I think as a rule, we should use epistemic humility and look to past instances of intervention unless we have really strong reasons to believe that this would be a uniquely successful intervention.