Return of the Boltzmann brains

Linde in J. Cosmol. Astropart. Phys. 01 (2007) 022 shows that some (though not all) versions of the global (volume-weighted) description avoid the ``Boltzmann brain'' problem raised by Page [Phys. Rev. D 78, 063535 (2008)] if the universe does not have a decay time less than 20 Gyr. Here I give an apparently natural version of the volume-weighted description in which the problem persists, highlighting the ambiguity of taking the ratios of infinite volumes that appear to arise from eternal inflation.