The second elephant is don't split training from inference. There is no such thing as inference in biological systems. All these efforts are futile as long as w

The second elephant is don't split training from inference. There is no such thing as inference in biological systems. All these efforts are futile as long as we do not get rid of the GD.