This is an excellent, and puzzling, question! Let me try to provide some insight into how physicists think about such paradoxes, by addressing the specific example you mention, of the presumed uncountable infinity of different possible photon energies in a finite range of frequencies. In the case of blackbody radiation, when physicists analyze the set of possible photon energies more carefully, they find that there is really only an infinite number of different possible photon energies (for a finite range of frequencies) if the volume of space containing the photons is infinite. In any finite volume of space, if we allow ourselves to place boundary conditions on the electromagnetic fields at the edges of that volume (for example, suppose we think of our volume as a cube with mirrored walls), we find that there are only a finite number of oscillating modes of the electromagnetic field in any finite range of frequency. In the case of a cube, there is a very lowest frequency of radiation whose wavelength will allow it to form a standing wave in the box, and all the other allowed modes are multiples of that lowest frequency. So the density of possible photon states per unit of frequency is actually proportional to the volume of space we allow to hold the photons. (By the way, this is also precisely related to the quantum complementarity of uncertainty between momentum and position. To confine a photon to a volume of space, the uncertainty in its momentum must be the same order as that carried by the lowest frequency standing waves which would be compatible with the container.)
I think part of my mistake was not thinking of the thermal energy packets (phonons?) as waves in boxes, with the box being the boundary of whatever thing has the energy. Which is still weird for a gas expanding into a vacuum I guess, but works for a hot solid object or a confined gas.
