Links 7/17: URL of Sandwich
Did you know: medieval Christians who didn’t understand Islam imagined Muslims as worshiping a god named Termagant; through a weird chain of events this became the modern word for an argumentative woman.
The fantastic Hungarian architecture of Imre Makovecz. Also: the Cologne Central Mosque and Michiel Schrijver’s cityscapes.
I was previously pretty convinced that lithium in drinking water was having a significant (and positive) effect on populations, but the most recent study is skeptical.
Ancient people believed the kidney was involved in conscience and deliberation, and according to the Talmud “one of the two kidneys counsels what is good, and the other evil”. What would they think of kidney donors? (h/t Elissa)
You know what nobody hates each other over yet? Quilting.
Study on economic vs. social politics finds that economically-conservative-socially-liberal people (libertarians?) are rarest, economically-liberal-socially-conservative people (populists?) are much more common than expected.
Another highly positive study on the connection between lead and crime, this one almost a true experiment. Children placed in a lead-reduction program, compared to children just over the cutoff for qualifying for the program, saw their risk of violent crime as adults drop by 66%! The reduction of lead in the experimental group of this study was about the same as the society-wide reduction over the past twenty years.
Vice presents a counter-narrative about the opioid crisis: pain patients prescribed opiates rarely get addicted, most addicts happen when the pills get diverted away from real patients. Haven’t really evaluated this to see how true it is but I agree with them that some of the statistics going around about how every single person prescribed a painkiller is at high risk of addiction are a little overblown.
Among the latest attempts to cut federal bureaucracy: ordering agencies to stop providing updates on their preparations for the Y2K bug.
That time Pepsi bought 17 submarines, a cruiser, and a destroyer from the Soviet Union as part of the Cola Wars.
Some context for Jon Ossoff’s loss in the recent Georgia special election: was the last Democratic candidate for that seat even a real person?
Daniel Lakens: Impossibly Hungry Judges. That famous study showing judges are more likely to convict just before lunch has such a high effect size that it can’t possibly make sense. Also a link to a more thorough critique of the study suggesting that courts schedule defendants without representation just before lunch, providing a more likely explanation than judges’ hunger.
Elizabeth Warren as synthesis of the Hillary/Bernie dialectic. I think she’s probably the Democrat closest to my own views right now.
Someone responded to my post using “murderism” as a reductio ad absurdum by pointing out the controversial police training classes by Dave Grossman, “the world’s sole authority on killology”.
Jonathan Kay discusses mob culture and attacks on free speech, but focuses on something important that isn’t mentioned enough. Yes, the PC-left are doing most of the attacking, but the PC-left is also most of the victims. Non-leftists can occasionally get in trouble if they’re Charles-Murray-level good targets, but generally escape unscathed (Murray’s conservative think tank unsurprisingly continues to support him). Leftists live in constant fear because they’re in social circles where this happens all the time and where all their friends will automatically side with the accusers. This isn’t just mean, it’s really bad strategy if you want people to stay on the left. I wonder if part of the success of the Bernie Sanders/socialist left is about it being a leftist space which is safe(r) from this kind of thing.
How much of effective altruism is about doing things directly, versus acting as a living advertisement to attract the attention of rich people with a thousand times more money available than everyone else? I think this is an important question insofar as it challenges the philosophy that doing good is always more important than looking good. Some form of weirdness which raises effectiveness 10% but turns off one otherwise-recruitable billionaire ends up being pretty costly.
Center For A Stateless Society has probably the best response to my cost disease post I’ve seen so far, which suggests the problem is something like oligopolies, plus weird accounting rules that treat “costs” and “revenues” in confusing and inappropriate ways.
Political Regime Type And Warfare: Evidence From 600 Years Of European History. Between 1200 and 1800, parliamentary regimes were more likely to get involved in wars than absolutist ones.
What Democrats mean when they say that AHCA is being “rammed through” Congress (compared to Obamacare).
Milton Friedman on how to change the world; relevant for almost everybody.
Vox on the sordid history of the COEXIST bumper sticker. Spoiler: the various people with financial stakes in the design aren’t very good at coexisting.
Washington Post: No One Is Paying Attention To The Worst Humanitarian Crisis Since World War II: “the danger [is] that about 20?million people in four countries will suffer famine in the coming months, and that hundreds of thousands of children will starve to death.”
Lots of discussion about the recent study finding that Seattle’s minimum wage increase backfired and hurt poor workers. The argument in favor of the study, as presented by the Foundation for Economic Education; the argument against, as presented by the Economic Policies Institute. But also, see the Seattle Weekly on how the city tried to cover up/muddy the waters on the incriminating data, the St. Louis Post-Dispatch on how St. Louis’ minimum wage is decreasing, Marginal Revolution on potentially relevant evidence from Denmark, Megan McArdle and Noah Smith‘s analyses, and Zvi (1, 2) on some ways the Seattle data don’t really add up. Luckily, there are enough other cities making large minimum wage increases (and Seattle plans to increase it further) that we should have much more evidence on this pretty soon.
Related: Maine Tried To Raise Its Minimum Wage; Restaurant Workers Didn’t Want It. “Some cried with relief, Buckwalter said, when the final vote ended at 110 to 37 — overwhelmingly [in favor of lowering their wages]”.
Popehat: “There are many very stupid ideas about free speech in academia. Perhaps the stupidest is this: free speech is a legal norm used to protect the powerful at the expense of the powerless, but exceptions to free speech will benefit the powerless. Nobody with a passing knowledge of the history of free speech takes this seriously.” Related (albeit old): Why I Think XKCD Is Wrong About Free Speech.
Related: Data On Campus Free Speech Cases. “Of the 77 cases, I coded 20 (26%) as censorship of liberals, 40 (52%) as censorship of conservatives, and 17 (22%) as apolitical censorship.”
Something I didn’t expect to see a serious argument for today: “The entire edifice of Western civilization – all the cultural, social, and philosophical structures that define the world in which we live today – can be traced back to a stupid loophole in Roman inheritance law.”
2,100 Australian public servants participate in a gigantic resume experiment to assess unconscious bias against women and minorities; finds that there is in fact bias in favor of women and minorities, and that gender-blind or race-blind assessments cause more whites and men to be hired. Concludes that this indicates “need for caution when moving towards blind recruitment processes”.
Everything about economics in India sounds like a mess, but there’s been at least one small step forward with the passage of a national sales tax. “The official schedule of rates runs to 213 pages and has undergone repeated changes, some taking place as late as on Friday evening…Adding to the complexity, businesses with pan-India operations face filing over 1,000 digital returns a year.”
Example-Based Synthesis Of Stylized Facial Animations, the movie – watch an AI convert a video to different artistic styles on the fly.
The US government can borrow money at about 1% per year. The stock market earns about 4% per year. “I expect the government should own a bunch of stuff.”
Roman concrete does outlast modern concrete, but it’s not a simple story about ancient wisdom so much as different solutions for different problems.
A very very thorough study not only finds no effect of birth order, but demonstrates some of the ways other studies that did claim to find an effect could have gone wrong. The only exception is a small effect on “intellect”, defined as whether people self-report as being “eager for knowledge”. Possibly related to this study on firstborn IQ and the very strong birth order effects on the LW survey?
NYT: How To Make Congress Bipartisan. Described by Jonathan Haidt as “the best single idea I’ve seen to reduce political polarization and dysfunction”. Make larger districts with proportional representation, so that there’s an actual fight between Democrats and Republicans everywhere, and nobody is more afraid of being primaried than of the general election.
Neural networks generate Harry Potter fan fiction.
New from OpenAI: Deep reinforcement learning from human preferences. Obvious AI safety implications.
Latest study in growth mindset shows decent effect sizes, persistence at least three weeks.
More evidence against corporate campaign contributions mattering: “We find no evidence that corporations benefit from electing their favored candidate, and we can statistically reject effect sizes greater than 0.4 percent of firm value…corporate campaign contributions do not appear to but significant political favors.”
Detecting polygenic adaptation in admixture graphs. Genes linked to educational attainment show signs of differential recent selection in different populations. Except if I’m reading it right, the only populations that show selection are East Asians and Peruvians, which is kind of a weird grab bag of groups. And the East Asian selection seems to have happened very early (10,000+ years ago?), which rules out explanations based on the Chinese civil service exams or any other historical selection pressures. Overall not sure what to think about this. [EDIT: See discussion in the comments]
Someone commenting on my perception/cognition post found me this paper, which tries the same thing and not only finds very little connection between illusion perception and personality, but even very little correlation between perception of different illusions. “The findings suggest that vision is highly specific; ie there is no common factor”.