You can get it to put out something a little more grimdark but you need to give it more context.
You can get it to put out something a little more grimdark but you need to give it more context.
I suspect that this relentless love of upbeat endings is a side effect of the RLHF training used to create the "assistant" persona. They wanted a friendly and harmless chatbot. And apparently it "likes" exactly the kind of stories you'd expect.
Our jobs are safe after all! The junior dev helped by ChatGPT will not take your job. ChatGPT cheerfully agreed with the junior when he didn't understand something and thought to point out how this can't be right and ChatGPT gave him a "fix" that brought down Production.
I always thought this a Hollywood thing. “Wrap up the loose ends and make it a happy ending.” Movies and series that end untidily are often my favourites.