Applying Neuroscience Methods to Microprocessors
arstechnica.com
arstechnica.com
But all of this is just data. Hoping all of this data + clever algorithms will help us understand how things work just seems like an attempt to throw stuff at the wall & see what sticks.
We still need a functional understanding of each of the components in & of themselves, and build from that. Just data, on its own, is as likely to lead to incorrect models as correct models.