On hiding the source of knowledge — LessWrong