It's the new fake EEAT thing. Pretend you have a quote from an expert and just make one up.

Option 2 is code, but I think the SF version is better.

A start...

Yeah! Exactly. I don't think you can say, here LLM, write the regex. But I do think it can act as a partner to help get through it faster.

I know you said no AI but I really think this could be done. Not with AI making the decisions but rather writing code to sort it for you.

If you're open I'd love to take a stab at it. Feel free to DM. But totally get it if you decline.

This wording has been there for a while. It's not referring to ChatGPT showing medical advice, as some X accounts have been claiming but rather your responsibility when using Openai's APIs.

Yeah, it's a bit confusing. I agree.

Ty Wil! Are you using it for your team?

Found them super complicated. Nothing I did worked! I see the potential though.

Is it possible this comes from something in your custom instructions? I had a response that said something like this at the end of one of my chats and it clearly was repeating something in my custom instructions.

Can you replicate in GPT-5?