I have been investigating the mass behavior of stochastic interacting automata. A "law of mass action" governs stochastic interactions; it is inspired by chemistry and incorporated in stochastic pi calculus in the form of the Gillespie simulation algorithm. The dynamics of even just finite state automata is very puzzling. This work is largely explorative: it is almost all simulation, with a tiny bit of process algebra, and no verification. But I think it points to the need for new (or renewed) verification efforts.