Search Revisited

I've been obsessed lately with a little site called Hunch. I got an invite from someone off some website some 21 days ago and have been hammering away at it ever since. Set to go public in a few weeks, the site is a clever mashup of Facebook quizzes, social bookmarking and the ubiquitous community-oriented website. But alas, it is so much more than that.

On the surface, it's just a collection of questionnaires that answer a fascinatingly broad set of questions: "Should I buy or lease a car?" "Which toaster oven should I buy?" "Where should I go on vacation?" Answer the questions, presented one-by-one in a cheerful, index-card style, and the site pays you back with a series of recommendations along with a clever organized comments section where Hunch users can weigh in with the relative pros and cons. Users write the questionnaires themselves, and other users can make their own amendments and additions to the surveys, wiki-style.

Screen shot from hunch.com; I did not get the tattoo (yet)

Hunch is one of a series of tools that are now emerging from the interwebs not to supplant search engines, but to supplement them. Rather than answer a keyword search with a laundry list of possibly relevant links and leaving it up to the user to sort through them, these tools attempt to organize the morass of the Internet into some form of taxonomy. One set of tools tries to do this using the same search paradigm (see the backlash-to-the-hype Wolfram Alpha or elements of Google-killer Bing); others have fun sorting through the social networks and presenting data in more organized, constrained and often compelling ways (SickCity, or We Feel Fine).

The tool also — so they say — gets better the more you use it. One of Hunch's most engaging features is a small gray box that appears on the home page, which rotates through a seemingly-inexhaustable set of questions. To what end? Just to get to know you better, and presumably, make more accurate recommendations. The queries are often sensible (engagement with technology, favorite types of books) and sometimes inexplicable (Do you usually think about matching the color of your belt with the color of your shoes?) Whether Hunch will ever figure out what to do with the reams of data is a good question. But it's a fun and compelling diversion for now.

I found myself coming back to Hunch for so many reasons. As Slate's Farad Manjoo points out in discussing Wolfram, testing the limits of the system's knowledge is a big draw. Hunch could give me a toaster recommendation, sure, but it could also educate me on the most inexplicable and incredible of topics: "Should I eat something?" (Yes); "What should I eat for lunch?" (Chinese food, apparently); "What conspiracy theory would I find fascinating?" (Big oil was behind the wars in Iraq)...the list goes on. In some future state, perhaps I could ask Hunch to run my entire life for me as a Web 2.0-enabled HAL.

It's worth mentioning that Hunch was often wrong; while it astonishingly recommended my favorite New York neighborhood (Brooklyn Heights), suggested a favorite indie rock band (Fleet Foxes, dead-on) and predicted whether I'd get a promotion (take a wild guess), it also told me I was born to be a physical therapist and that I should grow a goatee. Sorry Hunch, try again. But even when Hunch is wrong, it's oh so right; it combines the useful (vacation planning, product selection) with the entertaining (Which muppet am I?) and manages to provide us the emotional validation we crave when making life's decisions while reminding us subtly of who we are. A cybernerd's uber-sensei for the 21st century? Maybe, maybe not. But I've got a hunch.

Thanks to Theo for contributing.