I feel like GPT-3.5 has been nerfed in #ChatGPT. It's not like it has been very helpful to begin with, but its current output is straight up rubbish. I spent more time clicking ‘Regenerate answer’ than I would actually spend researching the problem I've had.