professor lee smolin (a theoretical physicist) makes a very interesting point regarding A.I.:
If we can't explain why our universe has the laws or initial conditions it does, we can invent a story in which a universe like ours arises randomly in a vast enough collection. Similarly, if we can't yet understand how natural intelligence is produced by a human brain, take the short cut of imagining that the mechanisms which must somehow be present in neuronal circuitry will arise by chance in a large enough network of computers. Neuroscience is advancing quickly; so sometime in this century we may understand how the several aspects of human intelligence arise. But why couldn't such progress require us to come to a detailed understanding of how natural intelligence differs qualitatively from any behavior that a present day computer could exhibit. Why should our early 21st century conception of computation fully encompass natural intelligence, which took communities of cells four billion years?i've said in class: a thinking computer is like an invisible brick rapidly approaching the face of the a.i. skeptic.