Thank you for putting my feelings into words. "Really weird" is exactly how I feel. AI output in our work should be measured in engineering progress, not in use metrics, right?