certain demographics associated with economic success. There's an overrepresentation of people in tech/IT. It's mostly male. When folks from other countries participate in English-speaking furry communities, you'll only be interacting with the people in those other countries who speak perfect English (a correlate of intelligence and economic success). And for what it's worth, the community is extremely autistic, something like 4x as likely to be autistic as the rest of the population (though I'm not sure what impact this has on their income).
Although...
I'm not actually sure if furries are that economically influential. It certainly seems that way based on the quantity of furry art, but is there actually data to support that idea? Sure, they appear to have a lot of money to burn on art, but furry art isn't always expensive. A lot of furry artists live in third world countries and make a living off doing commissions for people in first-world countries. Even if those commissions are only $10 USD, that can go a long way in a developing economy.
Plus, appearances could lead to a biased sample. I think a good analogy would be related to somebody hating gay people because they think they're all flamboyant and annoying. The issue here is that they don't notice the normal behavior of non-flamboyant gays because they don't realize non-flamboyant gays are gay at all. For furry art, we only see commissions from the wealthiest and most successful furries who can afford it. For every one of them, there might be 10 furries living in poverty.
Seems like a good question for someone like aella to answer via survey data.
Do you think there are good programming jobs at a non-profit in tulsa?
Have you read David Chapman's Better Without AI (https://betterwithout.ai/)? He thinks we don't need AI to accelerate progress (https://betterwithout.ai/radical-progress-without-AI), and even says that intelligence is not a real bottleneck in science (https://betterwithout.ai/intelligence-in-science), such that superintelligence may not result in superscience.
There are other possibilities here than AI salvation/destruction and collapse.
item 21 link doesn't work.
The discussion on DSL is uh...
> The Roman Emperors thought Christians not performing the relevant sacrifices to the Gods were an externality, too.
Yeah, that's some real high-level discussion there. They've definitely got a grip on what externalities are and have good ideas on how to handle them. This definitely isn't just throwing out the laziest possible argument to avoid having to consider the possibility that you're having a negative effect on other people. I wonder if the response is the same to the externalities imposed by homeless people on the sidewalk or drug markets in public places.
Re 5 on professors dating students - I expect that most people are actually much more fine with that than most people affiliated with universities would be. I suspect the American poll results would be pretty much in line with the British ones. You can find plenty of examples of it happening in TV shows (here's a list that includes a lot of *high school* students dating teachers, as well as college students, not too many of which are portrayed very negatively: https://www.refinery29.com/en-us/2017/08/168506/teacher-student-sexual-affairs-on-tv-shows - and it doesn't include cases like Jeff Winger dating his math professor in the first season of Community).
Meanwhile, within universities you have a lot of sentiment that is too far in the opposite direction. The Texas A&M University System (i.e., the thing controlled by the legislature that has authority over the College Station campus, as well as the others, so not the "woke" part, more the Republican part) a few years ago dropped a rule banning *all employees* from having "romantic, sexual, or amatory relationships" with undergraduate students, with only 24 hours notice to the campuses before the rule went into effect. Whoever wrote the rule realized they needed an exception for undergraduate student employees, but didn't understand that basically all graduate students count as "employees", and that the university is large enough that incoming grad students who are single are going to accidentally hook up with people from other parts of the university if they try to meet anyone in a forum other than carefully controlled online dating.
Re: the African birth rate decline: Scott’s gloss of it makes it sound like it’s consensus that Africa was terrifying everyone and pushing us towards overpop catastrophe.
This is the OPPOSITE narrative from what I’ve been hearing for the past several years in my circles: that overpopulation is mostly a eugenicist myth, that huge numbers of countries of developing countries are headed for UNDERpopulation crises, etc. So what gives?
Regarding #37, article is paywalled but here is the nut of it:
"....few have noticed a wealth of new data that suggest that Africa’s birth rate is falling far more quickly than expected. Though plenty of growth is still baked in, this could have a huge impact on Africa’s total population by 2100. It could also provide a big boost to the continent’s economic development. “We have been underestimating what is happening in terms of fertility change in Africa,” says Jose Rimon II of Johns Hopkins University. “Africa will probably undergo the same kind of rapid changes as east Asia did.”
The UN’s population projections are widely seen as the most authoritative. Its latest report, published last year, contained considerably lower estimates for sub-Saharan Africa than those of a decade ago....
Yet even the UN’s latest projections may not be keeping pace with the rapid decline in fertility rates (the average number of children that women are expected to have) that some striking recent studies show. Most remarkable is Nigeria, where a UN-backed survey in 2021 found the fertility rate had fallen to 4.6 from 5.8 just five years earlier. This figure seems to be broadly confirmed by another survey, this time backed by USAID, America’s aid agency, which found a fertility rate of 4.8 in 2021, down from 6.1 in 2010. “Something is happening,” muses Argentina Matavel of the UN Population Fund.
If these findings are correct they would suggest that birth rates are falling at a similar pace to those in some parts of Asia, when that region saw its own population growth rates slow sharply in a process often known as a demographic transition.
A similar trend seems to be emerging in parts of the Sahel, which still has some of Africa’s highest fertility rates, and coastal west Africa. In Mali, for instance, the fertility rate fell from 6.3 to a still high 5.7 in six years. Senegal’s, at 3.9 in 2021, equates to one fewer baby per woman than little over a decade ago. So too in the Gambia, where the rate plunged from 5.6 in 2013 to 4.4 in 2020, and Ghana, where it fell from 4.2 to 3.8 in just three years....
Demographers are divided over how much to read into these recent surveys, particularly since the data they produce can be noisy. “When you see a precipitous decline in fertility, your starting-point is that something is wrong with the data,” says Tom Moultrie of the University of Cape Town....Yet even comparing only within iterations of the same survey (as The Economist has done with the figures above), the trend is evident. Comparing across them in the case of Niger, which has the world’s highest fertility rate but few surveys, shows a decline from 7.6 in 2012 to 6.2 in 2021.
Others are also reducing their projections. In 1972 the Club of Rome, a think-tank, published an influential book, “The Limits to Growth”, warning that consumption and population growth would lead to economic collapse. Now it says the population bomb may never go off: it reckons sub-Saharan Africa’s population may peak as soon as 2060, which is 40 years earlier than the UN’s projections.
Even so, fertility rates are not dipping uniformly. Some countries, including Angola, Cameroon and Congo, are seemingly stuck at relatively high rates. And there are often big regional differences within countries such as Kenya. Almost everywhere in Africa, fertility rates are much lower for urban women, who typically have 30-40% fewer children than those in the countryside....
Family planning, especially when promoted by outsiders, has often caught the ire of religious leaders. Yet in some places that may be changing. Clerics talk more often about family planning these days, notes Amina Mohammed, a devout mother on the outskirts of Kano. “There is no verse in the Holy Koran where Muslims are forbidden from controlling, planning or restricting the number of children they have,” says Shuaib Mukhtar Shuaib, one such cleric. The Prophet Muhammad tacitly approved of the withdrawal method, he continues. These days Idris Sulaiman Abubakar, a gynaecologist in Kano’s biggest public hospital, is more worried about the impact of Nigeria’s film industry on contraception than that of religion. “They’ll bring a story-line that the woman’s reproductive system was damaged because she uses pills,” he explains.
Girls’ education also makes a big difference to fertility rates. In Angola, for instance, women without any schooling have 7.8 children, whereas those with tertiary education have 2.3. Educated women have a better chance of a job, so the opportunity cost of staying at home to look after children is higher and they are more likely to win arguments with their husbands over how many kids to have.
Research by Endale Kebede, Anne Goujon and Wolfgang Lutz of the Wittgenstein Centre for Demography and Global Human Capital suggests that a stall in Africa’s demographic transition in the 2000s may have stemmed from the delayed effect of cuts in spending on education in the 1980s, when many African economies were in crisis. The rapid falls in fertility rates that now seem to be taking place could be because of the huge push to improve girls’ schooling in the past few decades....
3:
> The plan, along with the Taiwan Strait Tunnel Project, has been widely mocked in Taiwan.
On the topic of right-wing policies becoming less popular once people are elected - "Reflections on the Guillotine" comes to mind, I think a mind-boggling percentage of people do not have the faintest understanding of what it means for someone to undergo any number of terrible things, from literally execution to homelessness to hunger. A policy that so many people agree with in theory that involves holding someone "accountable" or people "pulling themselves up by their bootstraps" or "giving criminals what they deserve" ends up horrifying a lot of people who were previously sheltered by their policies being mere ideas. It's easier to love "justice" when you're not watching someone's blood pouring out of their neck, just like it's easier to love "austerity" when local government functions so well you hardly know it exists.
Lost it at "gibber".
Man that North Dakota one is frustrating. It's just maddening to have local political change rugpulled on you. Like when the Supreme Court ruled states weren't allowed to put term limits on their federal representatives.
Tulsa has a long* history of oil billionaires using their wealth to indulge their particular whims and create cool stuff (see: anything beginning with "Phil," The History of Science collection (now in Norman), etc.)
*relative to the existence of Tulsa.
Matty: "Crackdowns don't work."
Also Matty: "Tee Hee, look at me snitch this BMW owner to the police!"
"In the 80s and 90s, Tulsa experienced a decline. It got gross and dangerous. "
As someone who lived there (and was a stage electrician), this is completely wrong. You could run around downtown after midnight. It had a symphony orchestra, an amazingly good youth orchestra, a ballet company, and at least three theater companies.
If you want to know what percent of Americans think college is worth the cost, you should just look at what percent of Americans go to college. Any delta between this and your poll results is cheap talk.
What most interested me about #6 was actually the magnitude of income's effect on happiness, rather than the non-linearities in the most + least happy groups:
> KD reported that the effect of an approximately fourfold difference in income is about equal to the effect of being a caregiver, twice as large as the effect of being married, about equal to the effect of a weekend, and less than a third as large as the effect of a headache.
Interesting to put this into perspective - I wish this had been emphasized! It was buried in a single paragraph in the middle of the paper.
#20: "In fact, I’ve updated so far that I’m starting to worry that the problem won’t be building a political coalition against unsafe AI, the problem will be not overshooting and banning all AI forever. " Wow, that seems really implausible to me. Here's why: (1) Government has done very little so far to regulate the big tech companies, even when they are doing things that are harmful in a ways that are much more immediate and easy to understand than Foom -- for example, providing powerful vehicles for spreading misinformation. Why would AI change that? (2) There are enormous amounts of money to be made from AI. That will motivate the makers, and also the many many industries that can benefit from it, to support its continued development. (3) The main danger is hard to understand. I am smart, well-educated and have spent a lot of time reading about and discussing AI risks, and I still feel unable to judge how great a risk it poses to humanity. (4) My impression is that most of the public's not a bit concerned. Even most of my professional friends sort of chuckle when the subject comes up, as though it's an urban myth like that killer bees are about to invade.
I can think of 2 things that might nudge the US in the directing of reining in AI development: One would be some well-publicized bad events resulting from widespread use of GPT4: Cybercrime greatly increases; or a bunch of business go under because they give AI too much responsibility and it makes a big mistake; or some AI use in the financial industry causes alarming events in the stock market. The other is the reaction of groups of professionals whose jobs are shinking because AI can do so much of them. Recently saw a study of accuracy of AI and radiologists at making a yes/no judgment from a medical imaging result relating to a certain pulmonary problem. It was a problem where after a few days you knew for sure whether correct answer was yes or no. Radiologists alone were about 80% accurate, AI alone was about 84%. If the images AI was unsure of were passed to the radiologist, combined accuracy of AI plus radiologist was about 86%. Seems likely to me that training AI a bit more on the difficult-to-judge cases will result in AI alone being the most accurate way to handle ALL the images. Of course there will still be things for the radiologist to do -- but a LOT fewer of them. Seems like GPT4 & similar, unlike other advances, is going to be taking away the jobs of some upper middle class professionals.
But still I doubt that even the 2 things above that might make the public more afraid of AI and more resentful of it are enough to overcome the very powerful forces that nudge events in the direction of continued AI development.
Apparently the Fargo/approval voting thing has resolved positively! https://www.kvrr.com/2023/04/19/fargos-approval-voting-system-lives-on-after-state-senate-fails-to-override-veto/
(8) has been a fun story to watch unfold from the inside - for a lot of physicists in condensed matter room temperature superconductivity is the ultimate holy grail and absolutely nothing gets them wound up more than a good old fashioned USO (Unidentified Superconducting Object). These happen occasionally but the fact that this one actually got published in Nature makes it totally unique - they are going to get a huge black eye at the end of this debacle.
This has basically devolved into the physics version of a slapfight between Dias and Hirsch, conducted in the traditional way via weekly combative back-and-forth rants very thinly disguised as arxiv papers (see: https://arxiv.org/abs/2304.00190).
https://transformer-circuits.pub/2022/in-context-learning-and-induction-heads/index.html
Referencing an earlier comment from the hidden thread but I knew someone had to be doing this! Don’t understand how this doesn’t solve almost the whole thing.
Why are there so few pairs of extremely successful identical twins?
Because they are very rare. One every 300 or so.
#6 matches something that I've been thinking for a while. Money doesn't buy happiness if you're a baseline miserable person, but it does if you're a baseline-happy person.
As a baseline-happy person, I tend to be content unless I've got some particular reason to be miserable. As I've increased my wealth, I've managed to use money to remove many of those reasons to be miserable. I used to be too hot in the summer, now I have air conditioning. I used to sleep less comfortably on a cheap mattress, now I sleep very comfortably on a fancy mattress. I used to be miserable doing housework (or miserable living in a filthy house), now I pay someone to do it for me. There are still some other miseries that I'm not rich enough to solve, like flying economy class or needing to go to work, but with more money I could solve those too and be even happier.
Scott should read stratechery. Yglesias' article on the media is a pale imitation of work Ben Thompson has been doing much better for years.
Regarding right wing populists: It's because they usually advocate some degree of "law and order." Everyone wants to have law and order, no one wants to do it. High effort, high resistance.
It's like people voting to clean their room, and then support dropping when it's time to actually clean their room.
(28)
>keep it a good place to live while the rest of the Midwest hollows out
I beg your pardon? First of all, Oklahoma isn't centrally Midwestern. Neither is rust-belt Ohio or barren Wyoming. The states in between have plenty of nice healthy towns and cities (sure the rural population is declining, but that happens everywhere) that are great places to live if you can handle a little snow. We like our rich farmlands and tight-knit communities; California can't have them.
(Yes, I'm still salty about the Wall Drug bit in Unsong literalizing "flyover states"; why do you ask?)
#2 : I have a hypothesis that numbered lists is just a good propaganda/educational tool from a purely memetic point of view. I don't have evidence, it's just a plausible just-so story that came to me when I thought about the question deeply.
As a starting point, I note that modern clickbait journalism and buzzfeed-style content mills have always loved the numbered list, "7 REASONS TO <insert random thing>, NUMBER <insert number between 1 and 7> WILL SHOCK YOU". Communism was also famous for its 5 year plans and 4 things that need to abolished and things like that. Judaism has the Ten Commandments, Islam has (off the top of my head, and translated non-professionally from Arabic) the Five Pillars of Islam, the Ten Companions of The Prophet Foretold To Enter Heaven, the Seven People Who Will Enjoy Allah's Shade On Judgment Day, etc.... Martin Luther upended Europe and Christianity with a 95-point list.
What are the common factors of clickbait journalism, Judaism, Islam, Communism, and East Asian philosophy ? I can find nothing except that they were all ideologies or families of ideologies that needed to indoctrinate, to preach, to capture attention and hold it. Indoctrination and propaganda are not necessarily bad things, some forms of education are and should be indoctrination, and lots of good Art is propaganda. Indoctrination and propaganda are just general purpose tools to capture attention and convince people, they are understandably favorite tools for ideologies like Islam, but they are neutral tools nonetheless.
Anyway, can we find reasons why numbered lists are a good indoctrination and propaganda tool ? Here are the 7 reasons why numbered lists are a good linguistic tool :
(1) The number serves as a simple checksum against data corruption or forgetfulness. You might forget how many The Pillars of Islam were or mis-remember the number to be 3 or 4 or 7, but you can never (or very rarely) make that mistake with The Five Pillars of Islam.
(2) The number serves as an attention grabber that builds tension and invites a listener or a reader to continue paying attention to get to the final thing. A classic template for a joke is "3 Xs are Y", where X is any general category and Y is any activity or action. Stories has evolved numbered chapters, television shows has numbered episodes nested inside numbered seasons, and movie series has numbers on every movie. Media players has a timer countdown to tell you how many hours and minutes and seconds remain. Etc..., Numbers are a "Progress Bar" service that tells you how much of the thing remains, and this is valuable from an attention regulation point of view. It's a simpe musical rhythm that starts at 1 and gradually builds up to climax.
(3) Lists are a good info organization structure in general, striking a sweet spot between free-form text on the one hand and elaborate diagrams or other overly-structured complex data structures on the other. Lists are how lots of writers outline essays, speeches, stories, screenplays, etc...[1]. When Dan Bricklin, the inventor of spreadsheet software, was thinking about a way to reference variables, he striked upon the grid system (i.e. a numbered, lettered, 2D list) as a way of organizing and storing the values in a spreadsheet[2]. Arguments in Logic and Analytical philosophy are written as numbered lists of premises and conclusions. High-level overviews of algorithms and code in Computer Science and Software Engineering are often written as ordered lists of steps. Laws and Constitutions are written as lists. The oldest surviving writings are lists of financial transactions. TODO lists.
(4) Numbered lists invite sorting along some natural or important axis. For example, the Pillars of Islam is sorted according to difficulty. First you simply have to profess that there is no God but Allah, just say a single phrase with your mouth. Then you have to pray 5 times a day, slightly harder but still within the realm of easy. Next you have to donate 2.5% of your money yearly, now we're talking money, but still a relatively small amount. Next you have to fast in Ramadan, spending anywhere betwen 12 and 16 hours without food or water or sex, for 30 days. It's hard, especially in the summer, especially in the near-equator regions that Islam is the main religion in. Finally, you have to go to mecca to perform the Haij, an ancient Arabic religious practice that Islam kept. The Haij is so difficult that it's optional, you can be a Muslim whose health or finances don't allow you to perform the Haij.
Other lists may sort according to importance or historical recency or any other topic-specific notion, numbered lists allow and suggest that to you by their very nature.
(5) The number can invite comparisons, metaphors, imagery and analogies, whether justified or not. Humans have built a lot of culture and mythology around numbers, and plenty of important things in our biology and in nature in general has distinct numbers. Two eyes, Two hands, Five fingers per Hand, Ten fingers in total per pair of limbs. 7 as a holy number, 13 as a lucky (or unlucky) number. Addition to mean conjunction or unity (e.g. "A man and a woman, 2 people unite in love to be 1"). Etc... If your numbered list can invoke a number with already a rich cultural imagery or history, it can be more effective and memorable. One such cultural imagery is the concept of the numbered list itself : you can piggyback off a more popular and established numbered list to get your numbered list to the audience quicker and more effectively. For example, if I name my list The Three Laws of X, Newton's and Asimov's (both Issacs, coincidence ?) immediately come to mind, boosting the attention my audience give me, even if its mocking or hostile (e.g. "Who this guy thinks himself to be ?").
(6) The list serves as a FAQ that followers of the ideology can always refer new comers or hostile opponents to in order to know the ideology better or to correct misconceptions or reiterate the ideology's core values.
(7) Lists are flexible, each point has a variable scope so you can always shorten or lengthen the list as you please by selectively splitting or fine-graining points and conslidating them. This point itself is such an example, I thought of it purely to get the list's length to 7.
tldr; numbered lists are a natural tool for storytelling and oration, they have plenty of educational and memetic advantages and disparate ideologies and propagandists have gravitated to them independently.
[1] https://www.lesswrong.com/posts/ZtMsyMP5F7zzP8Gvc/reader-generated-essays
[2] http://www.bricklin.com/history/saiidea.htm
re 35: Eddie Pepitone at 29:35
https://www.youtube.com/watch?v=ik-skhmSFLI
"I was listening to Stalin on, uh, Spotify, on the way over here...
And I'm talking about I listen to early Stalin. Not the later shit when he sold out. Early Stalin, when he was paranoid and he had the military behind him. Not when he started talking about wheat!"
29. The paper starts the clock on counting scientists in 1450. But what about before then? Before Francis Bacon was Roger Bacon. During the Renaissance there were lots of scientists in Italy, and then they went away. But were there many before the Renaissance? I don't know. Maybe it was just a return to normal.
Re: 5 I think what's going on is that the people most likely to talk about professor student dating are likely doing so either because:
1) They are in academia themselves in which case they have an incentive to signal that they totally aren't creepily into dating their students (and a lesser but still powerful signally concern for female academics).
2) Because it's an issue of concern for them. Few people who are just 'shrug it's not a big deal' are going to be very animated about it.
Overall this seems like an issue which suffers from a strong bias to overreport only one direction of concern. And that's bad because I think we are making things worse by imposing such strict rules for professor-student and, especially, TA-student dating.
Let's be honest, there is a great deal of temptation for this to happen. There are women who find that relationship sexy and men who are tempted to date younger women. So let's play out how this happens under a regime that says we fire professors who date their students and one that just requires reporting or class switches or mild punishment.
The student who dates under rule that says it's fine just give notice won't feel pressure not to ask to be transferred to a different class, advisor or even just to seek advice from someone else. OTOH, a relationship under the firing rule will necessarily start with the professor explaining that it's critical it be kept secret, that they'll lose their job if anyone ever finds out. That's going to make the student feel very reluctant to reach out (especially because they probably still have some feelings for the prof and just want get out of the situation not get them fired) for any help and absent that will have no way to extract themselves from the situation. And abusive relationships tend to make the victim doubt themselves. They, no matter how wrongly, may feel guilty for inducing the professor to put their career on the line.
Yes, you'll deter some professor student relationships from happening at all. However, that deterrence will be concentrated primarily among those who are best adjusted/most rule following and thus least likely to use psychological pressure, threats etc.. to manipulate a student. So the end result is you've probably denied some people a fun time and created a really awful trap for the most vulnerable students for the worst abusers of power to use against them.
The AI Risk phenomenon is a great example of why humanity will probably never implement most of the transformative technologies people have been theorizing about for centuries, if not millennia (i.e., nuclear energy - both fission and fusion, genetic engineering, gene therapy for common diseases, stem-cell therapy, space travel, etc.). Instead of working finding ways to properly integrate a new, transformative technology in ways that benefits our society, we consistently choose to regulate it either to death or near-death. Think about the number of problems that exist today because we have such low risk-tolerance and fail to organize enough cognitive capital to create ways to integrate these technologies in a sustainable way and make sure errors do not lead to a public outcry and subsequent bans.
As our world develops its communication technologies, any negative effect of a potential new technology will immediately be known to most of the world and will most likely be regulated to death, near-death, complete-uselessness, or near-uselessness.
Part of the problem is new tech is always on some pathway to worldwide use with time. Almost everyone will be using it given a reasonable amount of time. This propagates problems that otherwise would have been contained and effects which would be minimal. Part of the way to mitigate risk is to make sure the tech is only used in a constrained space where leakage is very low.
On BJJ and the related question of traditional martial arts vs. MMA, the explanation is simply that traditional martial arts have become stylized to fuck as it's become less acceptable *and* less needful to straight up murder a guy without using firearms. You can find video clips of traditional martial artists smashing someone's face in; they'll just look remarkably like e.g. a boxer or MMA fighter doing it. A fifty-year-old guy doing some feckless wavy-hands Tai Chi has just been wasting his time on ineffective martial arts because there was no pressure to be effective and he was probably too genteel to want to pay the cost of being effective in the first place.
23. Didn't Trumpism prevail on the anti-international trade issue? Biden signed that huge bipartisan bill with a lot of "America First" items on it. I doubt we would be here on this issue were it not for Trump making it hugely popular.
You could say that world events, not Trump, helped prod this attitude along. But I think it's hard to make the case that the USA is penalizing trade with England, Germany and France because of Russia, Covid or China.
In regards to El Salvador and "tough on crime policies" in general, I think a point that gets overlooked is that a lot of "tough on crime" rhetoric combines two policies that are often treated as a package deal, but in principle can be separated. These are:
1. Increasing the likelihood that a criminal will actually be arrested and punished.
2. Increasing the severity of their punishment.
"Tough on crime" policies like the one in El Salvador often do both at once. However, a lot of psychological research indicates that 1 is the policy that does most of the work. 2 isn't actually that effective. Criminals care much more about the odds of getting caught than what exactly happens if they are.
It seems to me that this might allow some kind of effective compromise between tough on crime activists and human rights activists. Give the human rights activists lighter sentences, nicer prisons, and more serious efforts at rehabilitation. Give tough on crime activists can get more cops on the street and more arrests.
"Daniel Kahneman and Matthew Killingsworth"
The thesis-antithesis-synthesis process really does work a fair amount of the time.
I do it on Twitter a lot: I say X, somebody responds with "What about Y?", I'm stumped for awhile because Y is true, and then figure out Z would account for both X and Y being true some of the time.
30. Your answer is government-sponsored media with a compulsory tax payment that everyone pays (like the BBC).
I think the "do less AI" movement is pretty doomed. To me, it feels analogous to degrowth arguments related to climate change. The incentives are just overwhelming in the other direction, and trying to regulate against those incentives is just going to cause unintended consequences. You'll just end up with shadier AI researchers making more progress relative to those that care about safety, at the margin.
Again analogizing to climate change and how solar, batteries, nuclear and so on are apparently the best solutions to the problem, I think more positive technological progress in both AI alignment, in tracking safety risks like bioweapons, and in progress on rapidly rolling out vaccines and whatnot should be what we emphasize. I'm glad that more of these concrete proposals are coming out along with the ultimately not very useful ones like the 6 month pause and the EY Time article.
I also think we need a ground-up rework of the concept of AI alignment. I'm not even necessarily aligned with myself from 8 hours ago, so how are we going to align a nebulous concept of AGI with a nebulous concept of humanity? In my opinion, a 21st century version of Asimov's laws seems more promising: for instance, do no more than acceptable harm X (e.g., hurting someone's feelings by disagreeing with them) as a fundamental AI rule.
On "21 - College no longer worth the cost" - it's just... If you keep rising the cost and the real margin of benefit against something that you can learn yourself lessen in all possible metrics - what exactly were the expectations here? Real "surprised Pikachu face" moment.
Pretty sure if the question was about "is going to college beneficial" - the results would've been different. But is it beneficial ENOUGH to also go into the personal bankruptcy for years? No, it isn't.
Okay, it's a bobcat not a Lynx but it's an interesting 2 minutes of safely removing a wild cat from behind a car grill.
https://bringmethenews.com/minnesota-lifestyle/watch-bobcat-pulled-out-from-car-in-wisconsin-by-officers
Regarding the linguistics article and the mysterious prevalence of 1s pronouns based on the phoneme /n/, I don't have time or energy to properly read the full article or to give a very carefully thought-out response. But throughout a couple of decades of being an amateur comparitive linguistics enthusiast I've always been extremely skeptical of hypotheses proposing language proto-families, and, while I appreciate how well the article thinks through multiple sides of each question it treats, quite a few things in it subtly triggered my crackpottery radar even apart from its speculations about the dawn of self-awareness and snake mythology and so on.
For one thing, while I don't have time to actually check through the hypothetical roots for 1s pronouns among the ancestors dozens of language families, from the few I do know about I'm immediately suspicious that the author (and linguists he cites) might be cherry-picking certain roots and certain debated forms of roots. For instance, Austronesian (okay, mainly the Malayo-Polynesian branch) is one of the few language families where I am familiar with a few roots and knew that the 1s pronouns come from a pretty clear *aku root, no nasal consonants involved. I just checked this on Wiktionary, and not only does it back me up on this, I don't think it lists any 1s pronouns in the living daughter languages with /n/ in them. So why does this article list *’a(ŋ)kƏn as the root and count it among the evident roots based on /n/? It looks like this alternate proposed root has only possibly a nasal before the /k/ and that it ends in /n/, but (and this may be just an amateur assumption) the fact that a two-syllable root ends nasally seems like very weak evidence of it being related to another proto-language's root which begins with or is based mainly on /n/. It also goes to illustrate one of the main reasons (I'm sure) why linguists typically don't like to give much credence to proto-family hypotheses: the exact roots of the proto-language are already very hypothetical and up to much debate (in this case, apparently someone thinks that the 1s pronoun root in proto-Austronesian may have been *’a(ŋ)kƏn rather than *aku; that seems to reflect at best some disagreement among similarly qualified experts!).
Relatedly, a language family may have more than one root for a pronoun for a particular person/number combination. Indo-European is a good example of this: one of the roots is *me, and this is the one that the author lists under the proto-pronouns that don't involve /n/ (but then uses this list to point out that a bunch involve /m/ and that /m/ is not far from /n/). But the other root, from which Latin and Greek ego, Germanic ich/ik/ek/jeg/etc. and English I are derived, is something like *éǵh; this is highly contested, and in fact some common form (like *éǵhóm) evolving into both *éǵh and *me seems to also have been proposed, but I think we can agree that at first blush, *éǵh appears to have no similarity to *me. (And the author even points out how our pronoun I doesn't reflect a root involving /n/, now choosing to ignore our object form me which does have a nasal that he's pointed out is not so far from /n/!) So, how many of the 1s pronoun roots listed may have existed alongside other unrelated-looking 1s pronoun roots? Once you acknowledge that a proto-language may have distinct, unrelated-looking roots for a particular person/number pronoun (and why shouldn't they? many modern languages do!), the waters have already been muddied quite a lot.
I also don't know what to make of all the discussion in the article about time scales and how slowly over centuries and millennia pronoun roots may have evolved compared to other types of roots, and what this might mean in terms of evidence of anything. Are the reconstructed roots listed even proposed to have been in use during the same millennia? This may be a case of my not having read the article carefully enough.
I notice that the author mentioned in one place the near-universality of certain consonants in roots for some kinship terms most likely arising from the sounds babies make first and mentioned in another place how pronouns may arise from kinship terms. (In fact, this is a huge phenomenon visible in the present: in many East Asian languages there is a blurring between pronouns and kinship terms.) So why not propose as a possible explanation for the prevalence of /n/ (and other nasals) in 1s pronoun roots that in prehistoric times they may have often arisen from parental kinship terms which in turn almost certainly were influenced by consonants prevalent in baby sounds? Think of the obvious origins of "mama" and then think of mothers talking to their babies who refer to themselves as "Mama" in place of using a 1s pronoun. It doesn't seem like an unreasonable hypothesis to me.
I hope this parent comment might inspire some people who are more knowledgeable and have more time for research (or more research behind them) to chime in and do a little picking apart of the article.
re Newspapers. US has just two parties, and is huge. so you can earn easily from polarised columns. any country with a multi party system will have more camps, so harder to capitalize.
the UK has badly two parties. papers are partly having a political identity. I think lots of entertainment is involved. do has durability maybe with the internet. in the UK the political commentary is part comedy IMHO
Re: African birth rates - that article annoys me. “Misguided western worries” okay how are they misguided? Yes Europe will have a shortfall of workers but not all migrants contribute equally. Previous articles by The Economist show this (see: https://www.google.com/imgres?imgurl=https%3A%2F%2Fwww.economist.com%2Fsites%2Fdefault%2Ffiles%2Fimages%2Fprint-edition%2F20211218_EUC232.png&tbnid=YyMuEsY00KBapM&vet=1&imgrefurl=https%3A%2F%2Fwww.economist.com%2Feurope%2F2021%2F12%2F18%2Fwhy-have-danes-turned-against-immigration&docid=zJinMKY6g2FP8M&w=608&h=662&hl=en-US&source=sh%2Fx%2Fim)
"That’s it, I’m cancelling my planned vacation in North Dakota and will not be buying any of their products." #MeToo
On meditation-
I've meditated off and on for a bunch of years. I do think it makes you more calm, makes attention more controllable, etc, etc-that's all fine. My problem with meditation is something in my head I call the doormat problem: the more I meditated, the calmer I became, and there were a bunch of things I didn't solve which I would have otherwise solved if I had the adequate emotional impetus pushing me along. For instance, for anger/sadness, I'd just sort of meditate it away, which is not great.
In general, it felt like I was turning down the volume control knob on my emotions as a whole, which is not very pleasing to me, even if it gives me reduced despair/sadness/anxiety. YYMV, etc.
> 24: 200 Concrete Problems In AI Interpretability. “You can note which you're working on, and reach out to other people doing the same.”
Thanks for signal boosting my sequence! This is a link to the spreadsheet to track which problems you're working on, you can also see the actual sequence (with the problems, resources + tips, and a bunch more context) here https://www.alignmentforum.org/s/yivyHaCAmMJ3CqSyj/p/LbrPTJ4fmABEdEnLf
In answer to the question of whether it should feel okay for some position of privilege to be offered to "Jews only", I think the question you need to ask yourself is whether you'd be okay with someone else offering a similar set of benefits but advertising it as "No Jews".
If you're okay with this then I admire your consistency. If you're not then you should reconsider being okay with the first one.
I'm always impressed with the large set of interesting links you are able to come up with.
29. Anything regarding Christian history is massively impacted by the belief status of the writer, particularly if they are Protestant, Catholic, or lately, secular. The writer’s institution is very relevant. The reason why you’ve heard so much from pro-Catholic side lately is because the internet has made it a lot easier to publicize and/or come across those opinions. English language liberal arts academia was an exclusively Protestant affair for a very long time.