Last week, one of Flickr's founders, Caterina Fake, launched a new 'decision-making engine' called Hunch.
Hunch differs from normal search portals in that you ask it a question, and bases its answers on what it already knows about you - plus the answers to additional queries that it throws at you.
To illustrate, when you first log on to the site, you are likely to be asked whether you live in the town or country, what sex you are, your age, plus a variety of more esoteric questions such as whether you believe in alien abductions.
If you then ask Hunch a question, such as "Where can I find out about SEO?", you will be given a list of topic matches (such as "How should I SEO my site?") and another list of result matches (eg "Hire a SEO firm"). If you choose a topic match, your search will be refined by further questions before leading you to a result that's personally tailored to you.
Using Hunch is simple and intuitive, but it is certainly too time consuming for people wanting instant answers. It also relies on attracting a huge groundswell of support to offer the breadth and precision it hopes for - at present it offers only 2,400 topics and somewhere in the region of 14,000 follow-up questions. This, according to Fake, results in accuracy of about 80% - something the people at Hunch want to raise to 90% or 95%.
Hunch's saving grace, though, may be its compatibility with social media. Users can share advice as well as ask questions, and the company is hoping to find a way of combinin the new engine with Twitter. If that happens, Hunch may carve out a niche as a useful adjunct to Google - but it certainly want knock it off its perch.
