Except it's nothing like that. As I've been saying, nothing "new" is being created.
I think you're confusing the MWI interpretation with other interpretations that have wave function collapse. In the MWI, there is no collapse at all. There is only the appearance of collapse. The system is always
in a superposition of many states. It's just that if certain interactions have occurred, then the different elements of the superposition cannot communicate effectively. That is all.
As for it not being legitimate to treat it as a macrostate in the thermodynamic sense, well, that's fine. A macrostate in thermodynamics is the set of macroscopic variables that are needed to fully-describe the macroscopic behavior of the system. Because different components of the wavefunction aren't observable in macroscopic systems, they don't belong in a consideration of thermodynamic macrostates. This doesn't mean that the MWI is wrong, just that the word "macrostate" is a poor word to use to describe the entire ensemble of configurations of the system.
That's not ducking the issue. It's solving the problem of wavefunction collapse in an extraordinarily simple and elegant manner. And given the difficulty in demonstrating precisely how the appearance of collapse arises from the MWI, I would consider a claim that it's "ducking the issue" to be an admission of ignorance of the difficulties involved.
And at small scales you can actually calculate how effective the collapse should be, and therefore compare the prediction of the MWI against observation. So far observations match the predictions.