@cshentrup but incorporating meta-models is a big choice! that is, you have a model of the world that is your best guess for how things work, but then you impose some probability that it's wrong and incorporate a variety of other models that you don't think are best, that in some sense you think wrong, with some probability, is a big deal. and then there's the problem of the choice of other models. how do you select & apply probabilities to worldviews you think not-best? /fin