The basic premise touted by the book Tipping Point as well as decades of work on the spread of epidemics, is that it is easier to start an epidemic by searching for a certain set of 'special' nodes and initiating your epidemic from there. These nodes are referred to as 'influentials' in marketing lingo.

Well apparently 'influentials' are useless according to Duncan Watts. While the arguments Watts makes are sound (the fact that the underlying society has to be more susceptible to trends for epidemics to start, as opposed to the idea that a trend can be made an epidemic if one finds the right set of 'influentials' ), I do have a problem with a couple of his other conclusions.

For instance - why should one assume that highly connected nodes are the most influential ones? Has anyone observed a positive correlation here? Probably not, as people have shown the power of lightly connected nodes in other contexts.

The main conclusion seems to be : if you want to engineer an epidemic, rely on mass-marketing tactics, in that do not worry about identifying special nodes, but build in special mechanisms for normal nodes to spread the epidemic faster. Cool.