THE FATE OF Cyg X-1: AN EMPIRICAL LOWER LIMIT ON BLACK-HOLE–NEUTRON-STAR MERGER RATE