What doesnt necessarily get selected for within the host is improved survival rates outside the body, ability to penetrate the body in the first place, etc. But bacteria are under the same evolutionary pressures.
Agreed.
There is a bucket of vulnerable ACE-2 receptors in everybody. Once a viral infection occurs in a single person, it undergoes an infection curve within the host bucket, analogous to the infection curve of the general population. The non-linear host cell infection curve can be greatly influenced by early intervention (immunity). It can be expected that some people have a big curve that results in a more complete viral infection than in others, but the number is limited until it can move to a new person, and run a new infection curve. So the mutation can occur in that once person, but it has to replicate in a new host to see the new infection curve of the viral prototype. If one never transmits the prototype, it is moot.
Bacterial resistances like MRSA are most likely when an infection that is able to survive treatment is able to re-establish. With most susceptible bacterial cells having now been eliminated in this case, the new infection is more formidable. There are of course other mutation vectors.