But if you're listening to an idiot.....

Well, that's not entirely true, in an a priori science like praxeology you can actually benefit from sounding things out and re-evaluating them. In neural networks ("AI") feeding output back into an earlier layers is commonly done to reinforce signals.