What Ever Happened to IBM’s Watson?

IBM insists that its revised A.I. strategy — a pared-down, less world-changing ambition — is working. The job of reviving growth was handed to Arvind Krishna, a computer scientist who became chief executive last year, after leading the recent overhaul of IBM’s cloud and A.I. businesses.

But the grand visions of the past are gone. Today, instead of being a shorthand for technological prowess, Watson stands out as a sobering example of the pitfalls of technological hype and hubris around A.I.

The march of artificial intelligence through the mainstream economy, it turns out, will be more step-by-step evolution than cataclysmic revolution.

A New Wave to Ride

Time and again during its 110-year history, IBM has ushered in new technology and sold it to corporations. The company so dominated the market for mainframe computers that it was the target of a federal antitrust case. PC sales really took off after IBM entered the market in 1981, endorsing the small machines as essential tools in corporate offices. In the 1990s, IBM helped its traditional corporate customers adapt to the internet.

IBM executives came to see A.I. as the next wave to ride.

Mr. Ferrucci first pitched the idea of Watson to his bosses at IBM’s research labs in 2006. He thought building a computer to tackle a question-answer game could push science ahead in the A.I. field known as natural language processing, in which scientists program computers to recognize and analyze words. Another research goal was to advance techniques for automated question answering.

After overcoming initial skepticism, Mr. Ferrucci assembled a team of scientists — eventually more than two dozen — who worked out of the company’s lab in Yorktown Heights, N.Y., about 20 miles north of IBM’s headquarters in Armonk.

The Watson they built was a room-size supercomputer with thousands of processors running millions of lines of code. Its storage disks were filled with digitized reference works, Wikipedia entries and electronic books. Computing intelligence is a brute force affair, and the hulking machine required 85,000 watts of power. The human brain, by contrast, runs on the equivalent of 20 watts.