Sorry for the triple-post, but I'm hoping to make my position a bit clearer --

I don't see why probabilities in many-worlds QM should produce any new mysteries that were not already present in ordinary functionalist philosophy. In functionalism, to turn a third-person view of the world into subjective anticipations, you need a criterion to determine whether the world implements a mind, and how many minds it implements, and/or to what degree. Once you have such a criterion, it should be straightforward to apply it to a branching quantum universe, and it doesn't seem obvious a priori that this criterion would say the branching quantum universe implements the same number of minds for every one of the structures that we happen to call a "world" (if there's even a reasonable way to decide when to call something one world and when to call something two worlds).

Mitchell Porter, I think, would say that the same problems of vagueness of existence/implementation/etc kill both functionalism and the MWI. I would say that the solution to these problems for functionalism will show us the way to a solution for the same problems in MWI, one that shows that we have to assign probabilities the Born way to begin with and not equally across worlds.