A post was merged into an existing topic: Tradeoff between generality and optimallity in regards to AI alignment issues