These are two different things. Make the case for the latter.
Could there be some advantage in having multiple populations with some positive and non-negligible but still not super high amount of mixing, over having a single population which, uh, mixes with itself "uniformly" ?
Like, what if random genetic variation over time in the different groups going in "different directions" but always mixing with the others a bit might allow for more diversity of things being "tried" than in the single group, and maybe this would result in "good ideas" being "tried"/"discovered" more often, and which could then spread to the other groups.
The "good ideas" would of course spread to the other groups more slowly than if it was all in one group, but maybe there could be some level of mixing where an improvement in the rate of "good ideas" being "found" was enough to outweigh the cost of "good ideas" taking longer to spread to different groups, so that all the groups benefit?
HOWEVER: note that I am not advocating making any decisions based on this argument. Even if such a level of mixing/not-mixing exists which would be better than having it all be one group, that doesn't mean that that level of mixing is less mixing than would happen if this argument were not considered, so this argument does not mean decreasing mixing would be good. If the level of (genetic) mixing has an impact at all (I don't know if it does), it seems equally possible that the optimal level is greater than the current level.
(Maybe if the level of mixing were "too low", then the mixing would become less effective? In the extreme (unrealistic) case, if the different groups became almost unable to mix, as a result of mixing too rarely, that might be very bad, because then "good ideas" couldn't spread from group to group?)
And even if some levels of mixing would result in better results (e.g. in health) over time than others, and even if a different level were better than the level that would be the result if no one tried to influence this level, that doesn't mean it would be good to try to influence the level towards that "better" level. Attempting to influence that level might be e.g. dehumanizing or un-virtuous, or bad for other reasons.
I want to make it very clear that I'm just talking about this as an intellectual exercise, and am not advocating any actions based on this argument. These lines of reasoning were ignoring the fact that the groups being considered are people, and really, the fact that these people are /people/ makes things relating to that much more important than any of the other considerations, and when they conflict, the fact that these are /people/ takes priority over any concerns about like, "gene flow" or whatever.
If you are concerned that the ideas in this comment might be repeated in harmful ways I am willing to delete this comment upon request, as long as this website lets me do that.