Hunch provides you with a recommendation to a question based on responses from the Hunch community based on you answering a dozen or so Myers Briggs style questions.

The New York based start-up from Flickr’s founder Caterina Fake launched today. It uses machine based learning to generate a personal user profile and pull recommendations from its historical community data.

If you’re seeking a quick answer use Twitter as Broadstuff’s Alan Patrick suggests. I had to answers 20 questions before receiving the recommendation that I should have cornflakes for breakfast.

Hunch is intriguing but its slow and hard going. And I guess this will be the case until it collects more user data.

But you can see that Hunch will quickly become a powerful source of consumer data for market analysis, testing and selling.

I have a hunch: a quick exit to an ad serving play.