But conversations evaporate. Tabs don’t scale. Threads don’t map to how humans manage multi-week commitments.
What’s missing is a persistent workspace layer:
— account-level, not chat-level
— structured but lightweight
— model-aware (so the AI can reference context)
— opt-in (never intrusive)
— supports memory, plans, tasks, notes, drafts, decisions, and coordination
This would transform ChatGPT from an assistant for “this moment” into a partner across time.
Curious if others feel this gap emerging — especially people using LLMs for long-term personal or professional workflows.
ofabioroma•6m ago
That's actually why I built UltraContext (Show HN is live today). It's infrastructure for this exact problem: persistent context with version history, so you can fork, rewind, and track how conversations evolve over time.
Not a ChatGPT plugin (yet), but the API could power something like what you're describing. Would love to chat if you're interested in exploring this.