Summary: Surreal Decisions — LessWrong