> if much higher rates of mutation were doing the virus a favor overall, then it would already be mutating at those rates.

This sounds reasonable, but is it actually true?

Edit: Long tail events are hard to reason about, and successful mutations are very definitely many sigmas out on the curve. And optimisation cost function is hard to think about: short-term survival has different parameters than long-term chance of becoming the dominant variant. “Punctuated” mutation rates might be the ideal nursery, versus a constant mutation rate?