I’ve been using ChatGPT heavily for large engineering and development projects for quite a while now.
The longer a conversation gets, the slower the interface becomes. Not the model itself necessarily, but the actual UI. Scrolling becomes laggy, message editing becomes annoying, loading takes longer, and the overall experience degrades over time.
I understand that rendering huge conversations isn’t a trivial problem.
What I don’t understand is how we’re still having exactly the same discussion years later. Every time the topic comes up, the answer is still some variation of “start a new chat”.
The second thing I don’t understand is why there is still no built-in option to hide or collapse old parts of a conversation.
Right now I’m using a browser extension for something that feels like it should be a basic feature:
-
Collapse old sections of a chat
-
Hide messages I no longer need to see
-
Keep the context available to the model while reducing what the browser has to render
-
Improve navigation in conversations that contain hundreds or thousands of messages
For people doing serious project work, long chats are not an edge case. They are the normal way of working.
I regularly see advice like:
“Start a new chat.”
But that’s not really a solution. That’s a workaround.
If a conversation contains months of project history, decisions, debugging sessions, design choices, documentation reviews, and project context, splitting everything into dozens of separate chats creates its own problems.
At this point I’d honestly rather have:
-
Conversation sections or folders
-
Manual collapsing of message ranges
-
Archived portions of a chat
-
A performance mode that only renders recent messages
than yet another UI redesign.
Am I the only one still frustrated by this in 2026?