Altruism makes sense for individuals when there is opportunity for payback. A model of this is the
iterated prisoner’s dilemma. Altruism with no payback is irrational for individuals. It can be rational for the group as a whole; that is called “
superrationality”. Superrationality is when you act as if doing your part of a centralized planner that optimizes the group’s total utility (loosely speaking, the “common good”).
Example: In the simple prisoner’s dilemma a rational agent defects; a superrational agent stays silent.