PERSON

Terry Winograd

The computer scientist who built the most convincing demonstration of machine understanding in history, then spent the rest of his career explaining why it was not understanding at all—and why that distinction has never mattered more than it does now.

Terry Winograd is the builder who dismantled his own achievement. In 1968, as a twenty-two-year-old graduate student at MIT, he created SHRDLU—a program that conversed in English about a simulated world of colored blocks with a fluency that electrified the artificial intelligence community. The demonstration appeared to prove that natural language understanding had been substantially solved. Winograd knew, from the inside, that it had not: SHRDLU did not understand English but operated in a closed domain so stripped of complexity that the gap between symbol manipulation and genuine comprehension was structurally undetectable. That gap—between processing and understanding, between output that looks like cognition and the real thing—became the question to which he devoted the rest of his career. His encounter with Heidegger, mediated through Hubert Dreyfus and Fernando Flores, transformed him from AI pioneer into AI’s most philosophically grounded critic, producing the 1986 masterwork Understanding Computers and Cognition. [YOU] on AI finds in

Keep reading with YOU ON AI