On a seriously neglected proof of the non-finite axiomatizability of PA