I’ve been working with ChatGPT 4.0 on a marketing project. Its been going well but its in a pattern of work ok for a week, then start creating a drama with endless loops of incorrect output and hallucinations. I use often new chats and break up the work that way and keep prompts simple.
But the fallout to never-never-land is so disruptive.
I then think its my prompts…questioning my reality…That I’m on a deadline means I’m more put out than advantaged with going back over the work to check. That it keeps occurring and that ChatGPT 4.0 keeps doing sickafant apologies, leads me up the garden path more times than not?
My solution - I’ve supplemented the process with using other AI’s to keep the course right.
Still…
Im having the same problem with both o3 mini-high and chatgpt o4. This is getting out of hand and they must take a look at this asap… Seems like alot of people have this issue so it must be something bigger.