IBM's Watson jeopardizes humanity in game 2
In the second of three nights of the IBM Challenge, supercomputer Watson dominates and even shows inadvertent humor.
I have an idea for a new TV show. Two men stand, almost in silence, while a machine talks in a monotone and shows the world how clever it is. Will it sell? Perhaps that depends on the ratings for Part 2 of IBM's Jeopardy Challenge. Hollywood loves to copy.
Tonight's episode showed that IBM's Watson supercomputer--and the highly paid engineers who made Watson what he is--had.
Those few errors IBM's Watson committed in the first of three "Jeopardy" shows against former champion Ken Jennings and Brad Rutter were merely like an underling giving his boss a chance on the golf course.
Tonight, Watson was like the know-it-all who turns up at a party and insists on telling everyone just how much he knows. And he knows a lot.
He finished with $35,734, the price of a Detroit house more than his two humanoid competitors.
I watched with another member of the human race and her first question, having watched "Jeopardy" many times more than myself, was about how Watson was getting the questions. I explained that they were texted to him when host Alex Trebek begins to ask them.
"Well, its obvious," she said dismissively. "He can just push the button quicker."
I told her that chess great Gary Kasparov, when he lost to IBM's Deep Blue, suspected foul play. There will clearly be some suspicion here.
However, the difficult truth may be that if you program a machine with enough information and enough logical patterns, it can take in language and interpret it very swiftly.
This show excited me into believing that Watson's understanding of language will lead to far better conversations with, say, the machines at Comcast.
It would have excited engineers more to see that Watson entirely dominated proceedings to the degree that the humans might as well not have been there. Which, I know, would have made some of the singularity souls so very happy.
However, perhaps even more interesting to us real people was that Watson proved himself a skilled bettor.
While he'd be something of a drag at the sports book in Vegas, he managed to calculate an early bet of $6,435 (on a question about my old college), which was amusing in its precision. At the Final Jeopardy question, one about airports, Watson managed to make an error. But, perhaps knowing that the question held a troubling kink for him, he staked a mere $947.
The airport question was in the category "U.S. Cities." But when Watson heard the clue: "Its largest airport was named for a World War II hero; its second largest, for a World War II battle," he seemed to forget that this had to be a U.S. city and answered "Toronto" instead of "Chicago."
One of his handlers, David Ferrucci, explained on the IBM blog that in training, the category subject was downgraded in importance, as answers might be more creatively related to the actual category.
This was, though, a crushing victory for the inanimate. The sheer decimation of the human race in this, one of its beloved game shows, might not work out so well for IBM.
A company that has created some great advertising over the years surely doesn't want to be seen crushing humanity's spirit--and so effortlessly. Worse, Ferrucci perhaps ought to have shown a little more grace in one of the show's little adlets, when he referred to Deep Blue as the computer that beat Kasparov, omitting to mention that it was actually crushed by the plucky human in the first match.
It seems that in Jeopardy, there will only be a one-way crushing. While Watson at least offered a scintilla of charm as he told Trebek "I'll take a guess," when he was only 32 percent sure of answer, not everyone will have been amused at this cataclysmic defeat for Jennings and Rutter.
The humanoids, who looked as if they had witnessed a ghost (perhaps their own), have one more day to fight for the human race. But if their demeanor tonight was anything to go by, they might not even make a buck tomorrow.
Oh, engineers, I understand your excitement. I understand your belief that all technical progress is good progress.
But surely the performance of the night goes to whomever placed an ad for the Dodge Charger in the first break of the show.
The announcer portentously spoke of the horrible events that some engineers are bringing to our lives. For example: "An unmanned car driven by a search engine company."
The worried announcer continued: "We've seen that movie. It ends with robots harvesting our bodies for energy."
It's hard for many of us real humans not to worry that this is how it will all end, even if Dodge will, as it promised in the ad, lead the resistance movement. Could IBM program Watson to understand that?