Reading these studies nowadays feels like reading horoscopes. Chatty, interesting topics but strictly meaningless without an associated prediction or set of pre

Reading these studies nowadays feels like reading horoscopes. Chatty, interesting topics but strictly meaningless without an associated prediction or set of predictions of what will be found. (Predictions derived from theoretical explanations.) Hopefully predictive coding will correct this as we learn that AGIs literally can't see things without learning how to see them first, and therefore neither can we.