Much of "artificial intelligence", as I understand it, gets ideas from the normal world. Take genetic algorithms for instance, which uses ideas from evolutionary biology like inheritance and mutation. I have this (perhaps false) picture of early MIT people combing through books on pretty much anything, getting inspiration for things like Selfridge's Pandemonium, which was populated by shrieking demons who recognized patterns.
And on the other end, many systems in nature seem to compute, or at least share nifty features with the structures a compsci person studies. Many enjoy _Gödel, Escher, Bach_'s description of these themes, and were inspired to join the computing world. http://en.wikipedia.org/wiki/Godel_escher_bach
Unfortunately, I get the impression that compsci is largely dead. Few interesting ideas are being pursued nowadays. A lot of researchers who once did really interesting stuff -- stuff that I use -- feel like they have to work on less interesting systems than they'd prefer. I assume that has to do with decreased funding for fundamental blue-sky research.
> Maybe I am missing something, but this whole thing smacks of old
> Platonic idealism (or the realist position on the universals
> http://en.wikipedia.org/wiki/Universals) wrapped in the computer
> jargon.
I wish there wasn't so much platonism. Alienated me from computing, during the time I thought programming was glorified accounting.
Tayssir