Links for December 2014

  • Steve Hsu notes that BGI has applied preimplantation genetic diagnosis (PGD) using next-generation sequencing techniques. Previous uses of PGD have only used simple screening methods, like searching for single-nucleotide polymorphisms, to avoid implanting embryos with a handful of well-understood genetic defects (which are then destroyed). Next generation methods have the potential to sequence most/all of the genome, allowing for traits influences by many genes to be selected for, like height and intelligence.
  • Crab housing exchange.
  • That the National Reconnaissance Office publicly releases the mission patches for their highly classified satellites is well-known on the internet, but I hadn’t realized that there was reason to think that classified details had been accidentally leaked through them in the past.
  • If correct, this is a surprisingly comprehensive and intuitive answer to many aspects of human bilateral (a)symmetry:

    Most of our asymmetry is due to just two organ systems: the GI tract and the heart. The concept that best explains the shape of both of these systems is the idea that a long organ that has to fit in a small body does so by being wound up.

    The heart could be composed of a linear arrangement of a pump, the lungs, and then a second pump. In some organisms like the worm, the heart is a linear pump. However the human body cannot accommodate a linear arrangement and thus we have what is effectively a tube curled up on itself.

    The GI tract is the same story. It would be hugely long if a linear, thus it has to be wound up inside of us. There is no symmetrical way to wind it up. Many organs like the pancreas and the liver actually bud off of the GI tract during development so the asymmetry of the GI tract explains the asymmetry of many of the other abdominal organs. However those organs not involved in the GI system like the ovaries in the kidneys tend to be relatively, although not perfectly, symmetrical. Likewise the lungs are not perfectly symmetrical because the left lung must accommodate the heart.

    The one interesting thing about this whole conversation is that the direction that things rotate in the human body during development is due to tiny molecular motors called “cilia”. If there is a genetic defect in just a single protein that composes the cilia, the cilia are no longer able to guide the process and there is a 50/50 chance that the organs will rotate the “wrong” way. This leads to the inversion of all symmetry in the human body called “situs inversus”. This leads to occasional moments of extreme confusion for doctors, seeing as patients often don’t even know they have reversed symmetry.

    (Needs sources though.) As you can check, it’s impossible to coil up a tube to fill a container in a bilateral symmetric way.

  • The always-excellent Scott Alexander on the perils of scientific research applied to politics: Beware the Man of One Study.
  • Video about a Tunnel Boring Machine in Seattle:

    The defensive comments at 3:40 are a reflection of the fact that work has been halted due to the ground shifting above the tunnel, damaging buildings.

  • Robin Hanson has argued that the conventional wisdom of the inevitability of democracy may be wishful thinking, and emphasizes that the economic success of dictatorships versus democracies is an empirical rather than normative question. The Wall Street Journal discuss some recent evidence against the inevitability of democracy.

    For now, the advance of democracy in Africa appears to have stalled. In 1990, just three of Africa’s 48 countries were electoral democracies, according to Freedom House, a Washington-based pro-democracy advocacy group. By 1994, that number had leapt to 18. Two decades later, only 19 qualify.

    H/t Tyler Cowen.

  • Two approaches to combatting crooked cab drivers: Nevada vs Uber.
  • Lyft is getting back to its ridesharing roots by introducing an easy way for commuters on their way to work to make extra money by picking up people who need a ride. This matches up folks with similar starting points and destinations in real-time, which is much more efficient and flexible than pre-smartphone techniques.
  • I was really disappointed to see the low degree to which this eminent group of scholars engaged with the best arguments concerning the risks of AI. (Luke Muehlhauser corrects some of their misunderstandings.) Straw manning is a common occurrence in all public debates, but it’s amazing that it doesn’t seem to be all that less frequent among very smart and respected people. (Contrast this with the fact that those people are less likely to use ad hominems, name calling, etc.). The cynical argument that I think is right is that people just don’t much care about the actual answers to these question in the same way that they care about the answers to questions that impact their well-being (e.g. “what’s the fasted way to drive home?”). The only reliable way to avoid the above pitfalls is for the arguers to really, actually want to understand the answer to the question (in that childhood curiosity sort of way), and to be at least somewhat uncertain.

    But this is a hard to induce in oneself, so we could really use mechanism that would force people to engage with the best arguments. Here I think collaboratively edited (wiki) documents have potential in the long term, but currently the tools, and the reliance on consensus for dispute resolution, are not sufficient for constructing quality “canonical position statements”.

    Still, this probably wouldn’t help much in this case because an approximation to such a document does exist in Bostrom’s Superintelligence, and it’s clear that most the commenters did not read it before commenting.

  • The NY Times and the Financial Times have popular pieces that are more sympathetic to people concerned about AI risk. However, the arguments are poor and, especially in the NY Times piece, unfortunately play directly into the hands of skeptics who claim that all of this is just Terminator-inspired speculation with no good support.
  • Nature is making their articles free to viewsort of.
  • It appears (arXiv) that a significant new records for high-temperature superconductors at high pressure has been set. The Physics arXiv Blog has an excellent popular-level post.

    More striking, it looks to be a conventional BCS superconductor rather than one of the (still poorly understood) ceramic high-temperature superconductors. However, see this story in Nature about the criticism their team has received about interpreting related experiments with metallic hydrogen.

  • The Science blog covers a new article in Nature about an advance in plasma wavefield acceleration.
  • Technology Review has an article on an idea I’ve seen popularized before: enhancing the passive-cooling abilities of novel materials by tuning the frequencies at which they absorb and radiate. These sorts of articles are always a little mysterious if you don’t know that the second law of thermodynamics induces has strong constraints on the relative degrees of absorption and thermal emittance of a material as a function of wavelength. (If you took an object that absorbed but didn’t emit at a given wavelength, then you take that object and a second body and seal them both in thermally isolated boxes that could only interface by a plane of glass that was transparent to the single frequency. The hypothetical object would keep absorbing without emitting, leading it to become arbitrarily hot and the other body arbitrarily cold.) The trick being exploited is that this restriction only applies at a given wavelength; you can be a perfect absorber at one wavelength and perfectly reflective at another, in principle, but finding material to do this is hard. The sky is transparent (and therefore effectively at order-unity Kelvin since you’re looking into space) at certain frequencies but reflective (and therefore ~ room temperature) at others, so by choosing your absorption curves right you can use space as a heat sink but be reflective (insulating) to solar radiation.
  • There has been recent progress on the prime-number gap, the companion problem to the twin-prime conjecture that is nearing proof thanks to the work of Yitang Zhang.
  • Hi-resolution photo of the Galactic center.
  • It came as quite a surprise to me a while back that nearby galaxies like Andromeda are quite large as viewed from Earth:

    Wish I could see this.

    The reason we don’t notice them usually is that they are very dim, although they are in fact visible to the naked eye under ideal conditions.

    Anyways, this is a simulation of what the upcoming Andromeda-Milky Way collision will look like from our front-row seat.

    Although some of the details are dependent on Andromeda’s transverse velocity (which is difficult to measure), it is generally believed that the two galaxies will pass through each other a few times before settling down and merging. Of course, due to the distance between stars, it is very unlikely that any two stars will collide, or even that the solar system will be measurably distorted by tidal effects.

  • Submarine nuke test.
  • I had always assumed that the Earth’s gravity was strongest on the surface because that’s the case for solid ball’s of uniform density. (A famous classical mechanics problem is to show that the gravity for such a ball starts at zero at the center and rises linearly until the surface, at which it point it falls off like distance squared.) However, the significantly higher density at the core of the Earth means that the gravitational acceleration peaks between the outer core and the lower mantle.
  • Reddit asks “Why is there so much clutter and loose wires on the International Space Station?”. Nasa Operations Support Officer in charge of maintenance on the ISS responds.
Bookmark the permalink.

Leave a Reply

Include [latexpage] in your comment to render LaTeX equations with $'s. (More.)

Your email address will not be published. Required fields are marked with a *.