dessalles

Sifting through SXSW panel notes…

The Supercollider: A Hero of the Social Network set out to understand what influence the well-connected have on the formation and usage of social web services. The intended conversation about “Supercolliders” wasn’t particularly interesting, but panelists Ben Cerveny and Matt Jones did drive a more compelling sub-plot about flow-based social networking. While neither attempted an explicit classification, both seemed to paint an impression of two types of social apps: The first class being services where the distribution of information is informed by pre-defined relationships – you receive photos I uploaded because we had previously declared each other as friends. And the second class of services are ones where the flow of information is what defines relationships – we are friends because we regularly send each other photos we’ve uploaded. The general consensus of the panelists was that the first, more “traditional”, model is proving increasingly ill-suited to support the activities of these extra-social, collision-prone users. Jen Bekman, qualifying herself as a Supercollider, described her network fatigue as a consequence of the inadequacies of this first class of services. She described the frequent problem she faces of forgetting what personal informational and media is being sent to whom as a result of having to define relationships long-before the need to share information arises. And she described the handicapping of services she uses as a result of tailoring her use to suit the least-common denominator relationships in her diluted, hyperextended social networks.

So, from where I was sitting, the $15 billion question was whether the second class of social networks would solve Jen’s problems. Are the mechanics of today’s social networks too limiting? If we instead define relationships based on the flow of information between people, would the value of more precisely defined relationships outweigh the transaction costs associated with specifying recipients every time something is to be shared? Are users willing to re-frame their relationships in a communications context? Would coherent social sub-networks emerge so that Bekman could eventually share what she wants with the appropriate audience easily and selectively? Could machine learning intelligibly encode enough meaning into relationships so users can forgo (which they already do anyway) the painstaking process of setting up privacy controls and filters for their growing social nets? Is the most effective means of gaining more precise control over who sees what personal information actually a matter of giving up control to some software? And so on…