williams@kirk.DEC (John Williams 223-3402) (10/28/85)
There is one critical point that is being missed here. A turing machine is a discrete model of intelligence. It should not be confused with intelligence. Any " real " machine would be operated in a continuous, or at least closely approximated continuous, reality. Could you build an intelligent machine? Yes. Could you accurately model intelligence using the turing model? No. As a discrete model, it neglects some of the continuous phenomenon that cause present computers to fail. It is precisely these uncertainties which are utilized in the structure of the brain to provide the phenomenon known as free will. We will be unable to duplicate intelligence using existing architecture. All current methods for indeterminancy, namely random number generators, repeat. The biggest problem is that a truly random process is unbounded, and that turing machines are bounded. It is hoped that the reader understands the difference between unbounded and infinite. The brain is finite, yet it's processes are unbounded. Current synchronous architecture is unable to duplicate this. John.