Solomon's reversal on the online safety meeting matters less as a scheduling event than as a signal about what Canada's AI regulatory apparatus will be built to see. A ministry that responds to pressure from parents and child safety advocates — and that raises the prospect of new AI safety regulations after a specific shooting event — is one organized around incident-response, not precautionary technical governance. The alignment researchers who have spent years arguing that frontier model risk deserves institutional attention are watching a regulatory moment close without them.