3 ms·
Networked services intrinsically reward monopolization/centralization, due to Metcalfe's Law [0]: users are more concerned with going where other users are, tha
by lukifer 6y ago
Networked services intrinsically reward monopolization/centralization, due to Metcalfe's Law [0]: users are more concerned with going where other users are, than the merits of the service itself. (A curious exception: Uber and Lyft, presumably because drivers can easily use both simultaneously.)
Facebook, for example, does not represent a monopoly, defined as a social media company generally; but they arguably have ~99% market share within their social niche (with some FOSS clones as a rounding error, and Twitter/Instagram/etc occupying different niches). Contrast with Netflix, which is more disruptable without social network effects, as we've seen with the proliferation of streaming services in the last 5 years.
I don't know what the threshold should be to regulate social media as utilities (mandating transparency, interoperability, and due process), but as such platforms become our de-facto political and conversational commons, I don't think we should allow those spaces to be entirely controlled by private capital, let alone manipulated by Blind Idiot God algorithms, which happily hijack our limbic systems for a quick CPM buck.
[0] https://en.wikipedia.org/wiki/Metcalfe%27s_law https://en.wikipedia.org/wiki/Metcalfe%27s_law
- dredmorbius 6y agoBen Tilly (@btilly here) and Andrew Odlyzko's "A refutation of Metcalfe's Law and a better estimate for the value of networks and network interconnections" gives a better estimator, n * log(n). But otherwise yes. http://www.dtc.umn.edu/~odlyzko/doc/metcalfe.pdf http://www.dtc.umn.edu/~odlyzko/doc/metcalfe.pdf My view is that there is also a costant-across-n (thouggh variable with t) cost function, so: v = n * log(n) - k[t](n) Where cost exceeds value, growth stops. Networks grow until log(n) = k[t](n) n == members t == time