OK, time to jump in to the screaming masses clamoring to give their opinion on last night's victory of Machine over Man. Watson, the super-computer designed by IBM to compete in a 2-night competition on the popular trivia show, Jeopardy, has defeated the best we can throw at it. This super-computer, so large it was housed in a separate building in New York, boasts such an impressive set of design specs and convoluted tech-jargon as to make the average lay person quake and tremble in awe. But personally, I'm not impressed. And before you start running for the hills clutching your copy of The Matrix to your chest like a Neo-Revelations, let me tell you why.
While it is indeed impressive that the good people at IBM were able to construct a machine that was able to best two of humanities best and brightest (at least as far as benign trivia is concerned), Watson's victory in no way alludes to the existence of any machine that possesses actual intelligence. Sure, the machine had access to enough raw information that they needed to store it on 4 terabyte drives, but knowledge in no way implies actual intelligence. Hell, I know more than a handful of living, breathing people that I would hesitate to apply that adjective to. No matter how impressive the machine's performance, it was still the group of 15 IBM programmers that spent the last 4 years designing and programming the beast who deserve the real accolades on this one.
Do you all remember years back when another machine, known as Deep Blue, defeated the Chess Grand Master Garry Kasparov? Now that was impressive. I don't wish to alienate all the legions of Jeopardy fans before I've hit the one week milestone of blogging, but if there is anyone out there that questions the nobility, integrity, and honor of a chess match over a round of Jeopardy, then you should probably just stop reading this blog now. I'll be saying a lot more things in the future that will piss you off more than any off-handed aspersions cast against Alec Trebek and his snooty coterie of trivia experts.
Originally defeated in 1996, Deep Blue returned in 1997 to be the first computer to ever defeat a seated champion in a regulation match. After suffering defeat, Kasparov cried "Cheat!" from the rooftops, and demanded a rematch, while Deep Blue and its IBM programmers went running back to their labs and denied the master his chance to reclaim his title. Not only did they refuse the rematch, they dismantled the machine.
Now, Kasparov had a few reasons to cry "Foul" after his match against Deep Blue. He claimed, especially after his defeat in '97, that the machine displayed creativity and intelligence in its selection of moves, leading Kasparov to assume that the programmers had let a human chess master take the helm of the machine during pivotal moments of the match. IBM denied these claims, stating that the only human interaction with the machine occurred between matches, as permitted by the official rules of the competition. Kasparov requested to see the machine's data logs and was again met by refusal from IBM, though they did eventually publish the logs on the internet several years later.
This is important for several reasons, the most glaring of them being the requirement of IBM's programming team to update the machine between matches to adjust itself to Kasparov's style. Kasparov, being a prominent public figure in the world of chess, had literally hundreds of transcripts of his games published for the IBM team to study. What this means is that while Kasparov's style of play was catered such as to make him a formidable opponent against any player of the game, Deep Blue was created and designed solely to defeat Kasparov. The Russian was denied the right to study his opponent's playing style, as the machine had no published matches to dissect. So right off the bat the parameters of the competition are stacked against the human champion.
My point here is this. While both Watson and Deep Blue give the illusion of being intelligent, both machines require the man behind the curtain to make the machine do what it claims to do. The real achievement here isn't in the fact that these machines defeated their opponents, but rather that their programmers were able to translate the rules of each engagement into a set of commands, no matter how convoluted, that the machine was able to navigate in order to claim victory. For example, Kasparov was able to defeat Deep Blue twice with the same trap in two of the earlier rounds of their match. Had the IBM programmers been forced to create something that intelligently adapted to the grand master rather than requiring the down time in which the team was allowed to update the code to compensate for his playing style, we'd still have mankind as the unquestioned champion of the game.
Even Watson's victory in my opinion was a victory of comprehending the convoluted nature of Jeopardy's question-answer format, rather than a victory over its opponents. The true challenge here was designing a machine that was capable of navigating its massive libraries of data in enough time as to be able to locate a valid response and buzz in before the competition. No easy task mind you, but still very different from creating something that mimics the cerebral process of man. It all comes down to the metaphysical debate of the difference between knowledge and wisdom. Sure, Watson has a treasure trove of knowledge at its disposal, from the Encyclopedia Britannica to the complete works of Bill Shakespeare, but does it have the intelligence to know what to do with all this data? Not without Old Man Oz behind the curtain pulling the levers. Does it have the creative intuition to be inspired by its collection of the Bard's works to compose a sonnet? Not even close.
To those of you still clutching to trembling knees and looking for the high-ground, I say this: while your fears and concerns may not be entirely unfounded, they are absolutely misdirected. I can't deny that the processing power and brute force capabilities of these machines are utterly astounding, but we are still a long way off from armies of T-1000 model Terminators marching to the beat of their own drums. A computer is nothing more than one fancy ass puppet twitching and dancing to the tune of whomever holds the strings. Pinocchio may have won this round, but he still needs Geppetto to dictate the moves.
Call me up when someone makes a processor that can write an original manifesto, or compose a unique piece of music. Then I'll be right with you lining the caves with pointy sticks and stockpiling canned rations. Until a machine creates something that unquestionably possesses real soul, you can find me doing the same thing I was doing yesterday, and will be doing tomorrow. Rocking out to that inner fire.