Links for February 2016

Just in the nick of time…

• Eliezer Yudkowsky has a large Facebook thread resulting in many public bets on the Lee Sedol vs DeepMind’s AlphaGo match.
  https://www.facebook.com/yudkowsky/posts/10153941386639228
  In particular, I have bet Carl Shulman $100 at even odd that Sedol will win. (For the record, my confidence is low, and if I win it will be mostly luck.) The match, taking place March 9-15, will be streamed live on YouTube.

  Relatedly, here is excellent (if slightly long winded) discussion of why the apparent jump in AI Go ability may be partially attributable to a purposeful application of additional computing power and researcher GO-specific expertise, rather than purely a large jump in domain-general AI power.

• SciHub has been in the news recently, and I guess they decided to upgrade their appearance.
• Victorian Humor.
• Want a postdoc doing theoretical physics, machine learning, and genomics? You’re in luck.
• Luke Muehlhauser has good quote from Bill Gates on AI timelines.

• “Assortative Mating—A Missing Piece in the Jigsaw of Psychiatric Genetics“.

  Why are psychiatric disorders so highly heritable when they are associated with reduced fecundity? Why are some psychiatric disorders so much more highly heritable than others? Why is there so much genetic comorbidity across psychiatric disorders?

  Although you can see assortative mating for physical traits, like height and weight, with your own eyes, the correlation between spouses is only approximately 0.20 for these traits. For personality, assortative mating is even lower at approximately 0.10. In contrast, Nordsletten and colleagues1 find an amazing amount of assortative mating within psychiatric disorders. Spouse tetrachoric correlations are greater than 0.40 for attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), and schizophrenia. The next highest spouse correlation emerged for substance abuse (range, 0.36-0.39). Assortative mating was significant but far less substantial for other disorders, such as affective disorders (range, 0.14-0.19).

  (H/t Tyler Cowen.)

• Make buses dangerous. A man after my own heart.
• Progress on understanding fast radio bursts.
• Zuckerberg and philanthropy by the ultra-wealthy:

  In the end, though, Zuckerberg’s greatest impact might be in the model he sets for other philanthropists. The Giving Pledge, which encourages billionaires to donate the majority of their wealth to charity, has attracted more than 142 commitments totaling more than $400 billion. The Founders Pledge has convinced 151 startup executives—most of them look about 19—to devote a portion of their exits to philanthropy. Charitable giving in the United States has nearly quintupled since 1994, and shows no signs of reverting back to opera houses and Harvard.

  (H/t Hauke Hillebrandt.) To be clear, I support folks giving to opera houses and Harvard; I just support them giving to AMF more. (And the argument that AMF is more efficient than Harvard is not airtight on account of positive externalities.) Until we reach the point where almost everyone in the developing world is giving away a large fraction of their income, folks who are harshly critical of any net-positive-but-not-optimal giving seem more driven by envy/mudslinging/tribalism than an attempt to improve the world.

• Google Docs now has Voice Typing and it’s pretty good. Google has been rolling out their high-quality dictation service to more of their products recently (i.e., beyond Google Now), and this will be most useful for composing emails and other documents on my phone. The potential speed advantage of dictating on my laptop is small, since I’m about as fast when typing, and you still need to correct mistakes about once a paragraph. But the advantage is huge on the phone; typing on a phone is much slowed than a laptop, but making small correction on a phone is pretty similar.
• Boston dynamics has a new humanoid bipedal robot:
  
• Anthony Aguirre on dealing with the dearth of experimental input in fundamental physics by pushing physicists to make more prediction on the record (in this case, on Metaculus). I strongly support this, although I am agnostic on the details of the implementation.
• Mir was gross:

  In the 1990s samples of extremophile molds were taken from Mir. Ninety species of micro-organisms were found in 1990, four years after the station’s launch. By the time of its decommission in 2001, the number of known different micro-organisms had grown to 140. As space stations get older, the problems with contamination get worse. Molds that develop aboard space stations can produce acids that degrade metal, glass and rubber. The molds in Mir were found growing behind panels and inside air-conditioning equipment. The molds also caused bad smell, which was often cited as visitors’ strongest impressions.

  Some biologists were concerned about the mutant fungi being a major microbiological hazard for humans, and reaching Earth in the splashdown, after having been in an isolated environment for 15 years.

  (H/t this article, which tells a more sensationalistic version.)

• When glaciers melt, surprisingly strong gravitational effects mean they don’t raise sea levels uniformly around the globe. This is easiest to think about in reverse: if you start with a spherical waterworld and then collect a bunch of the water and pile it into a block of ice, the average sea level lowers, but the sea level in the vicinity of the block rises because it is pulled toward the additional ice mass. When glaciers melt, global average levels rise, but within a thousand miles of the glacier they can actually fall substantially.
• The logistics of switching the egg industry to cage free. (H/t Jacy Anthris.) Insofar as you are satisfied with current cage-free standards (certainly debatable), I don’t think pessimism is warranted though. We should expect the average cost of cage-free eggs (which are more expensive, but not fantastically more expensive) to fall as farmers find new innovations and economies of scale.
• Interactive chart for a preferential vote system, in this case using the recent referendum on the New Zealand flag. (Reddit discussion.)
• The LIGO observation was obviously awesome. Here is the signal directly converted to sound:
  
  Here is some raw data. More discussion can be found here. John Preskill’s take on the history of the project here. Jester has a good discussion from the perspective of particle physics. Giant Reddit AMA from a bunch of LIGO researchers. 2000 comments.
• The extremely long development of LIGO is a case study for the point I first heard stressed by Robin Hanson: Innovation does not flow one-way from basic research to applications to the economy; rather, there is a significant back-and-forth. That seems uncontroversial, but it has the under-appreciated implication that a field like theoretical physics that initially races ahead can become maxed out, reaching a state of stagnation, while it waits for the economy (and therefore experimental physics capabilities) to catch up with it. The concept for LIGO developed in late 60s early 70s, and the basic idea has been unchanged since then. I would love to see an analysis of the LIGO technology development (which NSF has funded for decades), but my conjecture is that LIGO was mainly delayed by waiting for the external development various optical technology, and that LIGO-specific insights could have been postponed until relatively recently without greatly slowing the project.

  For speculation about the long term future (beyond LISA) see the BBO and DECIGO proposals (more), as well as Pulsar Timing Arrays
  (which relates to the only serious use of “nanoHertz” I know of, and the Einstein telescope.

  Who’s going to live long enough to see these?

• How touch typing differs from self-taught typing.
• How much noise is there in the criminal justice system? “Randomly Distributed Trial Court Justice: A Case Study and Siren from the Consumer Bankruptcy World“. (H/t Carl Shulman.)
• Benoît Lecomte was the first person to swim across the atlantic ocean. (He had a snorkel, and ate and slept in a boat the followed him, but did the entire swim himself without flotation.) His first words back on land?: “never again”. But now he’s swimming the Pacific, and he’s being studied to see how extreme endurance exercise affects the heart. There is some existing evidence from ultramarathoners that it causes nontrivial inflammation that can be worrisome, although not enough to keep extreme endurance athletes from outliving normal folks.
• Physical Review A finally admits that most of these quantum info papers don’t have much to do with atomic physics or optics.
• Very accessible article by George Church and collaborators describing the promise of CRISPR/Cas9 gene drives.
• Jason Ketola reviews New Omnivore, a different framing on attempts to reduce rather than completely eliminate animal products from people’s diets.
• Discussion with plasma physicist Thomas Klinger, who works on the recently opened Wendelstein 7-X stellarator, on what distinguishes a stellarator from a tokomak design like ITER:

  The crucial factor is how the magnetic field is twisted. For this purpose, in addition to the magnetic field of coils around the ring-shaped plasma vessel, in the tokamak, a component is generated through a current in the plasma, specifically through the current in a transformer coil inside the ring. Because the current must change over time in a transformer, it is regularly started and then switched off again. Then, the plasma is no longer enclosed and cools down again….

  [To avoid this in a stellerator, We] use only the geometry and arrangement of the coils in the plasma chamber to twist the magnetic field…

  The geometric characteristics of the plasma in a conventional stellarator make it very difficult to achieve good plasma confinement. It’s like having a limp: you can do as much training as you like, but you’re never going to be a 100-metre sprinter. However, our former Director, Jürgen Nührenberg, discovered a hidden symmetry characteristic of plasmas in the 1980s which makes it possible to also confine a plasma without plasma current. The shape of the plasma and the magnetic field resulted from this.

  The technical realization of a stellarator is very difficult. Many colleagues said that the idea was nice but no one would be able to build it. Today they say: nice idea but only the Germans can build it.

  This all smack strongly of PR, so not sure how insightful that really is…

• I am skeptical that cryonics technology will be useful in our lifetime. One thing that makes progress so difficult is that the end goal is far away, and it’s difficult to get funding for baby steps. That’s why studying cryogenic organ preservation is so interesting; you might actually achieve a useful intermediate goal. The Economist covers the recent increase in funding for this. (H/t Will Eden.) (See also.)
• Good breakdown of Google’s self-driving car anomalies.
• “Top 10 Replicated Findings From Behavioral Genetics“, R. Plomin et al.:

  In the context of current concerns about replication in psychological science, we describe 10 findings from behavioral genetic research that have replicated robustly. These are “big” findings, both in terms of effect size and potential impact on psychological science, such as linearly increasing heritability of intelligence from infancy (20%) through adulthood (60%). Four of our top 10 findings involve the environment, discoveries that could have been found only with genetically sensitive research designs. We also consider reasons specific to behavioral genetics that might explain why these findings replicate.

  (H/t Gwern.) Listing them:

  1. All psychological traits show significant and substantial genetic influence
  2. No traits are 100% heritable
  3. Heritability is caused by many genes of small effect
  4. Phenotypic correlations between psychological traits show significant and substantial genetic mediation
  5. The heritability of intelligence increases throughout development
  6. Age-to-age stability is mainly due to genetics
  7. Most measures of the “environment” show significant genetic influence
  8. Most associations between environmental measures and psychological traits are significantly mediated genetically
  9. Most environmental effects are not shared by children growing up in the same family
  10. Abnormal is normal

• The Center for Effective Altruism has started a Pareto Fellowship which, as far as I can tell, is like YCombinator incubation for EA projects. Of course, YCombinator accepts nonprofits (like 80k Hours) and claims that many of the same startup principle, but they aren’t focused on effectiveness.
• Microsoft developing underwater data center.
• The Meselson effect:

  The “Meselson effect” is the process by which the two alleles, or copies of a gene, in an asexual diploid organism evolve independently of each other, becoming increasingly different over time. In sexual organisms the processes of recombination and independent assortment allow both of the alleles within an individual to descend from a recent single ancestral allele. Without recombination or independent assortment alleles cannot descend from a recent ancestral allele. Instead the alleles share a last common allelic ancestor at or just preceding the loss of meiotic recombination

  It has reportedly been observed.

  [R]esearchers from the University of Glasgow have demonstrated the Meselson effect for the first time in any organism at a genome-wide level, studying a parasite called Trypanosoma brucei gambiense…T.b. gambiense is responsible for causing African sleeping sickness in humans, leading to severe symptoms including fever, headaches, extreme fatigue, and aching muscles and joints, which do not occur until weeks or sometimes even months after infection….In order to demonstrate the Meselson effect in T.b. gambiense, the research team, led by Dr. Annette Macleod, sequenced the genomes of 85 isolates of the parasite, including multiple samples from disease focus points within Guinea, Cote d’Ivoire and Cameroon, collected over fifty years from 1952 to 2004. The similarity of the genomes studied from these different locations, together with a lack of recombination in the evolution of the parasite, suggests that this sub-species emerged from a single individual within the last 10,000 years. “It was around this time that livestock farming was developing in West Africa, allowing the parasite, which was originally an animal organism, to ‘jump’ from one species to the other via the Tsetse fly,” says lead author Dr. Willie Weir. “Since then, mutations have built up and the lack of sexual recombination in T.b. gambiense means that the two chromosomes in each pair have evolved independently of each other, demonstrating the Meselson effect.” Dr. Weir adds that the parasites’ inability to recombine with each other prevents genes from being exchanged between strains. This could subsequently hamper the ability of the organism to develop resistance to multiple drugs. The team also uncovered evidence that the parasite uses gene conversion to compensate for its lack of sex.

  (H/t Fredrik Bränström.) Note that Wikipedia says “a number of putative examples of the Meselson effect remain controversial because other biological process, such as hybridation, can mimic the Meselson effect”, listing many references.