Hunch, the decision-making help site, scored a major publicity coup this week as Jimmy Wales, founder of Wikipedia, joined its Board of Advisors. I was an early beta user and initial fan of the site, impressed that my exploration of the question, "What credit card is best for you" led me 5 sub-questions later directly to my credit card of choice. Other questions were not quite so useful, but as the site is built by the community, it seemed reasonable that it would be weak in its early days.
Six months later, Hunch users have added considerable depth to the site, contributing 50,000 decision outcomes to help other guide other users of the site. Hunch combines this user feedback with machine learning tools to help you explore a wide range of questions from whether or not you should visit Russia to which military service you should join.
Wales described his reason for joining Hunch:
I’ve always been intrigued by the potential intersection of community-based, user-generated web platforms and algorithmic, machine-based ones. Wikipedia and Wikia have proven to do a pretty darn good job with the former. Search engines clearly do a great job with the latter. But until recently I hadn’t seen a great example of how the two approaches could come together, co-exist and truly complement each other to form something greater than the sum of the parts – which I believe is the future of the web. (Allow me to call it now: this is what we are going to come to call Web 3.0.)
This "intersection" has actually been well established by PayPal and Palantir, as we learned at Tap the Collective. And while I like the idea of applying it to decision engines, unfortunately Hunch still falls really flat. I just explored the site for the first time in months and the answers were just as random as during the beta period. The top "hunches" for what I should get my dad for Christmas were a shower radio and a margarita mix set. It said that 99% I should visit Russia. I put literally zero value on these suggestions.
As far as mechanisms to leverage the community to help you make decisions, I far prefer the model of Aardvark, which reaches out to your network to help you drive to very specific answers. Last month I got a chat from a guy in California asking me if I had an insights into a confusing part of The Angel's Game, a book that I had just finished days earlier and loved. We ended up corresponding several times. Now that's an amazing way to get a personalized response.