Thanks for all the ideas for things to write about. I had no idea people would actually make suggestions. Thanks!!!
But first, let me start with something that no one asked me to write about, but I just couldn’t pass up. I watched every episode of the computer vs. human Jeopardy! series. Watson (an IBM computer) played against the two most notable Jeopardy! Champions. For a computer nerd, it was riveting. I even found myself rewinding a few times to see why Watson messed up (sortof the nerd version of a football instant replay.)
By now, it is old news that Watson trounced his competitors. Yesterday, an old friend asked if I thought this was a profound accomplishment or something minor. I guess the answer is that it was inevitable, but I have to say that despite the major accomplishment of IBM in creating this computer, the more I looked at the problem Watson had to solve, the more I realized just how much Jeopardy! is suited to computer analysis.
I don’t mean to take away from Watson’s accomplishments, but basically, Jeopardy! gives you lots of hints. It makes it easier to “guess” at what the natural language in the clue actually means. Starting with the category. While some of them have wordplay or puns, most of them are pretty simple: Actors who Direct, Books of the Bible, Name the Decade. In other words, Watson gets a big hint about what kind of answer he is looking for from the category. One way to think about this is that Watson gets one more hint about what he is looking for than you give Google when you search for something online. It is then a bit easier to associate the hints from the clue to this category to find an answer.
That being said, its still damned impressive. Watson currently fills a large server room, but if history is any guide, that kind of processing power will fit in a server rack in six to eight years and it will cost less than 15% of what it cost IBM to build Watson. In other words, this kind of processing power will be available to all kinds of businesses and organizations across the globe.
So, how big of an accomplishment is Watson? I’d say it’s a pretty big mile marker on the road. But it will be even more interesting (although no doubt less reported) when we see machines with this capability pop up around the world over the next few years. What an amazing time we live in.