Your post seems to start out with the wrong assumption that human minds don’t (to good approximation) have the wrapper structure. See here for counter: https://www.lesswrong.com/posts/dKTh9Td3KaJ8QW6gw/why-assume-agis-will-optimize-for-fixed-goals?commentId=8E5DHLzXkdBBERx5B#8E5DHLzXkdBBERx5B
Your post seems to start out with the wrong assumption that human minds don’t (to good approximation) have the wrapper structure. See here for counter: https://www.lesswrong.com/posts/dKTh9Td3KaJ8QW6gw/why-assume-agis-will-optimize-for-fixed-goals?commentId=8E5DHLzXkdBBERx5B#8E5DHLzXkdBBERx5B