LLM’s are statistically great at inferring things? Pray tell me how often Google’s AI search paragraph, at the top, is correct or useful. Is that statistically great?