John hints at a conclusion I reached about twenty years ago when I ventured into the study of LISP and a couple of other approaches to what was then the glamourous and futuristic topic of AI. I spent several months surveying the field and I reached the conclusion that Artificial Intelligence was neither.

It is true IMO that neural networks and fuzzy-logic algorithms do a better job of emulating the kind of conclusions that intelligence comes to in situations that are largely repeating, such as well-established medical diagnoses or trouble-shooting a plumbing problem. I mean, better than more main-stream sorts of programs.

But that doesn't mean they approach the subtle high-speed network performance of a brain, let alone a mind.

A