In the past couple weeks, we’ve seen an additional couple of fundings in social networking. While I think that there is a lot of opportunity (i.e. ripe ad dollars for the teen/music demographic) that many of these hot new start-ups are chasing, the s…

I like the quote last week from Fox Interactive Media’s President Ross Levinsohn (via paidcontent.org): “[ FIM will focus on ] social portals… Why have one portal when we can have 70 million?” There is a lot merit to this strategy. Many of the  

Those are good thoughts. If you come up with an interesting result, I would be glad to publish it here.

I guess you are making your argument from the limits to the binomial coefficients:

http://en.wikipedia.org/wiki/Binomial_coefficient#Bounds_for_binomial_coefficients

If

C(n,k) 2k, C(n,k) > C(n,k-1), so the number of subsets with no more than k elements is

S = C(n,k)+C(n,k-1)+…+C(n,1)

Metcalfe: each mmeber can make (n-1) connections, namely to all but himself. There are therefore n(n-1) ~ O(n^2) connections possible

Reed: There are 2^n possible subsets of a set of n (different) elements. For each element, decide if it is in or not. Two options chosen n times = 2^n possibilities.

Aren’t you opening up yourself to Odlyzkoing? Clearly the cost of finding the right 150 people among the masses, and the diminishing value of overlapping groups wildly reduces the actual value of the network. Metcalfe’s case is at least plausible, since each interaction between two people is disjoint. But being a part of two subnetworks (cliques?) of 150 people that differ by only one person would be a minor annoyance, and being a part of n choose 150 such cliques would be maddening. You should head this off at the pass and figure out a tighter bound.

What’s the network’s value to an individual member, as opposed to as a whole? Is the value of the network as a whole bounded by the sum of these member valuations? What about leveraging synergies?

n is a big number. n is much bigger than c. n is the total size of the network. For example, n could be the number of the people on the Internet.

If you have a 151 people in a network, all the subgroups with 150 people or less are valuable. The 151 person subgroup is not valuable (by assumption).

ceo