Large language models are unconcerned with truth because they have no concept of it — and therein lies the danger ...