ABSTRACT
The reader will perceive that I cite this example not just because it is a possible empirical test of the hypothesis at issue, but also because the AI community has in effect been performing and re-performing this test for something close to a quarter century now. The results have been persistently negative in just the way feared. The results are indecisive, to be sure. But there is widespread acknowledgment of and celebrated disappointment in the decreasing cognitive returns generated in the classical fashion from machines of everincreasing speed and power. This is empirical evidence relevant to the hypothesis cited above, and it certainly isn’t positive.
