There's a few things here, outside of my usual complaints when someone says ChatGPT and doesn't say which model (4 is so much better than 3.5 it's really import

There's a few things here, outside of my usual complaints when someone says ChatGPT and doesn't say which model (4 is so much better than 3.5 it's really important). There's a question about whether gpt can be used, which is important because it's possibly a very powerful tool. This may require poking it to tell it it's supposed to ask followup questions, that its information may be incomplete, etc. Then the more important and immediate point in the article to me is people will use this right now to diagnose themselves. They won't be carefully constructing prompts and they'll probably be using 3.5, as that one is free. For good or ill it'll happen more and more. So with a new WebMD, how should doctors and public health messaging deal with this?