It turns out getting consistently quality results from ChatGPT requires some knowledge of prompt engineering approaches. Some people won't be willing to write

It turns out getting consistently quality results from ChatGPT requires some knowledge of prompt engineering approaches. Some people won't be willing to write a whole paragraph of information in their initial prompt, or go through further iterations, or check for errors using alternative sources, and so they end up disappointed with their results. At least ChatGPT isn't constantly trying feed you advertising content like Google does, and you don't have to wade through pages of SEO garbage, so it's still much better (although if OpenAI tries to monetize ChatGPT by inserting ads, that'll be the end of it I think).