A 100-to-1 cost asymmetry justifies ninety-nine false alarms to prevent a single catastrophic event. This LessWrong analysis argues against the social stigma of over-warning. Practitioners should prioritize risk mitigation over the fear of appearing alarmist. The logic applies directly to AI safety warnings where the cost of failure is extreme.