Tuesday, February 15, 2011

A modern day John Henry tale

We are, as a species, within a stone's throw of something fantastic.  It's difficult to say when it will actually show up in our lives but with technology expanding exponentially it’s a fair bet to believe that it will happen sooner rather than later.   As a lead in, perhaps you remember the story of John Henry, mighty railroader.  That story was the parable of mankind’s transition from the age of human powered production to the industrial age of machine powered production.   In the story, John Henry fights a valiant battle against a steam powered jackhammer in order to save the jobs of those work crews which would be replaced by the new technology.  In the story, the Herculean Henry pulls off a victory by swinging two hammers instead of one.  His effort was never a true depiction of what an average man could do and yet he only barely beat the machine.  As if to underscore the Pyrrhic victory, Henry dies of overexertion after the match.

Today, with the information age already in high gear, we are seeing the new version of the John Henry story which is ushering in the age of machine powered thinking.  The assault on man's long standing intellectual dominance actually started in 1997 when IBM created a computer they called Deep Blue whose goal was to beat the world's chess champion.  Many people thought the computer had no chance because chess is such a game of intense strategy that they believed no computer program could be written that could keep up.  Unfortunately for them they failed to understand that humans are biological machines that, when dissected logically, begin to closely resemble complex computer systems.  This means we have no inherent advantages over silicon based intelligence.  All it requires is evolution in order to catch up.  To underscore the point, Deep Blue beat the world chess champion of the day, Gary Kasparov.

Nowadays, tech savvy people widely understand that chess is a game which is well suited to massively parallel computing.  The computer tries every possible move that it can within a given time period and then runs the best outcomes through some probability filters in order to select the best next move.   That can mean looking hundreds of moves ahead involving billions of potential piece placements whereas Gary Kasparov indicated that his look-ahead limit was more like 12-14 moves.  Kasparov commented after the game that the computer showed what he perceived to be signs of "deep intelligence and creativity in the machine's moves" leading him to claim that IBM had cheated by having human intervention.  Of course that argument holds no water because it would take a chess genius as good as Kasparov in order to supply such intelligence and creativity to the machine.  In other words, had the computer programmers been stepping in to help in real time, their moves could not have impressed Kasparov in the slightest.  Kasparov was guilty of a common human shortcoming: he let ego cloud his judgment.  We are an ego driven species.  We are used to being at the top of the intellectual food chain and it violates our sense of stability and normalcy if something artificial can outperform us.

Enter the most recent John Henry competition: IBM's Watson computer vs. human champions in the TV game Jeopardy.  What an eye opener.  The computer understands not only the questions but the entire context of the game.  We are not simply talking about fast look up of obvious data here - a person sitting in front of a computer Googling the answers would not win.  Watson has to deal with puns and rhymes and clichés and in fact entire paradigms.  This is not just the rapid consideration of the billions of combinations of a finite set of possible moves as was the case with Deep Blue.  This is the thoughtful juxtapositioning of words and facts and implications and innuendo.  Surprisingly, Watson's not just winning, it's kicking ass!

As a software engineer I know that Watson is "just" a machine with software running on it.  I don't pretend to understand all of the advanced strategies and clever learning algorithms that Watson's programmers employed but I do know that it's not magic.  Still, watching Watson play that game against the best humans we can find for it to compete against is really something.  I also know that, as was the case in John Henry vs. the steam hammer, average people are not nearly as capable as the top Jeopardy players.  Watson is showing what we might call "intelligence" which is clearly genius level.

Even though the mythical John Henry won the day, steam powered tools won the war.   Anyone who tried to use human power to compete with steam power was fighting a losing battle.  Fast forwarding through time we now see that industrial operations and manufacturing have been pushed into the background by the information age and the new leaders are the ones who saw the trend in advance and employed it to their advantage.  Watson is now showing us that the day of Artificial Intelligence (AI) is very near.  I don't care how great a business person you might be, if you are up against someone who is using Watson-like AI against you then you won't stand a chance. 

Perhaps even more interesting will be the Watsonification of politics.  Wow.  Imagine participating in a presidential debate against 7 or 8 of the brightest political minds in the world.  Instead of coming empty handed and empty headed in a relative sense, Candidate X shows up with a Bluetooth earpiece which is connected to Watson - political edition.  Each time the other debaters say something, the computer analyzes their political thrust and then draws upon history, famous quotations, current events, facts and figures - you name it - in order to come up with a response that is better, more accurate and more human-resonant than anyone else on stage. 

As Candidate X gets the whisper in his/her ear from the Bluetooth device, he/she just parrots the information verbally - a huge advancement over Obama and his teleprompter!  The success of Candidate X would breed more use of AI in all forms of real time human decision making.  At some point we might be faced with computers defining the human experience!  Like it or not, this is coming no matter what we think or want.  As long as there is a short term advantage to be had, humans will exploit it regardless of the long term consequences.  If we outlaw the use of these things in the future then only outlaws will have access to them.  Watson is coming.
Twitter Delicious Facebook Digg Stumbleupon Favorites More