I like the ecosystem model. Organism vs ecosystem is just a matter of scale. I’ve been using the ecosystem model a lot when talking to people about the shortcomings of homogeneity in network design. Yes it makes your patch cycles simpler, but it also can lead to large areas of a network being prone to the same vulnerabilities / failures. (Think of the great potato famine, in a network sense.)
I’m not sure why Schneier chose to use the robot model.
I agree on the phase change, and that was the main thing that struck me as well.
