I feel like the AI has been storing generic answers on a lot of topics and referring back to them instead of taking into account the full context of the conversation.
I rerun the prompt every time this happens to get the kind of answer I’m used to, which takes my custom prompting and the rest of the conversation into account. One dead giveaway is preceding the answer with “I understand…”
It’s a minor annoyance. I’m not really complaining. Just observing.
Anybody else notice this?
I have noticed it as well and specifically noticed it more the past week. The ‘conversation/chat’ even within the chat seems to have a lot of generic responses that are not continuing the conversation and even giving information that was previously stated as not complete.
It seems like the ChatGPT just treats every single prompt almost like something new.
Previously I also noted how within a project that it “appeared” like the assistant had ease of access or knowledge/memory from ongoing chats within the project. It feels like not only is materials from one chat within a project to another chat is apparently “not possible” but even trying to revisit a component within a chat from earlier in that chat seems to be problematic.
at first it was just an annoyance. but now I’m starting to get tired of reloading the response because it takes two clicks