I used to think game theory was pointless. Daniel Davies has a funny bit about how game theory is only useful for predicting the behavior of other game theorists. But Herb Gintis warmed me up to it (the preface of his book here
and his lecture on the unification of the social sciences
. Keefe, I was supposed to remind you to watch that a while ago). If you think of game theory just as a notation for describing strategic interaction and behavior it's actually really useful.
A lot of the non-zero stuff Bob talks about isn't predicting that we're destined by the forces of history to choose cooperate but that if we had any brains we really should choose cooperate.
Fun fact: Rousseau is an acknowledge pioneer in game theory because of his illustration of the stag hunt
(also known as the assurance game and the coordination game). You hear a lot about the prisoner's dilemma but stag hunt dynamics are everywhere too. Somebody even wrote a book about it, which Gintis has reviewed here