Hi John! I don't see any criticism to my argument in your comment.

I agree neural networks loosely resemble a blank """brain""" trained afterward to do a certain task.

I never said we should pre-program all knowledge in a neural network to achieve AGI. I said Turing hypothesis to build AGI may be more sensible than what we're doing right now. Instead of making a NN learn from scratch, we may want to build a system with some a priori mechanisms to deal with the world (like a child), that would then learn from the environment.