Is the Boltzmann brain problem really clearly established as a problem? Starting from the assumption of a cosmology in which Boltzmann brains dominate over evolved ones, it is not immediately obvious to me that there is a real problem, since only Boltzmann brains indistinguishable from evolved brains will be posing the question in the first place.
In other words, the vast majority of Boltzmann brains will not be troubled by the paradox because their memories will not be consistent with a stable/large physical universe, and so they would never ask such a question in the first place.
So we must only consider the subset of Boltzmann brains whose experiences would be so consistent with a universe like ours that they would pose the question "why are we not Boltzmann brains?" And in such cases, even though we may be Boltzmann brains, our experiences would be indistinguishable from evolved brains.
Is this a criticism of the Boltzmann brain paradox that has been dispatched with, or is it a legitimate cause to be suspicious of whether it is really a paradox at all?
 A: There is some basic condition that must be satisfied by any scientific theory explaining what the world really is: the proposed structure of the world must allow for its scientific cognizability.
This is a practice requirement: if a scientific theory suggests that we cannot do science, it is useless in the very broadest sense, even if such theory is actually true.
And when we try to put forward a scientific theory that the existence of the universe is explained by a random process like Boltzmann's brain, we immediately get a contradiction with the above basic condition.
Namely, if the Universe appeared as a result of an absolutely random process, then it is much more likely that it has just appeared in its current state and, therefore, all cosmology is nonsense, all cosmogony is nonsense, and even the very reasons that prompted the hypothesis of the Boltzmann brain are nonsense.
Thus, on the one hand, the question makes sense. On the other hand, the proposing answer has as its consequence the meaninglessness of the question. This is a problem.
