Translating CFAR to Therapy — LessWrong