Daniel Lemire's blog

, 2 min read

Science and Technology links (August 29th 2020)

  1. In children, higher video game time is positively associated with cognition (i.e., kids who play more video games are smarter). Note that it does not follow that playing video games makes you smarter; it could be that smarter kids are more interested in video games.1. I always assumed that the introduction of antibiotics after world war 2 would be visible in the longevity statistics. But neither the introduction of vaccines nor the widespread introduction of antibiotics is visible in the longevity curves as a sharp upward discontinuity. Instead, we see a boring monotonic curve: people live longer and longer without any specific boost due to the introduction of a given technology. In fact, researchers tell us that antibiotics did not have a dramatic effect on the mortality of infectious diseases. I have never encountered a good explanation regarding our ever increasing longevity.
  2. Loeb discusses the effect of ever increasing longevity. He points out that, for the very old, there is no cake large enough to fit one candle per year on a birthday cake. Instead, he suggests, we should use a number of candle proportional to the logarithm of the age. So at year 2, you would get one candle, you might get 2 at year 4, 3 at year 8, 4 at year 16, 5 at year 32, 6 at year 64 and 6 at year 128. It scales much better.
  3. A single injection of a human gene believed to boost muscles and strength in old mice did just that: the old mice got stronger. Evidently, it suggests that the same could be done in elderly human beings.
  4. Computer speed has historically increased due to Moore’s law: processors acquired more and more transitors and got faster clock speeds. Some people have been pessimistic about the future. In There’s plenty of room at the Top, Leiserson et al. point out that we can do much to keep improving the performance of our computers. They conjecture that it will require more specific techniques, such as a better integration of software and hardware. If true, this would be in contrast with historical trends where we have increasingly abstracted out the hardware within the software, and tended to rely on generic hardware designs. It may help vertically integrated vendors like Apple while hurting horizontal players like Intel.