How does a Living Being solve the problem of Subsystem Alignment? — LessWrong