Hunch, a new Web-based service that is designed to help users answer questions, is now liive. But Hunch is not a search engine. Instead, the service wants to help users make smarter decisions.
Hunch is the latest project of Caterina Fake, the co-founder of Flickr. Along with a team of MIT grads and self professed "MIT nerds" with background in computer science, the Web service is a community-inspired effort that uses machine learning to answer specific questions.
Unlike search engines like Google or Yahoo or Microsoft's Bing, the idea behind Hunch is to deliver answers catered to specific questions with an individual in mind. The difference between it and other search engines may seem insignificant but, if the Hunch team can deliver on its promise -- and that's a big if -- the difference will be crucial.
Typing the query 'what kind of car should I buy' into Google turns up nearly 2 million results, with content focusing on price ranges, quizzes and tips for car purchasing. Hunch, on the other hand, wants to return results that say: Honda Civic, Volkswagen Jetta and Ford Focus.
Hunch delivers its hunches by gaining information about users. So, when asking the Web service about a new car, it will ask a series of questions to determine the best fit. Hunch will ask about price range, make and model before gathering more details. Electric or hybrid? Domestic or foreign? The idea is that as users answer more questions, Hunch will gather more data and be able to provide an answer specified for the individual.
The Hunch team is also straightforward about the role social networking, or community growth, will play in the service. In order to produce better results, Hunch asks users to log in while answering questions. The results each person types in will help shape and mold the answers that the service provides.
The Hunch algorithm is guided by a simple question: "What can I ask you next that will lead to the best possible result for this decision?" The goal of that question is to reduce the number of possible results from a lot to a lot less.
The hope of the new question answering service is that enough people will participate so that the hunches that are provided will deliver information that is useful and informative.
SHARE THIS ARTICLE
