I was primarily using ChatGPT & Perplexity. Then I started calling B.S. on them, and more often than not their replies were "You're right! Sorry about that!" Simply saying "that's wrong" reveals a terrifying amount of "hallucination" by AI. Far, far more answers stated confidently as fact, turned out to be completely made up.

If I want to play with ideas, I chat with AI. If I need facts, I use search.