Internal Inconsistency of Unrestrained Optimization: Why Dismissing Restraint Can Contradict Long-Horizon Optimization Logic | Synapse