Extensibility Argument For Greater-Than-Human Intelligence — LessWrong