Articles
从工作空间主体、远程资源版本与能力边界出发，讨论连接器进入 Agent 工作环境后，状态如何持续、信任如何建立、动作如何安全发生。
· 18 分钟阅读 · 中文
A product and systems design essay on what changes when connectors enter an Agent workspace: which state persists, how users can verify trust, and when proposed actions may safely become real changes.
· 20 min read · English
从 Ink & Memory 的工作空间设计出发，梳理 AI 如何跟随当前文档、区分 thread / workspace / editor session，并在没有打开文档时自然回到聊天。
· 6 分钟阅读 · 中文
A product design essay about shared attention in an AI writing workspace: current editor state, document switching, visible edit approval, and why the assistant should read the same page as the writer.
· 10 min read · English