Rohin Shah comments on What are the best arguments and/​or plans for doing work in “AI policy”?