IBM’s Watson project has recently beat two top human players in Jeopardy. Although this is not the first time that an IBM computer beat a top human player in a game, what Watson did actually scares me a little (more on that in a bit).
What’s the Big Deal About This Watson Character?
What makes Watson remarkable is that it is able to both understand natural language and to learn. Natural language comprehension is very difficult for a machine because there is so much more to understanding someone than just processing the exact words they say. Context and past experience can alter meaning in subtle but critical ways and machines don’t do well with subtle.
IBM engineers solved this problem by simultaneously applying hundreds of algorithms against huge amounts of reference material. Watson uses a 15 terabyte database of reference information (which equates to 200 million pages of text) to establish context and help it understand. It tears through all this data almost instantly using massive computational power of 90 Power 750 servers.
And did I mention that it learns? In fact, it took Watson 3 years to prepare for this match. When it first played it back in 2007, it got only 15% of the questions right.
Sounds Impressive, But Why Should I Care?
So here’s a gloomy prediction: if your job consists of applying human comprehension to information, you can probably be replaced in less than a decade (likely much less). Unfortunately, this means that knowledge workers across many different industries (healthcare, finance, academic research, etc) are in trouble. I’m certainly not alone in this view: Ken Jennings, the guy who got beat by Watson seems to agree.
Sadly, it’s not even that extreme a prediction. Even if the software (DeepQA) and hardware behind Watson are prohibitively expensive for most problems today, that won’t be the case for long. First, IBM will eventually recoup its development costs (which are said to run into hundreds of millions). Second, there is nothing preventing the cost of this type of service to be amortized across millions of users (Google does just that with its web search technology, right?). So, companies and people won’t need to own millions in hardware and software to use it. By the way, IBM is certainly not shy about its plans for Watson.
But What About Us Meatbags?
While I’m pretty sure that comprehension is no longer a competitive advantage for humans, I’m less sure about insight and creativity. Is the ability to come up with new ideas and see subtle links just a matter of number crunching or is there some divine inspiration that requires a human mind? I don’t know, but I suspect that it may be the former. And if that’s the case, I, for one, welcome our new computer overlords.
You may also like:
Did you love / hate / were unmoved by this post?
Then show your support / disgust / indifference by following me on Twitter!