I’ve been using ChatGPT in a long-term, high-context way—building and cross-referencing dozens of conversations tied together by personal, intellectual, and thematic continuity.
Right now, I’m doing everything I can to simulate a cohesive assistant experience:
- Manually summarizing past chats
- Re-uploading context into new threads
- Organizing everything inside Projects
- Using pinned instructions and memory settings creatively
But the per-chat memory model and siloed structure make this incredibly labor-intensive. What I really need is the ability to:
- Designate a “core chat” whose memory or tone other chats inherit
- Or opt into shared memory across all chats inside a Project
- Or at minimum, link a past chat into a new one and have ChatGPT actually access it in continuity
I understand OpenAI is building something extraordinary, and I respect that end-user structure will sometimes take a back seat to the massive complexity of creating the AI itself. But users who are trying to build something integrated—whether that’s a long-term research assistant, personal guide, or reflective partner—need more than isolated snapshots.
This kind of continuity wouldn’t just help power users—it would unlock new kinds of trust, depth, and creativity for anyone doing serious work with ChatGPT.
(Side note: This summary was written by ChatGPT, with edits from me. I think the team might appreciate that.)
Thanks for reading—and for building something worth investing this much effort into.