​In ordinary situations involving a small part of the universe, Born's rule seems to work well for calculating probabilities of observations in quantum theory.  However, there are a number of reasons for believing that it is not adequate for many cosmological purposes. Here a number of possible generalizations of Born's rule are discussed, explaining why they are consistent with the present statistical support for Born's rule in ordinary situations but can help solve various cosmological problems. One application is to Boltzmann brains, which can be kept at bay by Bayesian reasoning.