CaveJohnson comments on AI cooperation is already studied in academia as “program equilibrium”