> Early in my career as a LBOster I participated in an
> odd little discussion about artificial intelligence. I
> came down firmly against the idea's prospects because,
> as smarter people than myself have shown, no one's yet
> figured out a working bootstrap to cognition method or
> even created a believable ghost of an outline for such
> a technique.
I've always thought that the question is wrong. Someone said "Asking if computers can think is like asking if submarines can swim" ... who says they have to have human-scale (implied super-) intelligence? Wouldn't it be enough to have them perform the task at hand using faculties that _average_ (or even below average, frankly) humans have? Wouldn't that be enough to call it "artificial" intelligence?
They'll never get "there" because the idea is just stupid: who actually believes that the goal should be to have a "smart" computer?
/jordan