告诉HN:ChatGPT 需要一个持久的工作空间层

4作者: LostBeacon13 天前原帖
许多重度使用ChatGPT的用户已经从“对话”转向了“持续工作”:规划、研究、写作、协调和决策。 但对话是短暂的,标签无法扩展,线程也无法映射人类如何管理多周的承诺。 缺少的是一个持久的工作空间层: — 账户级别,而非聊天级别 — 结构化但轻量 — 模型感知(以便AI可以参考上下文) — 自愿选择(绝不干扰) — 支持记忆、计划、任务、笔记、草稿、决策和协调 这将把ChatGPT从“此刻”的助手转变为跨时间的合作伙伴。 我很好奇其他人是否也感受到这个缺口,尤其是那些将大型语言模型用于长期个人或专业工作流程的人。
查看原文
Many heavy ChatGPT users have shifted from “conversation” into “continuous work”: planning, research, writing, coordination, and decision-making.<p>But conversations evaporate. Tabs don’t scale. Threads don’t map to how humans manage multi-week commitments.<p>What’s missing is a persistent workspace layer:<p>— account-level, not chat-level<p>— structured but lightweight<p>— model-aware (so the AI can reference context)<p>— opt-in (never intrusive)<p>— supports memory, plans, tasks, notes, drafts, decisions, and coordination<p>This would transform ChatGPT from an assistant for “this moment” into a partner across time.<p>Curious if others feel this gap emerging — especially people using LLMs for long-term personal or professional workflows.