Claude Code：意外引发的软件开发革命

好的，我很兴奋地介绍下一位演讲者。请举手，谁在用 Claude Code？好，再举手，谁有 Claude Code 综合征？来吧，没事的。我的团队亲切地说我有 Claude Code 综合征，这可能是真的也可能不是。今天我们很高兴请到 Boris Cherny。Boris 是 Claude Code 的创造者、之父。在这个过程中，他正好亲历了现代软件开发方式的重新定义。我们非常感谢 Boris 今天抽出时间与我们交流。我们知道整个软件开发的重担基本都压在你肩上。所以，感谢你抽出时间来参加我们的活动。采访 Boris 的是我们团队的 Lauren Reader。谢谢。

Okay, I'm excited to introduce our next speaker. Show of hands, who here uses Claude Code? Okay, show of hands, who here has Claude Code psychosis? Come on guys, it's okay. It's okay. My team lovingly says I have Claude Code psychosis, which may or may not be true. We are delighted to have Boris Cherny with us today. Boris is the creator, the father of Claude Code. In the process of doing that, he has just had a front row seat to reinventing the modern way of software development. We're really grateful to you, Boris, for taking the time to speak with us today. We know that the entirety of software development kind of rests on your shoulders. So, thank you for taking time out of your day to be with us. And interviewing Boris is Lauren Reader from our team. Thank you.

你抢了我的开场白，Asia。我们问谁在用 Claude Code，很多人举手。太棒了。谢谢你加入我们，Boris。你能来非常特别。在座的都是创造者，我认为你正在彻底改变创造本身。所以我非常好奇想探讨你对软件、编程的未来以及我们应该把空闲时间花在什么上的看法。我再给你一点背景信息，让大家有更多了解。除了创造 Claude Code，Boris 是一位工程师中的工程师。你整个职业生涯写了大量代码，还编写了关于代码的教科书，包括《Programming TypeScript》。我记得上次我们聊天时，你说过去一年没写过一行代码，至少 2026 年至今是这样，这变化真大。还有一件少有人知的事：中学时，我写过一份关于在 TI-83 Plus 计算器上编写 BASIC 的指南。我搜了一下，它居然还在网上。非常尴尬，所以请别搜。但它确实存在。我们一定会找到它的。那么，我们先问几个问题。也许从 Claude Code 的历史开始，你是怎么开始的，然后我们会留大量时间给观众问答。所以，请在脑子里开始想问题，很快会把时间交给你们。是的。还有，快速问一下，用 Claude Code 的人，你们主要用 CLI 吗？好，大部分是 CLI？好，很多。大部分是桌面版？好。大部分是 VS Code 或 JetBrains IDE？好，其实不多。好。其他？我最近主要用 iOS。是的。

You took my opening line, Asia. We asked who here uses Claude Code. There's a lot of hands. That's awesome. Thank you for joining us, Boris. It's very special to have you here. As a roomful of builders, I think you are changing building entirely. So, I'm very curious to explore how you think about the future of software, coding, and what we should spend all of our free time on. I'll give you a tiny bit more background on you so that everyone has a little more context. Beyond creating Claude Code, Boris is very much an engineer's engineer. You were writing a lot of code through your whole career, writing textbooks about code, including Programming TypeScript. I think last time we chatted you hadn't written a single line of code in the last year, or at least so far in 2026, which is quite the change. There's also a little known thing: back in middle school, I wrote a guide about writing BASIC for TI-83 Plus calculators. I searched for it, it's actually still on the internet. It's extremely embarrassing, so please don't search it. But it exists. We will definitely be finding that. So, we're going to start with a few questions here. Maybe we'll start with a little bit of the history of Claude Code, how you started it, and then we're going to have a lot of audience Q&A for this one. So, start thinking about your questions in the back of your head, and would love to turn it over to you all soon. Yeah. Also, real quick, for people that use Claude Code, do people use the CLI mostly? Like okay, majority CLI? Okay. That's a lot. Majority desktop? Okay. Majority VS Code or JetBrains IDE? Okay. That's actually not a lot. Okay. Other? I'm like iOS mostly these days. Yeah.

好的，酷。是的，我以很多方式意外地开始了 Claude Code。我在 2024 年底加入了一个团队，那是 Anthropic 内部的一个孵化器，叫 Anthropic Labs。这个团队完成了它的使命。我们创建了 Claude Code、MCP 和桌面应用。团队只有几个人，很像一个创新团队。我们构建了我们想构建的东西，然后团队解散了。现在团队又重组了，进行第二轮。Mike Krieger，Anthropic 的首席产品官，曾是 Instagram 的联合创始人，现在领导这个团队。我开始做编程工作的原因是，我们觉得存在一个产品悬置。我猜在座很多人常用这个词。我们在实验室里也常用这个词。这个想法是，模型可以做所有产品尚未捕捉到的事情。在 2024 年底，当我们审视编程时，我们做编程的方式，当时最先进的是预输入。你打开 IDE，按 Tab，一次只能补全一行。这是 Sonnet 3.5 首次实现的功能。但我们感觉实际上可以走得更远。模型几乎准备好迈出下一步了。所以我们不再需要预输入，我们可以让智能体编写所有代码。于是我构建了它，但前 6 个月它真的不行。不是很好，几乎不可用。我写它的时候，可能只用了它来写我 10% 的代码。即使我们最初发布了 Claude Code，它也不是一个热门产品。很多人用了，但它没有像今天这样的指数级增长。这种增长是从 5 月份的 Opus 4 开始的。我记得很清楚。那是指数级增长开始的时候，然后随着每个模型发布而加速。从 Opus 4 开始，然后是 4.5、4.6，现在是 4.7。它一直在加速。本质上，我们试图构建一个处于产品市场匹配前的东西，我们知道它在 6 个月内不会有产品市场匹配，因为我们在为下一个模型构建。这几乎一直是我们的想法。对于 Anthropic 整体来说，我们一直非常专注。我们一直关心商业、企业、安全和编程。这始终是我们想要构建的方式。所以在某个时候，我们知道我们想构建一个产品，但不确定具体是什么。所以这最终成了产品赌注。

Okay. Cool. Yeah, so I started Claude Code kind of accidentally in a lot of ways. I joined this team back in late 2024. It was a sort of incubator within Anthropic called Anthropic Labs. The team kind of served its purpose. We created Claude Code, MCP, and the desktop app. It was just a few of us, very much like an innovation team. We built the thing that we wanted to build, we disbanded the team. Now the team's actually back together for round two. Mike Krieger, who is the chief product officer at Anthropic and used to be one of the founders at Instagram, is leading that right now. The reason I started to work on coding is we felt like there was this product overhang. I'm guessing people here use that word a lot. We definitely use this word a lot within the lab. There's this idea that the model can do all the stuff that no product has yet captured. In late 2024, when we were looking at coding, the way that we did coding, the state of the art at the time was type ahead. It was you open your IDE and you press tab and you can complete like one line at a time. That was the thing that Sonnet 3.5 enabled for the first time. But the feeling was we could actually go a lot further than that. The model was almost ready for the next big step. So, we don't have to do type ahead anymore, we can just have the agent write all of the code. So, I built it, and it just really didn't work for the first 6 months. It was not very good, barely usable. I wrote it from I used it for maybe 10% of my code or something like that. Even after we released Claude Code initially, it was not a hit. A lot of people used it, but it did not have this exponential growth that it has today. That started with Opus 4 in May. I remember that very clearly. That's when the exponential growth started, and then it kind of inflected with every model release. It started with Opus 4, then 4.5, then 4.6, now 4.7. It just keeps inflecting. Essentially, we were trying to build this thing that was like pre-PMF, and we knew that it wouldn't have PMF for 6 months because we were building for the next model. That was the idea pretty much the whole time. For Anthropic in general, we've always just been very focused. We've always cared about business and enterprise and safety and coding. That's just always been the way that we wanted to build. So, at some point we kind of knew that we wanted to build a product. We didn't know exactly what we wanted. So, this kind of ended up being the product bet.

这是一个不可思议的故事，尤其是它是个意外。你曾公开说过，你认为编程已经解决了。如果这是 Anthropic 三大最佳成果之一，你能告诉我们更多你的意思吗？还有什么可能尚未解决，或者可能出现的二阶问题？好的。我可以再问大家一个问题。谁 100% 手写代码？谁 100% 使用像 Claude Code 这样的智能体写代码？好的。谁介于两者之间？好的。所以，大概解决了 50%。

It's an incredible story, especially that it was an accident. So, you've said on the record that you think coding is solved. If this is one of the three best from Anthropic, can you tell us more about what you mean by that, and what might still not be solved, or what second-order problems might come? All right. I can ask another question for the room. Who writes 100% of their code by hand? Who writes 100% of their code using an agent like Claude Code? Okay. Who's like somewhere in between? Okay. So, like 50% solved.

对我来说，是 100%。Claude Code 的代码库泄露了，所以人们知道。它很简单，就是 TypeScript 和 React。没有大秘密，没什么复杂的。我们选择 TypeScript 和 React 的原因是它们对模型来说分布很广。所以当我们开始构建代码库时，模型没有现在这么智能，所以语言和框架很重要。如今，它可以写任何东西，可以学习没见过的语言和框架。但那时，你想用一些分布很广的东西。正因如此，我认为很早我们就达到了模型 100% 写代码的程度。对我们来说，这发生在去年 10 月或 11 月左右。所以对我来说今天，模型写了我 100% 的代码。我大概每天提交几十个 PR。

I mean, for me it's 100%. The Claude Code code base, it leaked, so people know. It's pretty simple. It's just TypeScript and React. There's no big secret. There's nothing really complicated. The reason we picked TypeScript and React is it's very on distribution for the model. So, when we started building the code base, the model was not as intelligent as it is today, so the language and the framework mattered a lot. Nowadays, it can write whatever, and it can pick up new languages, new frameworks it hasn't seen. But back then, you wanted to use something pretty on distribution. Because of that, I think fairly early we got to the point where the model just wrote 100% of the code. For us, this happened sometime in October, November last year. So, for me today, the model writes 100% of my code. I write somewhere, usually a few dozen PRs every day.

上周有一天我做了大概 150 个 PR，那是个纪录。我就是想试试看能推到多远。但对我来说，这已经解决了。不过并非处处如此。有些非常庞大复杂的代码库，还有一些模型还不擅长的奇怪语言。在座各位都知道，它正在进步。通常答案就是等下一个模型。

Uh, there was a day last week I did like 150 PRs in a day. That was a record. I was just trying to push to see how far I can get it. But yeah, for me it's just solved. But this is not the case everywhere. There are very big complicated code bases, kind of weird languages the model's not good at yet. And as everyone here knows, it's getting there. Usually the answer is just wait for the next model.

能跟我们讲讲你的个人设置吗？那天你给我们演示过，挺疯狂的。

Can you actually tell us about your personal setup? You walked us through it the other day. It is pretty wild.

是的。大概 6 个月前我在 Twitter 上分享过我的个人设置。有趣的是，我当时没意识到这会让任何人惊讶。那只是我编码的方式。后来它变了。现在我的大部分工作都是在手机上完成的。我有 Claude 应用，打开后左边有一个小小的代码标签页，我就在那里运行着一堆会话。

Yeah. So, I shared my personal setup like 6 months ago on Twitter. It's funny, I didn't realize it would be surprising for anyone. That was just the way I coded. And it's changed since then. Now actually most of my work I do from my phone. I have the Claude app, and if you open it, on the left-hand side there's this little code tab, and I just have a bunch of sessions going.

多少个会话？

How many sessions?

通常大概有 5 到 10 个会话。然后每个会话通常有一堆智能体，所以目前大概有几百个智能体在运行。通常每晚我有几千个在做更深层的工作。有几种管理方式。一种是让 Claude 使用一堆子智能体来工作。实际上，我发现自己越来越多地使用循环。就是 {slash} loop，这是最酷的东西。它是最简单但有效的东西。你让 Claude 用 cron 调度一个未来某个时间点的任务，而且是一个重复任务。它可以每分钟、每 5 分钟、每天运行，随你安排。现在我有几十个循环在跑各种事情。有一个在照看我的 PR，比如修复 CI、自动变基。另一个在保持 CI 健康，如果有不稳定测试之类的，它会去修复。还有一个每 30 分钟从 Twitter 抓取反馈并为我聚类。所以我随时都有一堆这样的循环在运行。我觉得循环就是未来。如果你还没试过，强烈推荐。我们还刚刚推出了 routines，和循环一样，但在服务器端运行。所以即使你合上笔记本电脑，它也会继续。

Usually maybe like five to 10 sessions. And then the sessions usually have a bunch of agents, so I think currently probably like a few hundred agents going. Usually every night I have like a few thousand that are doing kind of deeper work. There are a few ways to manage it. One is that you ask Claude to use a bunch of sub-agents to do work. Actually, the thing I've been finding myself using more and more is the loop. So, this is {slash} loop, and it's just the coolest thing. It's the simplest thing that works. All it is is you have Claude use cron to schedule a job for some point in the future, and it's a repeat job. And it can run every minute, every 5 minutes, every day, however often you want. At this point, I have like dozens of loops running for stuff. I have one that's babysitting my PRs, like fixing CI, auto-rebasing. I have another one that keeps CI healthy, so if there's a flaky test or whatever, it'll go and fix it. I have another one that grabs feedback from Twitter and clusters it for me every 30 minutes. So, I just have a bunch of these loops running at any time. I sort of feel like loops are the future at this point. If you haven't experimented with it, highly recommend it. And we also just launched routines, which is the same thing but on the server. So even if you close your laptop, it keeps going.

这是你的个人设置。跟我们说说你认为未来的团队会是什么样子。你如何从你正在做的工作中推断，让团队中的每个人都保持前进、理解上下文，或者你认为我们需要把更多事情交给智能体才能实现？

That's your personal setup. Tell us about what you think teams will look like in the future. How do you extrapolate from all the work you're doing to keep everyone on the team moving forward, understanding the context, or do you think we need to let go of a lot more to agents to make it work?

我想是的。做预测很难，但我就是来预测的，所以我会试试。我觉得事情的发展方向是，未来会有比现在多得多的通才。今天我们谈论通才时，很大程度上指的是仍然是工程师的人。所以他们还在写代码，但可能他们是产品工程师。比如我们说通才，就是他们同时做 iOS、Web 和服务器。那是工程领域的通才。但我认为我们会开始看到更多跨学科的通才。比如那些既非常擅长产品工程，又非常擅长设计，或者非常擅长产品、数据科学和工程的工程师。我们在自己的团队里已经开始看到了。Claude Code 团队里的很多人都是跨学科的通才。我们团队里的每个人都写代码。所以我们的工程经理、产品经理、设计师、数据科学家、财务人员、用户研究员，团队里的每个人都写代码。所以他们在某方面是专家，但现在每个人也都在写代码。我看到有人点头，但我猜这对在座各位来说其实并不意外，因为我相信你们也看到了同样的情况。

I think so. It's so hard to make predictions, but I'm here to make predictions, so I'll try. I feel like the way things are going, there's going to be a lot more generalists than there are today. Today when we talk about generalists, I think largely we're talking about people that are still engineers. So they're still writing code, but maybe they're kind of product engineers. So when we say generalist, it's like they do iOS and web and server, for example. That's a generalist in engineering. But I think we're going to start to see a lot more generalists that are cross-disciplinary. So engineers that are really good at product engineering, but also really great at design, or really great at product and data science and engineering. It's something we're starting to see on our team. A lot of people on the Claude code team are generalists across disciplines. Everyone on our team codes. So our engineering manager, product manager, designers, data scientist, finance guy, user researcher, every single person on our team writes code. So they're specialists in something, but now also everyone's just coding. I'm seeing some nods, but I bet it's actually not that surprising to people in this room because I bet you're seeing the same things.

我再问一个我最喜欢的问题，然后开放给观众。我们聊了一些关于编码正在发生的变化。我很好奇你看到软件或软件产品世界正在发生什么变化。随着 AI 让写代码的成本降低 10 倍或 100 倍，用软件生产出来的产品的价值会怎样？我们是否面临 SaaS 末日？你认为这会如何发展？

I'll have one more favorite question, then we'll open up to the audience. We talked a bit about what's changing with coding. I'm curious about what you see changing in the world of software or software products. As we see AI making writing code 10 or 100x cheaper, what happens to the value of the products that are produced with software? Do we have a SaaS apocalypse on our hands? How do you think this plays out?

SaaS 末日问题是我最喜欢的问题。我认为会发生两件事，而且我认为这两件事都不是人们一直在谈论的。第一件——在座有 Acquired 的听众吗？Acquired 播客？是的，它是最好的播客。前几周我和他们做了一期 unplugged，感觉就像见到了我的英雄，因为主持人是最棒的。他们有一个“七种力量”的概念，这是 Hamilton 的。他写了一本关于这个的书，这是商业中的七种模式。我认为因为 AI，其中一些模式会变得更重要，而另一些会变得不那么重要。例如，变得不那么重要的是转换成本，因为你可以直接使用模型，从一件事移植到另一件事。另一个变得不那么重要的是流程优势，因为对于那些模式是工作流和流程的公司，Claude 在理解流程方面变得非常擅长。尤其是 4.7，它可以爬山任何东西。所以如果你给它一个目标，告诉它迭代直到完成，它就会去做。我认为这是第一个这样的模型。所以这些会变得不那么重要，但之前的模式实际上仍然重要。比如网络效应、规模经济、稀缺资源——这些不会因为 AI 而真正改变。我认为第二件事是，如果你看看今天的初创公司数量，或者过去 10 年，我认为未来 10 年将会颠覆一切的初创公司数量会增加 10 倍。

The SaaS apocalypse question is my favorite question. I think there are two things that are going to happen, and I don't think either of them is the thing people have been talking about. One is — is anyone here an Acquired listener? The Acquired podcast? Yeah, it's the best podcast. I got to do an unplugged with them the other week, and I felt like I got to meet my heroes because the hosts are the best. So they have this idea of seven powers, and this is Hamilton. He wrote a book about this, and these are the seven modes in business. I think what's going to happen is because of AI, some of these modes are going to get more important and some are going to get less important. For example, one that gets less important is switching costs, because you can just use the model and port from one thing to a different thing. Another one that gets less important is process power, because for companies whose mode is workflows and process, Claude is getting really good at figuring out process. Especially with 4.7, it can just hill climb anything. So if you give it a target and tell it to iterate until it's done, it will just do it. I think this is the first model like that. So these are going to get less important, but the previous modes actually still matter. So network effects, scale economies, cornered resources — these are not really changing with AI. I think the second thing is if you look at the number of startups today, or in the past 10 years, I think the number of startups in the next 10 years that are just going to disrupt everything is going to increase like 10x.

因为现在你可以是一家很小的初创公司，却能打造出和大公司一样有价值的东西，而且你能正面竞争，因为大公司必须改变业务流程、改变工作方式、重新培训所有人使用技术，他们会面临很多内部阻力。但你知道，在座各位没有这个问题。如果你是从零开始，你就可以从底层原生地利用 AI 来构建。所以，我不知道，我认为这是最好的创业时机。这是成为初创公司的最佳时机。即将迎来巨大的颠覆。所以，我们终究还是有希望的。

Because right now you can be a tiny startup, you could build a thing that's as valuable as a large company and you can actually compete head-to-head because the large company has to evolve their business process, they have to evolve the way they work, they have to retrain everyone to use technology, they're going to face a lot of internal resistance to that. But you know, no one here has that problem. If you're starting fresh, then you can kind of build with AI natively from the ground up. So, I don't know. I think it's the best time to build. It's the best time to be a startup. There's so much disruption coming. So, there is hope for us after all.

谢谢 Boris。我想开放给观众提问，如果有人想问问题的话。Dan？

Thank you, Boris. I would love to open up to audience questions if anyone has anything they would like to ask. Dan?

我很好奇。你说过你在实现产品市场匹配之前花了 6 个月来构建，但现在模型已经足够好了，你认为 Claude Code 的成功有多少归功于模型，又有多少归功于产品决策？

I'm curious. You said that you built 6 months before there was product market fit, but now given that the models are good enough, how much do you attribute the success of Claude Code to the model versus like product decisions in the field of product?

我认为可能是混合的。是的，是混合的。我想如果你一年前问我，比例大概是 50/50。也许吧，我不知道。如果你 6 个月前问我，比例也是 50/50。

I think it's probably a mix. Yeah, I think it's a mix. I think if you asked me maybe a year ago, the ratio was maybe something like 50/50. Maybe I don't know. If you asked me 6 months ago, the mix would be 50/50.

那两年后呢？

What about in 2 years?

哦，两年后我不知道，老兄。我们计划只规划一周。几个月。未来的某个时候。

Oh, 2 years I don't know, dude. We plan in like we plan in 1 week out. Months. Sometime in the future.

顺便说一句，我认为之所以是 50/50，是因为我以前参加过 YC。我是 YC 公司的第一个员工，还做过很多初创公司。在初创公司里，他们反复灌输给你的，尤其是在 YC，就是打造人们喜欢的东西。所以产品是什么并不重要，模型是什么也不重要。你最终还是要打造人们喜欢的东西。我认为这就是产品重要的原因：我们非常关注细节，让你整天使用它时都能获得极佳的体验。我认为随着模型变得更好，工具本身就没那么重要了。我们现在思考的是如何改进工具？比如如何让循环成为更核心的功能？如何更容易地运行大量智能体？子智能体是一个想法。我们还在酝酿很多其他东西。但我认为一年后，模型的对齐会好得多。所以我们现在围绕提示注入、命令静态验证、权限模式、人在回路中等等所有安全机制都会变得不那么重要，因为模型会直接做正确的事。所以，这就是我的预测。

And by the way, I think the reason it was 50/50 is, you know, I did YC back in the day. I was like the first hire at a YC company and I did a bunch of startups. And in startups, the thing that they drill into you, especially in YC over and over, is build something people love. So it doesn't matter what the product is, it doesn't matter the model and all this stuff. You still in the end have to build a thing that people love. And I think that's why the product matters: we pay so much attention to the little details so that as you use it all day, it's a really great experience. I think as the model's gotten better, the harness kind of gets less important. And I think what we're thinking about right now is how do we evolve the harness? So, like how do we make loops more of a first-class thing? How do we make it easier to run a lot of agents? You know, sub agents is one idea. There's a bunch more stuff that we're cooking. But I think in a year, the model will be much better aligned. So all the safety mechanisms that we have today around prompt injection and kind of static verification of commands and permission modes, human in the loop, all this kind of stuff is just going to be less important because the model will just do the right thing. So, yeah, that's my prediction.

谢谢。你想把话筒扔给 Dan 吗？

Thank you. You want to toss the box, Dan?

很好。稍微从软件层面放大来看，我认为 Claude Code 在几个月前带来了一种文化变革，它让构建软件变得大众化。你可以看到店主为自己构建软件，甚至编程微控制器来控制开门时的灯光。你未来是否认为构建软件会成为一种像 Microsoft Office 那样的技能？也就是说，每个人都能做，而不仅仅是科技行业的人？

Great. To zoom out a little bit from software, I think Claude Code did a cultural change a few months ago where it democratized building software. You can see shop owners building their own software for themselves or even programming microcontrollers to control the light when someone opens the door. Do you see in the future building software becoming a skill like I know Microsoft Office? So, it's a thing that everybody can do, not just people in the tech industry?

哦，天哪，是的。是的。是的。我认为甚至不止如此。它会成为一种技能，就像我知道怎么发短信一样。我读了很多书——我的两个主要类型是科幻和技术史。在技术史中，有一件事我认为与现在正在发生的事情最相似。那就是 15 世纪欧洲的印刷机。在印刷机出现之前，欧洲大约只有 10% 的人口识字。他们知道如何读写。他们通常受雇于不识字的国王和领主。他们的工作就是读写，这不是每个人都会的技能。印刷机发明后，又多了两台印刷机，在第一台印刷机之后的 50 年里，欧洲出版的文学作品比之前一千年还要多。在同一时期，文学的成本，一本书的价格下降了大约 100 倍。然后，你知道，这花了几百年时间，因为学习读写很难。你需要教育体系和政府，而且不能所有人都在农场工作等等。但在接下来的几百年里，全球识字率上升到了大约 70%。所以，现在我们都能读写，你不需要一个读写学位就知道如何读写。尽管仍然有专业作家，这是你可以做的一件事。所以，我认为即将发生的事情——而且会比 50 年快得多——是软件将完全大众化，任何人都能做。这有很多推论。例如，假设你在写会计软件。我认为写会计软件的最佳人选，甚至今天，可能不是工程师，而是一个非常好的会计师，因为他们非常了解领域，而编码是容易的部分。了解领域才是困难的部分。我认为这显然就是未来。

Oh my god, yes. Yes. Yes. I think it's going to be even more than that. I think it's going to be a skill like, yeah, like I know how to send a text message. I think, you know, I read a lot—my two genres are essentially sci-fi and tech history. I think in tech history, there's one thing which to me is the clearest parallel for what's happening right now. And this is in the 1400s, the printing press in Europe. What happened was before the printing press, essentially 10% of the European population was literate. They knew how to read and write. They were often employed by kings and lords that were not literate. And their job was to read and write, and this is not something that everyone knew how to do. The printing press was invented, then there were two more presses, and in the 50 years after the first printing press, there was more literature published in Europe than in the thousand years before. And over the same period, the cost of literature, the cost of a book went down like 100x. And then, you know, it took a couple hundred years because learning to read and write is hard. You need education systems and government and everyone can't be working on farms and so on. But over the next few hundred years, literacy globally went up to like 70%. So, now we can all read and write and you don't need a degree in reading and writing to know how to read and write. Although still there are professional writers and that is a thing that you can do. So, I think the thing that's about to happen—and it's going to be much faster than 50 years—is software will be a thing that is fully democratized, that anyone can do. And there are a lot of corollaries to this. For example, let's say you're writing accounting software. The best person to write accounting software, I think maybe even today, is not an engineer, it's a really good accountant because they know the domain really well and coding is the easy part. It's knowing the domain that's the hard part. And I think this is just obviously the future.

Greg 说过的一件事是，你们有点生活在未来，因为你们能接触到模型和智能体。Claude Code 在发布之前是一个内部工具。你们在工程方面与外界之间的差距是一个月？三个月？六个月？这个差距是在变大还是变小？

One of the things Greg said was that you guys are living in the future a little bit because you get to have access to the models and the agents. Claude Code was an internal tool before you released it. Is the gap between where you guys are in engineering and the rest of the world, is that a month? Is it 3 months? Is it 6 months? And is that gap getting bigger or smaller over time?

是的，内部我们使用和其他人一样的模型。对我们来说，吃自己的狗粮非常重要。所以我们使用这里其他人都在用的东西。你知道，我们少量使用 Mythos 来测试，然后大量使用 Opus 4.7 来吃狗粮并编写大部分代码。我认为在模型方面，其实没有差距。你知道，基本上就是 Mythos，而它的一些后代版本最终会在某个时候对所有人开放。我认为在产品方面，差距可能要大得多。这仅仅是因为我们改变了所有流程。比如如果你和 Anthropic 的人聊天，我们几乎所有事情都用 Claude。

Yeah, so internally, we use the same models everyone else does. For us, the dogfooding is really really important. So, we use the thing that everyone else here does. You know, we use a little bit of Mythos to try it and then we use a lot of Opus 4.7 to dogfood it and to write most of our code. I think on the model side, there isn't really a gap. You know, it's like it's pretty much Mythos and that will become some version of some descendant of that will become available at some point to everyone. I think on the product side, there's probably a far larger gap. And that's just related to us changing all of our processes. Like if you talk to people at Anthropic, we use Claude for literally everything.

我们的 Claude 整天都在对话。就像我在编码时，我的 Claude 在循环中编码，它们会通过 Slack 与其他也在循环中运行的人的 Claude 通信，以解决未知问题。我们公司已经没有任何手动编写的代码了。所有 SQL 都由模型编写。一切都是由模型构建的。所以我认为我们领先的地方其实不是技术，因为同样的技术我们能用，在座的每个人也都能用，因为我们本质上是在构建一个平台。因此，让开发者能使用和我们一样的工具，并且我们对自己发布的所有东西都进行内部试用，这非常重要。但我认为实际上更大的问题在于组织结构和组织流程。而这是一个我们希望能在这样的场合讨论的地方，每个人都能从中学习并进化。

And our Claudes are talking all day. Like as I'm coding, as my Claudes are coding in a loop, they will communicate over Slack to talk to other people's Claudes that are also running in a loop to figure out unknowns. We have no more manually written code anywhere at the company. All of the SQL is written by models. Everything is just built by the models. So I think actually the place that we're ahead is not the technology, because the same technology available to us is available to everyone here, because fundamentally we are building a platform. And so for us, it's really important that developers can use the same thing that we're using and that we dogfood everything that we put out there. But I think there's actually a far bigger weed in the organizational structure and organizational process. And this is a place where hopefully we can talk about it in places like this and everyone can learn from it and evolve.

是的，我认为这是初创公司的优势之一。从那里开始要容易得多。Jared？

Yeah, and I think that's one of the advantages startups have. It's so much easier to start there. Jared?

是的，上次我们聊的时候，我记得你提到过我们谈了一点多智能体，当时在红杉的一次活动上，它还在代码阶段，你提到有一些东西正在酝酿中，你正在思考。现在显然有了 /batch、/loop、sub teams、teams。你能谈谈在模型层面和 harness 层面，你们是如何在 harness 层面注入先验知识，目标函数如何在模型层面变化，从而让委托工作、启动智能体这种体验变得更好吗？因为大部分工作都是可并行化的。你可以更快地做很多事情，而我感觉我不得不叠加自己的直觉来判断何时并行化，而不是模型理解你可以为某件事启动 10 个子智能体。

Yeah, last time we talked, I think you'd mentioned we talked a little bit about multi-agent and it was very in code at the time at a prior Sequoia event and you mentioned that there were some things going down the pipeline and things you're thinking about. Now obviously there's slash batch, there's slash loop, there's sub teams, there's teams. Can you speak some to either at the model level and at the harness level, how you're injecting priors in the harness level, how the objective function is changing the model level to make this experience around delegating work, spinning up agents better? Because so much of the work is parallelizable. You can do so many things so much faster and I feel like I have to overlay my own intuition for when to parallelize things rather than the model understanding that you can spin up 10 sub agents for something.

是的，在产品方面，这实际上归结为提示工程。就是这样。所以我们调整提示，帮助模型更多地并行处理事情。但说实话，随着模型变得更好，它自然就会这样做。像 loop 这样的功能，我发现实际上在 4.7 版本中，它就开始自动做了，这真的很酷。比如我告诉它：“去拉取这个数据查询。”它会说：“嘿，我注意到数据随时间变化。我会启动一个循环，每 30 分钟给你一份报告。”然后我说：“太好了。你能通过 Slack 发给我吗？”然后它就用 Slack MCP 来做到这一点。所以我认为实际上随着时间的推移，用户不需要去弄清楚如何更好地使用工具。如果是那样的话，那其实是产品设计问题，说明我做得不好。真正应该做的是让模型更好地完成这些事情，而我们通过提示让它自然做到这一点。

Yeah, on the product side, it really just comes down to prompting. That's how it is. And so we tweak prompts to help the model do stuff in parallel more. But also, honestly, as the model gets better, it just naturally does this. And so something like loop, I found actually 4.7, it just starts doing it, which is really cool. It does something like, I'll tell it, "Go pull this data query." And it's like, "Hey, I noticed that the data is changing over time. I'll start a loop and I'll give you a report every 30 minutes." And I'm like, "Great. Can you send it to me over Slack?" And then it uses the Slack MCP to do that. So I think actually over time, it's not on users to figure out how to hold the tools better. And if that's the case, it's actually a product design problem and I'm not doing a good job. It's really on the model to do this stuff better and on us prompting it so it naturally does this.

所以现在，我们很多人似乎都在使用 Claude 或 Codex 或这些云端工具来进行大量计算。但也有一些非常直言不讳的人主张让 AI 本地化。我可以想象，随着时间的推移，随着开放权重模型和其他东西的追赶，人们获得真正高质量编码辅助的可能性会更大。所以我很好奇你对未来几年的愿景。你认为大家仍然会依赖云端集中式计算，还是会转向我们都拥有本地智能体，它们不会受到限制，还有其他好处？

So right now it seems like a lot of us use Claude or Codex or these tools in the cloud to do a lot of our computing. But then there are some very vocal advocates of having your AI be local. And I could imagine over time as open-weight models and other things catch up that this could be more of a possibility for people to get really high-quality coding assistance. So I'm curious your vision of say over the next like years or something like that. Do you see the trajectory of everyone still really relying on cloud centralized compute or is there a pivot to oh, we all just have our local agents that we can rely on and they don't get throttled and other benefits?

是的，我想可能有几种方式来回答这个问题。我认为最根本的回答是：这并不重要。因为我认为我们现在已经到了模型能够自行解决这个问题的地步。所以我认为几年后，模型将完成所有代码编写。它将启动智能体。它将构建环境。所以如果它决定使用本地模型来做这件事，那它就会这么做。我认为这些将不再是工程师们需要做的决定了。

Yeah, I think there's maybe a few ways to answer that. I think maybe the most fundamental way to answer that is it doesn't matter. Because I think now we're getting to the point where the model is just able to figure it out. So I think by a couple years from now, the model is just going to be doing all the code. It's going to be starting the agents. It's going to be building the environments. And so if it decides like actually I'll use local models to do this, then that's what it'll do. I don't think these will be decisions that we are making as engineers anymore.

我们还有时间再问几个问题，所以我可以抛出这个问题。Jamie。Nester。谢谢。感觉 Claude Code 的一个伟大决策是利用了许多开发者的工具和工作流是本地的这一事实。但对于使用云端工具的一般知识工作来说，情况并非总是如此。我很好奇你们在 Co-work 方面是如何考虑这一点的：你们如何让 Co-work 有足够的权限访问我们使用的工具，使其像 Claude Code 对开发者那样强大？

We have time for a couple more questions, so I can toss this out. Jamie. Nester. Thank you. It feels like one of the great decisions with Claude Code was making use of the fact that a lot of developers' tools and workflows are local. But that isn't necessarily always the case for general knowledge work with cloud tools. I'm curious how you're thinking about this with Co-work: how do you give Co-work enough access to the tools that we use to be powerful the same way that Claude Code is for developers?

是的，这是一个非常好的问题。我知道当我在一家大公司时，我们花了大约 5 年时间把所有环境迁移到远程。工作量非常大，尤其是在大规模下。但对于知识工作来说，很大程度上已经实现了，比如 Salesforce 和 Docs 之类的工具。对我们来说，答案总是最简单的：就是 MCP。所以你在 Claude AI 中使用的同一个 MCP 连接器，你可以连接 Salesforce、Google Docs、Google Calendar。然后 Co-work 就可以使用它。Claude CLI 可以使用它。Claude Code 在任何地方都可以使用它。

Yeah, that's a really great question. I know when I was at a big company, we took like 5 years moving all the environments to remote. It's just so much work, especially at a big scale. But for knowledge work, largely, it's there already with Salesforce and Docs and things like that. For us, it's always just the simplest answer. It's just MCP. So the same MCP connector that you have in Claude AI, you hook up Salesforce, you hook up Google Docs, Google Calendar. And then Co-work can use that. Claude CLI can use it. Claude Code everywhere can use it.

对于那些没有 MCP 的系统，你认为这是不是计算机使用的一个巨大机会？

And for the systems that don't have MCPs, do you think that's where computer use is going to be a big opportunity?

是的，我认为计算机使用是一种万能方案。所以我认为目前，据我所知，Anthropic 在计算机使用方面相当领先。所以如果你通过 Co-work 使用它，效果相当好。它几乎可以使用你电脑上的任何软件。它非常慢，但现在做得相当好，尤其是 4.7 版本。但我认为除此之外，MCP 就是答案。所有这些都不那么重要。它可以是 MCP、API，或者某种程序化访问，因为模型并不关心。对模型来说，它只是 token。

Yeah, I think computer use is kind of a catchall. So I think currently, for as far as I know, I think Anthropic is pretty far ahead on computer use. So if you use it through Co-work, it's quite good. It's able to use pretty much any piece of software that you have on your computer. It's very slow, but it does it quite well now, especially with 4.7. But I think otherwise MCP is the answer. And all this stuff just doesn't matter that much. It could be MCPs, APIs, just some sort of programmatic access because the model doesn't care. To the model, it's just tokens.

好了，我们还有时间再问一个问题。Ryan。Sean，你想抛出这个问题吗？谢谢。你之前已经暗示过这一点，但如果说在某个时候你看到了产品过剩，并认为应该构建一个产品，一旦模型变得更好，它就会变得更有趣，你能用模糊的术语谈谈你今天会构建的产品形态吗？你认为随着模型在未来 6 个月到一年内变得更好，这个产品会变得更有趣。

All right, we have time for one more question. Ryan. Sean, do you want to toss the Thank you. You've kind of alluded to this, but if like sometime ago you saw the product overhang and thought to build a product that would then become more interesting once models got better, could you just talk even in vague terms about the shape of a product you'd build today that you think could become much more interesting as models get better in 6 months to a year?

是的，我认为 Claude Design 是一个很好的例子。它今天已经相当不错了。它会变得更好。我们还在为 Claude Code 准备一些东西，将在未来几周内推出。所以你会看到的。

Yeah, Claude Design I think is a really good example. It's pretty good today. It's going to get a lot better. There's also a few things that we're cooking up for Claude Code that are going to be landing over the coming weeks. So you'll see those.

然后我认为循环和批处理这类围绕大规模并行化智能体的技术会变得更好。计算机使用是另一个好例子。

And then I think loop and batch and things like this around massively parallelizing agents, that's going to get better. And computer use is another good one.

好的，Boris。非常感谢你参加我们的节目。如果大家有问题，我们还会再待一会儿。

All right, Boris. Thank you so much for joining us. I think we'll be here for a little longer if anyone has questions.

谢谢大家。

Thanks, guys.