Valdes comments on Latacora might be of interest to some AI Safety organizations