statements like "To break a warlord equilibrium, you need government" only confuse us.
I agree. I cannot think of any examples of societies that were stuck in a warlord equilibrium and then suddenly said, "Let's sit down and write a Constitution and all submit to the rule of law."
There is a tremendous amount of path dependence in political systems. See also Tyler Cowen's post on culture, where he points to the importance of priors, beliefs, peers, and stories.
It is confusing to talk about "breaking a warlord equilibrium," because that is not how governmental systems emerge. Similarly, I think it is confusing to talk about anarcho-capitalism--or any other political or economic system--as if we could just sit down one day with an institutional blank slate and adopt any system that comes to mind.
Government is not going to disappear in a flash. Trying to talk about what would or would not happen if it were to disappear in a flash is bound to be confusing.
I think that Bryan and I will have to agree to disagree on whether to characterize the U.S. system as one with a strong warlord. I still believe that checks and balances, institutional and otherwise, make a difference. That view informs my take on the latest flap over domestic spying.