[simonh]

The problem with this is that the vast majority of such brains would be a mess. Their experience, cognition, etc would be completely incoherent. Only a teeny, tiny, minuscule fraction of them would have coherent experiences that make sense.

So if the argument is we're most likely to be such brains, then we are most likely to exist in a haze of incoherence. We don't. Right now I have an experience of a coherent historic memory, intentionality, sensory experiences, all of which make sense in the instant. If I am a random Bolzmann brain none of that should be true.

So we'd have to have a reason to suppose that the minuscule fraction of coherent, consistent random Bolzmann brains are more likely than the occurrence of environments that generate 'actual' brains, each of which may generate many, many such brains.