Homo ex Machina

Computing machines have grown so efficient that the worst drag on their performance is the fallible human brain. Last week Engineering Consultant Stuart Luman Seaton told a Manhattan convention of the American Institute of Electrical Engineers that computing machines probably make less than one mistake in transferring 10^2DEG (100 billion billion) digits. Humans make one mistake in transferring only 200 digits. So the machine's accurate figuring often goes for nothing because it must depend for care and feeding on error-prone humans.

One way to get more efficiency out of human custodians, says Seaton, is by "tricks and dodges" such as printing numbers large and small, or in varied colors and type sizes. Another would be to spot and correct "psychic blindness" (habits and prejudices) in humans who feed information to computing machines.

Seaton does not expect very much from such measures. Says he: "The presence of humans, in a system containing high-speed electronic computers and highspeed, accurate communications, is quite inhibiting. Every means possible should be employed to eliminate humans in the data-processing chain." But Engineer Seaton feels that humans, however fallible, still have their uses. "The human brain," he concedes, "is a most unusual instrument of elegant and as yet unknown capacity." He favors "reserving to humans the unusual problems of judgment, moral and philosophical balances."

This file is automatically generated by a robot program, so reader's discretion is required.