From the 1993 "Kinetic Engineering" VHS tape. Audio replaced with CD version. Originally from the 1992 album "Man-Amplified".
Liner notes:
---
The most current research into the mind and computers has totally dismissed the idea that packaging raw data for bulk digestion by computers like the early artificial intelligence programs (used by Newell & Simon's Logic Theory Machine) that treated data in Alan Turing terms. These designers assumed that knowledge is purely information that could be formalized, codified and then programmed into computer languages, was the key to cognitive modeling. Instead recent work shows that the more complex and sophisticated pathways to problem solving such as the heuristics people use to latch onto relevant data and reject irrelevant are of paramount importance, indeed the operand in human intelligence. "Knowledge in expert systems is a melange of fact, prejudice and beliefs and most important heuristic strategies" says Edward Feigenbaum (computer science professor at Stanford University and one of the leading voices in California's artificial intelligence community). Boolean algebra is an axiomatic system like Euclidean geometry. The writing up of a system into axioms (or postulates) and theorems and rules of influence is much like describing a machine in blue print form. Methods of proof within such axiomatic systems are either algorithmic (exhaustive and precise) or heuristic (short cuts and uncertain) and these methods prove to be very important to our search for intelligent machines since much of the haphazard type of information processing and the making of off-the-cut plans and decisions are essentially heuristic in nature. Algorithms will never produce machine intelligence. Programmed heuristics on the other hands can make the computer provide results that are at once efficient and intelligent. Again intelligence is proving to be more that just syntactic or symantec.
---