Mathematician Cites Limits of Computers

Hao Wang, Reader in the Philosophy of Mathematics at Oxford, this week injected a note of caution into present optimism concerning the feats of computing machines.

Recounting his work with IBM computing machines in a series of two lectures given Tuesday and Wednesday, he first commented that "results have been slight so far, but better than I expected," but later pointed out limitations in machine logic. In his first talk, Wang concentrated on his personal experiences with the machines and his theoretic study of solving problems in logic; Wednesday he considered "artificial intelligence" in general.

Wang has developed a new method of programming information for his two computers, affectionately known as "704" and "Stretch." Stretch, for example, solved the first 376 theorems of the Principla Mathematics by Alfred North Whitehead and Bertram Russell, fed to it in code form, in nine minutes--it had taken Wang seven hours to code them.

He closed by commenting on the attachment of scientists to their machines. "People get as emotional about "thinking machines" as they do about politics. E. F. Moore compared them to the search for the philosopher's stone by alchemists: they are looking for something that they are not likely to get. This kind of effort sometimes does lead to useful results, as alchemy once did to chemistry; but too much should not be expected of them."