Someone please review the lastest update to the model. This is the worst consistent interaction I have experienced in the years of working with GPT.
I earlier created a new conversation and linked to Warp for debugging. It never registered being connected and gave the wrong instructions repeatedly.
It loses context of a new conversation within 1 reply.
It never wants to do research after saving instructions to memory or if instructed within the initial prompt using the browser extension.
Many false statements and only admits after stating that is wrong with my own research and validation. It even gave me attitude the other when I asked it there was a deficiency in one the areas that I was working. Stated it did not have a deficiency and it was something it overlooked if I recall it accurately.
Regardless, there is something strange with this variant. Previous revisions were not this inconsistent.
Also, please update the app version or web browser to allow deleting multiple conversations at once.
1 Like
I came here to express my frustration with the recent changes in ChatGPT. It seems like I’m not the only one who’s feeling this way. Give that you have some similar experience with the beloved AI Agent… It’s been a few months since I last complained, and apparently, when everything goes smoothly, we don’t feel the need to vent.
It’s frustrating because everyone else is talking about the AGI and how ChatGPT is way beyond… but it’s strange how for me ChatGPT is quite behind himself… far from all those advertised advancements in AI. I guess I’m not the only person who thinks the current version feels different from before. I don’t have a specific example because arguing with the AI Agent was just so exhausting, and I don’t think it’s very useful to demonstrate just one instance of a larger problem.
I don’t have the proper technical vocabulary to describe the feeling I have, so I use the term “quantized” to convey it. However, it could be any other word that describes the change.
It seems like the subtle changes that made ChatGPT so magical for a few months have been removed during the 12-day OpenAI experiment. Strangely enough, they did announce an upgrade a few days before the experiment began (probably around the 20th of November 2024). I’m currently using ChatGPT, and it’s quite frustrating because it was working perfectly well for a few weeks (for a few months I was happy and since a few weeks the last update didn’t affect my experience in a negative way).
I believe the people at OpenAI are more knowledgeable about the actual changes and will understand the actual implications of this thing i naively called “quantization”… Hopefully they don’t need a specific example to understand. However, since it was working perfectly fine for a few weeks after the last update and there hasn’t been any mention of further changes, I suspect it might be some kind of tweaking or other minor adjustment that made such a huge difference.
I’m unsure if the original poster, @EugeneBelford, mentioned that the problems with the AI Agent were with the API or ChatGPT itself. However, it seems like I’m not alone in this experience. I hope other members of the community will join us and express their concerns or share their experiences with the AI Agent. I’ve had my Karen moments with ChatGPT, but hey, it probably won’t be resentful. And I didn’t even ask to speak to its manager. Given that the toll-free number is only available for 15 minutes per month (or is it?), I didn’t call to complain either. 
1 Like