Rohin Shah comments on [AN #85]: The normative questions we should be asking for AI alignment, and a surprisingly good chatbot