Google Would Beat Bing At Jeopardy, Says Wolfram

destinyland writes "Stephen Wolfram, the physicist behind the Wolfram Alpha 'answer engine,' believes that Google would beat Bing in any contest based on questions from Jeopardy. 'Wolfram took a sample of Jeopardy clues and fed them into search engines,' explains one technology blog. 'When it came to the first page, Google got 69 percent correct, just beating Ask with 68 percent and Bing on 63 percent. ... To put that into context, the average human contestant gets 60 percent of answers correct, while champion Ken Jennings has a record of 79 percent.' Interestingly, Wikipedia came in last, scoring 23%, though they may have more to do with how Wikipedia handles searches. In two weeks, IBM's Watson computer will compete on Jeopardy against two of the show's all-time human champions."

Comments

Google results still much more accurate

[SuperKendall]

For the past few weeks I've switched over to Bing as my primary search engine.

Overall it works OK, but there have been a number of instances where Google has produced some dramatically better search results, as it in found something related to what I was looking for at all, on the first page. I've only gone over to look at Google when it seemed like the Bing results were not what I was expecting, but it has been interesting to find there still is a pretty large quality gap as I was thinking it might have been closed by now.

Re:Google results still much more accurate

[SuperKendall]

I switched for two reasons:

1) Because I wanted to see if other search engines could work as well

2) Primarily, because I differ too greatly with Google at this point philosophically on the killing of the video tag under the guise to move to an open codec, and I wanted to reduce support of Googles revenue stream, even if only a tiny fraction they will never notice - it just makes me feel better.

Mostly it doesn't matter much, but there are a few times a week at least I have to turn back to Google.

Standard Deviation?

[Kensai7]

Aren't these percentages too close to be meaningful? Of course it depends on the sample, but I think unless we get an all-winning AI it's interesting but nothing really special.