• swlabr@awful.systems
    link
    fedilink
    English
    arrow-up
    9
    ·
    1 day ago

    Possibly my favorite kind of article. “The numbers that get thrown about don’t mean what the people throwing them around think them to mean, here is what they actually mean”. It’s like someone telling you about the defcon numbers and that smaller is more serious, or that if they say 50% chance of rain, they don’t mean it is a coin flip on it raining, they mean that in 50% of the area they are talking about, it will definitely rain. Except this one is: “numbers used in polygenic embryo selection aren’t like base stats in a videogame menu, you turds”

    • scruiser@awful.systems
      link
      fedilink
      English
      arrow-up
      6
      ·
      14 hours ago

      The numbers that get thrown about don’t mean what the people throwing them around think them to mean

      That describes a common rationalist failure mode. They reach for a false sense of quantification by throwing lots of numbers at things, but the numbers are already approximations of much more nuanced, complex, and/or continuous things, so by overemphasizing the numbers, they actually get further from properly understanding. See for example… fixation on IQ; slapping probabilities everywhere; extrapolating trend lines (METR task length); and prediction markets.