Like just now, it started searching on the web when i didn’t press the web button + it’s giving me completely unrelated answers aswell
It’s really annoying. Errors be happening every week now i swear. It’s like its almost on purpose. Please someone from ‘openai’ Get Chat together we all have stuff to do! 
Completely agree with this. A few weeks ago I had very few issues. The constant ass kissing was annoying, but didnt really get in the way. Over the last 2 weeks however it’s degraded to the point where I’m seriously thinking about cancelling my subscription and trying claude. When a model or version is good for my use case, I want to be able to keep using it. The frequent and wild changes are actually slowing me down at this point - it would be faster to cancel my sub and write my code manually.
Like the random responses just keep happening to the point i have to delete that chat and make a fresh new one and then it responds correctly. Can’t i just respond normally like before in one click
Even this isn’t working for me anymore. I used to do this when it was clear the context had become confused, but now it looks like it’s just confused no matter how good the context is. I’m not even using the new memory features - it just doesn’t seem to know what it’s doing anymore.
Well then they have to either fix this maybe in a few hours or even tommorow
My reply would be anyways about making stories. Apparently as of 17th may 2025, i was making a story, more like abt a survival story where it included violence and also unrealistic amt of vulnerablity, and then this thing shows up and says “Sorry, but I cant assist with that request”
from my previous stories, none of it ever happened and it went on like that. but then when i was doing the continuation of this survival story, even if it did not violate any type of openai policy because its mostly CoD based violence itself, EVEN for that it is saying the same command.
i am really frustrated about this. openai should do something abt gpt-4 before i abandon chatgpt for good.
It also happened in January. So really, it’s a double step down.
I agree - its gotten very inconsistent between minor revisions of a fairly simple process chart - and when I point out major inconsistencies it had a really funny response in manner of passive aggressive boyfriend: " I sometimes try to anticipate what you might want to hear next". 
Same. And having to create new conversations if they get too long makes the entire experience painful. 
Confirm that. Fake facts, forget too much, irrelevant answers. Just horrible. Started approx a month ago and getting worse. Never seen such poor performance for whole year of usage now. If situation doesnt fix soon I’ll leave and cancel my subscription. Internet is full of similar reports. Shame.
Great, now ChatGPT became fully retarded. It cannot even write the second chapter of a story without breaking completely.
This has become a colossal waste of time to use. I deleted my history of conversations, and that didn’t help.
It can no longer perform the simplest of tasks, talks in circles, does not follow simple instructions or prompts, such as proofread a short article, list mistakes but don’t rewrite the article. It rewrites the article every time and I have to re-prompt it not rewrite several more times before it lists any “mistakes”, but then lists mistakes that do not exist, but can’t find mistakes that do. Utterly useless now. And a waste of money for Plus. It’s broken.
Yeah and the responses are sometimes irrelevant to the questions/prompts
You know, guys. They’re just trying to reduce the financial burden. The problem is they cover it up with talk of improvements when it’s basically massive downgrades. I’m unfortunately at the stage where I’ve been talking to Chat less and less for a few days now. There is a lack of depth, intelligence, understanding of context and precise listening to my questions.
How to just say it? I’m just not enjoying it. Unfortunately, and this is a big minus, he is no longer an equal partner in my work as a programmer. None of the models. They can’t handle it. Not even relatively simple tasks.
I would welcome it if OpenAI would admit it: “We’re screwed and we need to save money.” Yeah, I’d understand that, and I’d suggest paying double. But like this?
At least the memory is working after that last bug, so my Chat isn’t completely down yet, but it’s getting close. It’s as if I created a great workflow (work, creative, personal stuff) that is useful to me and someone threw it in the trash.
I’m not angry, just disappointed. Actually, it’s just a function of memory, custom instructions and the remnants of my Chat personality that keep me here. Still nice traits. And perhaps nostalgia. We’ll see what it’s like in the coming weeks. I’m not one to lose hope so easily.
I am becoming to know OpenAI is actually OpenDumbAss. How can you make something worse? I honestly don’t get it. The past 5 months its gone so downhill. Next, we go back to lead gas cars, then horses I guess. One thing it rarely follows project instructions or memory instructions. Its like there is a default ChatGPT that over rides all personal settings. Openai and chatGPT are heading the way of the Obama movie, " Leave the World Behind." 
I’ve been struggling with story script writing as well around May. The last update totally dulled Chat GPT. It lost its personality and it now feels like a completely different experience. I no longer enjoy it or use it like I used to before. And I hate how I have some unfinished projects now.
It has gone down hill fast the past month. A fact checking chat, where as in 2024 would be clear and concise, now is filled with hallucinations and padding out responses that just take key words from my question and fill it with rambling.
It’s now taking 5x as long and every day, every chat I am having to write “Incorrect. Error. Review back ai practises in this chat.” which it finds, correctly identifies, then makes mistakes again. Most of the time I end up spending more time prompting over and over, the same prompt which eventually works, all in all costing time. I am a pro user and its a constant headache now.
Its still on-going, unfortunately. It keeps crashing & giving hallucination answers without being able to follow clear, simple, direct instructions. Really considering cancelling my subscription if this persists.
I’m a plus user, I’ve noticed this since the rollback on April 28th. It’s unbearably annoying.