I assumed that the effect of the intervention is "small enough relative to the world" for your population ethics to be smooth. For average utilitarianism in particular, this corresponds to the percentage change in population being small. In your scenario, this isn't true, since there are only 2 people, but in the real world it holds up very well. Just 1% of the world population is 70 million people, and virtually no intervention (except for things like existential risk reduction) could cause such a large population change.