Comments by "चतुर्वेदी हर्ष" (@hershchat) on "Big Think" channel.

  1. A GPT neural network doesn’t have free will, but it does have the ability to be trained. If fact, it is nothing but its training. A part of “training” human neural-networks is also reward and punishment. The hatred of fellow humans is a powerful training input as to what not to do. The reason hatred exists is because it is an effective training input. Ditto love and adoration. Sapolski is 100% WRONG in claiming that the absence of free will means that the human organism need neither be seen a deserving reward or punishment. The opposite is true. If you are seeing a trainable automaton, which is what his formulation of human is, then you’re seeing a being that HAS to be rewarded and punished.
    1