IBM is pushing its revised AI strategy — a stripped-down, less world-changing ambition — to work. The task of reinvigorating growth was handed over to Arvind Krishna, a computer scientist who became chief executive last year after leading the recent overhaul of IBM’s cloud and AI operations.
But the great visions of the past are gone. Today, rather than being shorthand for technological prowess, Watson stands out as a sobering example of the pitfalls of technological hype and hubris surrounding AI.
The advance of artificial intelligence through the mainstream economy, it turns out, will be more of a step-by-step evolution than a catastrophic revolution.
A new wave to ride
During its 110-year history, IBM has introduced and sold new technology to companies time and again. The company dominated the mainframe computer market to the point that it was the target of a federal antitrust lawsuit. PC sales really took off after IBM launched in 1981 and adopted the small machines as essential tools in corporate offices. In the 1990s, IBM helped its traditional corporate clients adapt to the Internet.
IBM executives came to see AI as the next wave to ride.
Mr. Ferrucci first presented Watson’s idea to his bosses in IBM’s research labs in 2006. He thought that building a computer to tackle a question-answer game could advance science in the field of AI known as natural language processing, in which scientists programming computers recognize and analyze words. Another research goal was to develop techniques for automated answering of questions.
After overcoming initial skepticism, Mr. Ferrucci assembled a team of scientists—eventually more than two dozen—working from the company’s lab in Yorktown Heights, NY, about 20 miles north of IBM’s headquarters in Armonk.
The Watson they built was a room-sized supercomputer with thousands of processors running millions of lines of code. The storage disks were filled with digitized reference books, Wikipedia entries, and electronic books. Computer intelligence is a matter of brute force, and the hulking machine required 85,000 watts of power. The human brain, on the other hand, operates on the equivalent of 20 watts.