IrishOutlaw asked me to expand on the social graphs comments that I made in my previous entry; this is my first attempt of (maybe) several, as the topic is rich. 

Graph topologies have nodes and edges.  What is interesting about any graph is how many nodes it has, how many edges it has, how many edges each node has, which nodes are connected, and, for social graphs, how strong those connections are.  Unfortunately, in my previous entry I used the somewhat-misleading words “central” and periphery” to describe nodes; those terms usually imply (low-dimensional) geometric concepts.  In any interesting (i.e., complex) social graph, one may not necessarily be able even to identify where a “center” or a “periphery” lies.  (Just like our (expanding) universe.)

So my question about the importance of being a “central” or “peripheral” node in a social graph may be better expressed in terms of edge count, or edge strength, for nodes.  Thus my previous entry can be translated to imply that the higher a node’s edge count is, the more fit it is.  The longer the node will “survive” (whatever survival means for a social graph).

I received a slightly different perspective from an unexpected source this week: Ben Goertzel.  Ben is one of the most intelligent people that I have ever interacted with, and he posted a note on the SL4 reflector this week in reference to his blog.  I had never read his blog before, so I was amazed to find that he was writing about social graphs and “being a neuron” last fall.

Ben is somewhat horrified to find that the messages passed by those with the most social graph connections (teenagers that constantly use IM, MySpace, Facebook, etc.) seem to be the most trivial: who’s dating whom, who’s cute, etc.  He decries the fact that as social graphs merge into some type of “global brain,” the result is looking a lot like a bunch of morons in the middle.

So maybe those who are the most connected have the least value to add to the graph’s content, while those who are the least connected may have the most to add?  (The “Emily Dickinson” hypothesis, anyone?)

But is the value of the content the supreme measure of fitness?  When we consider communications across outer space, or across a computer chip, there seems to be much overhead, much protocol involved.  Some space communications contain many more correction bits than data bits, for example.  And on a computer chip, the clock keeps beating even if the chip is idle.  So maybe having a heartbeat of communications is extremely important to the social graph, no matter how trivial the content.

And in a market of goods and services, “middle” men are extremely important.  The best sales people constantly spend “face time” with their customers, whether on actual sales calls or not.  So while sales agents may not create much content, their mere communications capabilities are extremely important.  At least until they can be replaced.

Which brings me back to the survival idea: what happens when Google, Digg, MySpace, FaceBook, even artificial intelligences, replace (the functions of) “middle” men?  Will they ever?

A week ago, I was under the impression that smarter and smarter tools would creep in from the “edges” of the graph first, replacing the “peripheral” nodes with the fewest connections, but now I’m not so sure.  Maybe the immediacy and bandwidth of modern communications will render the “middle” superfluous first.  Of course, the answer probably lies somewhere in the “middle” (sorry!), as our global economy creates ever smarter machines, ever greater accumulations of capital, for many possible uses.