Except that in most practical scenarios, most models make basic mistakes that no human would ever make. In most cases, people have a very low tolerance for errors that humans would not make.