Claude Agent 平台的内部秘密：来自亲历者的深度拆解

A year from now, where do you think the platform will be? 一年后，你觉得这个平台会发展到什么程度？ We'd want to experiment with directions where Claude actually gets so good at understanding itself, it figures out what model you should be using, it figures out how to spin up all the sub agents. 我们希望探索一些方向——让 Claude 真正深入理解自身，能自动判断该用哪个模型，自动启动所有子 agent。 We don't have to think so much about what kind of architectures are there because Claude is actually able to understand itself enough that it can write itself on the fly. 我们不必再费心考虑用哪种架构，因为 Claude 已经足够了解自身，能在运行时即时编写适合自己的框架。 In that world, if Claude is on the fly or agents on the fly are becoming what they need to become in order for you to do what you're trying to do, the platform has to seriously scale. 在那样的世界里，如果 Claude 或 agent 能即时演变成完成你目标所需的样子，那平台本身就必须大规模扩展。 How close are we to Claude making a billion dollars? Claude 距离为我们赚到十亿美元还有多远？ This is this is really what I'm asking. 这才是我真正想问的。 Angela, Caitlin, welcome to the show. Angela、Caitlin，欢迎。 Thanks for having us. 感谢邀请。 Yeah, thank you. 是的，谢谢你们。 So, for people who don't know, you both work on the platform at Anthropic. 好，对于不了解的听众，你们两位都在 Anthropic 做平台工作。 So, Angela, you're the head of product for the Claude platform and Caitlin, you are the head of engineering for the Claude platform. Angela，你是 Claude 平台的产品负责人；Caitlin，你是 Claude 平台的工程负责人。 Um I'm I'm really excited to talk to you because A, you've been launching a bunch of stuff. 嗯，我真的很期待和你们聊，因为一方面，你们最近发布了很多东西。 You have Claude managed agents that came out recently, you've been launching new features for it. 你们推出了 Claude Managed Agents，而且一直在发布新功能。 And I think that it it comes at this really interesting time where it makes me think about what actually is a platform in AI for a model company because in the GPT-3 days, the platform was a completion endpoint. 我认为这个时机非常有趣，让我开始思考：对一家模型公司来说，AI 平台到底意味着什么？因为在 GPT-3 时代，平台就是一个补全接口。 You just like send a prompt to get a response. 发个 prompt，得到回复。 Um after that, it's like a completion endpoint with tool calling and a couple and like chat sessions, like that kind of stuff. 之后变成了带工具调用的补全接口，加上几个聊天会话之类的东西。 And now it like with Claude managed agents, you're essentially getting a Claude on a computer um with memory and all this other stuff. 现在有了 Claude Managed Agents，你基本上得到一个在云端运行的 Claude，带记忆等能力。 Um so, I'm just trying to I'd love to help I'd love for you to help me unpack that trajectory and like what it means to build a platform in AI. 所以，我很想请你们帮我梳理这条演进轨迹，聊聊在 AI 领域构建一个平台究竟意味着什么。 Yeah, um I think like, you know, your your characterization is like very accurate. 是的，嗯，我觉得你的描述非常准确。 I think like as we've kind of like as a lot of these kind of like technologies have evolved with the LM like for starting and then I think like uh putting that behind an API was very fun. 我认为，随着这些技术不断演进，从语言模型起步，最初把它放到 API 后面其实是件很有意思的事。 A lot of people were like, "Wow, I could like do some at the time." 当时很多人都在想：哇，我能用它做点什么。 I think it was very cool. 我觉得那真的很酷。 Now we'll probably look back at it and be like, "Oh, that was like really basic." 现在回头看，可能会觉得那时候其实挺基础的。 Um and then, you know, I think like we've moved more and more towards like a slightly more like stateful world as you kind of like want to persist the kind of like sessions, state um to be able to make sure that the kind of performance of the model is like better and better. 嗯，之后，随着大家越来越需要持久化会话状态，我们也逐渐向更有状态的世界迈进，以确保模型性能越来越好。 I think that that's probably like actually the the through line. 我觉得那大概是真正的主线。 Like as a lot of these kind of like um as we make improvements to Claude and as it continues to get better and like more autonomous, we find ourselves like basically needing to kind of like evolve the platform to be sort of like higher and higher order abstraction, but it's in the pursuit of like helping you get the best outcomes out of something. 随着 Claude 不断改进、越来越自主，我们发现自己需要不断演化平台，提供越来越高层次的抽象——而这一切都是为了帮你从中获得最好的结果。 Like I think in the very beginning, you know, we were very like everyone was very exploratory. 我觉得最开始的时候，大家都在非常探索式地前行。 It's like you have no idea what people are going to build with these LLMs, and you wanted to kind of have as much possibility out there as um as available. 没人知道人们会用这些 LLM 构建什么，所以你希望保留尽可能多的可能性。 And then as those use cases started kind of narrow down, like people started building products with it. 然后随着使用场景逐渐收窄，人们开始用它来构建产品。 People started now like building agents with it. 现在人们用它构建 agent 了。 And more and more of that is about, you know, like customers coming to us and being like, "How do I get the best out of Claude? 越来越多的问题是，你知道，客户来找我们问：我怎么才能最大限度地发挥 Claude 的效果？ Um how do I like set up my tools? 嗯，我怎么设置我的工具？ How do I run the loop?" 怎么跑这个循环？ And so on and so forth. 诸如此类。 And you have some people who are like really really experimenting, and they're on the edges, and that's great. 有些人真的在非常深入地探索，走在前沿，这很好。 And then you have like just a whole host of other folks that are coming in who are like, "I kind of want a lot of this stuff like out of the box." 然后还有大量其他人进来，他们想要很多东西开箱即用。 Um and in our pursuit for get make making sure that like Claude is basically producing the best outcomes, we find ourselves like enriching the platform to be richer and richer and richer, and that's in uh you know, contained in that is like both the state, it's like the tools that you start to see us adding. 嗯，在我们追求让 Claude 产生最佳结果的过程中，我们发现自己需要不断丰富平台，让它越来越完善——这包括状态管理、各种工具，以及很多云端基础设施组件。 Um it contains a lot of kind of like almost like sort of the cloud components of a lot of these types of things. 嗯，它还包含很多接近于云组件的东西。 Um but it's in pursuit of the same mission of like just making things literally as easy as possible. 嗯，但这一切都服务于同一个目标：让事情变得尽可能简单。 And I think in probably, you know, the forward state of a lot of these things in terms of maybe the philosophy of what a platform ultimately ends up doing, it probably ends up just being like whatever it's like the set of primitives and infrastructure that enables you to basically get the outcome as fast as possible with actually as little of work as possible. 我想，从长远来看，说到平台最终会做什么，答案可能就是：一套能让你以最少的工作量、最快的速度达成目标的基础原语和基础设施。 And I think that that tends to follow a certain form factor at least in this current stage. 我认为在目前这个阶段，这往往会形成某种特定的形态。 But yeah. 不过是的。 How would you characterize like what the primitives are today? 你怎么描述今天的基础原语是什么？ So maybe that's just asking what are the primitives in Claude managed agents? 或者更直接地说，Claude Managed Agents 里的基础原语有哪些？ Yeah, so Claude managed agents is built on all of our same primitives that you could otherwise build on directly. 是的，Claude Managed Agents 是建立在我们所有相同的基础原语之上的，这些原语你完全可以直接使用。 So the messages API. 就是 Messages API。 And within the messages API we've built a whole bunch of I guess maybe innovations around the API. 在 Messages API 之上，我们围绕这个 API 构建了很多创新。 Like you could just get tokens in and out if you really wanted to, but you know, you can use some of our built-in tools. 当然，如果你真的想，可以直接进出 token，但你也可以使用我们内置的工具。 You can build you can use stuff like code execution, spawn a sandbox and execute work. 你可以使用代码执行功能，启动一个沙箱来执行任务。 You can use I guess like you know, web search and all these sorts of different things. 你可以使用网络搜索等各种不同的功能。 And so I think we've taken what we see as all the most powerful of those things and put them together into a harness and a set of infrastructure that is you know, just the way to get what we think is the best outcomes out of Claude. 所以我认为我们把我们认为最强大的那些东西整合进了一个运行框架和一套基础设施，这套组合能让 Claude 发挥出最佳效果。 So I'm sitting here feeling this sense of I've been thinking of it as like time deflation. 所以我坐在这里，有一种感觉——我一直在把它想象成时间通缩。 Like my time gets more valuable in the future as opposed to the opposite. 就像未来我的时间变得越来越有价值，而不是相反。 What whatever the the opposite would be. 不管相反是什么。 My time gets less valuable in the future. 未来我的时间变得越来越没有价值。 Um 嗯 Uh and and and the reason is because we're 啊，而且……原因是因为我们—— So for example, internally for us we're building an agent. 比如我们内部正在做一个 agent。 We're building some agent products where it's like agents that do specific things for us internally and then hopefully for customers. 我们在构建一些 agent 产品，让 agent 帮我们在内部做具体的事，然后希望也能为客户服务。 And in order to do that we've like you know, we have a couple Mac minis with you know, Claude running in a loop on the Mac mini, right? 为此我们有几台 Mac mini，在上面跑着 Claude 循环。 And a lot of that and and it's like a thousand line Python file or whatever. 很多内容……就是一千行的 Python 文件之类的。 And a lot of that mirrors what you guys are building in Claude managed agents. 其中很多其实正好对应你们在 Claude Managed Agents 里构建的东西。 And so for for for me and I think for a lot of people building on Cloud or on the Cloud platform or ecosystem, there's there's at least I feel this, maybe we should just wait for you guys to build it. 所以对我、也对很多在 Claude 平台生态系统上构建的人来说，有一种感觉——或许干脆等你们把它做好再说。 Um, but then I don't know what the lines are and uh and I yeah, I'm sort of wondering if if I want to build an agent, like what is the best path to do that in a way that aligns with what you guys are doing? 嗯，但我也不清楚边界在哪里，所以我想知道：如果我想构建一个 agent，什么才是最合适的路径，能和你们的方向保持一致？ Yeah, um I think, you know, this is this part of the the kind of platform business is actually somewhat similar to any other form of the platform business where you do have customers like like yourself who are building and, you know, you're kind of thinking, should I go ahead and do it cuz 是的，嗯，我认为这块平台业务其实和其他任何平台业务很像——你确实会有像你们这样自己在构建的客户，他们会想：我是现在先动手呢， maybe I have this like immediate need, but at the same time I don't kind of want to like, you know, repeat the work per se and you could have just when you could have just gotten it in there for free um out of the platform. 还是因为我有某个紧急需求，但又不想重复造轮子——毕竟平台可能已经免费提供了。 Um 嗯 And also infrastructure sucks. 而且基础设施真的很头疼。 It's so it sucks so much to like spin up servers. 搭服务器真的太烦了。 I can't believe you do that all the time. 我不敢相信你们一直在做这个。 [laughter] [笑声] But I will actually say um part of why we ended up building Cloud Manager agents was because Anthropic ourselves had gone through enough of these iterations where we built products that were agents that you could run autonomously in the cloud and we did that stand up the infrastructure so that it works well sort of work enough times that um we ourselves were like, okay, we're done building this for ourselves. 但我必须说，我们最终开发 Claude Managed Agents，部分原因正是 Anthropic 自己经历了足够多的迭代——我们构建过能在云端自主运行的 agent 产品，也搭过这套基础设施让它运转，反复做了足够多次之后，我们自己说：好了，不再为自己重复搭了。 We're we're doing it once in a way that's going to really work from everything that we've learned, but also for all the people who are doing it, like you can run whatever you're running on a couple of Mac Minis maybe, right? 我们要把从这些经验中学到的东西，用一种真正可靠的方式一次性做好——同时也为所有像你们这样的人服务，你们也许可以在几台 Mac mini 上跑你们的东西，对吧？ And for a lot of people that could work, but I think if you're building agents into your product and you're running something really at scale, right? 对很多人来说这或许够用，但如果你把 agent 嵌入产品并真正大规模运行， Like that's where it really starts to become more and more challenging to get that infrastructure, right? 那基础设施就会变得越来越难做对。 That's really interesting. 这很有意思。 Yeah, and then maybe to answer the other part of your question, I think we have like two pieces of the philosophy here. 是的，还有另一部分回答你的问题，我觉得我们有两个设计理念。 One is is a bit in the way that we kind of design managed agents, which is that we try to have it be modular enough. 一个体现在我们设计 Managed Agents 的方式上——我们尽量让它足够模块化。 Like we want to be opinionated about some pieces that we feel like should be, you know, very well like married to the Claude model. 对某些我们认为应该和 Claude 模型紧密结合的部分，我们会有自己的主张。 Um but then we like often times like the way we want, for example, we want Claude to like very specifically use like file systems. 嗯，但在其他方面，我们想要的……比如我们希望 Claude 非常具体地使用文件系统。 Um that's like a very particular like Claude kind of style. 嗯，这是一种很特别的 Claude 风格。 or just file systems in general? 还是说只是文件系统本身？ systems in general. 文件系统本身。 We 我们 We also really want to lean into skills. 我们也真的很想发展 skills。 I know like a lot of folks like skills, but like that's something that we like we want to have our hearts be really opinionated about that. 我知道很多人都喜欢 skills，但这也是我们非常有主张的地方。 And so we're kind of particular about like those kind of primitives being the case. 所以我们对这些基础原语是有特定想法的。 So like use the file systems, use the skills. 比如用文件系统，用 skills。 They're really basic. 这些很基础。 Um but at the same time like we still find people who are like still trying other methodologies to go do that. 嗯，但同时我们还是看到有人在尝试其他方法。 And we want to kind of like help you, you know, when you build to start uh just kind of starting the best foot. 我们希望在你开始构建时，能帮你用最好的方式迈出第一步。 Um so that's one piece uh on some of the kind of more opinionated ones. 嗯，这是在一些比较有主张的方面。 But as each one of these kind of like, you know, endpoints or or APIs that we have as part of the suite, we try to like open them up a little bit um in certain areas. 但对我们套件中每个接口或 API，我们都尽量在某些地方留出开放性。 So there's like things that you know, we're looking uh kind of forward to and being like, you know, from maybe it's not available today, but in our design um we are trying to make it flexible enough for people to kind of like add in different pieces cuz we recognize that this API or suite of APIs is not necessarily going to solve like maybe everything in its original construct, and there're going to be pieces that need to kind of open up. 有些东西……可能今天还没有，但在我们的设计里，我们试图让它足够灵活，让人们能加入不同的部分——因为我们知道这套 API 在最初的形态下未必能解决所有问题，有些地方需要开放。 Um and then the second bit is like, you know, we're we're kind of public about this is like when we do design a lot of these things, we do put out like blog posts and sort of like reference implementations. 嗯，第二点是，这方面我们已经有些公开表态：在设计这些东西时，我们会发布博客文章和参考实现。 So if you did want to kind of at least be inspired by that construct, but still maybe make your own on the messages API, um you can definitely do that. 所以如果你想参考这套架构，但仍然想在 Messages API 上自己做，也完全可以。 I think that's to the to the point you just made, that's something that's that's coming up for us. 我认为这正说到你刚才提到的点，这也是我们正在面对的问题。 Again, we have, you know, Claude's running on a Mac mini with a Python file and a couple other like uh you know, bigger more serious implementations on like, you know, cloud infrastructure that we're trying to figure out what to do with. 我们有 Claude 在 Mac mini 上跑，有个 Python 文件，还有几个在云端基础设施上更正式的实现，我们还在摸索怎么处理。 And I think I I told the team that we were that we were talking today, and I think one of the uh one of the questions that they have or or one of the feelings of consternation that they have considering using Claude managed agents for this kind of thing for spinning up agents for our customers is just right now it's a like we have a playground. 我告诉团队我们今天要谈这些，他们有个疑虑或者说顾虑——在考虑用 Claude Managed Agents 来为客户启动 agent 时，感觉现在就像我们有个游乐场。 We have we just have like a little we have a server or Mac mini. 我们就有……一台服务器或者 Mac mini。