Tag Archive for 'gary kasparov'

WATSON OR WIZARD OF OZ?

Igor Yopsvoyomatsky, editor-in-chief of Paranoiaisfact.com,
answers readers’ questions.


Dear Igor,

I’m scared. First a computer  named Deep Blue beats the world champion of chess. Then another one called Watson beats the biggest winners of Jeopardy.  Rajiv, the IT guy at the office, says this is a positive: The thinking power of these machines will be harnessed to cure the ills of mankind, to improve our quality of life. But what hope is there for the Average Joe like me if the very best of our species can’t compete? I can’t sleep. I’m haunted by feelings of inadequacy. I ‘m convinced my worst scifi nightmare is  coming true. Machines are taking over. They’re going to turn us all into  fat, shuffling, overmedicated drones, servicing the giant , blinking Queen Bee computer. Is this paranoia or fact?
Al Angster
Cataract Falls, Pa.

Dear Mr. Angster,

This is fact. There is no hope for the Average Joe. But don’t blame the machine. Until there is a perpetual motion computer  that can reproduce itself the culprit will be the human holding the plug—the engineer.

Computer engineers  use the accretion of data to produce the illusion that a machine is actually thinking. But like the Wizard of Oz they are behind the curtain, pushing the buttons, doing their best to convince you that your mind is too slow and distracted to flourish in the high-speed world. It’s all about deception. The highest achievement in Artificial Intelligence will be to produce a “chatbot,” a computer that tricks a human into believing he/she is talking to another human. 

All human vs. machine contests are designed to show the superiority of the humans who run them. Computer engineers, the Average Joes of the science world, want to prove that middle-of-the-pack knowledge workers can dominate the smarter and more highly skilled. They want to wrest the zeitgeist away from the poets and give it to the Sudoku solvers. Whatever the task, they say, a machine (meaning us, its human masters)  will ultimately do it better.  

IBM is willing to invest millions of dollars to prove this point. Deep Blue creator, Feng-hisung Hsu, began to develop his chess-playing computer in 1985. He worked for 10 years with an elite team, including an international grandmaster, toward one goal—to defeat World Champion Gary Kasparov. When they lost in Philadelphia in 1996 they went back to the drawing board. In 1997 they returned with a new improved Deep Blue Two. “Going into the match I had some apprehension,” Hsu said…”but… we made history and knew we could compete.” 

Deep Blue might be an impersonal machine, but its creators sounded all too human. “After just an hour Kasparov realized how hopeless his position had become,” IBM flacks gloated.” We did not have to wait long for the killer blow from Deep Blow that ended the match…” 

Now Hsu could claim that “brute-force computation has eclipsed humans in chess.” 

But the victory was muddled by controversy. After losing a game the engineers discovered a programming glitch that allowed Kasparov to maneuver the computer into a trap. So they changed the rules to allow them to make a correction between games. To effectively coach the machine.

Kasparov cried foul, alleging Deep Blue had been given “human” guidance in violation of the ground rules. Program director C.J. Tan  spinned his defense in computer-speak.  “we developed a program to change the parameters in between each game…”

Kasparov demanded a rematch. IBM refused and dismantled Deep Blue. The episode ended inconclusively. Feng-hisung Hseu left IBM when he realized they were “not doing anything with the Deep Blue chess chip.” He tried to market it elsewhere, but couldn’t find a commercial application. So much for curing the ills of mankind and improving our quality of life.

Seeking a clear-cut victory IBM turned its attention to another form of competitive data accretion–the quiz show. It discarded the game-playing model and developed a “question-answering” program, which the corporate grovelers  named Watson, after IBM’s founder.  It spent millions on super computers  that  could “process the equivalent of 1 million books of information per second.” IBM challenged Jeopardy’s biggest winners,  Brad Rutter and Ken Jennings. A reporter claimed the stage was set for “humankind to either claim victory over machines or encounter a sobering wake-up call.” 

It’s been a PR bonanza. IBM is back in the headlines. Jeopardy’s ratings are the highest in years. Watson’s program director David Ferruci has become a media celebrity. Tweedy and tieless with a comfy salt and pepper goatee he is the perfect non-threatening representative for a system that IBM hopes will make billions. 

For IBM everything is a sales tool.  If Watson can get the right answers in Jeopardy  it can also pick stocks, diagnose disease, repair complex systems—since most of life is a question-answering process the possibilities are endless. The research center is already at work developing applications. 

Ferruci is calm and genial, but inside he must be churning. It is reported that he is in his office day and night and has to wear a retainer to keep from grinding his gums during the contest. He knows if Watson loses he will join Feng-hsiung Hsu on the IBM dead wood pile.

And again, the engineers behind the curtain are trying to deceive the public. It is not Watson’s knowledge that is carrying the day, but its speed. According to Richard Perez-Pena of  the NY Times success in Jeopardy is “all about timing, and the inherent advantage that chips and wires have over flesh.”  Factor in the nervousness of the human contestants who are in a death struggle against a machine that threatens to render the human species obsolete. “Well,” says Perez-Pena, “it should be obvious… that the computer’s timing edge would make a mockery of the contest.”

Win or lose Watson will not succeed in replicating human ability. Its high speed trading machines will cause markets to crater, power grids to crash, planes to collide. It will assuredly tell someone with lung cancer he has athlete’s foot.  And won’t even have  the good grace to send a wreath. Worst of all, it will put more of our daily activities at the mercy of some nose-picking, chain-smoking, Red Bull-swilling hacker in Guanduong Province, who can alter its questions and answers to sabotage our systems.

Yes, Mr. Angster, you will soon be a fat, shuffling, overmedicated drone.
Best wishes,
Igor