Thoughts / Mediated content, part 1 / August 8, 2007

We love social networks. Or we don’t.

There’s something interesting about social networks: they put us in contact with information and resources that are indiscernible for us without peer’s mediation.

Until now (in most of cases) peer mediation has existed as a side-effect, as a collateral. Most of social networks don’t let peers mediate information or resources, just share them.

Side-effect I mean since the fact that finding two peers that share the same information might actually influence (and inform) our vision of those resources and bits of information.

Currently some of us can benefit from mediation, as in “if [friend x] would recommend a film, I’d go and see it, cause he shares my taste”. Just an example.

Now, another (real-life?) example: I love TED talks, but I surely don’t have the time to see them all. Some of my friends (colleagues, mostly) have the same lack of time and the same passion for some of the talks. I get some from friends, some from browsing, some from blogs.

What if one could build a social sub-network for TED conferences? then one would have all one’s friends and colleagues that like the TED talks on one group, and they could post all their favourites, and one could browse their choices. Boring. I might not like the ones they like. Boring.

Why boring? Because it is accumulating data, and I think we have the desire and the need for mediated information. Someone that can make the decision for us on what we might like. Sort of a secretary, an agent (or better an intelligent agent)

Now let’s explore a scenario:

I have my facebook community. I select some people and add them to a channel I just created, the TED channel. I get some videos from TED they promote. I vote on every video I see. Voting goes directly to the person that suggest it, but also to what the video is categorised as, suppose “technology, sustainability”, and also (for the TED voting system), to certain keywords such as “dazzling, informative, mind-changing”. Server software analyses it and gives back an answer. The ranking of the person that published the video I voted gets raised or lowered, the keywords are weighted and added to a scrutinized, hierarchy-driven list, the theme gets promoted/demoted, and so on. At the end I start getting the TED videos I might like.

Where’s the magic? In mixing systems. My friends are my choice, and I’ll keep learning from them (“Chris posts too many psychology videos, not what I like” so Chris gets demoted in “TED importance” by the system) and the rest is the algorythms’ choice (user xxx has seen 12 “technology”, 4 “dazzling”, 9 “funny” videos, this “dazzling, technology funny” video might suit him), and video voting (this “dazzling technology funny” video was not that good, “technology and “funny” are still big, let’s keep them and throw “dazzling” down a notch). In the end I might get better choices.

Facebook already has the API, TED is easy to link and rather interesting, we might as well stop doing little “I hug/puke/bitchslap you” applications.

Is it too complicated? Perhaps. Four years ago we might have said the same to a page ranking system like Google has right now. “We’re not google”. Sure. But we might get there, eventually.