Thomas Hirschmann's Highlights on 'Why large language models aren’t headed toward humanlike understanding' | Glasp