Welcome to The Atlantic's liveblog of the epic match between man and machine that is Ken Jennings and Brad Rutter vs. IBM's Watson computer. The battleground is the game of Jeopardy!, which Jennings dominated for more than 70 days in 2004, launching him to international fame. Associate editor Jared Keller will be taking over the controls leading up to the 7:30 p.m. match.
I'm using Storify
to curate messages from around the Web on the Watson vs. Jennings showdown. Scroll to the middle of the post to see the most recent updates.
8:00: The major theme from tonight - both in IBM's infomercials and questions - was the linguistic difference between humans and machines. Watson nailed virtually every question he was asked, including a question about the Dana Carvey's SNL "Church Lady," whose telltale squeal is identifiable primarily because it's rooted in a specific vocal expression. Natural language is built on semantics, on constantly evolving aphorisms and idioms and other idiosyncratic verbal mash-ups that aren't easily constrained within a set verbal template. What Watson shows is a machine with intuition. I'm looking forward to seeing how the final round shapes up.
7:58: As end credits roll, Ken, Brad, and Trebek all look vaguely unhappy. Trebek is probably praying that Watson has better things to do than school his fellow humans. Like, say, travel back in time to protect John Conner.
7:56: Watson guesses wrong with Toronto! Obviously, he's never been to the Windy City. Good thing he only wagered $947. Trebek: "Oh, you sneak." Ken and Brad both got the question right. Neither looked that pleased.
7:54: Final Jeopardy is where legends are born. Category: U.S. Cities. Question: "Its largest airport is named for a WWII hero, its second largest for a WWII battle."
7:50: Watson sure knows a lot about hedgehogs. The computer ends the round with more than $36,000. Brad has something like $5,400. Jennings probably just wants to go home and take a cold shower. I don't see any way Jennings can claw his way back from this defeat, unless he suddenly becomes an art history expert overnight.
7:48: And we're back. Watson, with a brief pause in the beginning, is killing it. Ken gets an answer in edgewise. Everyone's stumped on a question about the King of Spain. Odd.
7:46: Another infomercial, this time about dealing with the torrents of data and information that pour down on everyone else. The idea of programming "insight" is intriguing...I wonder what that looks like in schematics.
IBM's making the comparison to using computer programs to read through massive amounts of medical information in the healthcare system. "We need the tools to glean the best [from medical data]...Most of that data is in natural language, not code."
7:43: The poor measly humans can't even get a word in edgewise. Watson has answered every question from the beginning of the round to the commercial break. Michelle Legro wondered last night if Watson's creators would have a chance to tweak the supercomputer's settings before this round. Maybe he just had a complete breakfast.
7:42: Watson hits another Daily Double, wagers $1,246...and nails it with only 32% confidence. Can a machine get cold feet?
7:40: Watson, Ken, and Brad are all stumped on an art question. Intriguing.
7:39: Watson is just killing it right now. He nailed seven questions in a row, right out of the gate. And a $6,435 wager on a Daily Double? Only a machine.
7:34: The speech problems that face robotics and AI are interesting (the "open answer" question). "Language is a human artifact." Perhaps I spoke too soon regarding this introductory video. "Jeopardy! is a real challenge. It requires regionalisms, nuances, and slang that a computer can't be programmed to do." Is language at the heart of human intelligence? Interesting question.
7:33: The introductory video for IBM is fascinating, but it feels like an infomercial. Space analogies, fellas? This isn't IBM's "Sputnik moment" (or is it? Let's read on).
7:32: Brad Rutter looks confident. Ken Jennings, not so much.
7:30: Goodbye, Vanna White: It's robot time. Jeopardy! is gearing up for the DC area. If you live in a city with consistent cable, feel free to leave your thoughts on this round in the comments section.
7:00: Washington, D.C. may home of the perpetual boredom of C-SPAN, but for some reason we can't even bask in the intellectual slugfest between evil supercomputer and mild-mannered homo sapian in real time. For this, Comcast will pay. Until then, I'll be aggregating insights from around the social web with Storify:
6:55: And...we're back. There are plenty of thoughtful takes on tonight's round of intellectual jousting between Jennings and Watson. At Technology Review, Henry Lieberman praises the contest, emphasizing the logical reasoning required of Watson on Jeopardy! as a tremendous advancement for AI. "Science is not a contest. Science advances by learning general problem-solving principles." At The New Republic, Ezra Deutsch-Feltman is skeptical that the dog and pony show of Watson's Jeopardy!
appearance really matters at all for science: "While "Jeopardy!" clues provide a good test of Watson's ability to decipher natural human language--and this ability really is quite amazing--the competition against the world's best "Jeopardy!" players is probably a misleading gimmick. A better comparison with human minds might be an individual version of "Jeopardy!" where Watson answers as many questions correctly as it can, as quickly as it can, and then its score and time are compared to those of other trivia whizzes."
4:55: Hm. Apparently our local Washington D.C. provider won't broadcast this episode of Jeopardy! until 7:30 p.m. EST. Tune in for our analysis anyway, if you dare. I'll be using Storify to track commentary from around the Web until we're able to provide our original analysis.
If you haven't already, read Brian Christian's March cover story in The Atlantic
on Mind vs. Machine
Hello, Atlantic readers! In case you're new to this much anticipated showdown between man and machine, visit the ineffable Michelle Legro's recap
of last night's clash of trivia titans. For a more technical take, check out Edward Goldstick's analysis
over at James Fallows' blog.