Adam Shai comments on The Geometry of Feelings and Nonsense in Large Language Models