Thanks, I really appreciate that.
One clarification: I framed this around Projects because that is where the issue becomes most obvious for creative users, but the pain point is broader than Projects specifically.
It is really a problem with the overall conversation architecture.
Even outside Projects, long-running threads eventually hit the conversation limit. When that happens, the user has to start a new chat, and the new chat does not reliably inherit even the basic working context from the previous one. For creative writing and worldbuilding, that means losing character relationships, established canon, tone, timeline decisions, terminology, emotional context, and the active direction of the work.
Projects need better cross-thread continuity, absolutely. But regular non-project chats need some kind of continuation path too.
One possible solution would be a “continue in new chat” or “branch with context” function. When a thread hits the limit, ChatGPT could create a new chat that automatically carries over a structured snapshot of the previous thread: key facts, active decisions, unresolved threads, important names, tone/style preferences, and any user-marked context that should persist. That would be much better than forcing users to manually rebuild the context every time.
So the core ask is not only “make Projects better,” though that is definitely part of it. The bigger ask is: give users a reliable way to continue long-running work across conversation limits, whether they are working inside a Project or just in a normal chat.
That would help creative users, but I think it would also help coders, researchers, planners, students, podcasters, and anyone doing complex long-form work with ChatGPT.
Really appreciate you passing this along.