集里怎么说它

• 《用一个智能体去修智能体:LangChain 如何打造 Engine v2》(00:11起):本集把它说成:LangChain 的『用于智能体工程的智能体』产品,自动扫描 trace、发现问题、写描述、提修复和 eval,今年 5 月发布后已扫描超 7000 万条轨迹、创建超 21000 个问题;本集讲它从 v1 打磨到 v2 的过程。

① 提到它的金句

24 条

如果 Google 是一个针对 ChatGPT 回答的搜索引擎，Google 就没有理由存在了。
If Google’s a search engine for ChatGPT responses, there’s no reason for Google to exist.
—— Ethan Smith · [55:57]

指向原始笔记的链接

在我看来:LLM 是发动机,而我们是车、是道路、是整个系统。
The way I see it: the LLM is the engine, and we’re the car, the roads, and the whole system.
—— Carlos González de Villaumbrosia · [16:37]

指向原始笔记的链接

他们是一个信心引擎。他们不是决策引擎。
They’re a confidence engine. They’re not a decision-making engine.
—— Adam Ward · [28:35]

指向原始笔记的链接

我想事实可能是，今天使用 AI 的人里 99% 只是把它当作一个搜索引擎。
I think the fact is probably 99% of people that use AI today just use it as a search engine.
—— Adam · [52:49]

指向原始笔记的链接

之所以难,是因为我觉得我们只是在两个世界之间搭一座桥,而这并不是最优的,你仍然需要知道该用什么词汇去问,用引擎的词汇、用代码的词汇。
It’s hard because I think we’re just building a bridge between two worlds, and it’s not optimal, and you still need to know kind of like what to ask in the vocabulary of an engine, in the vocabulary of code.
—— Arturo Nunez · [07:04]

指向原始笔记的链接

默认情况下,上下文是在游戏引擎上,而不是在游戏设计这部分,而我认为我们应该把这个理念翻转过来。
By default, the context is on the game engine rather than on the game design part, and I think we should flip that idea.
—— Arturo Nunez · [07:42]

指向原始笔记的链接

所以思考这个问题的一种方式是，你正在往公司里发射这些超级跑车，装着巨大的千匹马力发动机，却没有方向盘，没有刹车，也没有安全带。
So one way to think about it is you’re launching these supercars into your company with a giant, you know, a thousand horsepower engine without any steering wheel, without brakes, without seat belts.
—— Manoj Saxena · [09:48]

指向原始笔记的链接

如果你想一想火箭发动机的连续谱系,我们对打造一台本田思域火箭发动机,比对打造一台布加迪威龙火箭发动机更感兴趣。
If you think about the continuum of rocket engines, we’re much more interested in building a Honda Civic rocket engine than a Bugatti Veyron rocket engine.
—— Baiju Bhatt · [45:14]

指向原始笔记的链接

所以我们说，与搜索引擎在第一天相比，坐在搜索引擎上的人类似乎更容易竞争。
So we said, it seems like humans sitting on search engines are far way easier to compete with than a search engine on day zero.
—— Parag · [08:43]

指向原始笔记的链接

这就是为什么你的智能体需要一个上下文引擎才能完成真正的工作。
And that’s why your agents need a context engine to get real work done.
—— Peter Werry · [06:59]

指向原始笔记的链接

token 减少 50%,分类更快，答案更好。而这正是上下文引擎能做到的。
50% fewer tokens, faster triage, better answers. And that’s exactly what a context engine can do.
—— Peter Werry · [17:45]

指向原始笔记的链接

我喜欢用的一个类比是：如果模型是引擎，那 harness 本质上就是引擎周围的那辆车。
The analogy I like to use is that if the model is the engine, the harness is essentially the car around the engine.
—— Santi Garza · [03:50]

指向原始笔记的链接

为一架 30 座飞机制造喷气发动机的复杂程度，和为一架 70 座飞机制造是一样的，所以成本也相同。
It’s as complex to build a jet engine for a 30-seater as it is for a 70-seater, so it costs the same.
—— Anders Forslund · [02:52]

指向原始笔记的链接

一个是我们会有泛化能力更好的系统，因为每当你增加一个搜索引擎、或向量数据库(它就是一种搜索引擎)、或某种步骤间的路由逻辑，这种人工整理就真的限制了系统做很多不同事情的能力。
One is we will have systems that can generalize better because every time you add a search engine or a vector database, which is a search engine, or some conditional logic to route between the steps, this human curation really limits the ability of that system to do a lot of different things.
—— Alexander Whedon · [30:24]

指向原始笔记的链接

我不认为真的有人会把自己的职业生涯押在，比如说，用 vibe coding 写一个核反应堆或火箭引擎或某种后果非常严重的系统的控制系统上。
I don’t think anyone’s really going to kind of bet their career on like vibe coding a control system for let’s say a nuclear reactor or a rocket engine or something that’s very high consequence.
—— Scott Morton · [02:47]

指向原始笔记的链接

如果对人类最优的搜索引擎对 AI 系统同样最优，那就太疯狂了。
And it would be crazy if the same search engine that was optimal for humans was also optimal for these AI systems.
—— Will Bryk · [09:46]

指向原始笔记的链接

你没法在 VLLM 或 SGLang 上运行这些基于扩散的 LLM。我们必须构建自己的服务引擎，而且我们能处理真实生产工作负载的很多复杂性。
You cannot run these diffusion-based LLMs on VLLM or SGLang. We had to build our own serving engine, and we can handle a lot of the complexity of real production workloads.
—— Stefano Ermon · [13:01]

指向原始笔记的链接

我们实际上就是一个上下文引擎，或者说法务部门的 CRM。
we’re just a context engine or a CRM for the legal department.
—— Nick Fleisher · [11:10]

指向原始笔记的链接

今天思考视频模型的方式，就是把它当作现代渲染引擎。把它想成类似 Unreal Engine 或 Unity，只是输入类型不同。
The way to think about video models today, it’s just modern rendering engine. Think about this as like Unreal Engine or Unity, but just different types of inputs.
—— Alex Mashrabov · [25:16]

指向原始笔记的链接

我把那本质上视为引擎的一种失败状态，如果我到了出现 KV cache 抖动或者东西被刷到磁盘的地步，那说明我把引擎部署得不对。
I think of that as essentially like a failure state of the engine that we, like I’ve deployed my engine improperly if I get to the case where I have KV cache thrashing or like things being flushed to disk.
—— Charles Frye · [45:46]

指向原始笔记的链接

你能对推理引擎做的许多其他优化，都像是寸土必争的百分比游戏，每次百分之五、百分之十，而且你必须为每一个百分点做非常困难的性能工程。
So many other optimizations that you can do on your inference engine are like, it’s like a game of inches of a percent, five percent, 10 percent each time, and you have to do the very difficult performance engineering for each of those percentage points.
—— Charles Frye · [52:31]

指向原始笔记的链接

从那以后，Engine 已经扫描了超过 7000 万条追踪记录，并为今天正在使用它的众多客户创建了超过 21000 个问题。
And since then, Engine has scanned over 70 million traces and has created over 21,000 issues for the many customers that are using it today.
—— Ben · [01:31]

指向原始笔记的链接

但也许这种内部使用最有趣的地方在于，我们实际上是在 Engine 之上运行 Engine。
But maybe the most interesting of this dogfooding is that we actually run Engine on top of Engine.
—— Ben · [05:57]

指向原始笔记的链接

但我们对 Engine 的愿景是：发现问题、修复这些问题、测试那个修复方案，全部自主完成。
But the vision that we have for Engine is the finding of problems, the fixing of those problems, and the testing of that fix all autonomously.
—— Ben · [13:42]

指向原始笔记的链接

② 出现在这些集

1 集

• 《用一个智能体去修智能体:LangChain 如何打造 Engine v2》 — 作为概念

③ 关联

点进去有真内容 —— 本页主要出口

Ben · LangChain · LangSmith · 智能体 · trace · 评估 · 基准测试 · 红队测试