Paul Crowley comments on A one-question Turing test for GPT-3