Extending the stated objectives — LessWrong