Getting Feedback by Restricting Content — LessWrong