Stephen Baker

This is the Croatian version. I'd like to have a copy for my shelves, but am still waiting for the publisher to send one. It reminded me of a record from the '60, King Crimson. (Though now that I reacquaint myself with the original, the Croatian Numerati looks a little less energetic. Might be depressed, or perhaps just resigned.)

IBMers talk serious Watson with me

October 7, 2011

A month ago, I led a discussion with three members of the Jeopardy team about Watson. This was for an IBM High Performance Computing virtual event. The panel featured two important characters from Final Jeopardy, Jennifer Chu-Carroll, from the algorithms team, and Eddie Epstein, who headed up the hardware effort. They're joined here by Katharine Frase, who is leading the effort to find post-Jeopardy careers for the Watson technology. The 50-minute video is up in five Youtube segments, beginning with the one below:

Final Jeopardy: In Japanese

August 16, 2011

Coming out in September.

Final Jeopardy book trailer wins award

August 2, 2011

Back in January, my wife and I drove to Best Buy and spent more than $100 on a video camera. Then we came back and fooled around with the curtains and lights in the living room, trying to create an atmosphere which would render me as something between a shadowy droopy-eyed ancient and a ghost. It wasn't easy. Finally I delivered my lines about my upcoming book, Final Jeopardy, and we sent the video out to Sheila English, CEO of Circle of Seven Productions.

The next day I got an email. The lighting was bad. I was dressed too casually. I didn't deliver the lines all that well. Retake?

In the end, Circle of Seven mixed a lot of other material into the trailer, and my face was reduced more much of it to the size of a postage stamp. It was all for the good. And yesterday I learned that the trailer won an award for the best book trailer of the year. Congratulations to Circle of Seven!

We haven't used that video camera since its big day in January. It's sitting in my desk drawer, its mini-tripod still attached. I imagine someone will want to bronze it at some point. I'm willing to entertain offers.

The winning trailer:

Final Jeopardy in Asia

July 26, 2011

Final Jeopardy is coming out in Taiwan on August 1. It's been out for a couple of months in Korea (below), and will come out in September in Japan. I can only imagine how hard it must be to translate. After all, much of the challenge in programming Watson had to do with teaching a computer the nuances of language, including word play and puns. Those are brutal to translate.

1 comment [view]

Brad Rutter's Jeopardy comedy

June 28, 2011

Brad Rutter, the all-time leading money winner in Jeopardy, lost to IBM's Watson computer in February. (The match is the last chapter of the story I tell in Final Jeopardy.) A couple years ago, Brad moved from his hometown of Lancaster, Pa., to Hollywood, where he's trying to make it in television. Here's a short comedy riff featuring his recent history on the game show.

1 comment [view]

How to update The Numerati?

June 22, 2011

A Dutch publisher surprised me a couple months ago by buying rights to the Numerati--this almost three years after it was published in English, and a couple of years after most of the other languages. I immediately promised to update the book, and now I face the job of actually doing it. So here's my question to those of you familar with the book: What should I add or subtract?

I'm heading up to Nantucket tomorrow for a weekend with friends, and I'm going to take along a paperback Numerati to mark up. (We'll see how much progress I make; I usually make my way through six or seven pages on these weekends.) Here are some of the items I think I'll add:

* A reference to the data tracking of TomTom, a Dutch company
* More discussion of social networks
* More privacy
* More discussion of tracking cell-phone data
* A political update, incorporating the 2008 elections

Other ideas?

5 comments [view]

Arabic Numerati translation up for honors

May 20, 2011

I just received the Arabic version of The Numerati in the mail, and now I learn that the translation is up for an award. It is nominated for The Custodian of the Two Holy Mosques Abdullah bin Abdulaziz International Award for Translation. How about that? Congratuations to All Prints, the Arabic publisher, and the translator, whose name just has to be in the book--but is invisible to me. If the Numerati translation wins, I'd be very happy to fly over to Riyadh and join in the celebration. It should be a good one, considering that the winner gets 500,000 Saudi Riyals, or $133,000.

1 comment [view]

What does Microsoft think of IBM's Watson?

May 17, 2011

I'm in Seattle this week to promote Final Jeopardy. Tomorrow I'll be speaking at Microsoft. A little over a year ago, I was planning a reporting trip to Seattle, and I asked for interviews at Microsoft. I wanted to learn about their research in Artificial Intelligence and Question-Answering, and pick researchers' brains about IBM's Jeopardy computer, Watson. I didn't get the interviews. When I traveled to Seattle, I ended up talking to people at Vulcan Inc., Paul Allen's incubator, and at University of Washington--but never crossed to Redmond.

So here are some of the things I'll discuss tomorrow at Microsoft (Bldg 99, at 1:30):

1) Only a small handful of tech companies in the world can put 25 to 30 phds on a single project for four years. IBM does this regularly. Would this also be a good model for Microsoft?

2) IBM's David Ferrucci, chief scientist of the Jeopardy project, discouraged his researchers from publishing academic papers during the Watson project. Ferrucci criticizes the culture of producing academic papers, arguing that it focuses research on small and incremental advances, and distracts teams from larger goals--like building Watson. Microsoft Researchers publish lots of academic research. I'm wondering what they think about Ferrucci's points.

3) Watson represented an advance of natural-language processing for question-answering technology. Some people view it as a significant advance, others consider it a grandiose PR stunt. (As I wrote in the book, for many in Artificial Intelligence, Watson was "too dumb, too ignorant, too famous and too rich.") I view the Watson as both valuable research and a PR stunt, but will be interested to hear what people at Microsoft think.

***
I discussed these issues in an interview this morning with Todd Bishop of GeekWire. Here's his post. Also, Thursday evening at 7 I'll be speaking and signing books at Third Place Books, on Bothell Way in Lake Forest Park. If you're in the area, please come.

By the way, they have a very beautiful public library here in Seattle. It's just up the hill from my hotel. I took some photos there, including this one from a floor where everything is painted red.

2 comments [view]

Response to New Yorker article

May 6, 2011

In a New Yorker article earlier this spring, Adam Gopnik wrote about IBM's Watson and was kind enough to mention my book. I responded to the article with a clarification in a letter to the editor. Since it wasn't published, I figure I'll publish it here:

Dear Sirs:
It is true that Watson, the Jeopardy computer, is often clueless, and it doesn't "know" anything, certainly not the way we do. But I don't think it's quite right to call it "a server stuffed with answers." This is because the machine has to formulate its answers on the fly. For example, one clue asks for the northernmost country with which the United States has no diplomatic relations. The answer is not sitting there waiting to be retrieved. To come up with it, Watson must first make sense of the clue, and then dig through its data to identify the four countries on the outs with the U.S. Next, it must determine which one lies farthest north. It then runs through a few million final calculations to weigh its statistical confidence in that response (What is North Korea?), and to decide whether it is sure enough to justify a bet. As Mr. Gopnik correctly notes, Watson is woefully lacking in what we know as intelligence. But it is designed to overcome its cognitive handicaps with an immense amount of work.

4 comments [view]