Group items tagged

Hunch.com helps users search for answers -- but first, it performs a detailed search on the users themselves. Launching today after a year in development, Hunch aims to supply users with computer-generated advice on thousands of lifestyle and consumer questions: What kind of dog should I buy? What should I get dad for Father's Day? Which book by George Orwell would I like? Most important, though, Hunch is not a search engine. Rather than scouring the open Web for information, as Google, Microsoft's new Bing and scores of others do, or collating written opinions, as Amazon.com does, Hunch computes answers by comparing what it knows about you to what it knows about people like you. "Ultimately, what we're doing is providing a kind of shortcut through human expert systems," said Hunch founder Caterina Fake, who also started Flickr.com, the popular photo-sharing site that was acquired by Yahoo in 2005. By first inviting users to answer as many as 1,500 questions about themselves -- an addictive kind of personality test that involves such diverse questions as political orientation, relationship status and whether you believe in UFOs and keep your closet organized -- Hunch looks to assemble a demographic profile whose depth could rival anything in the commercial universe. The New York company also believes that users stand to benefit from this kind of large-scale data farming -- not just from getting better answers, but also from discovering the many microdemographics to which they belong. Hunch also says it will not sell user data to marketers. But this promise, written into the site's privacy policy, is not precisely a legal contract, said Siva Vaidhyanathan, a new-media scholar at the University of Virginia, and the difference leaves the data it collects in a fuzzy domain.