At the moment, IBM is at the top of my list of coolest companies in the world. In case you’re late to the party (as I often am), they’ve developed a system named Watson that’s scheduled to compete on the game show Jeopardy!. Based on mock competitions, it’s expected to do well. As you can see from the promotional video that I linked to, IBM decided to do this as a Grand Challenge. That’s just inspiring. While it’s not hard to imagine applications for this kind of technology, I think the greatest reason to do this is “Because it’s there.“.
They also point out that IBM has a history of taking on such challenges. I remember reading about Deep Blue about fifteen years ago, back when I was still growing up. I probably didn’t appreciate it back then as much as I would have, had I been as old as I am now. I knew it was a historic achievement, but in a way, I sort of expected it to happen eventually. Chess has nice neat rules, and while the possibilities might seem infinite, there’s only so many legal sequences of moves that two people can make on a 8 x 8 board starting with 16 pieces each.
I also knew that being good at chess had much to do with being able to “see” into the future, anticipating what moves your opponent might make in response to various moves one was considering. By playing chess, I knew the frustration of not being able to enumerate such possibilities in my mind, and having to face the limits of my mental abilities. Computers, I thought, do not have many of the failings that humans do, which would make them excellent chess players. Memory is perhaps the best example; we forget things all the time, but computers do not. Once a computer remembers something, it can keep a perfect record of it forever. Humans, on the other hand, tend to forget things over time. The only way I know to avoid this is to have constant reminders. For example, we tend to forget people’s names, unless we see and interact with them regularly.
I think Watson is another huge step forward by IBM. Hearing about it got me thinking about what makes a Grand Challenge. “Difficult” seems like an inadequate description; I find getting out of bed in the morning is pretty difficult. “inspiring” comes much closer to describing what makes a Grand Challenge. I think another part of it is how far back people have been considering the possibility that something could be achieved, even if they had no idea how to do it. Surely, people must have looked at the night sky and thought, “I could stand on that bright white sphere, if I had the power to get there” just as they must seen the graceful flight of birds and thought “If only I had wings with which to catch the wind, then I could soar above the earth”. Before a certain point in history, most people quickly forget such thoughts, disregarding them as impossible fantasy; however, for handful of others, the dream cannot be ignored. Somehow, it persists.
In time, we make progress. Slowly, all the necessary ingredients come into play, until one day, someone realizes that everything that has come before has cleared the way for us to make a bid for the summit, a place that had previously been beyond reach. I think it’s important to remember that if we achieve anything in this life, it’s because we had the good fortune to stand of the shoulders of giants. If this was true in Newton’s time, it is all the more true today.
Sometimes the giants are symbolic. By demonstration, they make us question what’s possible. If X can be done, then why not Y? In the case of the Space Race, the US saw that the Soviets had successfully put Sputnik into orbit. JFK was able to seize this moment, which could have been a cause for despair, and inspired a nation to take the next big leap. Of course, they may have also been motivated by the belief that technological superiority was an existential matter, with the specter of nuclear annihilation close to everyone’s mind. In any case, we had glimpsed the edge of space, and the dream was more alive than ever.
In the case of Watson, I believe web search plays a similar role. Despite Dr. Katharine Frase’s thinly veiled shot across the bow in the Watson video, Google has shown that a machine can provide answers that are stunning and uncanny in their accuracy (full disclosure: Google is my employer). If this is possible, why can’t a computer understand a question posed in natural language and be able to answer it? Perhaps the most amusing thing about the history of computers is how assertions that computers will never be able to do something are regularly proven wrong. Watson is the next big milestone in that history, whose seemingly inevitable end is artificial intelligence, indistinguishable or better than human intelligence.