NunoSempere comments on Why GiveWell should use complete uncertainty quantification