Pair programming is the work mode in development in which two people collaborate on the same task in split, but frequently reversed, roles (one thinking high level, the other implementing the changes). Compared to solo programming, pair programming may have multiple positive effects, for example, on code quality. Utilizing pairing correctly, therefore, is an important—but non-trivial—aspect of coordination within highly effective development teams. With the emergence of capable generative AI coding assistants in the last few years, pairing between human programmer and an AI coding assistant (sometimes called pAIr programming) has also developed into an established pattern. Since I personally enjoy traditional human/human pair programming but also use and see the benefits of AI coding assistants, I was curious about what the literature has to say on the pros and cons of both work modes. Here is what I found out.