jollybard comments on On decision-prediction fixed points