The IBM-"Jeopardy" Challenge was resolved on Feb. 16, 2011 with a clear victory for IBM's (NYSE: IBM) Watson computer. Should we care or worry that soon computers will battle it out on TV's Jeopardy?
An impressive win for Watson. But does designing a computer system that uses mostly brute force to win Jeopardy create any new knowledge? Given sufficient time and Google (Nasdaq: GOOG), I too could beat--or at least tie--most Jeopardy winners. So by using search algorithms and rules to pick the best answer and doing so in a few seconds before the time elapsed to push the button, Watson smashed its two human opponents. A technologcal victory, but not research--but it was a great public relations success for IBM.
Watson supercomputer consisted of 90 servers with a tremendous computing power of 2,880 CPUs operating at 80 teraflops and a storage capability of 15 terabytes of RAM. Watson is many orders of magnitude more powerful than the IBM 7090 computer that I programmed at Bell Labs in the 1960s to create its version of a painting by Piet Mondrian. People preferred the computer version to Mondrian's painting and actually believed the computer version was done by Mondrian! Like today's Jeopardy challenge, the computer also won decades ago.
In 1973, IBM offered me a job to work on a voice typewriter at its research laboratory. IBM hoped to have a real product in a few years from then. I declined, stating that speech recognition was just too primitive to make such a voice typewriter feasible for the forseeable future. Knowledge-- ot brute force--was needed to perform speech recognition. I notice that Watson did not use speech recognition and instead the actual text of the questions was used as input the same time the question was revealed to the human players.
Decades ago, IBM was synonomous with computers and innovation. Today most young people would have liitle idea about what IBM did--nor care. They also would not know about Bell Labs or the Bell System. How things have changed. And if Watson winning the Jeopardy challenge creates attention, I guess that is good for IBM.
A. Michael Noll is Professor Emeritus of Communications at the Annenberg School for Communication at the University of Southern California and regular FierceTelecom columnist. He is an early pioneer of digital computer art and 3D animation.