(These are excerpts from my book "Intelligence is not Artificial")
Brain Simulation and Intelligence
Behind the approach of neural networks is the hidden assumption that intelligence, and perhaps consciousness itself, arises out of complexity. This is a notion that dates back to at least to the British neurophysiologist William Grey-Walter who in 1949, before the age of digital computers, already designed early robots named Machina Speculatrix using analogue electronic circuit to simulate brain processes. More recently, David Deamer at UC Santa Cruz has calculated the "brain complexity" of several animals ("Consciousness and intelligence in mammals: Complexity thresholds", 2012).
In 1990 Carver Mead at Caltech described a "neuromorphic" processor, a processor that emulates the human brain.
All the "intelligent" brains that we know are made of neurons. Could the brain be made of ping-pong balls and still be as intelligent? If we connect a trillion ping-pong balls do we get a conscious being? What if the ping-pong balls are made of a material that conducts electricity? What if i connect them exactly like the neurons are connected in my brain: do i get a duplicate of my consciousness or at least a being that is as "intelligent" as me? The hidden assumption behind neural networks is that the material doesn't matter, that it doesn't have to be neurons (flesh), at least insofar as intelligence is concerned; hence, a purist of connectionism would argue that a system made of a trillion of ping-pong balls would be as intelligent as me, as long as it duplicates exactly what happens in my brain.
Back to the Table of Contents
Purchase "Intelligence is not Artificial"