LangChain 的智能体大会 Interrupt 又来了，3 月在旧金山办两天，12 月 7 日前买早鸟票能省一笔。
LangChain 宣布其智能体大会 Interrupt 将于 3 月 24 日至 25 日在旧金山举办。早鸟票开放至 12 月 7 日，报名站点为 interrupt.langchain.com。这是该大会第二次举办，聚焦智能体开发。
🌉 Interrupt, The Agent Conference by LangChain, returns to San Francisco March 24 + 25 🎟️ Early Bird tickets available until Dec 7th Info + RSVP interrupt.langchain.com xf 💬 2 🔄 9 ❤️ 34 👀 7675 📊 9 ⚡
LangChain 要开直播课了，教你用 Deep Agents 做智能体记忆管理，还会讲怎么让上下文窗口不膨胀，写 Agent 的可以去听。
LangChain 创始人 Harrison Chase 抛了个好问题：Agent 的组织记忆和 Skills 是不是一回事？评论区吵得挺热闹，做 Agent 的可以看看。
Browser-based agents that can log in without exposing credentials to the model address one of the biggest blockers to deploying agents against everyday business tools. Teams building automation on internal portals or legacy sites can test the beta without building custom browser integrations. It positions LangChain's managed offering against other agent platforms racing to own the browser-use layer.
Scaling up model size does not buy robustness against broken tools, so engineering effort belongs in framework-level error handling and fault-injected evals rather than bigger models. Teams shipping MCP-based agents in production should test against MCP-Injector-style failure mixes and compare their agent stack choices, since the framework mattered more than the model. The fault calibration is derived from real open-source MCP repository issues, making the findings directly reusable for benchmark design.
给做智能体的朋友：这篇测了 MCP 工具挂掉时各家 agent 框架的表现，结论挺反直觉，模型变大不顶用，框架的错误处理才关键。
Ramp 公开了他们自建 Agent 框架的思路，LangChain 创始人亲自转发讨论，做 Agent 基建的人可以对照看看自己的方案。
