The people from Aardvark wrote a research paper titled Anatomy of a Large-Scale Social Search Engine. It has just been accepted to WWW2010, the same conference where the classic Google paper Anatomy of a Large-Scale Hypertextual Web Search Engine was published 12 years ago.
The paper seems to describe and categorize the behaviour of Aardvark’s users which is interesting in itself.
But more importantly, the paper gives a confirmation of something I had been thinking a lot when I worked as the Chief Legal Officer of Wikipedia, and during my PhD:analyzing the online behaviour of people is easier, and more efficient, than actually trying to understand what they say or what they write.
To put it another way, many scientists and entrepreneurs have been trying for years to foster the semantic analysis of what people do online. This means heavy AI investments, lots of errors, less humanity.
A good example of this approach would be to try to understand what kind of goods to sell to someone based on the analysis of the content of his blog posts. Good luck with that.
On the other hand, a few people have been able to create hugely successful projects that rely solely on human interaction: Wikipedia, Meetup, Aardvardk, etc.
And if understanding human content is complicated, recording human interaction is easy, especially online.
And once you get these interactions right… you can predict the next interactions. For example, when you ask a question about restaurants in Paris on aardvardk, they will transfer the question to a friend of a friend of yours… because they connect to your facebook account, and who is actually interested in restaurants… because he already answered such questions.
Simple. Just the way Google did by deciding to narrow search to web links when everybody was trying to recreate an artificial mind that would be able to “understand” the content of web pages.
Good luck with that.
Get the paper and some more analysis here:
