Then why should we use category theory where set theory (with the same visual representation) would do? I.e. when your categories don't have morphisms between different objects, or have at most one morphism (thus being effectively posets)?