If many qualitatively different branches are consequentialistically optimizing a common goal, all these branches become better according to that goal, even if the specific situations and actions taken in them are significantly different. On the other hand, if these branches respond to your argument and abandon their optimization due to rarity of their particular situations and possible actions, all the branches would remain poorly optimized.
(More generally, an optimizer doesn’t care about the scope of their influence, only about comparison of available choices, however insignificant.)
If many qualitatively different branches are consequentialistically optimizing a common goal, all these branches become better according to that goal, even if the specific situations and actions taken in them are significantly different. On the other hand, if these branches respond to your argument and abandon their optimization due to rarity of their particular situations and possible actions, all the branches would remain poorly optimized.
(More generally, an optimizer doesn’t care about the scope of their influence, only about comparison of available choices, however insignificant.)