IBM has spent four years and untold hundreds of millions of dollars developing Watson, a computer that can play Jeopardy.
In fact, Watson answers questions so quickly and accurately that IBM challenged Jeopardy and two of the show's all-time champions to a match. Jeopardy agreed, and today it's taping three computer-vs.-human episodes that will air February 14-16.
So what impact will this have on the average IT executive or corporate data center? I attended a press conference on Thursday to ask that question. I witnessed a test round in which Watson squared off against Ken Jennings, winner of 74-consecutive-games during the 2004-2005 season, and Brad Rutter, the $3.26 million all-time-record money winner on Jeopardy. Formidable challengers to say the least.
If a two-minute Jeopardy round is any indication -- alas, no Double Jeopardy or Final Jeopardy in this press preview -- Watson is likely to win. But it was no blowout; the computer had racked up $4,400 to Jennings' $3,400 and Rutter's $1,200.
Whether Watson comes in first, second or third, very real computing benefits are likely to trickle down to the commercial sector within five years. But more on that a bit later.
Publicity Stunt?
The Jeopardy IBM Challenge follows in the tradition of several other "grand challenges" IBM Research has spearheaded. There was Deep Blue, the computer that took on and eventually beat chess champion Gary Kasparov in the mid 1990s. Then there was Blue Gene, a project launched in 1999 to study the folding of proteins -- crucial in studying certain diseases, but far less entertaining, from a publicity perspective, than taking on chess or Jeopardy champions.
(click image for larger view)
Is the Jeopardy IBM Challenge an enormous publicity stunt? As was the case with Deep Blue, the challenge will generate plenty of press coverage. There are already related clips on YouTube that have seen hundreds of thousands of visits. More than a dozen television news crews were at yesterday's event along with scores of journalists, analysts and bloggers.
That's just the beginning. PBS is planning an entire evening of "smarter" programming on February 9 -- just before the Jeopardy episodes air. The broadcasts will culminate with a Nova special entitled "Smartest Machine on Earth: can a Computer Win on Jeopardy?" Stephen Baker, author of the best seller The Numerati will release on February 17 (the day after the episodes air) a new book entitled Final Jeopardy: Man vs. Machine and the Quest to Know Everything."
And then there's the notoriety that the Jeopardy episodes themselves will generate, undoubtedly triggering yet another wave of news coverage.
None of this could have happened without methodical planning and access facilitated by IBM. If Watson does win, it will be a huge boost to IBM's prestige. It's something IBM can crow about in those commercials that run during the Superbowl and on Sunday morning talk shows. Here's proof enough for corporate decision makers why, as the old adage has it, nobody ever gets fired for choosing IBM.
That's how a cynic might see it, but if Watson doesn't play in the same league with two brilliant Jeopardy players, Big Blue runs the risk of exhibiting very red faces. And to play at that level, IBM had to significantly advance the state of the art in computing -- work that took four years to complete.
