Thanks! That is actually the plan, but hopefully I (or someone potential coauthor) prove my conjecture (and the stronger conjecture I mentioned in response to to cousin_it) first. I and the MIRIxLosAngeles group are also thinking about how to this approach changes when we give the system the ability to reason about its own WCB maximizing probability function. (Thus making the set of possible models an acausal function of the choices of probabilities)

I am kind of surprised and sad that according to my blog stats, only at most 27 people have actually followed the link and read it.