On harrod's model and instability once again