More Links For November 2014
Today’s thing which affects weight gain and is neither eating less or exercising more: exposure to ultraviolet radiation (warning: in mice). And it seems to work independently of Vitamin D, which has some relevance to the many studies showing sunlight has all sorts of good effects which just taking a Vitamin D pill can’t replicate.
Ross Douthat: how worried should we be about the decline in cults? I mean, there’s an obvious good side, ie fewer cults, but does it say something broader about a loss of creativity and nonconformity?
You probably all know about the phrase “turtles all the way down”, so I’m just posting this here as a reminder to myself to use its history and etymology next time I need an example of memetic evolution from a mildly amusing precursor to a nearly perfect version that goes viral.
A study confirms that global inequality is decreasing, an effect powered primarily by people in rapidly-developing countries growing closer to their developed counterparts. But before you celebrate too much, remember that at some point most countries will have caught up with each other and then global inequality will be driven by within-developed-country factors, which are pretty much all tending towards increasing gaps.
The kind of car they drive in Raikoth.
Things that exist: a Disney sitcom about a dog with a blog called Dog With A Blog.
Things I didn’t know: along with their infamous attempts to cast doubt on climate change, the Koch brothers also support gay marriage, cuts to the military, and the American Civil Liberties Union. It is nice that they are working to build a better society, but it would be even nicer if we could be sure we’d still have a planet to put it on.
Two genes have been found to have a significant association with violent crime, with an odds ratio as high as thirteen times the violent crime of the general population (!). The particular gene was already pretty well-known and has been discussed to death, but this acted as confirmation and gave an especially impressive picture of effect size. Since someone will bring up differing frequency in different ethnic groups, here’s a non-terrible discussion of that particular angle. Interestingly, there have already been court cases in which defendants have used a positive test for their gene to “excuse” their crime and decrease their sentence. The philosophical implications of this are confusing and probably too long to get into in a links post.
I always knew “anchorite” was something vaguely like a hermit, but I don’t realize how, uh, metal it was until I read Wikipedia’s anchorite article. After having the rites for the dead said over them by a priest, anchorites would entomb themselves in a tiny cell, with only a tiny opening through which food and water could be passed, and remain there without leaving for the rest of their natural lives, possibly decades (Ozy asks: “Can you have books? I think I would be okay with that if I got books”).
Sardinia is asking to be taken over by the Swiss, on the grounds that the Swiss seem better at running things than the Italians. Aside from the fact that it’s not going to happen, this sounds like a hugely important innovation in governance, adding a third prong to the ideas of competitive governance which so far consist mostly of charter cities and vague motions at running nations like corporations. It seems to keep the best features of colonialism (having corrupt areas with no history of effective self-government ruled by extremely competent foreigners) while throwing away the worst (because presumably if you invite the Swiss in, your association with them is voluntary and you can kick them out if they don’t do a good job). Add something where the Swiss get to keep 10% of whatever they add to Sardinia’s GDP and you’ve got a business model. Cowen memorably describes it as “competitive federalism on a world scale”. I hope the next US election includes a “forget politicians, just let the Swiss run America and see what happens” option.
Read Montague and team try to predict political orientation from fMRI correlates of disgust response. Not even anything obviously political, just how your brain reacts when you see a picture of a dead body. Now, I’m not super knowledgeable about ROC curves, but if I’m reading this right, they got 98 – 99% accuracy. Can that be right? Is this just one of those overfitting things where they’re doing machine learning on too little data and can explain anything they want? Or does some kind of neural disgust wiring explain almost all of politics? Somebody help me out here.
We know genetics causes 50% to 80% of variability in IQ. But no one’s ever been able to find a gene that explains more than a fraction of a percent of that. Is everything extremely rare mutations? Or is there some kind of very bad paradigm failure going on here? A study from 2013 that I just noticed finds that, no, common and easily testable genetic markers explain at least 50% (and probably more) of heritable IQ variance. That means we’re not likely to get some kind of huge breakthrough and we’ve just got to tediously go through each of a couple thousand genes and catalog the tiny contribution made by each and how they interact.
Speaking of IQ, Dalliard’s article Is Psychometric g A Myth? starts by accurately noticing that “as an online discussion about IQ or general intelligence grows longer, the probability of someone linking to statistician Cosma Shalizi’s essay g, a Statistical Myth approaches 1″ and goes on to try to refute the article. I had trouble understanding Shalizi’s original, but found Dalliard’s sketch of Shalizi much easier – which means either that he’s a vastly better writer or that he’s strawmanning him to something simpler and less compelling. I am pretty confident that Dalliard successfully refutes Shalizi’s argument as he understands and portrays it, but I’ll have to reread the original to make sure he gets the argument itself right.
I said last month that Leah Libresco won Halloween, but that might have been premature. Ben Hoffman dressed up as the book Superintelligence: Paths, Dangers, Strategies
I will always link my ingroup: ft.com: Artificial Intelligence – Machine Versus Man. Good profile of MIRI (though see Luke’s clarifications) and some other people involved in the same line of work. Some of whom are more clueful than others – Peter Diamandis is quoted as saying: “Why would machines bother to harm us when we are as interesting to them as the bacteria in the soil outside in the backyard?”, which comes off as less reassuring than he probably intended given that as I write this a dirt field outside my office is being paved over to build a new parking lot (with the consequences for its soil bacteria best left unstated). Much more interesting: Larry Page is reading Superintelligence: Paths, Dangers, Strategies. That might be the best AI risk-related news I’ve heard all year.
The Left learns how to play the “take an old party platform, see how it has changed, and use it to show that modern society has drifted in the wrong direction” game. And they’re doing a pretty convincing job. Compare section 3.5 here.
Media antics: CNN reports that a certain candidate “won 52-47 among women”, which he did. Salon states that this is racist, because the candidate lost among the subgroup of black women, so to say he won among women means that you’re claiming “women of color are some separate entity, some mysterious other, some bizarre demographic of not-women.” Has now turned into giant flame war, see eg Salon Writer Condemns Arithmetic As Racist, which seems like about the correct angle. Maybe we’ll get lucky and it’ll turn out to be that Salon parody site people keep confusing with the real thing?
Trouble At The Kool-Aid Point. Originally written about women in women-hostile fields, it compares harassment to the “Kool Aid Point” in consumer brands where a brand which is too successful starts to inspire backlash (eg “If you use Apple, you’re a sheep”). I think the idea is that women can do okay in these fields if they don’t stick out, but once they gain some measure of fame people start harassing them under the cover of trying to be the person bravely pointing out how undeserved their popularity is. Most of the commentary I’ve been reading has gone beyond the original gendered presentation to discuss how this happens to anyone popular. My personal go-to example would be that as soon as HPMOR became popular, it inspired all of these hate blogs and hate forums attacking it and Eliezer personally under the guise of “righting the wrong” of it being more successful than it “deserved”.
After the mid-term elections is as a good time as any to review the arguments that Obama is basically a Republican. Alternately, maybe the Obama administration has pursued a surprisingly conservative defense policy because the President has a lot less power than the “second government” of diplomats, military brass, and various levels of advisors.
First rationalist to get elected to a state legislature starts a blog for her constituents, name-drops Less Wrong and the sequences. What was I saying about always linking my ingroup? But I feel kind of bad because she’s getting more attention from rationalists than her actual constituents, so maybe don’t bother her too much.
Prisoner’s Dilemma tested among actual prisoners, find that they cooperate much better than the general population. I’m not too surprised. They’re in an environment where they feel like an oppressed group defined in contrast to a much larger group, and that tends to build cohesion (see: norms against ‘snitching’). I can’t access original paper to see if the study was anonymized, but if not that’s another factor – I’d hate to be the guy who defected against my cellmate.
Vox: Give the Democrats some credit for America’s economy recovering much quicker than any other developed nation, plunging unemployment, decreasing household debt, and other generally-ignored indicators of economic health.
Related: Asian-Americans are voting more Republican. This should probably be a bigger story for the Republicans than it is. First, the conventional wisdom is that Republicans are doomed because immigration will alter the future demographic makeup of the US in favor of minorities, who heavily lean Democrat. But Asian-Americans are one of the fastest increasing minorities, so if GOP can capture Asians while Dems capture blacks and Hispanics, they can stand their ground a little better. Second, the Democratic talking point will always be “GOP is the party of racist white people, Democrats are the party of vibrant diversity”, but if the Republicans can get a minority on their side, it will start looking like both parties are multiracial coalitions of different groups. That will confound the Democratic narrative and maybe it would force everybody to think about race and politics in a slightly more sophisticated way.
Related: the biggest lesson of the midterm elections is that unprincipled destructive obstructionism works.
Very related: Nick Land’s electoral strategy would be for the Republicans (or, I guess, the Democrats if they wanted to try) to try to get as much power as possible except the Presidency. Then use their power to obstruct things and ruin the country. The public will blame the (opposite party) President, allowing your party to gather even more power. Then rinse and repeat in a vicious cycle, gradually chipping off Presidential powers so that you control everything. The only downside is that you have to ruin the country for it to work. One may debate how much of a difference this represents from business as usual
Group selectionism has long been considered pretty dead in the evolutionary biology community. But I was recently clued in to a couple-year-old flare-up of the old debate. Biology titan E. O. Wilson published a paper The Evolution of Eusociality claiming that eusociality – the extreme form of cooperation found among insect colonies like ants and bees – could not have evolved through kin selection (as previously believed) but must have evolved through group selection (ie colonies where everyone cooperates beat colonies that don’t). The theory was met with very strong (and sometimes unnecessarily personal) opposition from other important biologists including Richard Dawkins and Jerry Coyne, although Coyne seems to unexpectedly admit the main point that kin selection can’t produce eusociality, which was news to me. Here’s another biologist who gives a good overview of the entire debate. Of interest to me because of the importance I place on this same process in human affairs; people will always be irresistably incentivized to defect, but this is held in check by a counter-incentive to form cooperative communities that spread by group selection.
Stuart Armstrong on explanations for unemployment. Most people familiar with economics know that in theory unemployment shouldn’t exist, since an oversupply of workers should lower salaries until the supply exactly matches demand. But it’s worth remembering how important this process is. If the employment market cleared, then every abled person could have a job in their field, the need for the social safety net would go way down, and people would be able to leave toxic workplace environments knowing there would always be another job they could go into. As such, “why does the employment market fail to clear, in defiance of classical economics?” becomes an important question.
If you’ve read the story of MsScribe, you already know how Internet harrassment + sock puppets + social justice can be a toxic combination. A science fiction author is found to secretly be the same person as a blogger called RequiresHate who uses social justice rhetoric and out-of-context quotes to rile up mobs, send them to harass and threaten competing writers, and damage their careers. She has since given a very partial apology, but her supporters have defended her by saying that it’s racist for white people to police people of color in how they respond to racism – meaning probably there will be no consequences and the same sort of thing will continue. I worry that this sort of thing seems to happen in any community that reaches more than a certain percent social justice people, and it’s one reason I get so paranoid about social justice memes entering communities I care about.
Leah Libresco was one of the first people to link to my old LiveJournal and so helped me get my start in blogging. She’s also put me in touch with some of the best parts of the Catholic blogosphere, given me a place to stay when I visited DC, and sung the part of Cerune in my version of “Philosopher Kripke”. So of course I will advertise her new book
But if you’re so irredeemably evil as to be beyond any hope of divine salvation, don’t worry: Nick Land also has a book out.
