Comments by "Science Daemon" (@sciencedaemon) on "" video.

  1. Correction: the output of an AI LLM corresponding to the real world is probabilistic, not coincidental. LLM output is based on predicting next most likely, highest probability, token. These might 100% correspond to the real world. Whether they will is a matter of other factors, like syntax or semantics.
    1
  2. This is wrong. AI predicts, it doesn't guess. Guessing can be random.
    1