Our New Computer Overlords

Posted by Matt at 11:25 PM Tagged with: ,
Feb 222011

So I’m sure all my readers are familiar with the exploits of Watson by now (for those of you living on a cave in Mars, Watson is a mechanical Jeopardy champion built by IBM).  I’m not going to rehash what everyone’s saying about what an amazing leap forward Watson represents (though it is pretty amazing).  I do, however, want to note a few things:

The Real Score

Watson not only beat two humans, but beat them in a way that more or less solidifies mechanical dominance in the field.  The fact that there were two human players likely worked AGAINST the human side, a fact which not many people seem to be aware of.  Saying it was 2-on-1 is misleading; in Jeopardy there aren’t any teams.  And I don’t think it’s a stretch to say that Ken and Brad had a lot more in common–in every way–than either of them did with Watson.  So it’s easy to imagine that some questions clearly fall in the “easier for humans in general” category, while others fall in the “easier for computers in general” category.  So Ken and Brad were competing for the same pool of dollars, while Watson had singular dominion over the low-hanging robot fruit.  Now, that said, even if both human players had wagered their full earnings during that final Final Jeopardy, Watson still would have beaten the SUM of their scores by over $10,000.  And that’s the real metric: how does Watson do against humans in general, not how does Watson do in a free-for-all where humans are more likely to beat each other out for points than beat out a computer.

The Real Challenges

There’s a lot of talk about how Watson will change the future of many industries.  And it might.  In fact, technology like this almost certainly WILL change the future, and it’s absolutely worth paying attention to.  Jeopardy was a great field test for Watson, but there are key differences between the game show and industry applications.  First and foremost, the buzzer.  Watson could get to a buzzer in 10 milliseconds.  Human players aren’t quite that fast.  I wonder what the score would have been like if every player had a full minute to answer the question, and everybody answered every question.  Would Watson’s silly answers have proven to be more of a handicap?  Because when people talk about revolutionizing fields like tech support and medicine, a high accuracy rate is MUCH more important than an extra minute of speed (if that’s not obvious, think about which you’d prefer in a medical diagnostician, which is the kind of application people are suggesting).  I’d like to see Watson’s next challenge be in a situation that more closely approximates the kind of applications its designers envision.

The Real Winners

Watson may have won the match, but Watson doesn’t really have the ability to understand that.  It’s not that kind of artificial intelligence.  So let’s look at the real winners:

  1. Engineers and researchers at IBM.  Best.  Resumes.  Ever.
  2. Humanity as a whole, minus the people who will lose their jobs to a Watson-like system
  3. Ken Jennings.  This guy finally beat his arch-rival Brad for second place, got $300,000 in prize money (50% more than Brad, though half of all human earnings went to charity), and proved himself a class act.  He’s also once again a major national figure.  I highly recommend reading his post-game piece in Slate.  It’s filled with good humor, keen observations, and some quite quotable quips.
  4. Fans.  That was awesome to watch.

The Real Losers

OK, those were the winners.  Who lost?

  1. Ken and Brad.  Sorry, guys.
  2. Everyone who’s going to get outsourced to a computer.  This is an especially important category of losers, because unlike past victims of technological advancement (like the typewriter mechanics who lost their jobs to computers), this revolution may not create that many jobs.  If you can replace an entire call center with a few Watson installs, you might lose 100 unskilled positions for every 5 computer engineering/computer science positions.  Economists are going to keep a close eye on this one, and it’s entirely possible that I’m wrong, but this has some society-shaking implications that I don’t think we’ve seen yet.
  3. Google.  Watson isn’t ACTUALLY competing against humanity, not yet.  This was a game.  It’s competing against technological providers of answers.  So watch out, online data sources.  Google, Bing, IMDb, Wikipedia, WebMD…they all just got a little wake-up call, whether they know it or not.  You’ve got a lot of lead time, but business as usual won’t be a viable business forever.

Bottom line: we got a glimpse of the future last week.  And it was pretty fantastic, bordering on fantastical.  I, for one–in addition to welcoming our new computer overlords–can’t wait to see what comes next.

© 2011 The Evitable Future Suffusion theme by Sayontan Sinha