February 23, 2011

Three Cheers for Watson

For a bunch of philosophical geeks, we sure gave short shrift (interesting digression on what the hell a shrift is and how long it should be) to a rather important milestone, namely a computer's kicking men's assess in their own game of Jeopardy. (Duuh, duh duh, d-d-duh duh duhnt...)

U Cal Berkeley Philosophy Professor John Searle has a guest editorial today in the WSJ, that claims no big. I'll credit his observation that Watson doesn't "know" he's won. And I'll avoid the Kurzweilian rush that we have seen the singularity. I'll nod to Brother AlexC's Facebook allusions to "SkyNet."

But at the end we saw a fantastic display of technology that holds incredible promise. I'll even accept a little Ludditism from the concerned wing. It is not the end of the world but it is a big deal.

I worked four years on a startup that dreamed of being Watson someday. We selected cutting edge AI and Natural Language Processing technologies from research organizations. The hope was to combine them into a useable and saleable toolkit to perform training and tech support. Even an incredibly stupid Watson, limited to a certain domain of material, with a good portion of the questions and answers available ahead of time was a huge challenge.

We ran out of investor patience just as we were starting to exploit synergies between different approaches. I don't have a ton of regrets in life, but I wish we would have had another year to play -- the system was just assembled as we closed shop.

Searle provides a philosophy professor's analogy.

Imagine that a person—me, for example—knows no Chinese and is locked in a room with boxes full of Chinese symbols and an instruction book written in English for manipulating the symbols. Unknown to me, the boxes are called "the database" and the instruction book is called "the program." I am called "the computer."

People outside the room pass in bunches of Chinese symbols that, unknown to me, are questions. I look up in the instruction book what I am supposed to do and I give back answers in Chinese symbols.

Suppose I get so good at shuffling the symbols and passing out the answers that my answers are indistinguishable from a native Chinese speaker's. I give every indication of understanding the language despite the fact that I actually don't understand a word of Chinese.


I think he badly misses the mark here. Watson provided answers that were not in "the database" and missed some that were. A pretty famous clip reveals that the programmers were often surprised.

Chess skills capitalize on the machine's ability to play out billions of scenarios and statistically score them. Impressive, but not Jeopardy.

Moore's Law has come back into currency, and reporters are dutifully noting that the massive server farm that was Watson will be small and cheap in the future. With the rush to the cloud, I think people are overestimating the time it will take by looking at 1990s mantissas.

I don't know that it's SkyNet, but it could well be the next Internet. The scene of an experienced Nurse or medical technician with a Watson-House-Doctor at her side is intriguing and game changing. Place that pattern across multiple industries and Misters Huxley and Shakespeare, our "Brave New World" is here.

UPDATE: Human contestant Ken Jennings has an interesting piece in Slate (HT: Jonathan Last) about the match:

In the final round, I made up ground against Watson by finding the first "Daily Double" clue, and all three of us began furiously hunting for the second one, which we knew was my only hope for catching Watson. (Daily Doubles aren't distributed randomly across the board; as Watson well knows, they're more likely to be in some places than others.) By process of elimination, I became convinced it was hiding in the "Legal E's" category, and, given a 50-50 chance between two clues, chose the $1200 one. No dice. Watson took control of the board and chose "Legal E's" for $1600. There was the Daily Double. Game over for humanity.

Technology Posted by John Kranz at February 23, 2011 11:35 AM

My allusions to SkyNet on Facebook were 98% tongue in cheek, but it doesn't take a Comp Sci PhD to admit that it would not need a "sentient" computer to unleash unstoppable autonomous digital assassins on the world.

Just a human mind and a willingness to give it directions and push the green button marked "go."

I couch that in hyperbole, of course, and perhaps over thinking it, but we're approaching a time where regular human tools of diplomacy and "human" reason might not be enough to stop a real problem.

All of our prior tools of war have always been human controlled. Whether it's an army marching across Europe, ships in the Pacific, a nuclear arsenal pointing at the bad guy or even drones flying in circles around the desert. Men have always had to say "do" before the wrath was unleashed.

See also: "Colossus: The Forbin Project", "Dr Strangelove", "Wargames" and "The Three Laws of Robotics" and probably the Matrix series.

... and no, I'm not a luddite. Just a computer nerd who has written buggy code I'm embarrassed of. :)

Posted by: AlexC at February 24, 2011 12:02 PM

Actually, I was giving you props for being the first to reference it.

Think of how many bugs are certainly there in Watson's code -- eeeeyikes!

Posted by: jk at February 24, 2011 1:01 PM | What do you think? [2]