Large language models are consistently wrong and to treat these things like tutors is to devalue our hard-working teachers.