I recently read The Quest for Artificial Intelligence by Nils Nilsson. Interrupting a fairly linear history of the AI field is an interlude on some more philosophical questions about artificial intelligence, minds and thought. Searle's Chinese Room, things like that.
I got to thinking about some of these questions on my daily peregrinations with Winnie, and I started wondering about the scales that are involved in most discussions of minds and AI. Searle talks about an individual person in his Chinese Room, which makes the idea of some sort of disembodied intelligence actually understanding the Chinese sentences it's responding to seem pretty absurd. But is that in any sense a realistic representation of a brain or a human mind?
In keeping with my upbringing as a baby physicist, I'm going to take a very reductionist approach to this question. I want to get some idea of exactly what the information storage capacity of a human brain is. I'm deliberately not going to think about information processing speeds (because that would involve too much thinking about switching rates, communication between different parts of brains, and other biological things about which I know very little). I'll treat this in the spirit of a Fermi problem, which means I'll be horrifyingly slapdash with anything other than powers of ten. There will be big numbers.