No, pair programming does not replace code reviews. That the reviewer(s) of a change is unfamiliar with the code is a *feature*, as that’s how *literally everyone* except for the author(s) will read it later.