Large language models are not ready for medical advice

https://archivalia.hypotheses.org/184258