I hope that eventually you would be able to swap who is talking, but that sort of thing seems like it could be quite tricky to implement with many possible edge cases which could be hard to account for.