Open access typically means authors pay a publication fee, which leads to the same result of the government paying twice and the journal profiting twice.

And most of those require ridiculous "article processing charges". Even non-profits. Elsevier is bad, but it's not much worse than other publishers.

Author (in practice author institution, in practice with public funds) pays open access is less bad than locking articles behind paywalls, but it's still a racket.

This CERN system is about diamond open access, meaning that neither authors nor readers pay.