Not all spontaneous orders are good. For example, obesity has been shown to spread through social networks. If you hang out with fat friends, chances are you'll gain a few pounds yourself. Smoking also demonstrates network effects. If a person or people start smoking in a given network, more in the network will start. But if a person or people stop smoking, more in the network will quit. The elements that affect network effects are also interesting. For example, those with more education were more likely to quit if their network peers quit.
Albert-László Barabási also has an editorial in which he discusses network medicine in light of the article on obesity. He points out that these kinds of networks -- disease networks, genetic/regulatory networks, neural networks, and social networks -- all have the same basic architecture, meaning if we understand one network, we understand them all at a fundamental level. He also points out that understanding networks allows us to see connections among things we may not otherwise notice. Diseases can be networked together, and thus understanding one may shed light on another. And as he says "drug side effects are inherently network phenomena."
Which is another way of saying there are unintended consequences/spillover effects in network processes. They can be positive, such as the "invisible hand phenomenon" of market economies, or they can be negative, such as drug side effects.
All of which shows that the content of the system matters as much as the architecture. Diseases spread through network effects, with emergent properties, as any epidemiologist will tell you. And there are tipping points. An epidemic is a spontaneous order, but I think we would all, from a human perspective (vs. the germs' perspective) agree it is a perverse one. But tipping points work both ways. Using epidemiology, we can also eliminate diseases. There is a tipping point for wiping out a disease as well.