Comments by "Peter deWolf" (@StoneShards) on "AI: The Beast or Jerusalem? | Jonathan Pageau \u0026 Jim Keller | EP 308" video.

  1. Jim's an extraordinarily difficult interview subject! Jordan seemed a bit perplexed by Jim's penchant for fizzling interjection. Jonathan tried to help Jordan's assertiveness by his example in persisting through Jim's interjections. As the interaction progressed exchanges became somewhat more gracious, efficient, though.
    1
  2. How mechanical are these "language models"? Is meaning, then, merely probabilistic?!
    1
  3. The human birth event is far more consequential than Keller's "error reduction" AI. Jordan gets to the point of AI's inherent inability to "desire" and the whole matter of motivation. The organic basis of human motivation is experience simply not available to AI. Imitating human behavior, the AI must ultimately fail. What will be the result of this failure!?
    1
  4. The old TV series, "Person Of Interest", painted a compelling picture of the possibility of surveillance AI--looking prophetic in the process!
    1
  5. AI's strength is the speed with which it can test ideas. But you can't just believe the results to be true--don't trust AI's powers of evaluation without verification. But people love to take shortcuts, and somebody is going to trust unwisely to terrible effect.
    1
  6. Jonathan's right: AI is the god of the scientists; a vision of maximized human intelligence whose realization renders AI a devil.
    1