揭秘 Anthropic 史无前例的增长：14 个月内从 10 亿到 190 亿美元 ARR

很多公司都声称自己是史上增长最快的公司，而 Anthropic 确实是。你们在 2025 年初的 ARR 是 10 亿美元。我看到的最后一个数字是 190 亿美元。也就是说，14 个月内从 10 亿增长到 190 亿美元。

A lot of companies claim to be the fastest-growing companies of all time, and Anthropic actually is. You guys were at a billion ARR at the start of 2025. The last number I've seen is 19 billion ARR. That's 1 to 19 billion dollars in 14 months.

从历史上看，我们一直是这个领域最小、资金最少的玩家。我们没有 Meta 或 Google 那样的自由现金流或分发渠道，也没有 OpenAI 的先发优势。我们能走到今天这一步，完全是个奇迹。

Historically, we were very much the smallest, least well-funded player in this space. We didn't have the free cash flow or the distribution of a Meta or Google. We didn't have the first-mover advantage of an OpenAI. It's a complete miracle that we've gotten to the stage that we have.

请给我们简单介绍一下在 Anthropic 领导增长是什么感觉。

Give us just a glimpse of what it's like to be leading growth inside of Anthropic.

这是我一生中最热门的职位。你来到 Anthropic，需要明白过去 50%、60%、70% 的运作方式，直接扔掉。

It's the hottest job I've had in my life. You come into Anthropic, you need to understand that 50, 60, 70% of how you operate in the past, just throw it out the door.

你们做的最聪明的增长举措之一就是从 ChatGPT 导入记忆这个想法。

One of the cleverest growth moves you all made was this idea of importing memory from ChatGPT.

激活在 AI 中是一个非常大的挑战。我们开始研究如何自动化增长？我们的增长平台团队正在推动一个名为 Cache 的项目，即 Claude 加速可持续超增长。我们如何利用 Claude 来自动化增长实验？它正在带来成果。

Activation is a really big challenge in AI. We are starting to look at how do we automate growth? Our growth platform team is driving this effort called Cache, which is Claude Accelerates Sustainable hypergrowth. How can we use Claude to automate growth experimentation? And it's delivering results.

你基本上就生活在未来。

You're basically living in the future.

我们总是谈论指数增长。两年后我们将交付的产品价值可能是今天的一千倍。最有趣的是，我注意到在公司内部，线性图表根本不酷。一切都是对数线性的，所以请用对数线性标度展示给我看。

We always talk about the exponential. The product value that we will deliver in 2 years' time is probably like a thousand X what it is today. The funniest thing is I've noticed internally, linear charts are just not cool. Everything is log-linear, so show me it log-linear scale.

今天我的嘉宾是 Amol Avasare。Amol 是 Anthropic 的增长负责人，该公司正经历历史上最前所未有的增长。在过去 14 个月里，他们的年经常性收入从 10 亿美元增长到超过 190 亿美元。就在过去几个月里，他们的收入翻了一番。他们一直以每年 10 倍的速度增长。这在如此规模上是闻所未闻的。到本期节目发布时，他们的收入还会更高。为了更直观地理解这个规模，像 Atlassian、Palantir 和 Snowflake 这些已经存在了 15 到 20 年的公司，各自的 ARR 大约在 45 亿到 60 亿美元之间。而 Anthropic 每隔几个月就能增加这么多 ARR。如果这还不够有趣，那么领导 Anthropic 增长的 Amol 是一个了不起的人。他之前曾在 Mercury 和 Masterclass 领导增长。在此之前，他是一名创始人和投资银行家。最有趣的是，大多数人不了解他的一点是，Amol 曾遭受严重的脑损伤。他花了 9 个月时间重新学习如何走路、工作，以及如何不一直感到恶心。他几年前在我的 newsletter 的一篇客座文章中分享了这个故事。我们实际上在对话中聊到了这一点。这是我最喜欢的一类对话，因为 Amol 和他的团队生活在未来，他来告诉我们事情的发展方向以及将会发生什么变化。在本期节目中，Amol 前所未有地分享了像 Anthropic 这样的公司如何运作和增长，包括他们如何看待增长、他们自动化了哪些工作、产品和增长角色的未来、Amol 如何通过给 Mike Krieger 发冷邮件获得这份工作，以及更多。Amol 非常棒，试着数数在这次对话中他让我震惊了多少次。在开始之前，别忘了查看 Lenny's Product Pass.com，那里有专门为 Lenny 的 newsletter 订阅者提供的超值优惠。话不多说，有请 Amol Avasare。Amol，非常感谢你来到这里，欢迎来到播客。

Today, my guest is Amol Avasare. Amol is head of growth at Anthropic, which is on the most unprecedented growth run in history. In the past 14 months, they grew from 1 billion to over 19 billion in annual recurring revenue. Just in the past few months, their revenue doubled. They've been growing 10X year-over-year. This is unheard of at this scale. By the time this episode comes out, their revenue will be even higher. To put this scale in perspective, companies like Atlassian and Palantir and Snowflake, which have been around for 15 to 20 years, each do something like 4 and 1/2 to 6 billion in ARR. Anthropic is adding this much ARR every few months. And if that isn't interesting enough to you, Amol, who leads growth at Anthropic, is an incredible human. He previously led growth at Mercury and Masterclass. Before that, he was a founder and an investment banker. And most interestingly, something that most people don't know about him is that Amol suffered a severe brain injury. He had to spend 9 months relearning how to walk and work and just not be nauseous all the time. He shared this story in a guest post in my newsletter a number of years ago. We actually chat about this during the conversation. These are my favorite kind of conversations because Amol and his team are living in the future, and he's come to tell us where things are heading and what's going to change. And in this episode, Amol shares an unprecedented look at how a company like Anthropic operates and grows, including how they think about growth, what parts of the job they've automated, the future of the product and growth roles, how Amol got the job first place by cold emailing Mike Krieger, and so much more. Amol is wonderful, and just try to count the number of times that he blew my mind during this conversation. Before we get into it, don't forget to check out Lenny's Product Pass.com for an incredible set of deals available exclusively to Lenny's newsletter subscribers. With that, I bring you Amol Avasare. Amol, thank you so much for being here, and welcome to the podcast.

很高兴来到这里。

Glad to be here.

Anthropic 的增长负责人，没什么大不了的。很多公司都声称自己是史上增长最快的公司，而如果你看增长轨迹，Anthropic 确实是。我这里有一些数字，让大家明白这有多离谱。你们在 2025 年初的 ARR 是 10 亿美元，然后在 2025 年中达到大约 40 亿美元，2025 年底达到 90 亿美元。我看到的最后一个数字是你们达到了 190 亿美元 ARR。这里提供一些背景信息：一是从 10 亿到 190 亿美元只用了 14 个月。我有很多问题。首先，你实际上是如何得到这个角色的故事非常有趣。谈谈你是怎么得到这个职位的吧。

Head of growth at Anthropic. No big deal. I've had a lot of people come on this podcast from companies that claim to be the fastest growing companies of all time, and Anthropic actually is if you look at the trajectory. I just have some of the numbers here just so people understand how absurd this is. So, you guys were at a billion ARR at the start of 2025, then hit something like 4 billion mid-2025, then 9 billion ARR at the end of 2025. And the last number I've seen is you guys are at 19 billion ARR. Which uh just to put a couple pieces of context here. One is that's from one to 19 billion dollars in 14 months. Um I have so many questions. First of all, the story of how you actually landed this role is really interesting. Talk about how you how you got this role.

是的，有点非正统。有趣的是，在我入职时，他们介绍了有多少比例的人是通过推荐、网站申请、猎头来源进来的，而我哪个都不是。我当时想，“好吧，这挺有意思的。”基本上，我进入 Anthropic 的方式是，我实际上是 Claude 的用户，而且用得很多。我想，“天哪，这些人产品好，公司好，但他们显然没有增长团队。”于是我就给 Mike Krieger 发了一封冷邮件。他是首席产品官。我发了一封冷邮件说：“嘿，喜欢你们做的事，喜欢这个产品。我觉得你们急需一个增长团队。想聊聊吗？”我没想到他会回复。然后他回复说：“嘿，是的，我有兴趣。我们聊聊吧。”有趣的是，我不知道他们当时并没有在招聘增长团队。没有任何增长团队的职位列出。但他们正好在那时开始考虑招聘增长团队。所以时机非常好。总之，我和 Mike 聊了聊，然后事情就顺理成章了。他说我是他唯一一个通过冷邮件招聘的 PM。我觉得非常幸运他决定回复我的邮件。

Yeah, it's a little unorthodox. So, it's funny in when I when I did my onboarding they they walked through what percentage of the cohort came through referrals, what percentage came through applying on the website, what percentage came through uh sourcing, and I was on none of those. And uh I was like, "Okay, this is interesting." Um basically the the way that I I got to Anthropic was that I was actually a user of Claude, and I was using a lot. I was like, "Man, these guys like great product, great company, but they they really they obviously don't have a growth team." And uh what I did was I just sent Mike Krieger a cold email. He was the chief product officer. I sent him a cold email saying like, "Hey, love what you guys do, love the product. I think you guys badly need a growth team. Want to chat?" And uh I didn't expect he would respond. And uh you know, he responds and says, "Hey, yeah, I'm interested. Let's let's talk." And it's funny, I didn't know I mean, I wouldn't have known they were not hiring for a growth team. There were no growth team roles listed. But they were just at that time starting to think about hiring a growth team. So, it was very good timing, but yeah, one one spoke to spoke to Mike, and and one thing led to another. He said I'm the only PM that he's hired from from cold email. And I I feel very lucky that he decided to respond to my email.

我之前不知道这个故事。这又是一个离谱的事实。显然你很擅长冷邮件。你在那封冷邮件里做了什么来吸引他的注意？

I did not know the story. That is that is another uh absurd fact. Uh clearly you're good at cold email. What did you do in this cold email to get his attention?

我想说，我基本上多年来已经完善了冷邮件。当我还是创始人时，我不得不变得非常擅长这个。我发了很多冷邮件，不断打磨主题行、信息和语气。基本上，在主题行里，从转化角度看，第一件事是有人看到邮件，他们需要点击它。所以我有一个经过测试的文案，打开率非常高。

I would say like I've basically perfected cold email over the years. So, I was a when I was a founder, I I had to get really, really good at this. So, I I sent a lot of cold emails out and just honed the the the subject lines and message and and the tone. And and so, basically, I have in the subject line, the first thing is like from a conversion standpoint, someone sees the email, they need to click on it. And so, I have a copy that I've tested that is like very, very high open rate. And so,

等等，这个文案是什么？还是说这是个秘密？

Wait, what is this copy? Or is this a secret?

这是个秘密。我觉得

It's a secret. I think

好吧，我们保密。

Okay, we'll keep it We'll kind of keep

是的，这是个秘密。嗯，所以第一点是让他们打开邮件。我认为第二点是策略：你需要了解人们在哪里收到 outreach。如果每个人都在一个领域收到 outreach，然后你在那里联系他们，那么你的回复率就不会那么高。

Yeah, it's it's a secret. Um so, that's one, getting them to open. I think the second is then the tactics of You need to understand like where are people getting outreach. And if you If everyone's getting outreach in one area and then uh you reach out to them there, then then you're not going to get as high of a response rate.

所以，你可以想想 LinkedIn，想想工作邮箱。这些都是大家都会发邮件的地方。所以，有办法找到人们的个人邮箱。这就是我做过的一件事。然后，好了，我有了他的个人邮箱，我知道有效的文案。然后就是保持简短：我是谁，为什么我适合，我们应该聊聊。你知道，这些事情通常不管用。然后你应该跟进几次。我的经验法则是，如果我真的在乎，我就应该一直联系他们，直到他们告诉我“请停止”。所以，我会一直这么做，但他第一次就回复了。一个优秀增长人才很擅长冷邮件和吸引注意力，这说得通。所以，这几乎就是一个面试步骤：我想不想读这封邮件？

So, you can think about like LinkedIn, you can think about like work email. These are things that everyone is emailing. So, there's ways to get people's personal emails out there. And so, like that's one thing that I did. And so, okay, I've got his personal email, I know the copy that works. And then it's just keeping it very short on here's who I am, here's why I'd be a good fit, and we should chat. And you know, these things typically don't work. And then you should always follow up a few times. I think my rule of thumb is like if I really care about it, I should just keep reaching out to them until they tell me like, "Please stop." And so, I would have kept doing that, but he responded the first time. It makes sense that a talented growth person would be very good at cold email and getting people's attention. So, that's almost like an interview step is just did I want to read this email?

本期节目由本季呈现赞助商 WorkOS 提供。OpenAI、Anthropic、Cursor、Vercel、Replit、Sierra、Clay 以及数百家其他成功公司有什么共同点？它们都由 WorkOS 提供支持。如果你正在为企业构建产品，你一定感受过集成单点登录、SCIM、RBAC、审计日志等大公司所需功能的痛苦。WorkOS 将这些交易障碍转化为即插即用的 API，拥有专为 B2B SaaS 打造的现代开发者平台。我投资的每一家初创公司，一旦开始向上市场扩张，都会与 WorkOS 合作。这是因为他们是最好的。无论你是种子期初创公司，试图拿下第一个企业客户，还是全球扩张的独角兽，WorkOS 都是实现企业就绪和解除增长障碍的最快途径。它本质上是企业功能的 Stripe。访问 workos.com 开始使用，或者直接联系他们的 Slack，那里有真正的工程师等着回答你的问题。WorkOS 让你通过令人愉悦的 API、全面的文档和流畅的开发者体验更快地构建。今天就去 workos.com 让你的应用企业就绪。

This episode is brought to you by our season's presenting sponsor WorkOS. What do OpenAI, Anthropic, Cursor, Vercel, Replit, Sierra, Clay, and hundreds of other winning companies all have in common? They are all powered by WorkOS. If you're building a product for the enterprise, you've felt the pain of integrating single sign-on, SCIM, RBAC, audit logs, and other features required by large companies. WorkOS turns those deal blockers into drop-in APIs with a modern developer platform built specifically for B2B SaaS. Literally every startup that I'm an investor in that starts to expand upmarket and support working with WorkOS. And that's because they are the best. Whether you are seed stage startup trying to land your first enterprise customer or unicorn expanding globally, WorkOS is the fastest path to becoming enterprise ready and unblocking growth. It's essentially Stripe for enterprise features. Visit workos.com to get started or just hit up their Slack where they have actual engineers waiting to answer your questions. WorkOS allows you to build faster with delightful APIs, comprehensive docs, and a smooth developer experience. Go to workos.com to make your app enterprise ready today.

好的。那么，给我们一个 glimpse，看看在 Anthropic 领导增长是什么感觉，这是迄今为止历史上增长最快的公司。到底是什么样的？

Okay. So, give us just like a glimpse of what it's like to be leading growth inside of Anthropic right now, the most by far fastest growing company in history. Just what is it like?

是的，我想说这很大程度上是全公司的努力，对吧？所以，我想我们是增长团队。我们做得很好。我认为我们带来了很多影响，但老实说，我们不能把公司的成功归功于自己太多。Anthropic 首先是一家模型公司和智能公司。所以，推动我们成功的主要部分是我们的研究团队。我们拥有我认为世界上最好的研究团队。我们在推理和算力方面有很棒的团队。然后还有很多其他团队，比如 Claude Code、Dodo Marketing 等，我认为他们比我们更值得称赞。我想你放大来看，就像你之前说的，增长轨迹简直疯狂。10 倍年同比增长趋势从一开始就存在。我认为 2023 年是从 0 到 1 亿。2024 年是从 1 亿到 10 亿。去年是从 10 亿到大约 100 亿。然后，我回想我 2024 年加入时，收入是数亿。然后到 2024 年底和 2025 年，我加入的第二周，我们进入 2025 年收入规划，我们有基准情况和激进情况方案，Darius 推动激进方案，人们吓坏了，说我们到底怎么才能达到那个目标，而 Darius 说我认为我们实际上可以比那高得多。我进来时觉得这地方疯了，绝对不可能。然后那发生了，对吧？然后到了 2025 年底，你会想，好吧，大数定律，基于 100 亿的基准率，这里会有相当大的放缓，你怎么能保持这个速度增长？但它就是没有放缓，你知道，这些数字是公开的。你引用的 190 亿是 2 月底的数据。所以那也已经过时了。而且，这绝对疯狂。最有趣的是我在内部注意到的一件事：线性图表根本不酷。没人在乎线性图表。一切都是对数线性。给我看对数线性尺度，那是我们思考的尺度。我认为总的来说，我们只是勉强支撑，努力管理增长，为我们的用户做到最好。

Yeah, I'd say it's very much a company-wide effort, right? So, like I guess we are the growth team. We have done great. I think we've driven a lot of impact, but honestly, man, we can't claim too much credit for the success of the company. We as Anthropic are really a model company and an intelligence company first and foremost. And so, the lion's share of what has driven our success is our research team. We have I think the best research team in the world. We have great teams on inference and compute. And then there's many other teams like Claude Code, Dodo Marketing, etc. who I think deserve much more credit than us. And I think you're zooming out going to some of what you said earlier that the growth trajectory has just been insane. That 10x year-on-year revenue growth trend has been there since the beginning. I think 2023 was zero to 100 mil. 2024 was 100 to one. Last year was one to roughly 10. And then like I look back to when I joined in 2024, revenue was in the hundreds of millions. And just that trajectory to the end of 2024 and 2025 like week two of when I joined, we're going into 2025 revenue planning and we have these like base case and aggressive case scenarios and Darius pushing the aggressive case scenario and people are freaking out being like how the hell are we going to hit that and Darius like I think we can actually go much higher than that. And I'm coming in like this place is crazy like there's absolutely no way. And that happened, right? And then you get to the end of 2025 and it's like okay, law of large numbers, there's going to be a pretty big slow down here based on your baseline rate of 10 billion, how are you going to keep growing at this rate? And like it just has not slowed down and you know, these numbers are public. The 19 billion number you quoted is from the end of Feb. So that is also out of date. And it's look, it's absolutely insane. Like the funniest thing is something that I'm noticing internally is like linear charts are just not cool. Like no one cares about linear charts. Everything is log linear. I'll show me it log linear scale and that's the scale we think and I think overall we're just, you know, really hanging on by the seat of our pants and we're trying to manage the growth and do the best that we can for our users.

我和 Anthropic 的某个人聊到你，他们说基本上任何时候他们想让某样东西增长，就会请你帮忙，而且很有效。所以你刚才说事情很神奇、很棒，Claude 和你们构建的所有工具天生就很棒，这是它们增长的重要原因。我想很多听众会想，你实际上做什么？就像一个神奇的微神，什么都能为你做？为什么我们需要一个增长人员？你到底做什么？谈谈你关注的事情，也许还有你的团队推出的一些有助于加速增长的胜利。

I was talking to somebody at Anthropic about you and they said that basically anytime they want something to grow, they ask you to help and it works. So you talked about just like things are magical and amazing and like Claude and all the tools you all built are amazing innately and that's a big part of the reason they grow. I think many people listening to this will be like what do you act even do a mole with like a magical micro god that just can do anything for you? Why do we need a growth person? What do you even do? Talk about just the stuff that you focus on and maybe like a couple of the wins that your team has shipped that has helped accelerate growth.

我想说他们并不完全错，对吧？我们非常幸运拥有世界上最好的模型。我们非常幸运拥有像 Claude Code 和 co-work 这样的产品。这突然让生活变得容易多了。话虽如此，我想说这是我一生中最难的工作，你知道，我当过创始人，当过投资银行家，还有其他类似经历。这很艰难，如果我看看我们作为增长团队在这里做什么？我认为最终和你在正常公司会考虑的类别是一样的。所以，我们关心获客，如何让更多人进门，进门的人的意图，我们关心激活，注册流程，将人们引导到正确的产品，让他们成功。你知道，我们关心变现、免费到付费转化、定价和包装等事情，所有这些工作类别都是一样的。我认为最大的不同可能是，我大约 70% 的时间花在我们内部称为“成功灾难”的事情上，那就是事情进展得太顺利，以至于其他东西现在正在崩溃。我认为任何在经历过快速增长的公司工作过的人，比如早期的 Facebook、Uber 或 DoorDash，他们都能切身理解这一点。

I would say like they're not fully wrong, right? Like we're very lucky to have the best models in the world. We're very lucky to have products like Claude Code and co-work. It suddenly makes life a lot easier. Having said that, I would say this is like the hardest job I've had in my life and that's, you know, having been a founder, having been an investment banker and other things like that. It's tough and if I look at what do we do as a growth team here? I think it's ultimately the same categories of things that you would think about at a normal company. So, we care about acquisition, how are we getting more people in the door, there are the intent of the people coming through the door, we care about activation, the sign-up flow, funneling people to the right products, making them successful. You know, we care about things like monetization, free to paid conversion, pricing and packaging, all of that stuff that the categories of work is the same. I think then the probably the big differences is I would say that like roughly 70% of what I spend my time on is what we internally refer to as success disasters and that is where things have gone so well that other things are breaking now and I think anyone who's worked at companies that have gone through rapid growth, like Facebook or Uber or DoorDash early on, like they understand this viscerally.

规模扩张到这种程度会带来很多挑战。如果你想想获客、激活、变现这几个环节，就会发现大量救火式的工作，从一个紧急事务跳到另一个，通常非常痛苦。有趣的是，你看着所有图表，全是绿色，一路向上向右，但每个人还是觉得……情绪上依然很艰难。所以你需要退一步，意识到我们很幸运能有这些问题。但大概 70% 的时间都花在这种成功灾难的救火上。剩下的 30% 是更常规的增长工作，更偏主动。所以你会思考：如果资源有限，我们有很多不同的产品，哪些产品值得投入？长期定价和包装策略是什么？尤其是技术变化很快，用户行为和参与趋势也在变。然后还有很多新产品要上线。比如，你推出了 Co-work，然后呢？增长团队什么时候该介入，开始优化 Co-work 的核心采用漏斗？所以大概 70% 是疯狂的救火，30% 是更常规的工作。

Where scaling this much just brings a lot of challenges. If you think about each of those categories on acquisition, on activation, on monetization, there's just a ton of firefighting, jumping from one urgent thing to another and it's often extremely painful. And it's funny because you look at all the charts, all the charts are green, fully up and to the right, and everyone's just like... it can be quite tough emotionally still. So, you need to sort of step back and just realize that we're very lucky to have these problems. But that's sort of 70% of my time, I'd say, is just this firefighting of success disasters. And I think the 30% remaining is just much more standard bread and butter growth work where it's more proactive. So, you think about, okay, if we have limited resources, which are the products — we have many different products — which are the products we want to put some juice behind? What does our long-term pricing and packaging look like? Especially given that the technology is changing a lot and behavior and engagement trends are shifting. And then, things like we have a lot of new products coming up. Like, okay, you shipped co-work. Now what? When is the right time that we should lean in as a growth team to start optimizing the core adoption funnel for co-work? So, it's very 70% just crazy firefighting, 30% more bread and butter stuff.

好的，我会深入探讨这些。你们最近最巧妙的增长举措之一是从 ChatGPT 导入记忆的功能，让用户很容易上手，并抓住了人们对 Anthropic 兴奋的趋势。你能分享一些这个功能背后的故事吗？

Okay, I'm going to dig into a lot of that stuff. One of the cleverest growth moves you all made recently was this idea of importing memory from ChatGPT, where you just made it really easy and kind of jumped on this trend of people getting really excited about Anthropic. Is there anything you could share about the behind-the-scenes story of that feature?

我们一直在思考如何改善冷启动问题和新用户体验。我认为激活是 AI 领域一个很大的挑战。所以那是一个例子，我们在特定时间点推出的功能。但归根结底，退一步看，关键是如何让注册用户更容易让 Claude 了解他们是谁，理解 Claude 能如何帮助他们，并引导他们到正确的地方。

We were always thinking about what can we do to improve the cold start problem and improve the new user experience. I think that activation is a really big challenge in AI. And so, that's one example of something that we shipped that was very specific to a moment in time. But ultimately, when you zoom out, it's like, okay, how do you really make it easier for people who are signing up to have Claude understand who they are and understand how Claude can help them and get them to the right place.

我想顺着激活这个话题聊下去。我和很多推动 AI 产品增长的人交流时，经常提到一点：现在有太多东西在争夺人们的注意力，要让用户达到“哇，这真是我想一直用的东西”的状态，非常困难。而且也不可靠，有时并不神奇。AI 是非确定性的。所以我想问：第一，专注于激活、让用户达到“顿悟时刻”有多重要？第二，关于如何用 Claude 或 AI 产品做好这一点，你学到了什么？

I want to follow that thread, activation. There's something that comes up a ton when I talk to people leading driving growth on AI products. It's just like there's so much stuff trying to get your attention these days for people to get to a place where they okay, wow, this is really going to be something I want to keep using. It proves to be really hard. And it's also just unreliable. Sometimes it's not going to be magical. It's AI. It's non-deterministic. I guess one is just like how important is focusing on activation, getting people to that aha moment with AI products. And two, what are some things you've learned about how to do that well with Claude or AI in general?

好问题。我认为激活至关重要。定义早期激活，比如第零天、第一天的产品体验。历史上，任何做过增长或产品的人都知道，这通常是提高长期留存率的最大杠杆之一。而且我认为它的重要性呈指数级增长。退一步看，我觉得行业最大的问题之一是能力开销，模型进步太快了。真正的挑战在产品端，如何将这些好处传递给用户。即使内部，新模型来了，你忙得不可开交，还得专门花时间想：“这个模型能做什么？我需要如何更新我的认知？”更广泛地说，对大多数人而言，你可能有一个 AGI 或能做各种疯狂事情的模型，但如果人们本能地来问“旧金山天气怎么样？”，那他们就不会充分利用产品。所以这很难，因为模型能力提升太快。比如，回想我们之前有 Opus 4，那时模型能做一系列事情，而 Opus 4.5 解锁了一大堆新功能。我们有了新模型 Opus 4，然后花时间跑测试，搞清楚新模型的能力，找到合适的入口引导用户使用这些功能。你跑测试，获得经验，然后推出新流程。但到那时，可能下一个模型已经来了，解锁了更新的能力，让之前的经验都过时了。所以这确实是个很难跟上的问题。我认为增长和激活中很多老规律仍然有效，最终，最高杠杆之一是找到合适的产品或功能给合适的用户。我在多个公司反复看到的一个经验是，适当的摩擦是有帮助的，如果方法得当，增加摩擦通常有效。我在这里也看到了这一点。所以我认为，关键是能够识别用户特征，从而推荐合适的功能或产品，并且不要害怕为此增加摩擦。这可能是最重要的一点。

Yeah, it's a good question. I think that activation is critical, right? And defining that is like early activation, call it day zero, day one product experience. I think that historically, anyone who's been in growth or in product understands that this is usually one of the highest levers that you have to actually increase even longer-term retention. And I think that the importance of that has just gotten exponentially higher. Now, zooming out, I feel like one of the biggest problems in the industry is capability overhead, where the models are just getting better so quickly. And the real challenge is on the product side of how do we start to diffuse those benefits to people where even internally, like there's new models coming internally and you're sitting there, you're so busy, and when a new model's available, you need to really carve out time to be like, 'What can this do? How do I need to update my priors?' And if you think about more broadly, for most people, you may have a model that is like AGI or some model that can do all sorts of crazy things, but if people's instinct is to come there and be like, 'Hey, what's the weather in SF?' then, you know, they're not going to get the most out of the product. And so I think that it's tough because the model capabilities are rising so much. So like if I think about, okay, back when we had, I don't know, say like Opus 4, at that point there's a series of things the model can do, and Opus 4.5 unlocked a whole bunch of new things. You can think about, okay, we sit there, we've got this new model Opus 4. The time to then go and run a bunch of tests, figure out, okay, there's the capabilities from this new model. What are the right on-ramps to guide people to those features? You run tests, you get the learnings, you then ship a new flow. By then, you may already have the next model which unlocks newer capabilities that makes all these learnings irrelevant, right? So it's actually just a really difficult problem to stay on top of. I think that many of the same things, same old trends in growth and activation remain accurate where to me it's like ultimately, some of the highest leverages come from finding the right product or the right feature for the right user. And I think that one learning I've seen time and time again across companies actually is that the right friction helps, and adding more friction usually works if you do it the right way. I think that's something I've consistently seen, that we've seen here as well. So to me, I think it's really understanding being able to identify what are the characteristics of the user that allows you to then recommend them to the right feature or product, and not being shy about adding friction to do that. I think is probably the single biggest thing that's important here.

我问联合创始人 Ben Mann（之前的播客嘉宾）该问你什么，他特别提到你在 Mercury 重新设计 onboarding 并让它变得神奇的经历。他和你观点一致，认为让人们理解 AI 工具的能力，帮助他们决定使用并坚持使用，非常重要。有没有一个你在 onboarding 中做出的改变，显著提高了激活率的例子？

When I asked Ben Mann, one of the co-founders, former podcast guest, what to ask you about, and this is what he highlighted is your experience especially at Mercury redoing onboarding and making it magical. And basically he's in the same place as you have just how important it is for people to understand what the AI tool is capable of to help people decide to use it and stick with it. Is there an example of something you changed in onboarding that helped significantly improve activation?

好问题，我很高兴他提到 Mercury。我经常谈论他们的产品。我在 Mercury 的增长团队工作过，我认为那是一个很棒的产品。

Yeah, it's a great question and I like that he brings up Mercury. I talk about their product a lot. So, worked on the growth team at Mercury. I think it's a fantastic product.

这是很多人使用的东西，他们使用它的原因是它提供了更好的银行体验。

It's something many people use and the reason they use it is because the better banking experience than you

是的，我是一个非常满意的客户，先说明这一点。我喜欢它。很棒的产品。强烈推荐。他们有个人银行业务。每个人都应该去用。对，他们刚刚推出了这个。我认为 Mercury 有趣的地方在于，其核心价值是更好的体验，对吧？这就是你使用它的原因。你并没有得到其他更好的东西，只是产品体验更好。所以这种理念在公司内部根深蒂固，它来自几位创始人。

Yeah, I'm a very happy customer just to put that out there. I love it. Great product. Highly recommend it. They have personal banking. Everyone's got to go and use it. Right, they just launched that. I think that the interesting thing about Mercury is like the core value is that it's a better experience, right? That's the reason you use it. You're not getting like better other things. It's just a better product experience. And so that ethos is very deeply held within the company. It comes from a number of the founders.

我认为我在那里的时候，有一个季度我们在注册流程上大力投入。因为银行机构和受监管实体的注册流程极其复杂。比如，我花了很多时间区分注册代理地址、法定地址和实际地址之间的差异。这些事情非常复杂，我们基本上审视了注册流程。我们想，‘好吧，我们在产品的其他部分投入了大量精力提升质量，但在这里却没有。而这是人们的第一体验。’于是我们说，‘忘记指标，忘记增长，忘记其他一切。’作为增长团队，负责转化率，我们要花整整一个季度来修复这个流程的质量。于是我们就这么做了。忘记指标，我们只是尽可能让体验变得更好，修复所有问题，比如从一个字段跳转到另一个字段，添加受益所有人详细信息。结果这实际上成了我加入这里之前，作为增长团队所经历的最有影响力的一个季度。我们从注册开始到完成的转化率显著提升，仅仅因为专注于质量。所以对我来说，这是一个更广泛的教训：质量驱动增长，我试图将这一点带到 Anthropic。

I think that we had a big push one quarter when I was there on the onboarding flow. So, onboarding flows for banking institutions and regulated entities are extremely complex. Like, the amount of time I spent on the difference between a registered agent address and a legal address and a physical address. These things are very complex and we basically looked at the onboarding flow. We're like, 'Okay, we invested so much in quality in the rest of the product, but we haven't really done it here. And this is the first experience that people have.' And so, we said, 'Forget metrics, forget growth, forget everything else.' As the growth team on conversion, we're going to spend a whole quarter fixing quality in this flow. And so, that's what we did. Like, forget the metrics, we're just going to make this as good of an experience as we can and fix all these like you go back from one field to the other field and adding in your beneficial ownership details. And it actually ended up being like probably until I joined here, the single most impactful quarter that I've ever had as a growth team in terms of the impact that it had. And so, we saw a significant uplift from basically our onboarding started to completion from just focusing on quality. And so, that to me is like a broader learning around quality drives growth that I think I've tried to bring to Anthropic.

我认为在 Anthropic，我们在注册流程中做的一些事情基本上是询问用户他们是谁、他们的兴趣领域是什么。然后我们利用这些信息推荐不同的产品和功能。很多人看到这个流程会说，‘你们的摩擦太多了。流程太长了。’而我的回答是，‘我们有数据。我们对它的表现挺满意的。’

I think for us at Anthropic, some of the things that we've done in the onboarding flow is basically like we ask users questions around who they are, what their interest areas are. And we then use that to recommend different products and features. And a number of people look at the flow and they're like, 'You have so much friction. It's such a long flow.' And I'm like, 'We have the data. We're kind of happy with how that's performing.'

你对摩擦有什么看法，好的摩擦与坏的摩擦？

What is your kind of a philosophy on friction, good friction versus bad friction?

我在每一份增长工作中都一次又一次地看到，增加摩擦、增加正确的步骤会带来更高的转化率和更高的漏斗完成率。所以，你想要消除那些不增加价值的烦人摩擦，但人们最简单的理解是，只要缩短价值实现时间就行了。砍掉所有步骤，直接让他们进入产品。但大多数时候这行不通。我认为如果你真的彻底测试过你的流程，看看我待过的公司 Masterclass。如果你现在走一遍 Masterclass 的购买流程，当你登陆并试图购买时，你会经历一个测验的所有步骤，你会想‘我是来买的，却要我回答这么多问题。你来这里是为了什么等等。’很容易认为‘他们为什么这么做？这太糟糕了。全部砍掉。’但事实并非如此，这经过了彻底测试，实际上它是一个重要的收入驱动因素，因为它通过了解用户的兴趣，然后推荐内容和课程，帮助用户感觉产品是为他们量身定做的。我们团队的一位增长产品经理离职去了 Calm，那个冥想应用。如果你去 Calm 的着陆页，走一遍他们的购买流程、登录流程，你会看到一个测验。这不是巧合。在 Mercury 我们也测试过，我想你可能在 Twitter 上发过，我们把注册流程中的一些步骤拆分出来，比如如果你有五六种不同的表单输入，你经常把它拆成两个屏幕，减少认知负荷，这表现很好。我们在流程中增加了一些步骤，实际上表现也很好。在 Anthropic 也一样。所以，我认为给我的启示是：当摩擦不能帮助用户理解产品为什么适合他们时，就消除它。但是，如果你能帮助用户理解产品为什么适合他们、如何使用以及什么对他们最相关，而这会增加摩擦，不要回避它。测试它，确认它对你有用，但我认为这是大多数增长从业者都深刻理解的东西。

I've just seen time and time again at every job I've been in in growth that adding friction and adding the right steps leads to higher conversion and higher funnel completion. So, you want to get rid of annoying friction that doesn't add value, but the most simple understanding people have is like just solve time to value. Cut all the steps and just get them into the product. And that doesn't work most times. I think if you've really thoroughly tested your flows, I look at the companies I've been at Masterclass. If you go through Masterclass's purchase flow right now, you'll go through all these steps in this quiz when you land on when you're trying to buy and it's like I came here to buy and it's taking me through all these questions. What are you here for etc. I think it's easy to look at that and be like why do they have this? This is a terrible thing. Just cut it all. And it's like no, that's been thoroughly tested and actually that was a significant revenue driver because it helps users feel that the product is for them by understanding what their interests are and then recommending the content and classes there. One of the growth PMs on our team left to join Calm, the meditation app. If you go to Calm's landing page and go through their purchase flow, their login flow, you'll see a quiz. It's not a coincidence. At Mercury we also tested I think you might post on Twitter like we broke out some steps in onboarding and just having one screen, if you have five or six different form inputs and you often break that into two screens and reduce the cognitive load, that is something that performs well. We added steps into the flow there that actually performed well. Same with at Anthropic. So, I think the takeaway to me is like cut friction when it doesn't add to the experience of helping a user understand why the product is for them. But, if you can help users understand why a product is for them and how to use it and what's most relevant to them and it's going to add friction, don't shy away from it. Test it, confirm that it works for you, but I think this is something that most growth practitioners deeply understand.

而“为他们”这一点非常重要，你描述为增加摩擦以更好地了解他们是谁，从而知道如何为他们推荐合适的东西。

And the 'for them' is really important there, which you described as adding friction to better understand who they are so that you know how to recommend the right thing for them.

没错。是的。如果做得好，这就会贯穿始终，对吧？所以它有助于激活，然后有助于生命周期。你对这些用户了解得更多，知道他们为什么在这里，在我们最成熟的业务中，即使有人流失，你也可以做相似人群定向，在广告层也能触达他们。所以，最初了解用户是谁这一部分就像是一笔持续回报的投资，如果你在漏斗下游正确使用它的话。

Correct. Yes. And done right, that just flows through, right? So, it helps you with activation, but then it helps you with life cycle. You know more about those users and why they are here and in our most sophisticated businesses, you can then even if someone drops off, you can do look-alike targeting and you can get them at the ad layer as well. So, that initial piece of how you understand who the user is is just like a juice that keeps on giving if you then use it right further down funnel.

大家都准备对 Cloud 的注册流程、MasterClass 的注册流程、Mercury 的注册流程做一大堆拆解分析。顺便说一句，我最近参加了一个产品经理晚宴，我问所有产品经理，你的角色在 AI 影响下变化最大的是什么？AI 对你工作的影响最大在哪里？其中一位产品经理回答说，实际上是做竞争分析，对别人的定价页面和注册流程做大量拆解。所以现在做这些很容易。只要说‘嘿，co-work。’我不知道，你会用 co-work 还是 Cloud 来做这个？你会用什么？好吧，这个不错。帮助人们选择使用哪个工具。

Everyone's about to go do a bunch of tear-downs of Cloud's onboarding, MasterClass onboarding, Mercury onboarding. Kind of as a tangent, I was at this PM dinner recently and I was asking all the PMs how has your role as a PM changed most with AI? Like where is AI most impacting what you do? And one of the PMs answered that is actually doing competitive analysis, doing a bunch of tear-downs of what other people are doing for pricing pages and onboarding. So, it's easy to do now. Just hey, co-work. I don't know, would you co-work for this or Cloud? What would you use for that? Okay, this is good. Help people pick which tool to use.

如果你想对竞争对手的注册流程做大量拆解，你会用什么工具？你可以用 Coda 来做，对吧？所以，如果你有带 Chrome 扩展的 Coda，让它去查找这些流程，然后给我展示哪些有效、哪些无效。这绝对是 Coda 能做的事情。酷。我觉得其中一个挑战是：你现在有这么多工具，不知道哪个适合自己。顺便说一句，那个 Chrome 扩展我一直在用，太棒了。我想进一步了解一下增长团队。前几天 Twitter 上有个梗，说 Anthropic 只有一个增长营销人员在推动所有增长，这听起来有点疯狂。你们有多少增长人员？增长团队的大致组织结构是怎样的？

If you want to go do a bunch of tear-downs of other competitors onboarding flows, what would you use? You can use Coda for this, right? So, if you have Coda with the Chrome extension, task it to go and look for these flows and show me a view of what's working, what's not. That's definitely something that Coda can do. Cool. I think that's one of the challenges: you have all these tools now and just which one's for me. By the way, the Chrome extension, I use it all the time. It's amazing. I want to drill a little bit further into the growth org. There's this whole meme on Twitter the other day of just like you have one growth marketer driving all growth at Anthropic and it's like, 'Okay, that's crazy.' What's like how many growth people are there? What's kind of the rough org structure of the growth team?

我们现在大概有 40 人左右。我认为我们的结构非常像传统的增长团队，我们有增长平台和变现的横向团队，负责考虑我们所有产品中的增长领域，然后我们有更多面向受众的增长小组。你可以想象 B2B 增长、Claude Code 增长、知识工作者增长、API 增长。所以这些受众需要保持狭窄的焦点，这是当你拥有这么多不同产品时必须做的事情。然后这些横向团队负责全面考虑问题。

We're roughly maybe 40 people now. I think we are structured very much like a traditional growth team in that we have horizontals of growth platform and monetization who think about the sphere of growth across the entirety of our products, and then we have more audience-focused growth pods. You can think about B2B growth, Claude Code growth, knowledge worker growth, API growth. So really these audiences to keep a narrow focus, which is a thing you have to do when you have so many different products. And then these horizontals that think about things across the board.

这是一个由工程师、设计师、产品经理组成的跨职能团队吗？增长组织内部有哪些职能？

And is it a cross team of engineers, designers, PMs? What's kind of the functions within the growth org?

是的，包括工程师、设计师、产品经理、数据人员。我认为整体组织形态与传统的增长团队非常相似。可能不同的地方在于，我们更倾向于大的赌注，而不是小的优化。比如，如果我想一个传统的增长团队，我可能会把 60-70% 的时间花在中小型赌注上，20-30% 花在大的赌注上，而对我们来说，我们翻转了很多。我们更多地走另一条路，大概是 70/30 或者更接近 50/50，而不是偏向小赌注。这可能是最大的变化之一。

Yeah, it's engineers, designers, PMs, data. I think that overall the shape of the org is quite similar to a traditional growth team. Probably the things that are maybe different is that we index a lot more towards larger swings as opposed to smaller optimizations. Like if I think about a traditional growth team, I would have probably done maybe 60-70% of my time on small to medium bets, 20-30% on larger swings, and I think that for us we flip it a lot. We do much more the other way, where it's sort of 70/30 or more like 50/50, rather than indexing towards smaller bets. That's probably one of the biggest changes, I think.

需要强调的是，有趣的是，在你们所处的规模上，1% 的胜利在整体上也是巨大的。所以，有趣的是，即使你们处于这个规模，你们也没有专注于这些微优化。这很容易，你可以轻松地专注于这些小优化，然后在季度末把它们加起来，然后说，‘看我们产生了多大的影响。’你可以这样做。

Just to highlight that, what's interesting there is at the scale you guys are at, a 1% win is massive in the scheme of things. So, it's interesting that even at the scale you're at, you're not focusing on these micro-optimizations. It is easy, you could easily focus on these small optimizations, and then you tally them up at the end of the quarter, and you're like, 'Look how much impact we made.' And you could do that.

再来十亿。没什么大不了的。

Another billion. No big deal.

是的，但问题是，我们一直保持着 10 倍的同比增长，这是我们牢记在心的。我认为这最终归结为我们公司对指数增长的执着。我认为如果你看 Anthropic 的人谈论任何事情，我们总是谈论指数增长。实际上，随着模型能力继续指数级增长，围绕它们的工具能够更好地将其扩散到有用的用例中，你基本上不断解锁新市场，这些市场的价值远远超过以前市场的价值。智能体式编码就是一个很好的例子。一年半前它还不存在。而现在，智能体式编码的价值已经超过了之前 AI 编码使用的市场。我认为这里的核心是，未来的产品价值比今天高一个数量级。我想一个普通的企业，比如一个交易应用或一个杂货配送产品。许多这些领先的公司都是伟大的企业，但如果我想一个标准的杂货配送应用，作为最终用户，你今天获得的产品价值与两年后相比如何？我认为两年后，即使你发布了所有这些新产品，作为最终用户，你从该产品中获得的价值可能只增加 30% 到 50%，如果公司在发布新功能方面做得非常好的话。但这并不是指数级的。所以，如果我想，好吧，你今天在这里，两年后你将拥有 30% 到 50% 更多的产品价值。那么作为一个增长团队，相对于今天，两年后产品价值的相对差异，我实际上可以通过中小型优化（通常具有更高的确定性）来捕获其中的相当一部分，而不是大的赌注。但对于 Anthropic 来说，情况并非如此，因为指数增长和我们的产品价值来自 AI，我们在两年内交付的产品价值可能是今天的 100 到 1000 倍。所以，如果我想这个，有太多的价值可供选择。你需要更多地转向，好吧，我们需要下更大的赌注，然后我们不能只见树木不见森林。这就是为什么我们仍然做所有的小优化。这真的很重要，而且没有其他人会做这些事情，所以我们需要做。复合价值并非微不足道，但我们也承担了更大的核心产品类型的赌注。

Yeah, but the thing is we've been tracking at 10x year-on-year, and that's the thing we keep in mind. I think it ultimately comes down to our fixation in this company about the exponential. I think if you look at anyone talking from Anthropic about basically anything, we always talk about the exponential. It's effectively as model capabilities continue to grow on an exponential, and the tools around them enable a better job of diffusing that into useful use cases, you basically just keep unlocking new markets where the value of those markets significantly dwarfs what the value of the previous markets were. Agentic coding is a great example. It didn't exist a year and a half ago. And now just the value of agentic coding is bigger than the previous market of AI coding use. And I think that is the core thing here, which is that the future product value is an order of magnitude higher than it is today. I think about a normal business, like a trading app or a grocery delivery product. Many of these leading companies are great businesses, but if I think about what is the product value that a company like a standard grocery delivery app gives you as an end user today versus two years from now? I look at it as again two years from now, even if you're shipping all these new products, as an end user the value you get from that product maybe goes up 30 to 50% if the company's done a really good job of shipping new features. It's not exponential though. And so if I think about okay, you're here today, in two years you're going to have 30 to 50% more product value. Then as a growth team, that relative differential of the product value two years from now relative to today, I can actually capture a decent percentage of that with the small to medium optimizations that typically have higher conviction as opposed to larger bets. But for Anthropic it's not really that way because of the exponential and our products being very much the product value coming from AI, the product value that we will deliver in two years time is probably a hundred to a thousand times what it is today. And so if I think about that, there's so much value on offer. You need to shift more towards okay, we need to take larger bets and then we need to not miss the forest for the trees. That's why we still do all the small optimizations. It really matters and no one else is going to do some of these things, so we need to do it. The compounding value is not immaterial, but we do take on much larger core product type of swings as well.

所以你提到 Chrome 扩展现在是支撑 Coda 和 Claude Code 上许多用例的东西。那是增长团队构建的。那是一个充满 AI 的产品，一个研究密集型的产品，但我们就是看好它。我们有一个工程师非常看好它，我们就说，‘嘿，没有其他人做这个。我们要做。’所以，这种事情我在其他公司是不会做的。哦，哇。这我还真不知道。

So you mentioned that the Chrome extension is now the thing that underpins a number of use cases on Coda and Claude Code as well. And that's something that the growth team built. That's a very AI-filled product that is a very research-heavy product, but we were just bullish on it. We had an engineer who was very bullish on it, and we were just like, 'Hey, no one else is doing it. We're going to do it.' And so, that's the sort of thing that I wouldn't have done at another company. Oh, wow. That's I did not know that.

那么，这里的一个要点是，你的建议中有些东西只在 Anthropic 成立，而其他公司可以从你的经验中学到什么。一是你的感觉是，如果你在 AI 领域工作，应该把更多的资源转向更大的赌注，因为未来的机会如此之大，你想尽快找到它们，而不是微观优化？

So, the takeaway here is that there's stuff to extract from your advice that is only true at Anthropic, and then there's what other companies can learn from this experience. One is your sense that if you're working in AI, shift more of the pie chart towards bigger bets because in the future, the opportunity is so large, you want to find those as soon as possible versus micro optimize?

更具体地说，如果你的产品的主要价值是由 AI 作为核心元素支撑的，那么你应该这样运作。像 Lovable、Cursor 这样的公司，随着指数级增长，它们的价值主张也会显著提升。如果你在构建一个 AI 优先的产品，我肯定会这样运作。如果你构建的产品并非 AI 优先，而只是附带一些 AI 功能，我不确定我会这样运作。这取决于产品的其他部分如何配备人员，以及增长团队如何与之相关。

To be more specific, I'd say that if the primary value your product delivers is underpinned by AI as a central element, then you should operate this way. Companies like Lovable, Cursor, all these great businesses, as the exponential rises, their value props are also going to continue to rise significantly. If you're building an AI-first product, I would definitely operate this way. If you're building a product where AI is not the core, and you have some AI features on the side, I don't know that I would operate this way. It would depend on how the rest of the product is staffed and how growth is staffed in relation to that.

然后关于你们的组织结构，我觉得很有意思。这是不同事物的组合。有 API 增长、Claude Code 增长，还有用户画像，比如知识工作者和 B2B 的垂直领域。这是有意为之吗？比如一些具体的赌注和一个广泛的市场机会？

And then in terms of the way you are structured, I thought that was really interesting. It's a combination of different sorts of things. There's the API growth, Claude Code growth, but then also personas, like a vertical of knowledge workers and B2B. Is that intentional? Like some specific bets and one just kind of broad market opportunity?

当你只有一个产品时，更容易有一个纯粹专注于漏斗的增长团队，对吧？你有转化、激活、变现。但当你开始有多个产品时，我认为这更难了，因为如果你只有一个激活团队，但你有 Code、Co-work 和其他东西，它们的受众非常不同，内部的跨职能利益相关者也不同。所以我们正在寻找一种能让我们保持专注的结构。专注在受众和问题上非常重要，与跨职能合作伙伴的联系也非常重要。我们 Claude Code 增长团队的人与 Cat、Boris 以及该团队的其他成员合作非常紧密。所以这种联系也非常重要。

When you have one product, it's easier to have a growth team that's more purely on the funnel, right? You have conversion, activation, monetization. But as you start to have multiple products, I think that's harder because if you have one activation team for example, but then you have Code, Co-work, and all the other things, they're very different audiences and very different sets of cross-functional stakeholders internally. So we're looking at what structure allows us to have as much focus. Focus is a really big thing on audience and problems, and also the ties to cross-functional partners is really important. Our folks in the Claude Code growth team work extremely closely with Cat and Boris and the others on that team. So that tie-in is really important as well.

你在很多不同的公司做过增长，基本上是在 Anthropic 之前的传统增长。随着 AI 产品和 AI 初创公司的兴起，增长作为一个职能和技能组合还在如何变化？

You've done growth at a lot of different companies, basically traditional growth before Anthropic. How else is growth as a function and as a skill set changing with the rise of AI products and AI startups?

我认为如果你的核心产品价值非常依赖 AI，那么它正在转向更大的赌注，而不是中小型实验。与此相关的另一件大事是，我们开始研究如何自动化增长。我们的增长平台团队有 Alexey Komissarouk，他在 Reforge 教授增长工程，他正在推动一个名为 CASH（Claude 加速可持续黑客增长）的项目。这不是我起的名字。但这是一项努力，研究如何使用 Claude 来自动化增长实验。它仍然非常小，非常早期。我们几个月前才启动。在 Opus 4.5 之前，这几乎不可能，但最近随着 Opus 4.6，我们看到它朝着正确的方向发展。我认为这将在整个行业越来越普遍。如果你考虑交付的生命周期，有四个部分：识别机会、构建功能、测试和确保质量，然后分析数据。我们从这四个方面思考，并评估 Claude 在每个方面的表现。我们一直在小规模测试，涉及文案更改和微小的 UI 调整。它确实产生了结果。你可以按下播放键，它就能赚钱。胜率大约相当于一个 2-3 年经验的初级 PM，还不到高级 PM 的水平。但几个月前这根本不可能，所以它正在快速进步。我认为它将改变你进行更大规模实验的方式。但我没有提到的一个部分是跨职能利益相关者管理。仍然需要人类的大脑。

I think if your core product value is very backed by AI, then it is shifting where you skew more towards larger bets as opposed to smaller medium experiments. Another big thing related to this that I think will accelerate is we are starting to look at how to automate growth. Our growth platform team, we have Alexey Komissarouk who teaches growth engineering at Reforge, and he is driving an effort called CASH — Claude Accelerates Sustainable Hacker Growth. I didn't come up with that. But it's an effort to look at how we can use Claude to automate growth experimentation. It is still very small and very early. We kicked it off only a couple of months ago. Before Opus 4.5 it wasn't really possible, but more recently with Opus 4.6 we're seeing it head in the right direction. I think this will happen more and more across the industry. If you think about the lifecycle of shipping, there are four parts: identifying opportunities, building the feature, testing and ensuring quality, and then analyzing data. We think about it in those four ways and score how good Claude is in each area. We've been testing this at small scale with copy changes and minor UI tweaks. It delivers results. You can press play and it prints money. The win rate is like what I'd expect from a junior PM with 2-3 years of experience, not quite at senior PM level. But it wasn't available at all a couple of months ago, so it's getting better rapidly. I think it will change where you can do this for larger experiments. But the one piece I didn't mention is cross-functional stakeholder management. There's still a need for human brains.

我认为这意味着，在我看来，产品经理的工作短期内不会消失。对于较大的项目，你不需要像小项目那样做那么多，对吧？你可以跳过，但对于较大的项目，这部分工作不会消失，直到其他利益相关者都变成他们自己的小智能体在四处运作。

And I think that one is going to mean that in my eyes, the work of PMs is not going away anytime soon. That piece, especially for larger projects, you don't need to do as much of it for smaller stuff, right? You can skip it, but for larger stuff, that piece is not going away until the other stakeholders are their own little agents running around.

没错。我认为那才是改变的时候。有趣的是，几周前我们开了一个很困难的会，之后我和我的设计主管 Joel 进行了复盘。他认为，而且他直接说：“哦，就算有了 AGI，让六个人在一个房间里达成一致仍然是不可能的。”我回答说：“是啊，我觉得我能理解。”这就像是，哪个对齐问题更难？

That's right. I think that would be the point where that changes. It's funny. We had a difficult meeting a couple of weeks ago, and me and my head of design, Joel, we were debriefing afterwards. He thinks, and he's just like, "Oh, we will have AGI and it will still be impossible to get six people in a room to align." And I'm like, "Yeah, I think I can see that." That's like, what's the harder alignment problem?

好的，这太有趣了，这正是我感觉事情发展的方向。所以澄清一下你分享的内容，基本上有一个工具会提出要运行的实验，帮助增长 Claude 和所有工具。它提出想法，有人查看并批准，好，我们来做这些事，构建、发布、测试结果，看看效果如何，然后它反馈说，这些是有效的东西。大致是这样吗？

Okay, this is so interesting and this is exactly where I feel like things are going. So just to be clear what you shared here, there's basically this tool that comes up with experiments to run to help grow Claude and all the tools. So it comes up with an idea, somebody looks at them, approves, cool, let's do these things, builds it, ships it, tests the results, sees how it's doing, and then it comes back with like, here's things that are working. Is that roughly right?

大致正确。目前我们有一个人工审批环节，但我无法想象以这种方式进行 Scaling——每周看各个领域是否在进步？人们花在每个领域的时间是否减少？每个领域的结果是否变好？只要每周都在变好，你就会觉得，好吧，整个计划正在 Scaling。所以大致正确，但你可以认为其中很多可以自动化，不需要人工审核。我们非常在意品牌，对吧？所以目前我们确实会检查这一点。我们不想发布违背品牌的东西。但你可以有一个技能，包含你的品牌指南和非常明确的品牌注意事项。所有这些配套的东西，我认为会越来越好。模型也会越来越擅长使用它们，所以随着时间的推移，我认为人工审核的需求会显著减少。

It's roughly right. Right now we have a human in the loop approving, but the amount of time... I can't think about scaling this way of just week on week: are things getting better in each of the areas? Are people spending less time on each of the areas? Are your results getting better in each of the areas? And as long as week on week that's getting better, then you're like, okay, this whole initiative is scaling. So that's roughly right, but you can think about it as a lot of this can be automated where human review is not needed. We care a lot about brand, right? So that is something we do look at right now. We don't want to be shipping something that goes against the brand. But then you can have a skill that contains your brand guidelines and very clear do's and don'ts on brand. All of these types of accompaniments, I think, are going to get better. The models then get better at understanding how to use them, so over time, I think the need for human review decreases significantly.

是的，而且你随时可以撤回，如果觉得，哦，那其实不是个好主意。这一点说得很好，我们以为永远需要人来做这些事，结果发现，一个技能就能做得很好。这是我们的品牌指南，这是我们的愿景，这是我们的使命，这是我们的目标，这是我们在乎的东西。好了，别发布那个东西。

Yeah, and you could always unship it if it's like, oh yeah, that was actually not a great idea. And that's such a good point that we think we need people to do these things forever, and it turns out, okay, a skill could do this really well. Here's our brand guidelines, here's our vision, here's our mission, here's our goals, here's what matters to us. Okay, let's not ship that thing.

所以我认为这非常有趣，就像我一直在观察 AI 在产品开发过程中承担更多工作，在增长过程中，从帮你写代码，到写所有代码，再到审查代码。现在感觉另外两端是什么？它似乎是从中间向外扩展。顶端是决定做什么，对齐仍然很难。另一端是审查代码、发布和获取分发，这完全是另一回事，我想和你聊聊。所以我在这里听到的，正是我认为将要开始发生的事情：AI 现在非常擅长告诉我们该做什么，而不仅仅是接受指令并构建。感觉增长版本是起点，因为它更简单——不是说增长容易，而是它数据驱动，有你提到的那个循环。所以我认为这是一个非常有趣的信号，预示着产品领域的未来。

So the reason I think this is so interesting, this is like I've just been watching the expansion of AI doing more of the product development process, in the growth process, of just, okay, it from helping you write code to like writing all your code to reviewing your code. Now it feels like what are the other ends of this, the two ends around this? It's kind of going from the middle out. The top of that is coming up with what to do, and then there's like the alignment stuff still very hard. And then on the other end it's reviewing the code and then shipping it and then getting distribution, that's like a whole other thing I want to talk to you about. So what I'm hearing here and this is exactly what I thought was going to start happening is AI is now getting really good at telling us what to do, not just taking our orders and building it. And it feels like the growth version of this is where it starts because it's so much simpler, not that growth is easy, but just like it's data driven, there's this loop that you talk about. So I think this is such an interesting sign of things to come across just generally product.

是的，确实如此。只是提一下。

Yeah, definitely. Just putting that out there.

好的，我一直在思考的一个问题是产品、产品经理、工程角色的未来，这些角色如何随着我们讨论的内容而演变。你们作为三人组是如何协作的？你认为这些角色会走向何方？这三个角色中，哪个变化最大？

Okay, so something that I'm constantly thinking about along these lines is just the future of product, PM, engineering, how those roles shift over time based on the stuff we're talking about. How are you working together as a triad? And where do you see these roles going? What's most going to change do you think across these three roles?

这是我们经常讨论和思考的事情，情况变化很快。过去当事情出问题时，通常是你需要招聘更多工程师、更多设计师等，这更像是团队和特定小组的生命周期问题，需要找出瓶颈所在。但现在当事情出问题时，你仍然需要审视：是比例问题，还是也有潜在的技术转变导致问题？所以这很有趣。我认为在较小的公司，比如一个 15 或 20 人的公司，只有一个产品经理与一些设计师和工程师合作，你会看到最大的融合，产品经理会做各种事情：设计、发布等。我认为会出现极端分化。在较大的组织、更规模化的组织中，我认为尚无定论。即使在这里内部，不同团队的人对角色融合与分离的程度也有不同看法。我想说的是，甚至很多产品经理也在发布代码、提交 PR 等。但根据我的观察，很明显，虽然产品经理和设计师从 AI 中获得了更多杠杆，但工程目前获得的杠杆最大。我看看像 Claude Code 这样的工具，工程师从中获得的杠杆比设计师和产品经理今天获得的要高。虽然这正在迅速变化，但这是我今天的看法。所以如果你考虑一个默认团队，比如五个工程师、一个设计师、一个产品经理，有了 Claude Code，那五个工程师的效率提高了 2 到 3 倍，对吧？产品经理和设计师的效率也提高了，但他们现在实际上是在管理一个更大的工程师团队。

This is something we talk about and think about frequently, and the picture changes rapidly. Sometimes when things break, historically in the past, it's like, okay, now you need to hire more engineers, you need to hire more designers, etc., and it's more just a lifecycle thing of where the lifecycle of your team and that specific pod is to identify where the bottleneck is. But now when things break, you still need to look at it: is it the actual ratio, or is there also underlying technological shifts that are causing this to break? So that's an interesting thing. I think that smaller companies, if you find me a 15 or 20 person company with the only PM there working with some designers and engineers, I think in the smaller companies you'll see probably the biggest blend, where the PM will be doing all sorts of things: designing, shipping, etc. And I think you just have extreme bifurcation. At larger organizations, more scaled organizations, I think the jury is still very much out. You speak to people even internally here, and different people in different teams have different views of how much these roles are coming together versus how much they will be separate. I think that it's not to say that even the PMs, a number of PMs are shipping and pushing PRs, etc. But if I look at what I'm seeing, I think it's clear that while PMs and designers are getting more leverage from AI, engineering is getting the most leverage right now. I look at tools like Claude Code, and the amount of leverage engineers are getting from them is higher than the leverage that designers and PMs are getting from them today. Now this is rapidly changing, but that is my view today. So if you think about the default team, which is say five engineers, one designer, one PM, with Claude Code, that five engineers is like 2-3x'd, right? And the PMs and designers have also increased, but they are now managing what is effectively a much larger group of engineers.

所以尽管组织架构中的人数没有变化，但现在你面对的情况是：过去可能有 15 到 20 个工程师、1.5 到 2 个产品经理、1.5 到 2 个设计师。我们看到这给产品经理和设计带来了很大压力。并非所有团队都这样。我看像 Claude Code 这样的团队，因为产品非常技术化，产品经理本身基本上都是工程师，所以可能还好。但前几周我们有一个产品负责人来现场，大家都在讨论，普遍感觉产品经理和设计被压得喘不过气来。我们就在想，是不是应该大量招聘产品经理？这可能是最终的解决方案。

And so even though the head count in the org structure hasn't changed, you're now dealing with a situation of maybe 15 to 20 engineers in the old world, 1.5 to 2 PMs, and maybe 1.5 to 2 designers. So we're seeing that putting a lot of strain on PM and design. It's not everywhere. I look at teams like Claude Code and I think that because that product is so technical, it's probably the right thing where the PMs are all basically engineers themselves anyway. But we had a product like a PM lead on site the other week and we were all talking about this where across the board their feeling is PM and design is just squeezed. It's just absolutely squeezed. And we're like, is it the right thing here that we just need to hire a ton more PMs? That could actually be where we land.

关于增长，我是这么想的：第一，我们正在招聘一批增长产品经理。我们迫切需要非常优秀的人才，我们正在招人。所以如果你对我们做的事情感到兴奋，并且了解增长，请随时申请。我很乐意聊聊。也许可以给你写一封很棒的冷邮件。是的，尽管写那封邮件。所以这是第一点：我认为我们会招聘很多产品经理。但第二点是，我们非常注重招聘有产品思维的工程师。我认为这在增长领域一直是最好的做法。你总是希望工程师能提出想法。所以我们特别需要那些能够在产品经理缺席时充当迷你产品经理的人。由于我们目前人手非常紧张，我们正在更正式地利用这一点。我们的框架是：如果一个项目需要不到两周的工程时间，那么工程师就要负责有效地充当该项目的产品经理。这意味着要跟安全、法务和跨职能利益相关者沟通。工程师要主导这些。产品经理会被拉进来，在需要时提供建议。如果事情严重偏离轨道，他们会介入，但更多是顾问角色而非执行角色。如果一个项目超过两周工程时间，那么默认产品经理应继续负责确保项目顺利进行。他们会更多地向工程授权，但他们对结果负全责。这并非完全一刀切。要动脑子。如果是一个一周的项目但极具争议，产品经理可能仍应主导。但我认为更多公司会开始采用这种方法：让工程师充当迷你产品经理。不是每个人都能做到。所以那些更有产品思维的工程师，他们的价值会显著提升，比如一个数量级。而且我认为我们可能仍然会招聘很多产品经理。

You know, on growth how I think about it is like one, we are hiring a number of growth PMs. We desperately need people who are very very good and we are hiring. So if you are excited by what we're doing and you know growth, please feel free to apply. I would love to chat. And maybe craft an amazing cold email to you. Yeah, feel free to craft that email. So that's one: I think we are going to be hiring a number of PMs. But the second thing we do is we very much hire product-minded engineers. I think this has always been the best thing to do in growth. You always want to have the engineers coming up with ideas. So we especially want people who can really step in as mini PMs if the PM is absent. So we're basically more formally leveraging that right now because we are so stretched. The frame we have is that if a project is less than two weeks of engineering time or less, then the engineer is on the hook to effectively be the PM for that. That means things like talking to security, legal, and cross-functional stakeholders. The engineer is very much driving that. The PM will get looped in and advise if needed. If something is wildly off track, they'll step in, but they're much more in an advisory capacity versus execution. If a project is more than two engineering weeks, then the default is that the PM should continue to be on the hook for making that go well. They'll delegate more to eng, but they're squarely accountable. It's not fully clean cut. Use your head. If this is a one-week thing but extremely controversial, the PM should probably still drive it. But that's the approach I expect more companies will start to do: deputize the engineers to be mini PMs. Not everyone can do it. So the engineers who are more product-minded, their value goes up significantly, like an order of magnitude. And I think we will probably still be hiring a lot of PMs.

这里有很多有趣的东西我想跟进。好，第一个是关于两周这个想法。我作为产品经理经常开玩笑说，你可以去度假离开团队几周，一切都会没事的。有 momentum，有计划，人们会继续运作。感觉这有点像你用的经验法则：如果是两周的项目，没有产品经理也没问题，你能搞定。我喜欢这两者的联系。好，另一个点非常有趣。你说因为工程师速度大大加快，这很合理，产品经理和设计就有点跟不上工程的速度了。你说要么需要越来越多的产品经理来跟上，要么就是工程师本质上能当产品经理。这很有趣。就像说，好吧，产品经理的用途很好，直到更多产品经理的工作能被 AI 完成。但这是一个非常有趣的趋势。我不知道，还有别的吗？比如，哇，我们可能真的需要更多产品经理。更多产品经理、更少工程师的比例可能是未来。

There is so much interesting stuff I want to follow up on here. Okay, so one is this idea of two weeks just briefly. I always joke as a PM, you can go on vacation and be away for a couple weeks from your team, and things are going to be all right. There's momentum, there's a plan, people will keep operating. And it feels like that's kind of a rule of thumb you use: if it's a two-week project, you'll be all right without a PM, you can handle it. I love that those two connect. Okay, the other here is so interesting. So, you're saying here that because engineers are so accelerated and this all makes sense, PMs and design are kind of just like holy sh— there's like hard to keep up with the space of engineering. And what you're saying is you need more and more PMs to keep up. That's one route, or it's engineers that can PM essentially. Which is so funny. It's just like okay, great use for product managers until more of the PM-y stuff can be done by AI. But that's a really interesting trend. I don't know, is there anything else there just like oh wow, we actually may need more PMs. The ratio more PMs, fewer engineers might be the future.

是的，我认为这完全取决于行业和公司规模。任何构建更偏开发者产品的公司，你会更多地依赖工程师。早期公司没有那么多跨职能协调、利益相关者对齐这些需要产品经理的琐事。所以你可以用更少的产品经理。但随着公司规模扩大，假设比例从 1 个产品经理变成 2 个，5 个工程师变成 20 个，1 个设计师变成 3 个。那么产品经理的时间最佳用途是什么？我认为这是一个非常有趣的问题：产品经理应该花多少时间自己交付东西，多少时间做其他事情。在工程资源有限的世界里，产品经理肯定应该自己交付东西。在当今世界，这是了解工具的好方法，非常重要。所以产品经理应该为此而交付。但如果我是一两个产品经理，面对 20 个工程师，我会想我的时间能带来多少增量价值。是交付第 21 个产品经理功能，还是更好地引导团队找到正确的机会？所以我认为在这个世界里，你可能有很多像迷你产品经理一样的工程师。这种情况越好，我就越愿意多做。但即便如此，如果你能招到一个非常优秀的产品经理，他能把“为什么”和“做什么”提升 5%，那将是一个高杠杆的招聘。

Yeah, I think this really depends on the industry and the size of company. Any company where you're building something that's much more developer focused, you're going to rely on the engineers a lot more. Earlier companies don't have as much of this cross-functional coordination, stakeholder alignment nonsense that you need all these PMs for. So you can get by with less. But as a company scales, then if I think about okay, now you have this ratio where maybe the one PM became two PMs from productivity, the five engineers became 20 engineers, the one designer maybe became three designers. If you think about what is the best use of time for that PM? I think this is a really interesting thing: how much should PMs be actually shipping things themselves versus everything else. In the world where you're limited on engineering, the PM should definitely be shipping things. In today's world, this is a good way to get an understanding of the tools, which is really important. So the PM should be shipping for that reason. But if I'm one PM or two PMs and there are 20 engineers, I think about what is the incremental value I can add with my time. Is it actually shipping the 21st PM feature, or is it saying how am I getting a little bit better at guiding the team on what the right opportunities are? So that's where I think in this world, you may have all these engineers who are like mini PMs. And the better that happens, that's where I would love to be doing more of. But still, if you then get a really good PM who can come in and improve the why and the what particularly by 5%, that is such a high leverage hire.

你提出的这个见解非常有趣。这与很多人对产品经理角色演变的看法截然相反。我听到的是，因为产品经理落后太多，工程师完成的工作太多，很多人认为产品经理需要做原型、提交 PR。而你说的，我完全同意，是你的时间最好花在帮助产品经理本身，以及帮助工程师成为更好的产品经理上。这种杠杆作用在大多数情况下远高于你花时间写代码提交 PR。

This is such an interesting insight you're making. It's so counter to how a lot of people are thinking PM is evolving. What I'm hearing here is because PMs are so behind, because engineers are just getting so much done, there's like many people here, okay, you need to be prototyping, you need to be shipping PRs as a PM. What you're saying, which I completely agree with, is your time is much better spent helping PM basically, and helping the engineers become better PMs themselves. And the leverage there is a lot higher than you spending time coding shipping PRs in most cases.

我认为在某些情况下确实如此。但作为一家较小的公司，我不确定情况是否相同。在一个人人全力以赴的小公司里，你可能需要不断交付产品。如果你在一家预算非常紧张、工程师人手不足、无法因为资金无限而随意招聘的公司，那么你需要做必要的事情来加速团队的影响力。因此，在很多情况下，作为产品经理，正确的做法是说：不管了，我要交付，我要坚持到底。但如果我在这里谈论的是更大的公司、更规模化的业务，如果你有 20 名工程师，那么你最高杠杆的时间是利用是交付一个额外功能，还是想办法提升我们正在做的一切，获得更好的用户洞察等？而且我认为还有通过交付来学习的元素，构建一个原型以便你能有更好的判断。很多人说他们只是尝试三件事，看看效果如何，这有助于制定路线图。现在的 PRD 是用来查看的，而不是用来讨论的。所以这仍然很有价值。

I think that's true in certain circumstances. I think that as a smaller company, I don't know that that's the case. In a smaller company where it's all hands on deck, you probably need to be shipping. If you're in a company where budgets are very tight and engineers are very tight, and you can't just hire because money's unlimited, then you need to do what is needed to accelerate the impact your team is going to have. So there will be a number of cases where as a PM, the right thing to do is to say, screw it, I am shipping, and I am going all the way. But if I'm talking here more about larger companies, more scaled businesses, if you have 20 engineers, is it the highest leverage use of your time to ship an extra feature, or figure out how to up-level everything we are doing, get better user insights, etc.? And I think there's also an element of shipping to learn, building a prototype so that you can have a better opinion. A lot of people talk about just trying three things, seeing how it goes, and that will help inform the roadmap. Now the PRD is to look at it instead of talking about it. So there's a lot of value there still.

这非常正确。我认为这是一个很好的观点。所以即使对我现在来说，我们有很多产品经理和工程师，有时候我想表达我脑海中的想法，对我来说更好的方式就是做一个原型并展示出来。所以我认为这非常重要。而且我们非常灵活。所以我们名义上和估值上是一家大公司，但我们极其灵活，内部的重点就是尽量减少官僚主义，直接行动。所以我们交付的产品中大概 70%，可能 60-70-80% 没有 PRD。我反感 PRD。我讨厌文档。我就是说：做、做、做，砍掉障碍。20-30% 的事情很重要，必须做对，文档应该非常好，人们应该花很多时间在上面。但总的来说，我认为 PRD 现在已经过时了，你可以直接和一个好的团队启动项目，而不需要做那种事情。

That's very true. I think that's a great point. So even for me now, where we have a number of PMs and many engineers, there are certain times I want to articulate the idea I have in my head. It's just better for me to prototype it and show it. So I think that is really important. And we are very scrappy. So we are like a big company by name and valuation, but we are extremely scrappy, and the internal focus is just to minimize bureaucracy and just go. So probably 70%, maybe 60-70-80% of what we ship does not have a PRD. I am averse to PRDs. I just hate documentation. I just say go, go, go, cut the blockers. 20-30% of stuff where it's important and really important to get right, the documentation should be really good and people should spend a lot of time on it. But by and large, I think PRDs are outdated at this point, and you can just kick things off with a good team without needing to do that sort of thing.

再多说说这个。你如何帮助确保——因为人们可以构建得很快，或者花很多时间走错方向，交付的不是你要求的东西？你如何启动一个项目并明确我们正在做什么？只是对话还是有其他方式？

Say more about that. What do you do to help make sure, because people can build so fast or spend a lot of time going in the wrong direction, ship things that are not what you are asking? How do you kick off a project and clarify here's what we are doing? Is it just a conversation or is there anything beyond that?

这真的取决于事情的大小。这又回到了两周的事情。我们之所以有那个两周的事情，更像是一个过滤器，用来判断如果我们对某件事投入很大，我们应该在背后投入更多思考；而如果是一组较小的投入，就直接去做。作为一个增长团队，你确实有很多这样的小事。所以对于非常小的改动，比如某个功能即将上线，我们需要一个升级销售。我们在做什么？这只是在 Slack 上，纯粹在 Slack 上。就是来回发消息。我认为这也取决于工程师的水平。工程师能理解，‘嘿，我知道你说了这个，但那个对用户来说怎么样？’所以我们很幸运，我们有很好的产品思维工程师。但所有这些小事基本上都是在 Slack 上你来我往，这就是你做的。对于更大的事情，我非常坚信要有一个正式的启动会。这个地方事情太多了。没人有时间。没人知道发生了什么。所以做一个跨职能的启动会，让法务、安全保障、所有人都参与进来，然后说：这是我们计划做的。你关心什么？你关心什么？你关心什么？那个 30 分钟的会议对于更大的事情仍然非常重要，可以简化如果你不早期做这些工作后面可能出现的所有混乱。

It really depends on the size of it. This goes back to the two-week thing. The reason we have that two-week thing is more like a filter to say if we are investing heavily into something, we should apply more thinking behind it, versus if it's a smaller set of investments, just go for it. As a growth team, you do have a decent amount of these smaller things. So for very small changes, like there's a thing coming, we need to have an upsell for it. What is the thing we are doing? This is just on Slack, purely on Slack. It's just messages back and forth. I think it also depends on having a good caliber of engineer. Engineers can understand, 'Hey, I know you said this, but what about that for the audience?' So we are lucky we have good product-minded engineers in that sense. But all these smaller things are very much just on Slack back and forth, and that's what you do. For the larger things, I very firmly believe in a proper kickoff. There is so much going on at this place. No one has time. No one knows what's going on. So doing a cross-functional kickoff, get legal, get safeguards, get everyone in the room and just say, this is what we plan to do. What do you care about? What do you care about? What do you care about? That 30-minute meeting for larger things is still so important to streamline all the mess that may happen later if you don't do that work early on.

在这些情况下，你如何撰写 PRD？是像对 Claude 随意说一通？你有模板吗？

What's your approach to crafting that PRD in those cases? Is it like ramble into Claude? Do you have a template?

即使在这些情况下，有时我也不写 PRD。因为一切都在快速变化。所以仍然有一些情况，我只是安排会议。然后在会议前 5 分钟，我会把它放到 Coda 里，比如我的一些想法，我需要考虑的事情。快速创建一个基本文档，我们用这个来讨论。其他时候我甚至没有文档。但如果我要创建 PRD，我基本上有一个技能，我认为他们创建了，然后里面有所有以前的 PRD 项目。这很简单。我有固定的格式。所以我会说：这是我关心的事情。这是为什么。这是问题。然后详细说明关键考虑因素。详细说明跨职能的利益相关者，诸如此类。但同样，我的默认做法是，如果我能避免文档，如果我们能直接行动，那么我们就应该这样做。而且我越来越倾向于直接跳到原型制作。

Even in those cases, there are times I don't craft a PRD. Because everything is moving so quickly. So there are still some of those cases where I just set up the meeting. And then 5 minutes before, I'll put it into Coda, like here are some of my thoughts, here are the things I need to think about. Spin up a basic doc and we use it to talk. Other times I won't even have a doc. But if I am creating a PRD, I have basically a skill that I view they've created, and then there are projects with all the previous PRDs in there. It's pretty simple. I have the format down. So I'll just say, here are the things I care about. Here's the why. Here's the problem. And flesh out the key considerations. Flesh out the cross-functional stakeholders, those types of things. But again, my default is if I can avoid the doc and if we can just jump to action, then that's what we should do. And increasingly I just jump to prototyping the thing.

是的，我认为一旦我们能自动化更多这方面的工作，就会变得非常有趣。就像你的助手与法务助手对话，解决所有这些小问题，你知道，什么对你重要，什么对……

Yeah, that's where I think it'll get so interesting once we can automate more of that. Just like your assistant talking to the legal assistant, ironing out all these little, you know, what's important to you, what's important to the...

是的，而且它正在到来。我认为法务在这方面做了很好的赋能。我们研究这些版本，思考如何模拟某人可能说的话并设置一个引用。所以我们现在有类似的东西。然后随着 Claude 获得越来越多的上下文，并且更擅长传递长上下文，我认为 Claude 会越来越擅长知道这些。我使用 Claude 最有效的方式之一，或者说最有趣的方式之一，是帮助你识别不一致。我发现这真的非常非常有帮助。所以通过 Coda，你有这个 Slack MCT。

Yes, and it's coming. And I think the legal has done good enablement around this. We work in these versions of how you can think about how you can mimic what someone might say and set up a quote. So there are things like that that we have now. And then as Claude gets more and more context and gets better at passing long context, I think these are things that Claude will just get better at knowing. One of the most effective ways I use Claude, or one of the most interesting ways I use Claude, is to help you identify misalignment. This is something I have found really, really helpful. So with Coda, you have this Slack MCT.

然后你可以告诉 Coda 去查看 Slack，了解我正在做的项目，这些是当前最重要的事情，然后找出潜在的不一致之处。它做得非常好。这是我每周都会安排的任务。Claude 会查看各种信息，然后回来告诉我：‘嘿，我觉得这些事你应该注意。’所以在发货的背景下，类似地，Claude 可以查看公司里发生的事情，然后说：‘你正在考虑发货这个东西，这是你需要沟通的人，这是你需要记住的事情。’

Then you can tell Coda to say basically look across Slack, you know the projects that I'm working on, these are the things that are top of mind, go and find me areas of potential misalignment right now. And it does a really, really good job. So, this is something that I have scheduled, runs every week. Claude's like looking at things and coming back to me with and saying like, 'Hey, I think these things you should be aware of.' And so, you can think about it in that shipping context, it's a similar thing where Claude can basically be looking at what's happening across the company and say, 'You're thinking about shipping this thing, here's who you need to talk to, here's what you need to keep in mind.'

这太酷了。

That is so cool.

我非常兴奋地告诉你本季的赞助商 Vanta。Vanta 帮助超过 15,000 家公司（如 Cursor、Ramp、Duolingo、Snowflake 和 Atlassian）赢得并证明客户的信任。得益于 AI，团队比以往任何时候都更快地构建和发布产品。但因此，产品和业务中引入的风险也比以往任何时候都高。我交谈过的每一位安全负责人都感到保护组织、业务以及客户数据的压力越来越大。因为事情发展太快，他们不断被动应对，不得不猜测优先级，并凑合使用过时的解决方案。Vanta 通过超过 35 个安全和隐私框架（包括 SOC 2、ISO 27001 和 HIPAA）自动化合规和风险管理。这帮助公司快速合规并保持合规。信任比以往任何时候都更能成就或摧毁你的业务。了解更多请访问 vanta.com/lenny。作为本播客的听众，你可以获得 Vanta 的 $1,000 优惠。网址是 vanta.com/lenny。

I am so excited to tell you about this season's supporting sponsor, Vanta. Vanta helps over 15,000 companies like Cursor, Ramp, Duolingo, Snowflake, and Atlassian earn and prove trust with their customers. Teams are building and shipping products faster than ever thanks to AI. But as a result, the amount of risk being introduced into your product and your business is higher than it's ever been. Every security leader that I talk to is feeling the increasing weight of protecting their organization, their business, and not to mention their customer data. Because things are moving so fast, they are constantly reacting, having to guess at priorities, and having to make do with outdated solutions. Vanta automates compliance and risk management with over 35 security and privacy frameworks, including SOC 2, ISO 27001, and HIPAA. This helps companies get compliant fast and stay compliant. More than ever before, trust has the power to make or break your business. Learn more at vanta.com/lenny. And as a listener of this podcast, you get $1,000 off Vanta. That's vanta.com/lenny.

我觉得这就是产品经理的工作，以及它们如何逐渐被 AI 支持或完成。所以存在这种不一致，就像细微的不一致，然后也许有一天它会像最初对齐人们一样。你谈到了这种用于增长的现金自动化。比如，我如何增长这个产品？这已经开始发生了。另一个我自己在构建东西时使用的就是问：我如何让这个产品更好？我如何让用户体验更好？然后让 AI 给我一堆想法，它们实际上非常好。所以有趣的是，这些小部分开始被整合起来，承担越来越多的这个角色。

I feel like there's this, what are the jobs of a product manager and then how are they slowly going to be supported or done by AI. So, there's this misalignment, just like fine misalignment, and then maybe one day it'll be like aligning people initially. You talked about this cash automation for growing. Like, how do I grow this product? That's starting to happen. Another that I use myself when I'm building stuff is just asking, how do I make this product better? How do I make this a better user experience? And asking the AI just give me a bunch of ideas, and they're actually really good. So, it's interesting how these little pieces are starting to be put into place to do more and more of this role.

我很好奇你和你的团队还有哪些有效的自动化。我喜欢这些对话的地方在于你基本上生活在未来。你在最前沿的公司工作，与最有才华的人一起，使用最先进的工具，你可以看到事物的发展方向，并开始真正生活在未来，构建别人甚至没有想过或做不到的事情。我很好奇还有什么其他有效的方法，你的团队还做了什么来帮你节省时间、提高效率。

I'm curious what other automations you have and your team have that are effective. What I love about these conversations is you're basically living in the future. You're working at like the most bleeding edge company with the most talented people, with the most cutting-edge tools, and you can like see where things are going and start to actually live in the future, build things that nobody else has even thought about or can do. I'm curious just what else is working, what else your team has done to help save you time, be more productive.

我们广泛地使用它，对吧？所以有标准的 PM 工作，比如写文档、头脑风暴、查看数据。对于数据，我个人有一个同事按计划运行，每天早上查看大约 20 到 25 个不同的图表，然后我早上来的时候，有太多图表、太多产品需要跟踪。Cowork 会告诉我：‘好的，这些是你应该关注的事情。这些是令人担忧的，还有一些有趣的见解。’

We use it pretty extensively across the board, right? So, there's the standard PM stuff, like writing docs, brainstorming, looking at data. For data, I personally have like a coworker runs on a schedule and looks at sort of 20 to 25 different charts every morning and then when I come in the morning, there's just so many charts, so many products to track. Cowork will tell me, 'Okay, here are the things that you should pay attention to. Here's like what is concerning and here are just some interesting insights.'

它是在 Slack 里给你发送更新吗？工作流程是怎样的？

And it sends you the update in Slack or what? How does the workflow work?

对我来说，它就在 Cowork 里显示。Cowork 有一个桌面应用，你可以创建定时任务，所以我有一堆 hex 链接，它会去查看。它用 Chrome 扩展做某些事，用 MCP 做其他事，然后给我一个摘要。我知道，我还是有一些图表喜欢自己看，因为我就是那种人，我喜欢图表。所以它们都在右边。我喜欢看图表，但还有长尾的东西，甚至中尾的东西，你每天没时间看。如果 Claude 主动查看，随着时间的推移，你会感觉良好，Claude 带给你的误报率在下降，漏报率也在下降，这样你就更有信心和安心了。

For me it just shows it comes up in Cowork. So Cowork has a schedule desktop app. Yeah, in the desktop app you can have a scheduled task that you create and so I have a bunch of hex links that it will go and look at. It uses the Chrome extension for some things and uses MCP for other things and then it will just give me a summary and then I know, I still there's some few charts that I just like to look at because I'm this guy, I just like charts. So they're all up to the right too. So I like to see the chart but then there's a long tail of things and even the medium tail where it's like you don't have time to look at it every day that if Claude is proactively looking at it and you start to feel good over time of okay, it's like the false positive rate is going down, the false negative rate is going down of things that Claude brings to you then you just get a little bit more confidence and peace of mind there.

这触及了我一直以来的一个想法，团队会有一个策略机器人，想象一个 AI 不断观察指标、市场、路线图、什么有效什么无效，然后说：‘嘿 Amol，这是我认为我们现在应该做的。这是我们应该采取的转向。这是我们将要获胜的地方。’感觉我们离那很近了。

This touches on an idea I've always had that teams will have is a strategy bot which just imagine an AI that's constantly watching metrics, the market, the road map, what's working and what's not and just like, 'Hey Amol, here's what I think we should do now. Here's the pivot we should take. Here's where we're going to win.' Like it feels like we're very close to that.

我认为我们很接近了。我觉得今年晚些时候我们就能达到那个点，变得非常非常有效。这是我的直觉。我认为那种主动性和跨上下文查看并提炼洞察的能力，就像我之前提到的对齐部分。那是它的一个版本，对吧？现在你只是查看更多的数据源。

I think we're close. I think we'll get there later this year to the point where that's very very effective. That's my gut. I think that that level of proactivity and getting looking across a bunch of context and distilling insights, I think it's like the thing I mentioned earlier, Lenny, about the alignment piece. That's like a version of it, right? Now you're just looking at across more data sources.

这是我用的东西。我谈过一些标准的 PM 工作，比如头脑风暴、数据、用户研究。有很多行政事务我讨厌，比如生活行政和文书工作。我讨厌它。所以我让 Claude 帮我订会议室。Claude 归档我的邮件，初步清理收件箱。我不做任何报销和费用。Claude 会去 Brex 提交报销。他会去 Brex 提交费用。所以所有那些事情，我就交给 Cowork。摆脱它们。然后我觉得很有趣的是管理视角，我谈到的对齐是一方面。我也查看我的直接下属，Claude 可以查看他们这周做了什么？查看我们的团队目标和 OKR。查看我们讨论的转录，我基本上可以问 Claude 我应该记住的关键要点和观察，以及你认为我应该给他们什么反馈。这也是你可以每周设置的事情。但质量目前时好时坏。有些方面还不错。

This is something that I use. So I talked about some of the standard PM stuff with it, you know, brainstorming, data, UXR. There's a lot of admin stuff I just hate, like life admin and paperwork. I just hate it. And so I get Claude to book my meeting rooms. I get him to book meeting rooms. Claude archives my email, first pass at clearing up my inbox. I don't do any of my reimbursements and expenses. Claude will go to Brex and file the reimbursements. He'll go to Brex and file the expenses. So all of that side of things I'm just like just hand it to Cowork. Just get rid of it. And then the stuff that's quite interesting, I think, is the manager lens where I talked about alignment is one thing. So I look also across my direct reports like Claude can look into what have they done this week? You look at our team goals and OKRs. Look at the transcripts from our discussions and understand I can basically ask Claude like what are the key takeaways and observations I should keep in mind and what feedback you think I should give them. And that's something that again, you can just set that up weekly. But the quality is like hit or miss right now. It's like it's decent on some things.

有时候你会想，“天哪，我真高兴我发现了这个。”这非常有帮助。我自己也这么做。基本上，我的一位经理 Amol Avasare 是你播客的嘉宾，对吧？所以我会说，“嘿，根据你对 Amol 的了解——她公开写了很多关于产品的文章，以及内部讨论——基于我这周所做或未做的一切，你作为 Amol 会给我什么反馈？”我每周都会得到这样的反馈。所以很多事情今天已经可以做到了。我认为随着模型的改进，这些事情的准确性和信号会继续快速提升。我不觉得你意识到你分享的这些有多棒。每一个都让人惊叹。好，所以这个 Amol 的例子。你和她有一对一的会议。你让 Claude 阅读她所有的文章，建立她的模型，然后问根据你对 Amol 的了解，我应该做哪些不同的事情？这就是区别吗？

Sometimes you're like, "Holy, I am so glad that I caught this." And that's very helpful. And I do that for myself as well. So basically, one of my managers, Amol Avasare, was a podcast guest of yours, right? So I say, "Hey, based on what you know of Amol, both publicly—she's written extensively about product—and internally, and our discussions, what feedback do you have for me as Amol?" And I get that every week. So a lot of these things can already be done today. I think as the models improve, the accuracy and signal of these things will continue to improve rapidly. I don't think you realize how much awesomeness you're sharing here. Every one of these is like, "What?" Okay, so this Amol example. You have one-on-ones with her. You ask Claude to go through all her writing, build a model of her, and ask what you should be doing differently based on what you know about Amol? Is that the difference?

是的，基本上是这样。有很多方法可以做到这一点。如果某人有公开资料，写了很多文章，那很有帮助，因为 Claude 可以获取所有信息。否则，你可以创建一个项目或技能，但 Claude 越来越擅长理解，因为你可以告诉 Claude，“在 Cowork 上查一下。”你可以说，“使用 Slack MCP，查看这个人上周说的所有内容，等等。基于此，最重要的优先事项是什么？我的经理根据他们的时间安排有哪些我不知道的优先事项？”所有这些层面，我把它看作是软性指导。我认为这已经可以实现了。就像和一个有时有点醉醺醺的教练一起工作——不是醉——但有时他说了些什么，你会想，“你为什么要提这个？这明显是错的。”但其他时候你会想，“哇。”我们企业团队的负责人 Scott 有几次发现了重大的对齐问题，这些问题本会导致团队白费力气或做重复工作。想想影响：在这个世界里，在一家大公司，交付往往受到跨职能协调的限制。我认为我们现在开始看到，在跨职能协调层面，AI 确实可以用来减少这种辛劳。六个月前这还不可能，我不知道六个月后可能实现什么。

Yeah, effectively. There are a number of ways to do this. If someone has a public profile where they've written extensively, it's helpful because Claude can get all that information. Otherwise, you can have a project or a skill, but increasingly Claude is better at understanding because you can tell Claude, "Look on Cowork." You can say, "Using the Slack MCP, look at everything this person has said in the last week, etc. Based on that, what are the top priorities? What are my manager's priorities based on how they're spending their time that I don't know?" All this layer of stuff, I think about it as soft coaching. I think that is unlocked in my opinion already. It's like working with a coach who's kind of drunk at times—not drunk—but sometimes says something and you're like, "Why would you bring that up? That's clearly wrong." But other times you're like, "Wow." There's one of the guys, Scott, who leads our enterprise team. There were a few cases where he found major areas of misalignment that would have caused teams to spin their wheels significantly or do overlapping work. You think about the impact of that: shipping in this world, in a bigger company, is often constrained by cross-functional coordination. I think we're now starting to see at that cross-functional coordination layer that AI can really be used to reduce that toil. Six months ago that wasn't possible, and I wonder what will be possible six months from now.

所有这些数据都在那里——Slack、Granola 或人们使用的任何工具——所有这些来自对话和讨论的笔记确实是让这一切发挥作用的关键。

And the fact that all this data is there—Slack, Granola, or whatever people use—all these notes from conversations and discussions are really key to this working.

是的。

Yes.

那么对于想要做类似事情的人，如何设置？步骤是什么？

So for someone who wants to do something like this, how do you set it up? What are the steps?

去 Cowork，下载 Cowork 的桌面应用。连接 Slack MCP。根据组织的大小，你可能需要团队或企业管理员权限来启用它。但一旦你在 Cowork 上连接了 Slack MCP，你只需要问 Claude 就行。就这么简单。

Go on Cowork, download the desktop app on Cowork. Connect the Slack MCP. Depending on how big the organization is, you may need team or enterprise admin permissions to enable that. But once you have the Slack MCP connected on Cowork, you just ask Claude. That's it.

太棒了。好的，我要换个方向。我想谈谈 Anthropic 多年来一直保持的专注。看看你们取得的数字，真正令人难以置信的是你们的专注。我认为这是它如此成功的原因。市场上某个竞争对手正在意识到他们本应这样做，并开始转向类似的方法。作为一个外部观察者，感觉 Anthropic 非常擅长做很少的事情，但做得非常深入。例如 B2B，只是深入 B2B，然后深入编码用例。Claude Code 就是一个例子。而且效果非常好。是谁在推动这种专注？是谁从一开始就帮助保持了这种专注？

Amazing. Okay, I'm going to go in a different direction. I'm going to talk about the focus that Anthropic has had over the years. If you look at the numbers you're putting up, what's really incredible is the focus you've had. I think this is the reason it has worked out so well. There's a certain competitor in the market that is realizing they should have done this and are starting to shift to a similar approach. As an external observer, it feels like Anthropic has been very good at doing very few things but going super deep. B2B, for example, just going deep on B2B and then going really deep on coding use cases. Claude Code being an example. And it's worked out really well. Who has been driving that focus? Who has helped keep that focus from the beginning?

我认为这是公司的基础部分。从很早开始就存在了，而且真正来自领导层。他们在提炼这一点上做得非常出色。我最近看到一份文档——是我们的一位创始人 Dan Mad 写的。日期是 2021 年，公司成立几个月后。内容大概是，“为什么我们应该专注于 AI 编码。”这是五年前，远在任何人知道实际市场机会之前。我认为这是我们内部从一开始就对编码和 B2B 重要性的深度专注。有两个视角：一个是这可能在商业上有利可图，另一个是加速研究。现在这已经是一个相当主流的观点——我听到各个实验室的人都在谈论编码对加速研究的重要性。但这是我们内部一直非常坚定的观点：如果你有最好的模型，那将加速你的研究循环。我认为 Dario 多年来一直非常清楚地看到了这一点。所以我会说很多来自 Dario，来自领导层，来自我们的 DNA。第二个可能是必要性。现在情况变了——我们是一家更知名的公司，筹集了很多资金等等——但历史上，我们是这个领域最小、资金最少的参与者。在很多方面，我们能走到今天这一步完全是个奇迹。我们没有 Meta 或 Google 的自由现金流或分发渠道，也没有 OpenAI 的先发优势。

I think it's a foundational part of the company. It's been there from the very early days, and it really comes from leadership. They've done a phenomenal job of distilling that. I saw a doc come up recently—it was something Dan Mad, one of our founders, had written. It's dated 2021, a few months after they started the company. It was like, "Here's why we should just focus on AI coding." This is five years ago, long before anyone knew the actual market opportunities around this. I think this is a deep focus we have had internally from the start on the importance of coding and B2B. There are two lenses: a view that this could be commercially beneficial to tackle, and also the side of accelerating research. It's a pretty mainstream view now—I've heard people from various labs talking about the importance of coding to accelerate research. But that has been a very firm view we have been laser-focused on internally: if you have the best models, that's going to accelerate your research loop. I think that's something Dario has seen very clearly for a number of years. So I would say a lot comes from Dario, from leadership, and from our DNA. The second is probably necessity. It's now changed—we're a more well-known company, raised lots of money, etc.—but historically we were the smallest, least well-funded player in this space. In many ways, it's a complete miracle that we've gotten to the stage we have. We didn't have the free cash flow or distribution of Meta or Google, nor the first-mover advantage of OpenAI.

所以，你该怎么办呢？我认为生活中有一个更广泛的原则——通过约束获得自由。当你身上有很多约束时，无论是个人生活还是工作，我认为这能带来很多自由，因为它释放了所有多余的选择。你会想，好吧，这显然是该走的路。对我们来说也是如此：你没有大量资金，你是小玩家，没有分销渠道。你只能选择一个非常狭窄的焦点，即使对于一项非常通用的技术，也要最大化你达到逃逸速度的机会。我认为这也与历史如何展开有关，对吧？所以，那远在我入职之前，但 Anthropic 有一个 Claude 的版本。我们在 ChatGPT 发布之前就有一个聊天机器人。但我们最终出于安全原因选择不发布它。我认为团队不想引发一场全球 AI 军备竞赛。而 ChatGPT 发布了，他们获得了惊人的增长，对吧？这自然就把他们拉向了消费者方向。在另一个世界里，如果 Anthropic 先发布了 Claude，也许情况会反过来，即使有那些专注的因素。所以，谁知道呢？回顾这些事情总是很难说。

And so like what do you do, right? I think there's this broader principle I have just around life of freedom through constraints. When you have a bunch of constraints applied on you, whether that's in personal life or at work, I think that can bring a lot of freedom because it just frees up all this excess choice. You're like, okay, this is clearly the path. And I think for us it's like, okay, you don't have a ton of funding, you're a small player, you don't have distribution. You just have to really pick a very narrow focus and even for a very generalizable technology to maximize your chances of getting to escape velocity. And I think that's also relating to how history played out, right? So, it was well before my time, but Anthropic had a version of Claude. We had a chatbot before ChatGPT was launched. And we had ultimately chosen not to launch it for safety reasons. I think the team didn't want to kick off effectively an AI global arms race. And ChatGPT launched and they got insane traction, right? And that just naturally sort of pulled them towards consumer. There's another world where if Anthropic had launched Claude first, maybe it did the other way around even with all that focus stuff. So, who knows? It's hard to say looking back at some of these things.

哇，我不知道这个。我认为人们也没有意识到 Anthropic 当时落后了多少。现在当然觉得他们很棒。但我记得当人们谈论 Anthropic 融资时，我想，他们不可能在这一点上与 OpenAI 竞争。已经结束了。他们领先太多了。而这显示了专注的力量，以及你们所做的一切。你们取得的进步、成功和变化之快，简直不可思议。

Wow. I didn't know that. I think also people just don't realize how far behind Anthropic was. Like right now it's like, of course they're amazing. But I just remember when people were talking about Anthropic raising money. I was like, there's no way they're going to compete with OpenAI at this point. It's like so over. They're so far ahead. And it shows the power of focus and just all the things you all did. It is absurd how far you all have come and how successful and how things have changed so quickly.

是的，我非常同意。我认为这很大程度上归功于我们的领导团队。我们有很多人曾在所谓的世界最佳公司工作过，都是非常资深的人。几乎所有人都一致认为，这是所有公司中最强的领导团队。所以我认为很多都来自他们，而且我们很幸运拥有优秀的人才，这也很有帮助。

Yeah. And I would say I very much agree. I think a lot of that comes down to our leadership team. We have a number of people who've worked at what I call the best companies in the world. Very senior people. And I think almost uniformly everyone's like this is the strongest leadership team out of any of these companies. So I think a lot comes from them and then we're just very lucky to have incredible people. I think that helps a lot as well.

关于编程部分，为了确保这一点清楚。我从未想过，在编程上押注如此之深的原因不仅仅是市场规模巨大，而是这将形成一个反馈循环，不断加速我们前进。所以如果我们成为编程最好的，编程将帮助我们进行研究，帮助我们构建更好的模型，并加速越来越快。

On the coding piece just to make sure that part is clear. I never thought about this that the reason the bet was so deep on coding is not just that's a huge TAM, but it's that this will excel this is a feedback loop that will accelerate us further and further. So if we get the best at coding, coding will help us to research, it will help us build better models and it will accelerate faster and faster.

是的，没错。我认为 Dario 公开谈论过这一点，所以我说出来没问题。

Yeah, that's correct. And this is something I think Dario has talked about this publicly so I'm fine to say it.

安全部分真的很有趣。所以我花点时间在这里。众所周知，Anthropic 的官方名称是 Anthropic 一家 AI 安全研究公司。作为一个负责增长的人，我想你在增长和使命之间取得平衡，使命是不计一切代价地增长。我们的目标是 AI 安全和对齐。你如何平衡这两件事？这对你的工作有什么影响？

The safety piece is really interesting. So I'll spend a little time here. So famously Anthropic, I think the official name of Anthropic is Anthropic an AI safety research company. As a growth person there's this balance I imagine you strike between growing and the mission being don't grow at all costs. Our goal is AI safety and alignment. How do you balance those two things? How does that impact your job?

听着，我认为我们非常非常重视这一点，这也是公司存在的全部原因。如果你仔细想想，从他们离开创办公司的原因开始，这一点就深深植根于我们的公司结构中。所以，历史上每个人融资时都会说，去创建一个特拉华州 C 型公司，这就是你要做的结构。而对于一家公司，你有信托责任去最大化股东回报，最大化股东价值。但从一开始，他们就走了一条不同的路。我们创建了一家公益公司（PBC），这让你可以合法地说，最大化股东价值并不是这家公司的首要目标。你可以为公共利益而优化。所以，我认为这从那里开始，并层层向下。对我们来说，我们的目标、我们的使命最终是确保向强大 AI 的过渡顺利进行，并对人类产生净收益。我们非常兴奋，老实说，对它的前景非常乐观。但我们也理解风险所在。所以对我们来说，这个首要目标——这必须对人类有利——是我们愿意为此承受重大商业打击的。而且我们一次又一次地这样做，对吧？所以，当时我们有 Claude，但我们不想发布它，因为存在安全风险。所以我一次又一次地看到，我们乐于承受这种打击，而且实际上它在其他方面对我们有利。从增长的角度来看，我认为增长团队常常会推动良好用户体验等的边界，因为他们有时会试图挤出指标。当我考虑是否要做一个有争议的测试时，我把它看作两种类型的有争议测试。第一种是测试如此有争议，以至于你根本不应该运行它，因为结果不重要，你不会因为品牌、客户友好度和价值观的综合原因而发布它。第二种是它有争议，哎呀，我不喜欢它，我当然不爱它，但它不是红线。所以如果有人带着信念来找你，说我对这个有一个非常好的假设，而你说我不喜欢它，但你可以运行测试，看看它有什么影响。如果它让人非常尴尬或糟糕，那么我希望看到结果有很高的回报。我基本上把所有事情都归为第一类或第二类，对每家公司来说，第一类和第二类是不同的。我认为对我们来说，AI 安全绝对属于第一类。这是我们存在的理由。我们就是不会做这件事。

Look, I think it's something we take very very seriously and it is the whole reason the company exists. And if you think about it, it's all the way from why they left to start the company. It's deep in our corporate structure itself. So, historically everyone raising money is like go create a Delaware C-corp and that's the structure you do. And with a corporation, you have a fiduciary duty to maximize returns to shareholders, maximize shareholder value. And from the beginning they went a different way. We created a public benefit corporation PBC, which allows you to legally say that maximizing shareholder value is not the overarching umbrella goal of this company. And you can optimize for public benefit. So, really I think it starts from there and it ladders down from there. For us, our purpose, our mission ultimately is to make sure that the transition to powerful AI goes well and is net beneficial for humanity. We are very excited, honestly, very optimistic about where this can go. But we also understand what the risks are. And so for us that top line objective of this just needs to go well for humanity is something that we are happy to take a significant commercial hit for. And we've done that time and time again, right? So, back then we had Claude, but we didn't want to release it because there was a safety risk. And so time and time again I've seen we've been happy to take that hit and it's actually worked out well for us in other ways. From a growth lens, I look at it as growth teams can often push the boundaries of what is good user experience etc., because they're trying to eke out metrics at times. When I think about if a controversial test is brought to me, I sort of look at it as two types of tests that are controversial. One is when that test is so controversial that you just should not run it because the results don't matter because you would not ship it for a combination of brand, customer friendliness, and values. And the second is where it's controversial, yikes, I don't like it, I certainly don't love it, but it's not a red line. So if someone comes to you with conviction and says I have a really good hypothesis around this, and you're like I don't love it but you can run the test and see what impact it has. And if it's a high level of cringe or yikes, then I want to see a high level of return for the result. I kind of think about everything in terms of is it in one or is it in two, and for every company that one and two is different. I think for us the AI safety is very much in one. It's why we exist. We're just not going to do this thing.

我觉得还有一些事情属于第二种情况，就是高度敏感，但我们可以试试看会发生什么。放大来看，Lenny，我觉得我看到增长团队，尤其是那些硬核增长实践者，犯的最大错误之一就是试图榨干每一分钱。这其实也是生活中的一个普遍原则：如果你是一个创始人，在融资时试图榨干最后一分钱，你其实不应该这么做，因为你希望别人下次还会回来，这是我的看法。我认为在增长中，你通常需要接受“留点钱在桌上”这一点，这非常重要。这是我们增长团队的一个核心原则：我们非常愿意为了优先考虑安全、品牌、保持高质量标准和维护出色的用户体验而放弃指标上的影响。如果你超越短期，问自己“这个季度的数字是多少？”，然后放大来看，想想那些最优秀的产品，你会发现它们都是这样运作的。而这实际上也是推动长期增长的关键。我认为这又回到了安全问题上：随着风险和利害关系越来越高，我们在安全上采取立场对我们所做的事情至关重要，而且实际上会成为我们一个显著的竞争优势，长期来看会帮助我们。

I think then there's other things that fall into two where you're like heightened sensitivity but we can try it and see what happens. I think zooming out, Lenny, one of the biggest mistakes I feel like I see growth teams make, and particularly just hardcore growth practitioners, is just trying to squeeze every last dollar. This is like a general principle in life too: if you're a founder raising money, just trying to squeeze out the last dollar, you don't want to do that because you want people to come back next time as well, in my view. And I think in growth, it's really important that you usually need to be okay leaving money on the table. That's a core principle for us as a growth team: we are very comfortable forgoing metric impact in order to prioritize safety, our brand, hold a high-quality bar, and maintain a great user experience. If you look beyond the short-term and ask, 'What are the numbers for this quarter?' and zoom out and think about what the very best products out there are, you realize this is how they all operate. And that's actually the thing that's going to drive more growth long-term as well. I think that ties back to safety: as the risks get higher and the stakes get higher, the fact that we are taking a stance on safety is critical to what we do and is actually going to become a significant competitive advantage for us that will help us in the long run.

是的，当你分享这些时，很明显它奏效了，Anthropic 做得非常出色。所以我喜欢那些例子，当有人用这种方式处理问题时，它真的能成功。这和我对我 newsletter 的想法是一样的。我可以做更多事情来增长它，但我的理念就是专注于创作好内容，其他什么都不管。所有这些微优化最终都不重要。如果内容对人们有用，它会通过分享来增长。所以在很小的规模上，我有类似的理念。

Yeah, and as you share all this, clearly it's working and Anthropic is killing it. So I'm loving those examples when someone approaches a problem that way, it actually works out. It's the same way I think about with my newsletter. There's so much more I can do to grow it. My philosophy is just focus on creating good content, nothing else. All these micro-optimizations are not going to matter in the end. It'll grow through people sharing it if it's useful to them. So on a very small scale, I have a similar philosophy.

在我们谈话时，人们可能正在想一件事。你知道，我们开玩笑说 AI 正在取代部分工作。这对很多人来说相当可怕。就像，‘我工作的未来是什么？我会有工作吗？我如何在这个未来中保持相关性？’所以这里有几个问题。一个是，对于那些想在即将到来的 AI 未来中茁壮成长的产品经理、增长人员，你有什么建议，他们现在应该做什么？

One of the things people are probably thinking about as we talk. You know, we joke about AI replacing parts of jobs here and there. It is pretty scary to a lot of people. Just like, 'What is the future of my job? Will I have a job? How do I stay relevant in this future?' So there are a couple questions here. One is just for folks that want to thrive in this approaching AI future as a PM, as a growth person, do you have any advice, things that they should be doing right now?

对我来说，有几件事。第一件事是每个人都在说的：使用工具。你需要掌握这些工具。我认为你需要使用 Claude Code，你需要使用 Cohere，并理解每个模型发布后你能用它做什么新事情，如何将其应用到你的工作中。它在某些事情上会表现很好，在其他事情上会很糟糕，然后一个模型之后，你会说，‘哦，那件事现在可以了。’但如果你没有回去尝试，你就不会知道，而几个月过去了，你都不知道那是可能的。所以掌握工具非常重要，既能提高你自己的生产力，也能培养对 AI 产品的产品直觉，我认为这只会变得越来越重要。然后，放大来看，我倾向于把它看作是利用你拥有竞争优势和不公平优势的地方。所以对我来说，就像：有些产品经理非常擅长工艺，而另一些产品经理，你把他们扔进一个充满各种利益相关者、每个人都有强烈意见的情况中，他们无法调解，但最终每个人都能朝着正确的方向前进。所以如果你想想你最大的技能集是什么，你在哪里特别突出，并且这可以转化为在产品角色中为公司带来影响，我会加倍投入在那上面，几乎忽略弱点。就专注于你能做什么来成为那方面最好的人？因为那非常有价值，而且我认为这与跨学科的理念有关。回到我们之前提到的：在这个工程师是迷你产品经理的世界里，一个高度产品思维的工程师是独角兽，绝对的独角兽。我认为产品经理也是如此：如果设计师非常紧张，而你是一个能设计的产品经理，你也是独角兽。公司解雇你的可能性大大降低，因为你现在更有用了。所以我认为这非常重要。如果我看是什么让我受益，可能就是这个版本。我有创始人背景，所以创始人背景、投资银行的金融背景，然后是销售——我差点成为客户经理而不是进入产品。我当时就在纠结是去销售还是产品？我认为这些组合加上增长，可能导致了在某些领域和情况下，我能比其他人产生更大的影响。理解对你来说那是什么非常重要。看看我们的金融服务产品。我们推出了 Quip for Sheets、Quip for Excel。它们会占领市场，顺便说一句，当 Quip 推出时，每个人都知道。但负责那个产品的人，Nick Lin，他来自投资银行，来自私募股权。他就有这样的竞争优势，他在构建那个产品时，就像‘我懂这个。我懂这个。’他就是为此而生的。所以我认为理解你可以利用哪些跨学科领域来让自己产生更大影响非常重要。最后一点就是保持适应性。任何试图继续应用旧剧本的人，我认为你会让自己的生活更加艰难。当你来到 Anthropic 时，最重要的事情之一就是你需要明白，过去你运作方式的 50%、60%、70% 可能都要扔掉。

To me, a couple of things come up. One thing that everyone says is use the tools. You need to be on top of the tools. I think you need to be using Claude Code, you need to be using Cohere and understanding each model release, what new things you can do with this, how you can apply this to your job. It'll work well in some things, it'll work terribly in other things, and then one model later it's like, 'Oh that other thing worked.' But if you didn't go back to try it, you would not have known, and now many months have passed and you didn't know that was possible. So being on top of the tools is really important both for improving your own productivity, but also for getting product sense around AI products, which I think is just going to become increasingly important. I think then, zooming out beyond that, I kind of look at it as leaning into where you have a competitive advantage and an unfair advantage. So to me, it's like: there are some PMs who are really good at craft, for example, and there are others who you'll throw them into a situation with all these stakeholders who have strong opinions, and there's no way they're going to mediate this and they come out and everyone's kind of swimming in the right direction. So if you think about what is the major skill set you have where you spike that can be tied to delivering impact for a company in a product role, I would just double down on that and almost forget the weaknesses. Just what can you do to become the best person at that thing? Because that's very valuable, and I think that ties to the notion of leaning into being interdisciplinary. Going back to some of what we mentioned earlier: in this world where engineers are mini PMs, the engineer who is highly product-brained is a unicorn, an absolute unicorn. I think the same thing is true for PMs: if now the designer is really stretched and you're a PM who can design, you are also a unicorn. The chances of a company letting you go have gone down dramatically because you are now so much more useful. So I think that is really important. I think if I look at what's benefited me, it's probably a version of this. I came from a founder background, so the mix of founder background, finance as an investment banker, and then sales—I almost became an account exec instead of going into product. I was right on the edge of should I go into sales or product? I think a combination of those along with growth is probably what's led to certain areas and situations where I can have outsized impact compared to other people. Understanding what that is for you is really important. I look at our financial services product. We've launched Quip for Sheets, Quip for Excel. They're going to take the market by the way, everyone when Quip launched. I know. But the guy running that, Nick Lin, he came from investment banking, he came from private equity. And he just has such a competitive advantage where he's building that product and he's like, 'I know this. I know this.' He's built for this. So I think just understanding what those interdisciplinary areas are that you can lean into to make yourself higher impact is really important. And then the last one is just being adaptable. Anyone who is trying to just keep applying old playbooks, I think you're going to make life a lot harder for yourself. One of the biggest things when you come into Anthropic is you need to understand that probably 50, 60, 70% of how you operated in the past, just throw it out the door.

那不会相关。如果你坚持那样做，你会遇到很多摩擦，而且不会有帮助。所以保持适应，理解工作已经变了方向，我就朝那个方向走，我认为这非常重要。这是很棒的忠告。这和 Jenny 在播客中分享的很多内容一致，她是 Claude 和 Claude 工作的设计负责人。哦，就像深入钻研、成为某个非常具体领域的最佳或最佳之一的想法。并不是每家公司都需要，你也不需要全部十项，但深入某件事。我忘了她是怎么描述的。Marc Andreessen 也说过同样的话。就像我们称之为“侧向 E”而不是 T 型或单一技能。如果你能有两三件非常擅长的事，那会很有力量。

It's not going to be relevant. And if you try to stick to that, you're going to have a lot of friction and it's not going to be helpful. So just being adaptable and understanding that okay, the job's changed this way. I'm going to go that way is I think it's so so important. That is awesome advice. It matches a lot of what Jenny shared when she came on the podcast, the design leader on Claude and Claude work and all these things. Oh, just like the idea of going deep, becoming the best or one of the best at a very specific thing. And that not every company will need and you don't need all 10, but just going deep on something. I forget what how she described it. Mark Andreessen said the same thing. Just like I think what we call it a sideways E instead of just like T-shaped or one thing. If you could have a couple things you're really really good at. There's a lot of power to that.

哦，天哪。好的。在我谈到一件我认为会让很多人震惊的事情——我们实际上是如何认识的以及你旅程中很重要的一部分之前，让我先问你这个问题。关于 Anthropic，还有什么值得分享或谈论的吗？

Oh man. Okay. Before I get into something that I think will blow a lot of people's minds about how we actually met and kind of a big part of your journey. Let me just ask you this. Is there anything else about Anthropic that might be worth sharing, might be worth talking about?

可能想到的是我们的文化和这里的人。我真的认为这是我们的秘方。我认为这是最难以被复制的东西，是别人无法复制的。而且我不认为这是偶然的。领导层在这方面投入了很多，Daniela 和 Dario 非常相信这一点，我认为这创造了一种非常特殊的文化。所以，我想说，这是一家真正使命驱动的公司。我加入时并不完全确定。我当时想，这是一家令人兴奋的公司。我非常同意他们的原则，但我不认识公司里的任何人。我对内部情况没有任何参考。所以我加入时有点怀疑。我想，至少他们在谈论这个，但我不知道他们是否认真。然后我很早就进来了。我意识到，哦，他们内部可能比对外说的更认真。所以，这是一家使命驱动的公司，人们发自内心地理解这项技术的好处和坏处，因此理解这对人类可能产生的不同走向以及未来可能有多么不同。我认为当你理解这对我们所有人、我们的孩子和孙辈的未来可能有多么不同时，它会对我们所做的事情产生巨大的热情，并带来巨大的信念。所以我回顾过去的所有工作，公司里总有一些人只是混日子，比如“我在这里，我厌倦了，但我没有更好的选择，或者我薪水太高离不开”等等。这里完全不是这样。我没有遇到一个人——我真的没有遇到一个混日子的人。每个人都全力以赴，倾尽所有，完全投入其中。我认为这释放了一种难以描述的能量。你有这种能量，有这种使命驱动的本质，还有如此开放的文化。领导层对我们非常透明。Slack 就像一个迷宫。很多事情都在 Slack 上发生。我们非常开放。每个人都有笔记本频道，有点像你自己的 Twitter 信息流，你在上面谈论你的想法，所以你可以加入研究人员的笔记本频道，学习任何你想学的东西，也可以花很多时间沉浸其中。这种开放性甚至鼓励人们与 Dario 争论。有一次全体会议，他说了一些话，有人不同意，然后那个人就进入 Dario 的笔记本频道说：“嘿，我不喜欢你这样说，等等”，然后引发了一场大辩论。但这种鼓励人们去领导层那里提出异议、公开挑战他们的做法，我认为这带来了信任。所有这些加在一起，意味着我们有非常深厚的团结感，我从未经历过类似的事情。然后还有人才，对吧？我认为人才密度很高，有时我感觉自己像是在为皇家马德里踢球。我环顾四周，心想，天哪，我在为皇马踢球，这里有世界上最好的人。我认为在研究方面尤其如此。我们有世界上最好的研究人员，但即使在产品方面，像 Amie Vora，她非常出色。我们有 Mike Krieger，他随便就创办了 Instagram，他在这里。在增长方面，我们有 Johnny，他是我的工程对应方，他是增长工程领域的 OG，他很棒。我们有 Alexa，他在 Reforge 教增长工程，他只是团队里的另一个人。所有这些都非常特别。我最喜欢的是，几个月前，在十月份的全公司现场活动中，我四处走动，看到一个家伙独自吃着爆米花。我走过去说：“你是 Jeff，对吧？”他说：“是的。”我说：“你实际上是美国驻澳大利亚大使，而你只是这里的一名员工。”我说这太疯狂了，和你谈论我们国家的总理等等。我认为人才加上这种文化就是秘方，也是我们如此成功的原因。

The thing that maybe comes to mind is just like our culture and the people that we have here. I really think it's our secret sauce. I think it's the thing that is the most defensible, the thing that no one else is going to be able to replicate. And I don't think it's an accident. Like leadership has really invested in this a lot and Daniela and Dario, they really believe in this a lot and I think it's just created a very special culture. So, I would say that, you know, this is like truly a mission-driven company. And I was not 100% sure of that when I joined. Like I was like I think that this is an exciting company. I very much agree with their principles, but I didn't know anyone at the company. I didn't have any references on what it was like inside. So, I was a little skeptical when I joined. I'm like, at least they're talking about it, but like I don't know. Are they serious about this? And then I came in very early. I was like, oh okay, like they are maybe they're even more serious about this internally than they talk externally. And so, it's just like it's a mission-driven company where people viscerally understand both the upsides and the downsides of the technology and therefore understand the divergent ranges of how this may go for humanity and how different of a future that could be. And I think when you understand how different this could be as a future for all of us and like our children and our grandchildren, etc. I think it leads to a lot of passion for what we do and it just leads to a lot of belief in what we're doing and so I kind of look at every other job I've had in the past, there's some degree of people at the company who are just checked out where it's like, you know, I'm here. I'm sick of this but I don't have a better option or I'm getting paid too much to leave, etc. That is just not the case here. I have not met a single person — I'm saying I'm really not a single person who's checked out. Everyone is putting everything they have on the table. Everyone is pouring it out and leaving nothing behind and is just fully in it. And so I think that leads to this releases energy that is just very hard to describe. I think you have that energy, you have that mission-driven nature and then it's such an open culture. So leadership is very transparent with us on things. We Slack is like a it's a whole maze. Every There are so many things that play out in Slack. We're very open. Everyone has these notebook channels where you kind of have your own like Twitter feed in a way where you're just talking about your thoughts about things and so you can go and join the Slack channel the notebook channels of people on research and all these other areas and you'll you can learn whatever you want and you can spend so much time getting lost in that as well. But that openness where we even encourage like people can just argue with Dario there was an all-hands he said something where someone didn't agree and then the person goes into Dario's notebook channel and just says, "Hey, I didn't appreciate how you said this and this and that" and then it sparked a whole big debate but that sort of thing where it's encouraged like go to leadership and disagree with them, challenge them publicly and I think that just leads to a level of trust and all of that together I think it just means that we have this very deep sense of togetherness that man I have just never I've never experienced anything like it and then you get to the talent, right? That I think that is the thing where the talent density is like I feel like I'm playing for Real Madrid at times. I look around and I'm like man I'm playing for Madrid where it's like you just have the best people in the world. I think it's most the case on research. We have the very best researchers in the world but even you look on product like Amie Vora she is phenomenal. We have Mike Krieger you're like okay casually started Instagram he's here you know on growth we have Johnny again who is my engineering counterpart he's the OG in growth engineering he's great. We have Alexa who's like guy teaches growth engineering at Reforge he's just another dude on the team and all of that I think is just very special. My favorite here is like in the linear a couple of months ago we had our on-site company-wide on-site in October and I'm walking around and I see this guy he's just walking around eating popcorn by himself. I go up to him and I'm like you're Jeff, right? And he's like I am and I'm like you are literally the US ambassador to my country Australia and you're just an employee here. I'm like this is insane and talking to you about prime ministers of our country and all these things and it's just the talent combined with that culture I think is this secret sauce that is the reason that I think we are as successful as we are.

笔记本频道作为技术工具非常有趣。所以这个想法是 Dario 分享它，就像一个内部 Twitter 信息流，人们分享他们的想法、优先事项等等。

That's notebook channel is so interesting just like as a technical thing. So the idea there is Dario just shares it's like a little Twitter like internal Twitter feed where folks just share what they're thinking about and what their priorities are and things like that.

是的，基本上就是这样。嗯，这是一个内部信息流，不只是 Dario，每个人都有。你基本上分享你的内心想法。这是一种让人们了解最新情况、了解什么在起作用的方式。

Yeah, that's basically it. Well, it's an internal feed where it's not just Dario, everyone has one. And you basically you share your internal thoughts. It's a way to keep people updated on things on what's working.

这是一种人们分享挑衅性观点的方式。从领导层面来看，我们也认为这是在组织快速扩张时，将你的信念和观点规模化传播的方式。如果你在增加很多人，你需要思考我们想要树立什么样的行为，以及哪些原则是最重要的。你可以开会，也可以以身作则，但如果你有一个渠道来分享你的首要想法，比如我上周发的一篇关于“乐于把钱留在桌上”的重要性的帖子，新加入的工程师看到后就会意识到，这和他们以前知道的不一样。所以这对开放性有好处，对领导者来说也是将自己的观点规模化、让团队快速了解做事方式的好方法，这对于避免漂移非常重要。当有很多新人加入时，漂移会导致战略和认知上的偏差。这有助于更严格地管理公司。

It's a way where people share provocative things. I think from a leadership level, we also think about it as a way to scale your beliefs and views across an org as it grows quickly. If you're adding a lot of people, you need to think about what behaviors we want to model and what principles are top of mind. You can have meetings and model behavior, but if you have a channel where you share your top of mind, like a post I did about the importance of being comfortable leaving money on the table, new engineers see that and realize it's different from what they knew before. So it's good for openness, and also for a leader to scale their views and get people up to speed, which is important to avoid drift. With many new people, drift can lead to strategy and perception drift. It helps run a tighter ship.

更重要的是，这些内容为运行的智能体提供了数据，帮助它们与所有人类协作。

And more importantly, it's data for the agents that run, to help them work with all these humans.

是的，非常正确。这些内容会进入 Claude。在入职文档中，HR 团队会注明：“在编辑此文档前，请先与某人确认，因为这是 Claude 引用的关键文档。”所以越来越多这类事情——增长团队怎么想，安全团队怎么想——你是在用这些上下文武装 Claude，让它未来能更好地提供帮助。

Yes, that is very correct. It goes to Claude. There are certain documents in onboarding where the HR team has written, 'Before editing anything on this document, please check with this person because this is a document that Claude references as a key thing.' So more and more, these types of things—how the growth team thinks, how safeguards think—you're arming Claude with that context to get better at helping in the future.

这很有意思。感觉每家公司都得开始这样做：以结构化的方式分享想法，这样我们运行的所有智能体都能获得它们需要知道的信息。

That's interesting. It feels like that's something every company will have to start doing: sharing their thoughts in a structured way so that all these agents we have running have what they need to know.

另一个有趣的旁注是 Slack。大家都在说 SaaS 工具会被 AI 取代，而你们还在用 Slack。有一条著名的推文说你们还在用 Workday 和所有这些 SaaS 工具。所有人都觉得它们会被消灭，但你们这些站在代码构建前沿的人仍然在用 Slack 和其他工具，这在我看来是一个好迹象，说明 SaaS 公司未来可能没事。不知道你对此有什么看法。

Another interesting side note is Slack. There's all this talk of SaaS tools being replaced by AI, and you guys use Slack. There's a famous tweet that you still use Workday and all these SaaS tools. Everyone thinks they'll be wiped out, but the fact that you, at the cutting edge of what can be built with code, still use Slack and other tools, is a good sign that maybe SaaS companies will be alright. I don't know if you have anything to share.

是的，情况非常复杂。我们内部自己构建了一些东西。随着时间的推移，Claude 在这方面做得更好的能力在提升。同时，我们大量使用 Figma、Slack、Workday 等许多前端工具，短期内我看不到变化。所以有些说法可能有一定道理，但其他部分被夸大了。这些产品中有很多是我们的客户；我们非常重视它们，并且大量使用它们的产品，我看不到这会发生改变。而且你有更重要的事情要做——谁会花时间在 Anthropic 内部构建一个 Slack？那不是价值会积累的地方。这也有助于你看到这些工具有多复杂；团队已经思考这些问题很长时间了。总之，这有点跑题了。

Yeah, it's a really complicated picture. There are a number of things we build internally ourselves. As time goes on, Claude's ability to do that better increases. At the same time, we use Figma a ton, Slack a ton, Workday, a lot of these front ends, and I don't see that changing in the immediate future. So there's probably some truth to some of this, but other parts are overblown. Many of these products are customers of ours; we value them a lot and use their products very heavily, and I don't see that changing. And you have better things to do—who's going to spend time building a Slack at Anthropic? That is not where value will accrue. It also helps you see how sophisticated these things are; teams have been thinking about these problems for a long time. Anyway, that's a tangent.

回到价值观这件事，我想强调一个重要的点。很多人批评 Anthropic 谈论对人类的危险，抛出关于工作岗位消失的数字，制造恐惧。人们认为我们是在试图筹钱、吸引注意力或者制造头条新闻。但我内部看到的一切始终是：这是我们的信念，我们希望人们知道可能会发生什么。即使结果不坏，我们也希望人们了解可能发生的情况，并且我们正在努力避免它。你可能会问，既然这么危险，为什么 Anthropic 还要构建 AI？Ben 分享的理解是，我们认为与其什么都不做、指望坏事不会发生，不如主动应对并尝试以正确的方式构建它。

Just coming back to the values thing, I want to highlight something important. A lot of people criticize Anthropic for talking about dangers to humanity, throwing out numbers about jobs going away, creating fear. People think we're trying to raise money or get attention or create headlines. But everything I've seen internally is always just: this is what we believe, we want people to know what might be coming. Even if it isn't bad, we want people to understand what might happen and we are trying to avoid it. You might ask, why is Anthropic even building AI if it's so dangerous? Ben shared the understanding that we think it's better to go at this and try to build it the right way rather than just hope nothing bad happens.

嗯，我想到三点。第一，回到我之前说的：我们以指数视角看待事物。如果你以线性视角看，你会看到完全不同的世界。你看看今天的情况，然后问，两三年后能好多少？如果你以指数视角看，并理解指数增长如何运作，你就会意识到，很多事情的到来会比人们用线性视角想象的要快。如果你明白可能有正面和负面结果以及一系列可能性，你就需要讨论负面结果以避免它们，并推动正面结果。公司里大多数人都是乐观主义者；我们对未来非常乐观。但我们理解风险——我们并不保证最终会处于一个好的境地，而且没有足够多的人以富有成效的方式讨论风险，尤其是那些了解风险内情的人。有些人可能谈论风险，但他们不在局中，所以并不完全了解具体的风险。我们在局中，我们了解正在发生的事情。第二，我认为我们实际上比对外表达的更坚信这些。这在我们内部思考中是非常关键的一部分，以至于有时我们需要重新措辞，因为人们可能觉得我们太夸张了，但内部我们就是这样想的。我们对外发布的是经过软化的版本。

Yeah, three things come to mind. First, I go back to something I said earlier: we think about things from an exponential lens. If you think from a linear lens, you see the world very differently. You look at where we are today and ask, how much better can it be in two or three years? If you look from an exponential lens and understand how exponentials work, you realize that a lot of this stuff will happen sooner than people think if they're looking from a linear lens. And if you understand that there could be upsides and downsides and a range of outcomes, you need to talk about the downsides to avoid them and push towards the upsides. Most people at the company are optimists; we're very optimistic about the future. But we understand the risk—it's not guaranteed that we end up in a good place, and not enough people are talking about the risks in productive ways, especially those who have the know-how. People may talk about risks but aren't in the game, so they don't fully know the specific risks. We're in the game, we understand what's happening. Second, I think we actually believe in this stuff more strongly than we say externally. It's such a key part of how we think internally that sometimes we need to reword statements because people might think we're being too over the top, but internally that's how we think. What we put out is a softer version.

嗯，然后我认为第三点是本说的，我们非常注重推动“竞逐顶峰”。如果你不在游戏中，只在场边喊叫，没人会在意。这就是世界的现实。如果你在游戏中，是领先玩家，并且你的做法有效，你就可以影响其他也在游戏中的人采取正确的行动。这就是核心。如果我们打包回家，你对这件事就没有影响力。如果你留在游戏中，商业上做得很好，那么你就有办法影响别人，比如“这些是原则”，并将其带入对话，让更多人相信这些。我可以和你聊很久，Amol，但还有两个问题来收尾，并触及我之前多次暗示的一些事情。一个是我想带我们去“失败角落”，因为听的人可能会想：“好吧，看看这家伙。在所有这些了不起的公司工作过，给 Anthropic 的 CPO 发了冷邮件，得到了工作，加入了这艘火箭船。一切都在上升，一直在成功。”你职业生涯中有没有事情不顺利、失败的经历？你从中学到了什么？

Um and then I I think the the third piece is that is is going to what Ben said, where we we think a lot about driving the race to the top. And so it's like if you're not in the game and you're shouting from the sidelines, no one cares. It's just the reality of of how the world works. Like no one no one really cares. If you're in the game and you're a leading player and what you're doing is working, you can influence people in who are also in the game to to take the right actions. And so that is like the core of it. Like if we just pack up and go home, you have no influence on this thing. And if you stay in the game and you're like commercially you're you're doing great, and then you have ways to influence the that okay, the these are the principles and put that into the conversation and and make more people sort of believe in those things. I could talk to you forever, Amol, but um two more questions just to round out the conversation and touch on some stuff that I've hinted at a number of times. One is I want to take us actually to failure corner because someone listening to this may be like, "All right, look at this guy. Just worked at all these amazing companies, cold emailed the CPO at Anthropic, got a job, joined this rocket ship. It's all been up into the right. Just killing it constantly." What's a story from your career where things didn't work out, when you failed, and what did you learn from that experience?

我有几个。我觉得完全不是那样。我感觉来到这里经历了很多曲折。最大的一个就是，创办一家公司，筹集资金，雇佣员工，然后不得不关闭它，告诉投资者你亏了钱，你曾经有一个愿景，但它不会实现了。这可能是最大的一个。我们花了 3 年时间。这不是兼职尝试，而是全力以赴。我们花了 3 年，筹集了几百万美元，最多时有 7 到 10 名员工，我们真的相信这件事。它围绕心理健康，如何量化心理健康以帮助理解或早期预测诸如广泛性焦虑、重度抑郁等问题。我们非常相信它，但最终，我们当时才二十出头，完全不知道自己在做什么，很多事情我会做得不同，但那是一个非常非常痛苦的过程。好的一面是，我们一直让投资者了解情况。所以对于任何挣扎中的创始人，请发送每月的投资者更新。事情顺利时很容易发送，但事情糟糕时很难。你坐下来，给投资者发更新，解释为什么上个月说的所有事情都没有按计划进行，但这是正确的事，能让人们了解情况，避免意外。但当你打电话说“我们要关闭了”或做出那个决定时，仍然非常艰难。每个人都相信你，这成了你身份的重要部分。非常非常艰难。那真是残酷。我花了好几年才真正走出来。那很艰难，但你从那段经历中得到了很多，当时很难看到。没有那段经历，我不会进入现在这个领域。我以前不是 PM，没有任何这些技能，从未做过产品，也不知道如何发冷邮件。正是通过那份工作，我学到了很多技能，使这条职业道路变得可行。在某个时间点很难看到这一点，回头看更容易画出那条线。但我的收获是，记住这是一场长跑。有些事，我现在很感激那段经历。我很感激它失败了，走向了那个方向。很痛苦，因为它不会导致我现在做的事情。所以，这很艰难，但积极的东西可以从中而来。

I have a couple. I think it's very much not that way. I feel like it was a lot of squiggly lines to to to get here. The biggest I'd say is just you know, founding a company, raising a bunch of money, having employees, and then having to shut it down and and tell your investors that you you've lost the money and that you you had a vision of what you're going to do and that it's not it's not going to happen. And so that I think is is probably the the biggest one. You know, we spent 3 years on it. It was not like a we tried something part-time. It was like no, we we went we went for it. We had spent 3 years, raised a couple of mil. We had maybe seven to 10 employees at our biggest and it was something that we really believed in. It was around mental health and how you can quantify mental health to help understand or get early predictors of things like generalized generalized anxiety, major depression. So like it's stuff that we really really believed in, but ultimately you know, we were like you know, our early 20s. We had like no idea what we're doing at the time and many things I'd do differently, but that was just a very very very painful process. Um I think the good thing was like we kept our investors in the loop the whole time. So to any founders who are struggling out there, like send those monthly investor updates. It's really easy to send those when when things are great. It's really hard to send those like when when things are bad. You just bat it down. You sit down and send an update to your investors about why everything you talked about last month like did not go go to plan, but it's just the right thing to do and it it keeps people in the loop and it avoids surprises, but it's still so tough when you're then calling to be like hey, like we are we're shutting down or making that decision. Everyone had believed in you and it becomes such a big part of the identity. It's very very tough. So man, that was just brutal. Like it's brutal. And it took me I think like a number of years to truly get over. I think that it's it's it's a tough thing, but you get so much from that experience that is hard to see in the moment. And I think without doing that I would not have gone into this. I was not a PM before. I didn't have any of these skills. I'd never worked on products. I didn't know how to cold email really. And And so, it was the through that job that I learned a ton of the skills that made this career path viable for me. And it's just really hard to see that in the moment when you're looking at a point in time. It's much easier to to like draw the line looking back, but I think that'd be my my like takeaway is just keep in mind that it's a long game. And some of those things like I'm so grateful for that experience now. I'm so grateful that it failed and went that way, actually. It's very painful because it wouldn't have led to what I'm doing now. And And so, it's just that it's it's a tough thing, but the yeah, positives can come from it.

我喜欢这一点，很多人在职业生涯中都有这样的时刻，觉得一切都完了。比如“我失败得这么惨，我的名声毁了，人们指望我，现在他们看到我是个冒牌货。我根本不知道自己在做什么。现在他们终于看清了。”看到这个，你花了 3 年，我看你的 LinkedIn，你有 7 个员工之类的。然后又是 MasterClass、Mercury、Anthropic。我觉得听到这样的故事很鼓舞人心，知道你可以经历这么大的失败，事情仍然可以成功。那不是终点。好的，2023 年，我本来要休陪产假，因为我妻子怀孕了，我想：“等等，我的 newsletter 怎么办？我需要休息一段时间。休一两个月假会很棒。”所以我的计划是安排一系列客座文章。我发出了客座文章征集。人们申请。我选了几个，提前安排好，这样我就可以休息了。所以我发出征集：“嘿，谁想为我的 newsletter 写客座文章？”我收到了 500 多份申请。其中一份来自你。你的提案基本上是“创伤性脑损伤如何让我成为更好的产品经理”。我还记得读你发给我的初稿时，我想：“天哪，这个故事太不可思议了。你能感受到那种情感。”而且它在战术上非常有趣和有用。比如，“哦，哇，这真的会让我成为更好的产品经理。”分享一下你经历脑损伤的故事和那段旅程，因为这会让人大开眼界。

What I love about this is a lot of people have these times in their career where they're like it's all over. Like I'm such a I failed in such a big way and my reputation's screwed and people were counting on me and now they see I'm I'm an impostor. I never knew what I was doing it after all. Now they finally see. And just seeing that And this was 3 years you had I'm looking at your LinkedIn. You had seven employees. Something like that. And and then it's like MasterClass, Mercury, Anthropic. Um I think it's inspiring to hear a story like that to know that you can have a big failure like that and things can work out. That is not the end. Okay, so in 2023, I was uh I was going to go on pat leave because uh my wife's pregnant and I was just like, "Wait, what do I do with the newsletter? I have I need to take some time off. That would be really nice to take a month or two off." And so, my plan was I'm going to I'm going to do a bunch of guest posts where people line up. I put out a call for guest posts. People apply. Uh I pick a few and then I kind of slot them in ahead of time so I could take time off. So, I put out this call, "Hey, who wants to write a guest post for my newsletter?" I got 500-plus applications. One of those applications was from you. And the pitch was essentially how a traumatic brain injury made me a better product manager. And I still remember reading the first draft sent me and I was just like, "Holy This story is incredible. It's just like like you feel such feelings." And it was just so tactically interesting and useful. Like, "Oh, wow, this is actually going to make me a better product manager." Share the story of this of the the brain injury that you went through and just the journey that you went on there because this is going to blow people's minds.

那是我人生中最艰难的时期。它让关闭公司这件事变得像“哦，那其实没什么”。视角就是这样有趣。早在 2022 年初，我遭受了创伤性脑损伤。我练 MMA 很多年了，也练泰拳很多年了，从未出过问题。但就是那种情况，一次错误的训练，一次错误的……只是普通的一天对练，没什么特别的。你头部被错误地击中了一下。然后我的整个生活都变了。基本上，我休了 9 个月的假。头几个月非常残酷。我花了大约半年时间才重新舒适地走路。

It was the toughest time in my life. It made shutting down a company be like, "Oh, that was nothing, actually." It's just funny how perspective works that way. Uh back in uh early 2022, I had a traumatic brain injury. So, I've done MMA for many years. I've done Muay Thai, which is a type of martial arts, for many years. Never had a problem and just It's just like the one of the things that happens, like the wrong session, the wrong It's just a normal day of sparring, nothing crazy. You get the wrong hit to the head in the wrong way. And uh my whole life changed. And and basically, I spent 9 months I was off work for 9 months. The first couple of months were brutal. It took me roughly half a year till I was comfortable walking again.

那段时间非常非常艰难。头两三个月，除了洗澡和上厕所，我妻子为我做了一切，包括帮我给朋友发短信。我可能听音乐 20 秒就会觉得想吐。我完全不能看屏幕。恢复过程非常漫长。有一段时间，我甚至不确定自己还能不能继续工作。我们甚至和妻子讨论过，如果真那样了该怎么办等等。我们不得不在那个层面上去思考。通过大量的坚持和真正的自我努力，你必须慢慢提高自己对事物的耐受度。这是一个残酷的过程。但你需要慢慢接触不同的事物，在每一件小事上不断进步。要主动去努力。不要逼得太紧，否则会大倒退。基本上到了那个点，8 个月后我恢复了工作，然后慢慢好转。但 Lenny 你不知道的是，2023 年中我们发了那篇 newsletter，一个月后我实际上又受伤了。

It was very very difficult. The first 2-3 months beyond just showering and going to the bathroom, my wife did everything for me, like including texting my friends. I would listen to music for maybe 20 seconds and feel like I needed to vomit. I couldn't look at screens at all. And it was a very very long recovery. It was not clear to me that I would ever work again for a while, actually. And we'd even discuss with my wife like what would we do in that case, etc. And we had to think even on those levels. And through a lot of pushing and really working through myself, you have to just slowly increase your tolerance to things. It's a brutal process. But you need to slowly expose yourself to different things and just get better and better at each little thing. Actively work on that. Don't push it too far, otherwise you have a big setback. They basically got to the point then where it was an 8-month recovery, so after that I returned to work. And then slowly got better. And the part Lenny you don't know is, in mid-2023 we posted that newsletter and a month later I got reinjured actually.

哦，哇。

Oh, wow.

因为当你脑部受伤后，在完全康复之前，再次脑震荡或脑损伤的风险会升高。当你完全康复后，风险会降到与未受伤人群相同。但在那之前，你需要达到 95% 的恢复，而不是 100%。那只是一次日常生活中的小事，下飞机时行李撞到了我的头。然后我休了两个月假，当时我刚加入 Mercury 一个月。我说，抱歉各位，我需要离开三个月。然后那又是一次非常漫长的恢复。实际上我到现在还没有 100% 康复。所以大部分时候还好，但有时会头晕、头痛，还有其他问题需要应对。

And because when you have a brain injury until you're 100% healed, your risk of another concussion or brain injury is elevated. When you're 100% healed, your risk falls down to that of a non-concussed population. But until that time, you need to be 95% healed, you're not 100% healed. And it was just an innocuous sort of everyday life thing, getting off a plane and bags that hit me on the head type thing. And I was off work for 2 months, like 1 month into joining Mercury. I'm like sorry guys I need to peace out for 3 months. And that then it was a very very long recovery from that. And I'm actually still not 100% healed. So I'm mostly good but I have times when I have dizziness and headaches and other things that I need to work around.

总的来说，我觉得这是发生在我身上最好的事情之一。我认为保持这种心态很有帮助。当然，有时候你可能会走极端，当那种看法脱离现实时，但我觉得这让我作为一个个人变得更好、更高效。很多习惯，比如我不喝酒、不喝咖啡因，我必须为了身体健康做这些事情，所以我一直坚持。我休息。这是非常重要的一点。你可能会想，即使在 Entropia 这样的地方，你怎么能这样生存？即使在最疯狂的日子里，从一天开始到午餐之间，以及午餐到一天结束之间，我都会短暂休息。即使在模型发布等最疯狂的日子也是如此。幸运的是，我们办公室有一个冥想区，我会去那里。还有我在那篇文章中谈到的关于冥想的方方面面。我想你和我都去过 Spirit Rock 的静修营，参加冥想静修改变了我的生活，现在我每年至少做一次。我很快又要去一次。我认为所有这些都有助于更好地管理身体方面，因为这份工作非常耗费精力，然后在情感上给自己留一点空间。就像你一边有觉知，另一边有现实，而现实很疯狂，伙计。现实太疯狂了。每天都有太多事情发生。噪音太多。令人难以置信。所以觉知与现实之间的关系，就是你有选择的地方，你从深度冥想中学习如何改变这种关系，并在那里运用你的选择。我认为这极大地帮助我保持冷静的头脑。我认为在这种高强度的游戏中坚持下去，关键就是保持头脑清醒。在疯狂的时刻不要失去理智，我认为我所经历的一切帮助我做到了这一点，而没有诉诸不健康的应对机制。

Overall I feel like it's one of the best things that could have happened to me. I think that keeping that mindset helps. And there's a point when you can take that too far, when that view is devoid of reality, but I think it's just made me so much better and more effective as a person. A lot of the same habits like I don't drink alcohol, I don't drink caffeine, I have to do a bunch of these things for my physical health, I just have to do them. So I keep doing them. I take breaks. This is a really big one. You might think even in a place like Entropia, how do you survive in this way? It's like even on the craziest days, between the start of the day and lunch, and between lunch and the end of the day, I take a short break. Even on the craziest days with model launches etc. Luckily we have a meditation area in the office that I'll go to. And then the whole side of things around meditation that I talked about in that post. I think both you and I have done a retreat at Spirit Rock, and doing a meditation retreat changed my life and it's something that I do at least once a year now. I have one coming up relatively soon as well. And I think all of that has helped with better managing the physical side of it, because this job is very taxing, and then emotionally having a little bit of space to what happens. Like you have sort of awareness on one side, you have the reality on the other side, and reality's crazy here, man. Like reality's insane. There's so much happening every day. There's so much noise. It's mind-blowing. And so that relationship between awareness and reality, that's where you have choice that you sort of learn from deep meditation on how to shift that and apply your choice there. I think that is the thing that has helped me significantly with just keeping a more level head. I think a lot of staying in this game at this level of intensity is just you keep your head. Just don't lose your head in the crazy times, and I think all that I've gone through has helped me do this without resorting to unhealthy coping mechanisms.

哇。你在很多方面都太鼓舞人心了。有太多理由说明你本不会成功，事情本不会顺利。从你分享的故事和你走过的旅程中，有太多教训可以学习，帮助人们克服他们面临的挑战。比如，如果你能从一次疯狂的……是踢到头吗？是那样吗？

Wow. I'm all you're such an inspiration in so many ways. Just like there's so many reasons that you would not have been successful and things would not have worked out. And there's so many lessons to learn from just the stories you've shared in the journey you've been on to help people overcome challenges they're having. Like if you can come back from an insane... Was it like a kick to the head? Is that what happened?

是踢到头。没错。是的。

It was a kick to the head. That's correct. Yes.

就像感觉你可能无法走路，无法听音乐，到领导历史上增长最快的公司的增长。这里也有一个关于约束的教训。这种约束的力量，比如你被迫休息，被迫去冥想，这真的很有帮助，对我们所有人都是一个教训。这实际上对我们所有人都很好。

Just like feeling like you may not be able to walk, not being able to listen to music, to just leading growth at the fastest growing company in history. There's also just like a lesson here of constraints again. This idea of just like the power of constraints, like you're forced to take breaks, you're forced to go meditate like in a way that's really helpful and it's like a lesson for us all. This is actually really good for us all.

是的，我认为通过约束获得自由是那段时间我最大的收获之一。因为你有这些约束，你被迫适应，它创造了……在商业环境中，你必须更专注。在个人环境中，如果约束是我什么都不能做，我受伤了，不知道会发生什么，不知道我是否会好转。你有两个选择。一是，你是要抗拒、与现实抗争，并真正因此受苦？还是接受现状，不让它影响你的情绪？不是说我把所有事都做了，伙计。我做了我能做的每一件事，饮食、锻炼。我全力以赴。我能采取的每一个行动，我都做了。但然后你有一个选择：你是否让那件事的影响定义你的幸福？我认为当影响可能没有到来时，你必须适应它。好吧，我希望我的幸福是什么？正如冥想老师所说，生活中真正的自由是学会在得不到想要的东西时感到满足。不是说当你得不到想要的东西时你不应该做些什么，但我觉得这个收获就是：你能不能感到满足，不让你的幸福依赖于得到某样东西？我认为这是一个挑战。我做得并不完美，但我认为这可能是整件事中最大的收获之一。

Yeah, I think that freedom through constraints is one of the big takeaways I've had during that time. Because you may have these constraints, you're forced to adapt, and it creates... In a business setting, you have to focus more. In a personal setting, if the constraint is I can't do anything and I'm injured and don't know what's going to happen, I don't know if I'm going to get better. You have two choices. One, are you going to resist and fight with reality and really suffer from that? Or are you going to accept the situation and not let that affect you emotionally? It's not to say I did everything, man. I did every single thing I could do, diet, exercise. I was like, I'm on it. Every action I can take, I'm on it. But then you have a choice of are you going to let the impact of that define your happiness or not? And I think that's the thing when the impacts maybe are not coming, you have to then adapt to it. Okay, what I want my happiness to be? And as the meditation teachers said, the true freedom in life is learning how to be content when you don't get what you want. And it's not to say you shouldn't do something when you don't get what you want, but I think that takeaway of just can you be content and not have your happiness rely on getting something? I think it's a challenge. I'm not perfect at it, but I think it's probably one of the biggest takeaways from that whole thing.

太棒了。我们会链接到那篇文章，如果大家想读的话。

Incredible. And we'll link to the post if folks want to read this.

在我们进入非常激动人心的快速问答环节之前，你还有什么想分享的吗，还是我们直接开始？

Before we get to our very exciting lightning round, is there anything else you wanted to share or should we just jump right in?

我觉得我们可以直接开始。

I think we can jump into it.

我们聊了很多内容。好的，开始吧。我有五个问题要问你。

We covered a lot of ground. Okay, here we go. I've got five questions for you.

第一个问题，你最常向别人推荐哪两三本书？

First question, what are two or three books you recommend most to other people?

这可能和我刚才聊的很多内容有关。我经常思考冥想和我的情绪状态，这是我在工作之外花最多时间思考、阅读、研究和努力的事情。所以我的推荐也与此相关。第一本是佛教僧侣咏给·明就仁波切写的《喜悦之道》。这本书讲的是如何以不同的方式看待你的生活经历，并提供了一些策略来改变你对事物的看法。我向很多人推荐过这本书，他们都非常喜欢。另一本是安东尼·德·梅洛的《觉醒》，内容类似但角度不同。这两本书我一直向别人推荐。第三本在某种程度上与产品更相关：安妮·杜克的《对赌思维》。我认为能够分解一个情境，当有人说‘我不知道能不能按时完成’时，你可以说‘好的，你能给个数字吗？大概有多大可能性？’这类事情在产品工作中非常有战术价值。

It probably ties to a lot of what I just talked about. I think a lot about meditation and my emotional state. That's what I spend most of my time thinking and reading and researching and working on outside of work. So my recommendations are really to that. The first is a book called Joy of Living by a Buddhist monk named Yongey Mingyur Rinpoche. It's just really about how you can start to think about your life experience in a different way and have tactics of what to do to change just how you think about things. This is one I've recommended to a lot of people and they've really enjoyed it. I think another one is Awareness by Anthony De Mello, which is a similar thing but from a different angle. Those two I consistently recommend to people. And then the third one is more relevant to product in many ways: Thinking in Bets by Annie Duke. I think just being able to break down a situation and someone says, 'I don't know if it'll get done in time.' Okay, can you put a number to that? What percentage likelihood is it? Those types of things are so tactically helpful in product.

最近最喜欢的电影或电视剧是什么？

Favorite recent movie or TV show?

电影的话，可能是《狂欢节至尊》。这是我最近看的为数不多的电影之一。我觉得很棒，简直疯狂，绝对疯狂的电影。虽然很荒谬，但我很喜欢。电视剧我很久没看了。如果算的话，可能是奥运会。

I think movie is probably Mardi Gras Supreme. It's one of the only ones I watched recently. I thought it was great. I thought it was just insane, absolutely insane movie. It was ridiculous but I loved it. TV show I have not watched TV in a while. I think it's probably the Olympics if that counts.

有没有你很喜欢的产品，比如最近发现的或者生活中一直在用的，让你觉得‘哦，这很酷，人们可能想知道它’？

Is there a product that you love, like something you discovered recently or just have in your life that's like, 'Oh, this is very cool. People might want to know about it.'

嗯，这有点好笑。我上周在日本，住在酒店里，非常喜欢酒店里的那个枕头。这很偶然。我有时脖子和上斜方肌会疼，而且因为有时仰睡有时侧睡，枕头高度总不合适，让我很烦恼。但这个枕头就很好。它里面装满了珠子，所以你可以在睡觉时自然地调整枕头高度，甚至无意识地就能做到。所以我看了标签，心想这是什么？我在亚马逊日本站下单了，然后带上飞机带回来了。名字在这里：叫 Maruhachi Shinsui Maruhachi Pro Pillow。只有日本有卖，但他们可以寄到美国。我没有利益关系。我觉得这枕头在过去一周改变了我的生活。

Yeah, so this is a funny one. I was in Japan last week and I was sitting in this hotel and I just really loved the pillow that I had in this hotel. It was very random. I've had at times neck and upper trap pain and sometimes I've been frustrated with my pillow because I sometimes sleep facing up, sometimes sideways and the height's not quite right. And I was just like this pillow is just great. It's full of beads so you can naturally shift the height of the pillow while you're sleeping, even unconsciously without thinking. So I looked at the tag, I was like what is this? I ordered it on Amazon Japan and brought it back with me on the plane. I have the name here: it's called the Maruhachi Shinsui Maruhachi Pro Pillow. You can only get it in Japan but they ship to the US. No affiliation. I think that has changed my life in the last week.

你有没有一个在工作和生活中经常想起的人生格言？

Do you have a favorite life motto that you often come back to in work or in life?

最主要的一个是‘她会没事的’。这是澳大利亚很常见的说法，就是‘她会没事的’。当我们面临困难或棘手的情况时，你通过说‘是的，会好的’来淡化它的严重性。我认为在很多情况下这是一个很好的策略。另一个是‘有时候在生活中你只需要放手去做’。这是 Discord 的前 CMO Eros Resmini 经常推动我的，他是我作为创始人的非常亲密的顾问。他经常说‘放手去做’。我觉得这非常有帮助：你可以坐在那里想‘我该不该做’，但有时候在生活中你只需要放手去做。

Main one I'd say is just 'she'll be right.' This is a very common Aussie saying. It's like 'she will be right.' When we are faced with a tough or difficult situation, you're kind of dismissing the severity of it by saying 'yeah, it'll be fine.' I think that's such a good tactic in many cases. And then the other one is 'sometimes in life you just have to go for it.' That is something that Eros Resmini, who was the CMO of Discord and a very close advisor to me as a founder, would push me on. He would often say 'just go for it.' I think it's very helpful: you can sit there thinking 'should I do this, should I not,' and sometimes in life you just have to go for it.

我很好奇，你以前喜欢武术，但不太顺利。你现在有没有新的爱好？你会为爱好腾出时间吗？

I'm curious, you used to be into martial arts, that didn't go great. Do you have a new hobby that you find yourself loving? Do you make time for hobbies?

我觉得我的时间很紧，没什么时间给爱好。基本上就是工作、保持身体机能，然后照顾身边最重要的关系。我想唯一的爱好就是我很喜欢体育，尤其是橄榄球，看球但不打球。我妻子来自密歇根，她带我去 Big House 看了一场球赛，那改变了我的人生。从那以后我就成了狼獾队的粉丝。我是狼獾队的大粉丝，也是 49 人队的大粉丝。我就是喜欢橄榄球，所以这大概就是我能想到的爱好。

I think I'm just quite tight on time for hobbies. I would say it's largely just work and have my body function and then take care of the most important relationships around me. I think maybe the one hobby I'd say is just I'm really into sports and I've gotten more into football in particular, watching, but certainly not playing. My wife is from Michigan and she took me to a game at the Big House and it just changed my life. From then on I was like, I'm a Wolverine fan. Big Wolverine fan, big 49ers fan. I just love football, so that's probably the one that comes to mind.

Amol，这次访谈在很多层面上都太棒了。非常感谢你在百忙之中抽出时间。大家可以在网上哪里找到你？如果他们想申请你团队的工作，应该去哪里？听众怎样才能帮到你？

Amol, this was incredible on so many levels. I'm so thankful you made time for this considering how much you have going on. Where can folks find you online and say they want to apply for a job maybe on your team or where can they find that? And how can listeners be useful to you?

你可以在 LinkedIn 上找到我，就是 Amol Avasare。我很无趣，就这样。你也可以在我们的招聘页面上在线申请。我们有增长工程、增长产品、增长设计等职位。我们正在为增长团队寻找优秀人才，所以请加入我们。至于如何帮我，只有两件事：试用我们的产品，给我们反馈，尤其是严厉的反馈，告诉我们哪里可以做得更好；另外，请把优秀的人才推荐给我们。我们正在寻找最优秀的人加入团队，如果你认识适合我们某个职位的人，我们很乐意听到他们的消息。

You can find me on my LinkedIn. It's just Amol Avasare. I'm boring. That's it. You can look online to apply on our jobs page. It's roles across growth engineering, growth product, growth design. We are looking for great people on the growth team, so please come and join us. And being helpful to me is just two things: trying our products, giving us feedback, harsh feedback in particular on what we can do better, and please send great people our way. We are looking for the best of the best to join the team and we would love to hear from anyone you know who could be a fit for one of our roles.

对很多人来说这是梦想的工作。Amol，非常感谢你来做客。

Dream job for so many people. Amol, thank you so much for being here.

谢谢，Lenny。大家再见。

Thanks, Lenny. Bye, everyone.

非常感谢你的收听。如果你觉得本期内容有价值，可以在 Apple Podcasts、Spotify 或你喜欢的播客应用上订阅本节目。另外，请考虑给我们评分或留下评论，这真的能帮助其他听众找到这个播客。你可以在 lennyspodcast.com 找到所有往期节目或了解更多信息。下期再见。

Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify or your favorite podcast app. Also, please consider giving us a rating or leaving a review as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lennyspodcast.com. See you in the next episode.