For such a bold claim, I'm not really impressed with the model. Basically, the authors take the well-known SIR model [1] (susceptible -> infectious -> removed)

For such a bold claim, I'm not really impressed with the model. Basically, the authors take the well-known SIR model [1] (susceptible -> infectious -> removed) for simple modeling of infectious diseases and modify its removal term to be infectious as well (rather than the infected population just decaying exponentially). The big problem I have with their model is that the population flow is still one-directional, so ANY infection modeled with these equations will eventually go to zero. There is no term in the equations to account for social adhesion that would prevent people from leaving, and there is no pressure for people who have left Facebook to join again. The authors have validated their model by curve-fitting it against MySpace's decline after the fact, so there's no real prediction here. It's an interesting model, but take any predictions based on it with a grain of salt. For anyone interested in reading on similar subjects, here's a paper I co-wrote in which we incorporate a reinfection term into the SIR model and use it to model viral memes [2,3]. [1] http://en.wikipedia.org/wiki/Compartmental_models_in_epidemiology#The_SIR_model http://en.wikipedia.org/wiki/Compartmental_models_in_epidemi... [2] http://www.sciencedirect.com/science/article/pii/S0307904X11002824 http://www.sciencedirect.com/science/article/pii/S0307904X11... [3] http://stash.synchroverge.com/files/viral_memetic_model.pdf http://stash.synchroverge.com/files/viral_memetic_model.pdf