The Suicide of Jeopardy

George Rebane

[This is the transcript of my 18feb11 commentary broadcast bi-weekly on KVMR-FM 89.5.  A related post 'Epochal My Dear Watson' was posted on 17feb11.]

We Americans have always loved winners, and see ourselves as winners, especially in this age of universal self-esteem.  Oh yes, there is that phase we sometimes go through when we root for the underdog.  But that lasts only as long as we can see the underdog eventually coming out on top, or at least making good.  For popular champions, we seldom stick with the perennial loser – all of us look forward to Lucy’s annual snatch of the football that once more lands Charlie Brown on his back.

Something momentous happened on TV’s Jeopardy this week.  And I’m talking about ‘momentous’ as in momentous for Mankind.  But let’s go back a little bit before we examine what a computer called Watson did to the best that we humans could deliver.

Way back when, there was a very popular and long-running TV quiz show called ‘Let’s Make a Deal’ run by host Monty Hall.  In it the contestant wound up facing three doors behind only one which was a valuable prize.  The contestant picked a door to open, then Monty Hall – who knew behind which door the prize was – opened one of the other doors that did not have the prize.  Then came the moment of truth when Monty asked the contestant whether he still wanted to open his originally chosen door or switch to the remaining closed door.  The audience shouted advice until the contestant decided to switch or stick with his first choice.

For years, the game was lots of fun for national audiences until someone took a look at the analytics behind the contestant’s choice, and discovered that it was always better to switch because it doubled your odds of winning the prize.  The game, long thought to be a 50-50 toss up, literally died overnight as all contestants switched doors.  Since the optimum strategy or pat solution was now known, nobody cared.  Uncertainty in the outcome was gone.

Years later more analytics were applied to the game of chess.  This time the question was ‘can a machine be taught to play better chess than a human?’.  Computer giant IBM developed a machine called Deep Blue that beat world champion and chess grand master Gary Kasparov.  Today it is acknowledged that machines can play better chess than humans.  And suddenly no one cared about world championship chess between humans.  Why?  Because everyone now knows that watching humans play chess is like watching the B Team instead of a varsity game.  When was the last time you saw the current champion – whoever that is – challenge a machine?

But a lot of people discounted the machine after Kasparov’s defeat by saying that chess was too simple of a game, and its mastery required just a brute force algorithm to search for the best among future possible moves.   These same people said that humans would continue to dominate in areas where natural language was required with all of its colloquialisms and grammatical quirks, and the domain of required knowledge was unlimited.  In such arenas humans would continue to dominate.

Enter a computer called Watson.  Again developed by IBM, Watson is a super-computer that was built to answer natural language queries over an unlimited knowledge base.  In fact, it could even do it backward – you tell it various parts of a complex answer, and Watson will come up with the question.  That sure sounds a lot like the ever popular TV game Jeopardy.  And sure enough this Monday, Tuesday, and Wednesday Watson was pitted against the two all-time Jeopardy champions.  And in front of a studio audience while ‘standing’ behind its podium between the two human contestants, Watson beat them hands down.

So from now on, everyone will know that when you tune in to Jeopardy and watch humans compete, you know you’re watching the B Teams at play.  Not only that, but the varsity player is too busy to even waste its valuable time with such a game which remains difficult only for mere humans.  My guess is that Jeopardy has lost its sizzle, and with no sizzle there will be no steak – period, end of story.

But it also the beginning of a brand new story about what Watson, and its cheaper and even more accomplished successors, will do with their valuable time.  IBM is already in talks with leading firms in retail, medicine, finance, insurance, law, and, yes, even governments for applications and jobs that were unheard of, especially for machines.

Watson is the latest quantum leap in general artificial intelligence during these pre-Singularity years, and it heralds that another huge segment of human talent will be displaced by machines.  But now they have entered our last bastion of defense, they have penetrated our supposedly unique preserve, our cognitive redoubt.  Never mind the camel’s nose under the tent, the damn thing just walked in through the front door.

My name is Rebane and I also expand on these and other themes in my Union columns, on NCTV, and on georgerebane.com where this transcript appears.  These opinions are not necessarily shared by KVMR.  Thank you for listening.

Comments

[Russ Steele]

I would welcome Watson in answering my technical question at a national call center for a major manufacture. He/she can do better than what we now have to endure from some foreign born person that does not even know where California is.