No, language models hallucinate because they’re language models.

→ View original post on X — @pmddomingos