The problem of killing bacteria, as I see it, will increasingly become something akin to the demarcation problem of philosophy. Why would it not be? It is natural that the optimal situation for pathogenic bacteria to be in is one in which they are indistinguishable from necessary cells. It may take an extraordinarily long time, but until we are able to completely eradicate a pathogenic bacteria from the earth, we will remain in that arms race.
Humans are creative at killing bacteria, though it's already not difficult to "annihilate them at will". It's much harder to annihilate only the ones we don't like, only the ones that harm us, and only the ones that pose the greatest risk.
I feel a more likely future scenario is one in which we create cell-sized robotics that we can use for very narrowly targeted operations in the human body, like treating bacteria. As the bacteria mutates, so does the targeting mechanism in the software that programs the tiny robots.
But as I said, they are ultimately bound by the laws of nature. Evolution can only happen so quickly; they can only evolve their genome so much in a given period of time.
What you're suggesting is they'll evolve to look like human cells. But that doesn't make sense. They would need to evolve that way in a single generation. Something that's "almost like" a human cell would just be wiped out.
Just consider how long it took for the last common ancestor (of humans and bacteria) to evolve into a single celled organism.
Humans are creative at killing bacteria, though it's already not difficult to "annihilate them at will". It's much harder to annihilate only the ones we don't like, only the ones that harm us, and only the ones that pose the greatest risk.