Comments by "Wiczus" (@wiczus6102) on "The Tesla Bot: BUSTED!" video.

  1. People sometimes wear lingerine, or leather during sex, but spandex???
    1
  2.  @zerosumgame5700  Majority of the work for an AI engineer is to build generalized models with scarce data. It's like science. You observe few objects falling down, measure it and you come up with a formula for gravity. Then you can predict situations you've never experienced yet and you'll be wrong in some of those situations, if your generalisation (9.81m/s^2 in this case) is not very good and you build different models that might train differently or use different data to come with a better generalisation (Gmm/r^2 or Einsteins field equations). In neural networks (as well as many other ML types) the AI learns such generalisations on it's own, the engineer doesn't program them, he only primes the ML model to learn them. In this sense, the machine learns more than the engineer does. There are people who specialize in interpreting the outcomes of ML, because it's not even remotely easy to do so. tl;dr "You can't just use neural networks without understanding them. " "haha tensorflow goes BRRR"
    1