Here's hoping that the average HN commenter will actually read the paper and realize that the study was performed using GPT-3.5.