Comments by "Pete Venuti" (@petevenuti7355) on "Wes Roth" channel.

  1. I was really hoping "pieces" meant some kind of factored vector framework where "pieces" of the neural network can be trained independently, good for distributed computing both in training and operation with the possibility that one of these pieces could be installed as an overlay on top of an existing neural network after being separately trained much like neo getting skills downloaded to his brain when he connects to the matrix...or like getting someone else's memories upon receiving a brain tissue graft.
    1
  2. This looks like something I would write as a practical joke if I worked there.
    1