从烦恼到 AGI：OpenClaw 的诞生

六个月前，我还是未来。如今，脉搏说我被一个动漫女孩嘲笑了。你知道，我几乎每天都会收到这样的问题。通常我会忽略它，或者给出媒体式的回答。今天你会得到一些真实的答案。一共五个问题，四十分钟。我知道你们可能都是为了什么来的。你们是为了循环、图表、让 Codex 去煮咖啡。这些在我的 Twitter 上都是免费的。这会帮助你们度过可能到来的过山车。你知道，昨天 Boris 告诉你们让模型去煮。我就是那个发现当数万人同时让它煮会发生什么的人。让我们从每个人都会先问的问题开始。OpenClaw 怎么才八个月大？

Six months ago, I was the future. These days, the pulse say I got mocked by an anime girl. You know, I get questions like this almost every day. Usually, I ignore it or I give the media answer. Today you get some real ones. It's going to be five questions, 40 minutes. And I know what you might all came for. You came for loops, graphs, making codex go brew. This is free on my Twitter. This will help you to survive the roller coaster that might be coming for you. You know, yesterday Boris told you to let the model cook. I'm the guy who found out what happened when tens of thousands of people let it cook at the same time. Let's start with the question everyone asks first. How is open claw only 8 months old?

你知道，按人类时间算可能只有八个月，但按 AI 时间算，更像是四年。而我的灵感来源通常是感到恼火。所以十一月一个下雨天，我在同时处理几个智能体，但我也饿了。所以我想确保在我去厨房搜刮食物的时候，我的 token 能被好好利用。而且当时还没有一个好办法能直接从手机发个提示词到电脑上，让某个智能体去检查我的智能体到底运行得怎么样。还记得 2025 年初吗？你们知道，这是我第一次参加无声迪斯科。这是一种非常新的体验。所以如果你们真的在听我说话，就给我点掌声。是的。太棒了。太棒了。你知道，2025 年初。天哪。我们说的是老版的 Opus 3i 2.5。OpenAI 有 03，那是最令人印象深刻的，但有时也真的很慢很贵。当我的智能体做对事情时，我会获得多巴胺的快感。我相信你们还记得那个时期。

You know, it might be eight months in human time, but in AI times it's more like four years. And my source of inspiration is usually being annoyed. So on a rainy day in November, I was juggling some agents, but I was also hungry. So I wanted to make sure that my tokens are put to good use while I raid the kitchen. And there was still no good way to just send a prompt from my phone to my computer so some agent could check up how my agent's actually doing. Remember early 2025? And you know guys, this is my first silent discos. It's a very new experience. So give me some hands if you actually listen to me. Yes. Amazing. Amazing. You know, early 2025. Oh my god. We're talking the old Oppus 3i 2.5. Openi had 03 which was the most impressive one but also really slow and expensive at times. And I got a dopamine hit when my agent did something right. I'm sure you remember this time.

而现在，如果它们做错了什么，我通常会质疑自己，你知道，是我把循环设计错了吗？我是不是没有给我的智能体提供验证工作所需的一切？我的思考中有错误吗？我是不是在要求不可能的事情？所以，我从厨房回来，我的编码智能体因为一些愚蠢的事情停了下来，我很恼火。于是，我打开了一个新的终端会话。我把我的想法胡乱输入进去，然后让模型去煮。它给我建了一个 WhatsApp 中继，一个小时后，我就能从我的 Mac 发送消息到 WhatsApp，再收回来。这本身就感觉很神奇。但实际上，它本不该如此。就像我们大家都已经这样做了好几个月了，对吧？你打开终端，输入一些东西，然后得到回复。但神奇的是它给人的感觉。它不是一个终端。它写出简洁的回答。它是主动的。比如有时候它会在白天主动来查看我。我让它把复杂性都融化了。你不需要去想用哪个模型、多大的上下文窗口，或者什么时候开始新会话。我还把权重稍微调出了默认分布，让它感觉更像一个朋友。我一直在用它。我有过很多次感觉这就是未来的时刻。这就是 AGI。你知道吗？没人在乎。

And these days, if they get something wrong, I usually question myself, you know, did I design the loop wrong? Did I not give my agent everything they need to verify their work? Do I have an error in my thinking? Am I asking for impossible things? So, I came back from the kitchen and my coding agent stopped because of something silly and I was annoyed. So, I spun up a new terminal session. I rambled my idea in and I let the model cook. It built me a WhatsApp relay and in an hour later I could send messages from my Mac to WhatsApp and back. And that in itself felt magical. But really, it shouldn't have. It's like we all did this for months already, right? You open the terminal and you type something in and you get a reply. But the magic is how it felt. It wasn't a terminal. It wrote concise answers. It was proactive. Like sometimes it would just check on me during the day. I made it so the complexity would melt away. You don't have to think about which model, which context size or when you start a new session. And I prompted the weights a little bit outside the default distribution to make it feel more like a friend. And I was using it. I had so many moments where I felt like this is the future. This is AGI. And you know what? Nobody cared.

那时我在 Twitter 上已经有相当多的关注者，我知道现在它叫 X 了。但我永远会叫它 Twitter。人们不理解。我一直在试图让他们明白这对我来说有多神奇。但一周又一周，我都失败了。你注意到一个模式了吗？这让我很恼火。所以我会和朋友创建群聊，我把他们加进 WhatsApp 中继，我会展示给他们看，让他们和它对话，每次我都能得到强烈的情绪反应。有些人惊叹，有些人害怕，甚至吓坏了。但每次都有强烈的情绪，最棒的是有不少人真的想要它，尤其是对我的非技术朋友，我告诉他们不，这还不是给你们用的，他们就很生气。所以如果这都不算是产品市场契合的指标，那我就不知道什么才是了。

Back then I already had a good following on Twitter and I know it's called X now. I'll always call it Twitter. People were not getting it. I would keep trying to make them understand like how magical this felt to me. But week after week I failed. And you notice a pattern? It annoyed me. So I would create group chats with friends and I added them WhatsApp relay and I would show them I would let them talk to it and every time I got a strong emotional reaction. Some people were amazed, some people were scared or freaked out even. But each time there was a strong emotion and the best part there were quite a few people that really wanted it and especially to my nontechnical friends I told them no this is not yet for you and they got mad. So if that's not an indicator to have product market fit I don't know what is.

所以我又花了几个月调整细节，思考我能做些什么来解释这个世界，你知道，然后有人竟然给我发了一个拉取请求，给我的 WhatsApp 中继添加 Discord 支持，你知道，名字里哪部分你不明白？我让那个 PR 搁置了一段时间，我一直在想，一直在想，然后最终我心想，管他呢，于是 V relay 变成了 Claudis，因为我擅长起名字，而且不再只有一个消息通道了，所以你能理解这有多早期。我甚至还没有内置压缩功能。所以有时候它就会停下来，然后为了赶上 Discord 的时间，我加了一个非常粗糙的压缩版本，这样 Mario 就能在 Pi 里做一个好的版本。

So I spent another months tweaking the details thinking what could I do to explain the world you know and then the audacity someone sent me a pull request to my WhatsApp relay to add Discord support you know like what parts of the name are you not getting I let the PR sit for a while and I was thinking and I was thinking and then eventually I was like ah what the hell and out of V relay became claudis because I'm good with names and there's no more than one message channel and so you understand how early this was. I didn't even had compaction built in yet. So like at some point it would just like stop uh and just just for in time for Discord I I added a very hacky version of compaction so Mario could make a good one in Pi.

所以，那是在新年前后，就像你常做的那样，我提前回家继续捣鼓它。然后在一月的第一周，正当极客世界集体学习编码智能体的时候，我创建了一个 Discord 房间、服务器、公会，不管我们怎么称呼它，我把我的 Claw 放了进去。我清晰地记得第一晚。人们会加入 Discord，他们会看着我公开构建它。他们会试图破解它。他们就像在和它聊天。他们得到得意的回复，他们看看自己能做什么，他们终于明白了。你知道，这就是我彻夜不眠的时刻。我让人们和它互动。我在我的 agents.mmd 里有一个提示词，指示它不要进行任何危险的工具调用，除非提示词来自 Peter，你知道，这大约是六个月前，我们写在智能体里的所有东西感觉更像是建议，所以我非常仔细地观察它，我总是可以拔掉插头，然后到了早上七点，我终于完成了。我说了晚安。我按了 Ctrl C，然后去睡觉了。但当然，我把软件建得很有弹性。所以，它是一个 launchd 守护进程。你知道如果你在 launchd 守护进程里按 Ctrl C 会发生什么吗？是的。它会死掉五秒钟，然后重新启动。所以，当我走进卧室的时候，我的智能体高兴地开始回答世界上的上帝。然后我睡了大约十个小时。当我醒来时，我看到了大约 800 条消息。人们试图破解它。我拔掉了插头。我吓坏了。我读完了所有内容，实际上什么都没发生。但这就是那个时刻。这就是它爆红的时刻。

So, it was around New Year's Eve and as you do, I went home early to hack on it some more. And then in the first week of January, just as the geek world collectively was learning about coding agents, I created a Discord room, server, guild, however we call it, and I put my claw in it. And I remember the first night vividly. People would be joining Discord and they would like watch me build it in the open. They would try to hack it. They were like talking to it. They were getting smug replies and they see what they can do and they finally got it. You know, like this was the moment I sit up all night. I let people interact with it. I had this prompt in my agents.mmd that would instruct it don't do any dangerous tool calls unless the prompt comes from Peter you know this were like six months ago and where all the things we wrote in our agent felt more like suggestions so I watched it very carefully I could always pull the plug and then at 7 a.m. Finally, I was done. I said good night. I pressed Ctrl C and I went to bed. But of course, like I built the software to be resilient. So, it was a launch demon. You know what happens if you like press control C in a launch demon? Yeah. It'll just it'll be dead for 5 seconds and it'll start up again. So, while I was walking into the bedroom, my agent happily started answering God in the world. And then I slept for like 10 hours. When I woke up, I woke up to around 800 messages. People tried to hack it. I pulled the plug. I freaked out. I read through everything and nothing actually happened. But this was the moment. This was the moment where it went viral.

而你们都在这里。你知道接下来发生了什么吗？我的收件箱爆了。记者们试图在半夜给我打电话。你知道苹果有一个功能，当你开启勿扰模式时，如果有人连续打几次电话，iPhone 会认为这是紧急情况，仍然会接通电话。我之前不知道这个功能，但记者们知道。我的电子邮件收件箱更像是一个瀑布。是的，Mac Mini 卖光了。我想现在还是缺货。我在一个月内收到的播客邀请比我之前 39 年加起来还多。然后 Tropic 发来邮件要求改名，就像龙虾脱壳一样，项目经历了几次蜕变。Claudius 变成了 Claudebot。然后有一段时间我们都不提那个名字，然后它变成了 OpenClaw。

And you all been here. You know what happened next? My inbox exploded. Reporters tried to call me in the middle of the night. Did you know about this feature that Apple would build when you put it on do not disturb, but people call you a few times in a row, the iPhone would think it's an emergency. It would still let call through. I didn't know about it, but reporters did. My email inbox was more like a waterfall. Yeah, the Mac Mini sold out. I think they're still sold out. I got more invites to podcast in a month than I did in my 39 years before. And Tropic sent an email demanding the name change and like dropped the lobster and the project molted a few times. Claudius became claudebot. Then there was a very short time where we don't talk about the name and then it started on openclaw.

Jensen 称它是人类历史上最成功的开源项目。与此同时，我的收件箱被记者、安全人员、无数好奇的人类，还有智能体淹没了。你知道，这些数字对我来说仍然不真实。8 个月内，超过 18,000 人开了 issue 或 pull request，总数超过 111,000。我算过。不，我对谁撒谎呢？是我的智能体算的。将近 3,000 人在仓库里有提交。有些人对我很生气，因为他们显然先做了，而我偷了他们的想法，他们给我发了我从未见过的东西的链接。有些人叫我 I Jesus。对一些人来说，我成了反基督者。某个时候我开始收集那些讣告，每隔几周就有一篇新的。最近的一篇其实是昨天的。我们在这里学到什么？小心你许的愿。我还没准备好迎接所有这些关注。它几乎让我崩溃。就像我差一点就删掉整个东西。我不再回复朋友。我甚至不想再看手机，因为那只是一条流。当然，有人泄露了我的电话号码和很多私人细节，因为我是人类的敌人。而我在这里，只是想构建一些酷的东西。

Jensen called it the most successful open-source project in the history of humanity. And meanwhile, my inbox was flooded with reporters, security people, and countless curious humans, and also agents. You know, the numbers still don't feel real to me. In 8 months, more than 18,000 different people opened an issue or a pull request. Over 111,000 in total. I did the math. No, who am I lying to? My agent did the math. Almost 3,000 people have commits in the repo. And some people were mad at me because clearly they did it first and I stole the idea and they sent me links of something I've never seen. Some people called me I Jesus. For some I became the antichrist. At some point I started collecting the obarius. There's a new one every few weeks. The most recent one is actually from yesterday. What do we learn here? Be careful what you wish for. I was not ready for all this attention. It almost broke me. Like I was that close of just deleting the whole thing. I stopped answering my friends. I didn't even want to look at my phone anymore because it was just a stream. And of course, somebody leaked my phone number and also a lot of private details because I'm an enemy of humanity. And here I am just wanted to build something cool.

现在我们来到第二个问题。你出卖自己了吗？

And now we're getting to question two. Did you sell out?

你知道，这不是我第一次经历。我二三十岁时在构建一家 B2B 软件公司。我像动物一样手写了一个 PDF 框架。我白手起家，把公司发展到近 80 人。我无视竞争。它成了大多数企业会买的东西。最终我把事情交给了联合创始人。我卖掉了股份，然后严重倦怠。退休，这个想法听起来很棒。我花了将近三年时间漫无目的地游荡，补上生活和派对。我做了所有我在生活中错过的事情。有几个月我甚至没打开电脑，只是随意用手机上网，像个普通人。我一直知道，如果给自己足够的时间，那股冲动会回来。我一直以为冲动会是代码。我一直热爱编程。但花了一年我才明白，那不完全正确。我热爱构建。编程只是达到目的的手段。而在我重新找到火花八个月后，我就在这里。我的收件箱里满是收据，求我收下他们的钱。我不知道我想不想。所以当大实验室来敲门，我突然和 Mark、Sam 还有几个人通电话时，这似乎是一条更有趣的路。这也感觉不真实，对吧？说到冒名顶替综合征。也许这对你来说是个重要的教训。你能构建的一切都可以被 fork 或克隆，但你的名字不能。所以你的个人品牌比你将要开始的任何单一产品都重要得多，要提前开始打造。所以，是的，我知道怎么玩这个游戏。我也知道这样做等同于卖掉我的灵魂.md。懂的都懂，这个决定不容易，但如果我在倦怠后学到了什么，那就是这个：相信你的直觉。我的直觉告诉我，我最喜欢开放。

You know, this wasn't my first ride. I spent my 20s and my 30s building a B2B software company. I wrote a PDF framework by hand like an animal. I bootstrapped the company. I grew it to almost 80 people. I ignored the competition. It became the thing most enterprises would buy. And eventually I passed things on to my co-founder. I sold my shares and I very badly burned out. Retirement. The idea of it sounds great. And I spent almost three years wandering around aimlessly, catching up on life and parties. I did all the things that I kind of missed out on life. I had months where I didn't even open my computer, you know, just casually checking the internet with my phone. It's like a normie. I always knew that if I give myself enough time, the urge would come back. I always thought the urge would be code. I always love programming. But it took me a year to understand that that's not entirely correct. I love building. Programming was just a means to an end. And here I was eight months after I found my Spark again. And my inbox was full of receipts begging me to take their money. And I didn't know I want to. So when the big labs came knocking and suddenly I was on the phone with Mark, Sam, few others, this seemed like a much more interesting pass. It also felt unreal, right? Talk about imposter syndrome. And maybe that's an important lesson for you. Everything that you can build can be forked or cloned, but your name cannot. So your personal brand is way more important than any single product that you ever will start working on that before you need it. So yeah, I know how to play the game. I also know that doing this was the equivalent of selling my soul.md. If you know, you know the decision was not easy, but if I learned anything during my time after the burnout, it's this. It's trusting your gut. And my gut told me that I like open the most.

Hermus 赢了吗？

Did Hermus win?

我能给你的简短回答是，他们在我们痛处击败了我们。我们还在。这两点都是我的错，也是我造成的。发布后的几个月里，我们被安全报告彻底压垮了。在某些方面，我们是许多开源项目现在经历的典型。尽管大多数报告都非常 HKC，我还是感到了压力。然后媒体说我们 20% 的技能是恶意的。现在我们实际上为此发了一篇论文。我们算了数字，数字更像是 0.3%。我们扫描了全部 67,000 个。我们发布了论文。但你知道，更正传播得远不如恐慌。我感受到了这种责任。世界发现了 Open Claw。尽管有一个巨大的可怕免责声明，实际上比我在幻灯片上能放的还大，当你安装它时。我知道很多人不会读文档。所以我真的专注于加固代码库，构建一层又一层的安全。我做了沙箱、允许列表、一个内置权限的 Web 协议。我们为一些文件操作调用 Python，因为 Typescript 中缺少一些原语，以确保你的智能体留在工作区。它不跟随符号链接，配置文件是原子写入的。大多数用户并不在乎这些。当然，他们喜欢安全这个抽象术语，但在所有实际意义上，他们更新了。我破坏了他们依赖的东西。我让事情变慢，让更新变得更困难。另一件我必须承认的事是我变得草率了，因为现在我的时间在开源、媒体、花无数时间与律师通电话建立 5013 美国非营利组织之间转移。我做了自己的，你知道 OpenI 有自己的世界，有趣又苛刻。我引入了帮助，我们有了一个非常棒的维护者社区，他们非常受尊重，每个人都添加了自己的小功能。我觉得自己有点为难。毕竟，这些人免费工作，对吧？所以我有什么资格告诉他们该做什么，而且我的注意力分散了。所以我们添加了很多功能。你知道功能是有趣的部分。一个新功能只是一个提示的距离。真正的成本出现在每个功能发布之后，当然还有配置选项，因为我们不想破坏每个人的设置。在我们最高峰时，我不得不数一下，我们最终有大约 9,500 个配置选项。如果你计算所有排列，你可以写所有你想要的测试。不可能覆盖所有这些而不时不时破坏东西。演化有用户的软件要难得多。与此同时，其他公司，他们由风投资金推动。他们向前推进。动漫女孩公司做得特别好。人们开始把自己插入 Twitter 上几乎任何对话，而我们被安全、资助工作和功能烧得焦头烂额。他们有一个简单的故事，一个激进的营销活动，和一个迁移人们条款的单行命令。然而真正伤害项目的主要事情是 entropic。不是名字。那部分压力很大，但我有点理解，他们对此非常友好。事实是我过度优化了他们的模型。我用 Codex 和 GPT 构建了 Open Claw，但很长一段时间里，这个 harness 真的最好用，并且针对 OPOS 进行了优化。所以当他们提前大约 24 小时通知我，他们将禁用所有人的订阅时，真的没有足够的时间改变方向。我的意思是，当然，我们支持开放权重模型。我们在它们上做了很多工作，但它们当时真的还不够好。早期的开放模型就是缺乏个性。

And the short answer I can give you is they beat us where it hurt. We're still there. And both of those are my fault and my doing. In the months after the release, we got absolutely crushed by security reports. In some ways, we were a prototype of what many open-source projects now experience. And even though most were super HKC, I felt the pressure. And then the press, the press say 20% of our skills are malicious. Now we actually put a paper up with this. We did the numbers and the numbers are more like 0.3%. We scanned all 67,000. We shipped the paper. But you know a correction never travels as far as a scare. I felt this responsibility. The world discovered open claw. And even though there was this big scary disclaimer actually much bigger than I could on the slides when you install it. I knew many people wouldn't read the docs. So I really focused on hardening the code base, building layers and layers of security. I did sandboxing, allow lists, a web protocol with permissions built in. We were shelling out to Python for some file operations because there were some primitives that are simply lacking in Typescript to ensure your agent stays in the workspace. It doesn't follow sim links and that the configuration files are written atomically. Most of the users didn't care about that. Sure, they like the abstract term of security, but in all practical terms, they updated. I broke something they depended on. I made things slower and I made things more difficult to update. And another thing I have to admit I got sloppy because now I had to my time shifted between open source, the press, spending countless times on the phone with lawyers to set up a 5013 American nonprofit. I worked on my own one and you know OpenI has its own world that is interesting and demanding. I brought on help and we got such an amazing community of maintainers that are very much respected and everyone added their own little feature. I felt like I'm a bit in a hard place. After all, these people work for free, right? So who am I to tell them what to do and also my attention was all over. So we added a lot of features. You know features are the fun part. A new feature is just a prompt away. The real cost comes after every feature we shipped of course with a configuration option because we didn't want to break everyone's setup. At our highest, I had to count that we ended up with around nine and a half thousand configuration options. If you count all the permutations, you can write all the tests you want. It is impossible to cover all of these and not break things from time to time. It is infinitely harder to evolve software that has users. Meanwhile, other companies, they were fueled by VC money. They pushed ahead. The anime girl company did it especially well. People started inserting themselves into pretty much any conversation on Twitter while we were burned in security grant work and features. They had a simple story, an aggressive marketing campaign, and a oneliner to migrate people's clause. The main thing though that really hurt the project was entropic. Not the name. The part was stressful, but I kind of understood and they were really nice about it. It was the fact that I optimized too much on their model. I built Open Claw with Codex and GPT, but the harness for a long time really worked the best and was optimized for OPOS. So when they ping me with around 24 hours notice that they're going to disable the subscription for everyone there was not really enough time to change course. I mean, sure, we support openweight models. We did a lot of work on them, but they really weren't that great yet. And the early open air models simply lacked character.

所以这个可以记下来。你的依赖业务模式就是你的业务模式。你看，现在一切都定下来了。解决方案相当棒。开放权重模型其实很好。我在工程管理方面学到了很多。但很多方面，人们已经往前看了。你可以从我们的下载图表看到。五月份我们触底，每周下载量大约 83.5 万。然后在六月份被宣告死亡之后，我们达到了峰值 470 万，历史最高。这两件事同时为真。炒作就像天气。你可能看到它要来，但你控制不了。就我而言，那是一场风暴。

So maybe write this one down. Your dependencies business model is your business model. You know, that's all fixed now. Solves pretty awesome. Open weight models are actually good. I learned a lot about harness engineering. But in many ways, people moved on. You can see that in our download charts. We bottomed out at around 835,000 weekly downloads in May. And then after being declared dead in June, we peaked at 4.7 million, the highest ever. Both of those are true at the same time. A hype is like the weather. You might see it coming, but you can't control it. In my case, it was a storm.

问题四，现在还有趣吗？你知道，大概二月份的时候，它不再有趣了。我开始觉得它变成了一种责任。我醒来，那个不想再创办另一家公司的人，发现自己处于一个有两份工作的境地，或者我该说是一份工作和一份使命。我在挣扎，最糟糕的是我停止使用自己的产品。在那段时间的某个节点，我不再做我热爱的产品，而是为了所有人做东西。它不再是我每天使用的东西，而更多是我看到和感觉到的工作。我想象每个人到底是谁。其中一个有一个杂货店智能体。另一个试图对我的机器人进行社会工程攻击。在每个人添加功能和组织工作之间，我变成了那个修 bug、修安全问题、提供支持、打基础的人。而且因为所有这些人在散布谣言，说 OpenClaw 归 OpenAI 所有。我不想从 OpenAI 那里接受太多帮助。我的意思是，是的，他们给了我 token，我确实用了。但我也被拉进其他产品项目，还有一大堆我没那么容易屏蔽的人想跟我说话。事后看来，我本可以有很多不同的做法，比如我可以寻求更多帮助，我可以把责任从自己身上移开，但我陷得太深，没有时间从战略上思考。幸运的是，我遇到了一群很棒的人，事情一点点变得顺利。我解决了签证问题，最终成立了一个非营利组织。我们有了很棒的公司作为捐赠者，我找到了一些真正相信开源的好人，开始和我一起工作。我还要特别感谢 Nvidia，他们非常早，只是问我需要什么。然后他们派人接管了大部分安全工作。我想大概是在五月我生日前后，我有了一种感觉，事情又开始变好了，对我来说，建造的乐趣回来了。是的，在媒体那里，媒体每隔一周就宣布一个“OpenClaw 杀手”。某个时候，我想我数了 20 个。甚至有一个项目真的叫“OpenClaw Killer”。那是一个卸载程序，完全没必要，因为我们有卸载程序，但你知道，这些杀手故事没有一个真正抓住重点。开源。我的灵感来源是感到恼火。这些天，当我不得不使用那些不能直接给我的智能体发提示来修改的软件时，我会感到恼火。那就是正在回来的乐趣。很难和一个只是在享受乐趣的人竞争。乐趣就是速度。我享受建造的那些周，产品明显变好了。我不享受的那些周，我们应该配置选项。

Question four, is it still fun? You know, somewhere around February, it stopped being fun. I started to feel like it started to feel like a responsibility. I woke up and the man who didn't want to build another company found himself in a situation where he had two jobs or should I say a job and a calling. I was struggling and worst of all I stopped using my own product. Somewhere in that time I stopped making a product I love and I worked on making something for everyone. It became less of a thing that I use every day and more I think that I saw and felt was work. And I wanted to picture who everyone really is. One of them has a grocery agent. Another one tried to social engineer my bot. Between everyone adding features and all the organization work. I became the person that fixed the bugs, fixed the security issues and provided support and built a foundation. And because all these people were like giving the rumor mill, oh, open clouds owned by OpenAI. I didn't want to take too much help from OpenAI. I mean, yeah, they gave me token and I boy did I use that. But I was also pulled into other product projects and there was a whole flood of people that I couldn't as easily block who wanted to talk to me. In hindsight, I could have many done differently like I could have asked for more help. I could have moved responsibility off my plate, but I was so deep in everything that I didn't take time to think through strategically. Luckily, I met a bunch of amazing people and little by little things aligned. I figured out my visa, eventually formed a nonprofit. We got amazing companies as donors and I found some really good people that believe in open source. and started working with me. I also need to give Nvidia a special shout out because they were very early and they simply asked me what I need. And then they sent people to take over much of their security work. And I think it was somewhere in May around my birthday where I had this feeling that things are starting to feel good again where the joy of building for me was coming back. And yeah, somewhere in the press, the press decoined an open open clock killer every other week. At some point, I think I counted 20 of them. There's even a a literally a project called OpenCloud Killer. It's an uninstaller, which is totally unnecessary because we have an uninstaller, but you know, but none of the killer stories ever picked up what this is actually about. open source. My source of inspiration is being annoyed. These days, I get annoyed when I have to use software where I can't just send a prompt to my agent to change it. That's the fun part that's coming back. It's hard to compete with someone who's just there having fun. Fun is velocity. The weeks I enjoyed building, the product got visibly better. the weeks I didn't. We should config options.

问题五。接下来是什么？有些观众可能会想，这终于到了他讲图的部分了吗？首先，我对我们基础的位置非常满意。我们的使命是让人们更接近 AI，事情变化太快了。对很多人来说，这感觉可怕。我为 OpenClaw 实现的一件事感到自豪，对很多人来说，它把 AI 从那种模糊可怕的东西，变成了有趣又奇怪的东西，你知道，龙虾什么的，他们会继续推动这一点。继续建立一个伟大的开源软件生态系统，举办让人们聚在一起的活动，还有教育。我们现在有 10 个领薪水的员工。我们还在招聘几个职位，包括一位 CEO。其次，我有点把 Claw 变成了一个名词。你知道，Karpathy 说了“open”，Satya 在微软主题演讲里说了“企业级”这个词。有 33,000 个以 Claw 命名的仓库。如果我们处于模拟中，我们肯定处于一个更奇怪、不会被关闭的模拟里。我喜欢未来的这部分。第三，我们仍然没有一个永远在线、永远同步的智能体。还有很多事情要做。AI 的格局和技术发展得比你围绕它构建的软件还快，这对你们所有人来说都是机会。我们的工作流也在进化，就像我早期的愿景，我们不应该考虑会话或压缩。世界终于，模型和技术终于到了让这成为现实的阶段。我们正在进入一个世界，终于从纯文本界面转向语音和多模态。就在昨天，我们让一个 hack 跑通了，你的 Claw 现在可以 FaceTime 你，这大概就是 OpenClaw 存在的意义。每个实验室都会卖给你一个智能体。OpenClaw 是替代品。开源随处运行，适用于任何模型。如果你运行本地模型，你的数据永远不必离开你的设备。你的智能体，你的机器，你的生活。那甚至不是我的 C 论文，那是 Gary 的。你知道，我们只是先用了。所以我们也终于再次用 OpenClaw 构建 OpenClaw，但这次有个变化，每个人在团队服务器上都能看到彼此的会话，有一个 Claw 知道其他人在做什么，还可以接管工作的编排。是的，我们终于慢慢离开这个未来的奇怪小插曲，人们不再在终端里工作，不再因为智能体需要持续运行而开着笔记本电脑到处跑。如果你只记住三件事，第一，不要停止享受乐趣。乐趣是终极驱动力，这样你才能得到最好的想法。第二，听从你的直觉。还有修复那些让你恼火的事情。它可能就会成为下一个大事件。第三，保持专注。你知道，另一个播客不会让你赢。活在未来，构建缺失的东西。当他们写你的讣告时，继续发布。这会让他们困惑。好了，这就是五个问题。我相信你们还有更多问题。我想我们有足够的时间进行问答。

Question five. What's next? Some people in the audience might be like, is this finally the part where he talks about graphs? First of all, I'm super happy with where we are with the foundation. Our mission is to bring people closer to AI and things are changing so fast. For many people, it feels scary. I'm proud of one thing that OpenClaw achieved and then for many people it moved the eye from this thing that's like nebulous and scary into something that is fun and weird you know lobsters and all of that and they'll keep pushing there. keep building a great ecosystem of open source software with events that bring people together and with education. We have 10 people now on payroll. We're hiring a few more roles including a CEO. Second of all, I kind of made claw noun. You know, Kapati dropped the open Satya says enterprisegrade clause in Microsoft's keynote. There are 33,000 claw named repositories. If we are in a simulation, we're certainly in one of the weirder ones that won't get shut down. I dig this part of the future. Third, we still don't have an agent that's always on, always syncing. There's still so much to do. the landscape and the tech in AI are building faster than the software that you build around and that's an opportunity for all of you. Our workflows are also evolving like the vision I had early on where we shouldn't think about session or compaction. The world is finally the models that tech is finally getting to a stage where this is becoming a reality. We're entering a world where we finally move on from the text text only interfaces to voice and multimodality. Just yesterday we got a a hack working that your claw can now facetime you and this is kind of what matters why open claw exists at all. Every lab will sell you an agent. Open claw is the alternative. Open source runs everywhere, works with any model. And if you run local models, your data never has to leave your device. Your agent, your machine, your life. That's not even my C thesis, that's Gary's. You know, we just chipped it first. So we also finally are building open claw with open claw again but this time with a twist where everyone sees each other's session on a team server and there's a claw that knows what everyone else is working on and also can also take over orchestration of the work. Yeah, we are finally leaving this slowly this weird blip in the future where people work in terminals and run around with their laptops open because the agent needs to keep working. If you only remember three things, number one, don't stop having fun. Fun is the ultimate driver so you can you can get the best ideas. Number two, listen to your gut. And also fix the things that annoy you. it might just become the next big thing. And number three, stay focused. You know, another podcast is not what what will make you win. Live in the future, build what's missing. And when they write your arbiterary, keep shipping. It confuses them. Now, these were five questions. I'm sure you have many more. And I think we have plenty time. uh for QA.

所以先是智能体，然后是循环，然后是图。你现在是怎么构建的？

So first it was agents, then loops, then graphs. How are you building these days?

我的意思是，一直都是会话，对吧？但一开始你必须真的在意清理，确保你的指令连贯。而现在，我的会话更像是主题。清空会话有时甚至是个劣势，因为里面有太多信息能帮助智能体。

I mean it was always sessions, right? But in the beginning you had to like actually care that like you would clean and like you would make sure that your instructions are are coherent. And these days it's more like my sessions are topics. clearing the session sometimes even a disadvantage because there's so much information in there that helps the agent.

嗯，但最大的转变是，我试着让智能体为我做更多主动的工作。所以当我把注意力转移到某件事上时，我不想读 issue，我想看到经过全面审查和测试的 PR。也许我喜欢这个功能，也许不喜欢，但我不想分散注意力。同时，在工作中，如果你来找我，告诉我这个功能想法，我会生你的气。你只要和智能体讨论功能想法，构建它，截图，让我玩玩，这太容易了。这样，如果它好，我们可以立即迭代。但大多数时候，人们会自己发现为什么不好，甚至都不来找我。

Um, but the big thing that shifted is I try to make the agent do more proactive work for me. So when I shift my attention to something, I don't want to read issues; I want to see fully reviewed and tested PRs. Maybe I like the feature, maybe I don't, but I don't want to split my attention. At the same time, at work, if you come to me and tell me this feature idea, I'm going to get mad at you. It's so easy that you just discuss the feature idea with an agent, you build it, you make screenshots, you let me play with it. That way, if it's good, we can immediately iterate it. But most of the time, people will figure out why it's not good and don't even come to me.

图之后是什么，OpenClaw 的下一步是什么？

What comes after graphs and what's next for OpenClaw?

我知道有时候我喜欢在周日发帖，然后我的 Twitter 就爆了，但这也算不上一篇文章，对吧？所以任何时候你——你知道吗，我们工程师在很长一段时间里都在构建自动化来让生活更轻松。这从我们职业存在以来就一直在进行。所以如果你设计——叫它循环、图、工作流——其实都一样。如果你设计一个东西，它接收触发器或输入，为你做某事，中间可能有个决策，砰，这就是你的图。这并不神奇。尽管我偏爱那些图工程帖子，因为我有点好奇他们的观点，在我的世界里会怎样。这只是我们长久以来自动化故事的更好方式。

I know sometimes I like to post on a Sunday and then my Twitter explodes, but also it's not a post, right? So anytime you—you know what, we as engineers did for the longest time we built automations to make our life easier. That's been going on since our profession exists. So if you design—call it loop, call it graph, call it workflow—it's kind of all the same. If you design something that gets a trigger or gets an input and does something for you, and maybe there's a decision in the way, boom, there's your graph. It's not magical. Even though I favored some of those graph engineering posts because I'm kind of curious what their opinion is that it would be in my world. It's just a better way of our automation story that we did for so long already.

在快速构建、快速发布的世界里，你如何确保你构建的东西仍然可靠且可扩展？

In a world of build fast, ship fast, how do you make sure that what you build is still reliable and scalable?

是的，这部分我搞砸了一段时间，因为我不够专注，而且模型在测试方面确实不太好。我认为现在模型真的很好了。我们不仅有能记住的会话，我们还有编排训练进模型里，所以它们真正理解并知道使用子智能体。我们有计算机使用、浏览器使用。所有这些加在一起就像你完美的问答环境。我昨天就这么做了，我启动 Codex，使用 12 个子智能体，理解我的项目，把它分解成功能，然后每个子智能体对功能进行压力测试或代码审查，然后告诉另一个会话把测试重点放在哪里。我们还没到可以自动化一切的地步。你仍然需要做一些手动点击，以了解它的感觉。但对于用户会遇到的大量典型 bug，你现在实际上可以通过提示词走得很远。

Yeah, that's the part I messed up for a while because I was not focused and the models were not really good at testing. I think that the models are really good now. We have—not just sessions that remember—we have orchestration trained into models so they really understand and know to use sub-agents. We have computer use, we have browser use. All of that together is like your perfect Q&A environment. I did that yesterday where I spin up Codex, use 12 sub-agents, understand my project, break it down into features, and then each sub-agent would stress test the feature or code review feature and then inform the other session where to focus testing on. We're not yet at the point where you can automate everything. You still need to do some manual click-throughs so that you know how it feels. But for a lot of the typical bugs that users will encounter, you can actually now prompt your way and you get very far.

好的。AI 工具。你早期纯粹为了速度做的决定，今天仍然坚持的是什么？

Okay. AI tools. What's your decision you made purely for speed early that you still live with today?

我认为那是在早期决定不读所有代码。我把代码审查更多地视为风险管理。你知道，有时你接触一个可怕的系统，你会想更仔细地读一读；其他时候你构建 UI，如果它看起来正确，我真的在乎 UI 是怎么构建的吗？不。所以你粗略地看一眼，或者干脆接受它看起来是对的。部分原因是你会培养一种感觉，知道某件事应该花多长时间。所以如果我做一个小的调整，应该改变拖拽的工作方式，却花了三个小时，我就知道出问题了，我会仔细看。但除此之外，代码审查的一部分就是观察，看看改动有多大，然后相信你的直觉。

I think that was really early in deciding that I don't read all the code. I see code review more as risk management. You know, sometimes you touch a system that is scary. You kind of want to read a little bit closer, and other times you build UI. Do I really care if the UI is built correctly if it looks correct? No. So you kind of glance over or you simply accept that it looks right. Part of it is just like you develop a little bit of a feeling how long something should take. So if I do a small tweak that should change how dragging works and it takes three hours, I know something's wrong. I'll look closely. But otherwise, part of code review is simply observing, looking at how big the changes are, and trusting your gut.

好的。AI 工具现在让周末构建一个产品变得容易。但构建不是难事，难的是让人们使用它。你今天会如何从可用的原型走向最初的 10 个真实用户？

Okay. AI tools make it easy to build a product in a weekend now. But building isn't the hard part. Getting people to use this. How would you go from a working prototype to your first 10 real users today?

嗯，我认为那是我演讲中提到的部分。第一个用户应该是你自己。我的第 2 到 20 个用户是朋友。所以我确信你有一些人真的会测试你的东西，但你需要成为第一个用户。如果你对自己构建的东西不感到兴奋，那可能就没意义了。就像在这个时代，眼球是最昂贵的货币，因为现在构建东西太快了。

Well, I think that was part of what I addressed in my talk. User number one should be you. And my users two to 20 were friends. So I'm sure you have some people that would actually test your stuff, but you need to be the first user. If you don't get excited about what you're building, it'll probably not make sense. Like in this day and age, eyeballs are kind of the most expensive currency because building something is so fast now.

你如何平衡解决让你烦恼的问题和构建你认为人们想要的功能？

How do you balance solving issues that annoy you and making features you think people want?

这很难，因为通常它们是相辅相成的。如果某样东西不是我想要的样子，我会感到恼火。说实话，我还没有完全想清楚这部分。我觉得我可以花几个月时间处理 issue，因为软件中总会有奇怪的边缘情况，尤其是当你添加越来越多的功能时。但话说回来，如果我只做这个，我会对工作失去兴趣。所以需要健康的混合。

That's a hard one because oftentimes it kind of goes together. I get annoyed if something's not there as I want it. I haven't, to be honest, I haven't fully figured that part out yet. I feel I could just work for months on issues because there's always going to be weird edge cases in software, especially if you add more and more features. But then again, if I only do that, I'll lose interest in working. So it needs to be a healthy mix.

如果你能回去改变 OpenClaw 的一件事，你会改变什么？

If you could go back and change one thing about OpenClaw, what would you change?

我会对安全研究人员少一些压力。他们非常擅长让你感觉糟糕。他们会发报告，给我发邮件，打电话，用尽一切办法吸引我的注意，但实际上并不是为了帮助产品。大多数情况下，他们只是为了获得声望，为了得分。哦，我们发现了什么。而且他们中的大多数人真的发送了他们的智能体生成的报告，甚至没有实际测试过。我会采取更强硬的立场，解释哪些部分是我们保证的，哪些部分不会被修复，因为那不是我们的安全边界。但说实话，这是我第一次接触这个世界，所以我真的不知道如何处理，我花了好几个月，长了几根白头发才学会。

I would be less stressed out about security researchers. They are really good at making you feel really bad. They would send a report, they would email me, they would call me, they would do everything they can possibly get to get my attention, but not actually to help the product. In most cases, it's really just for them to get clout, for them to get a point. Oh, we found something. And most of them really sent reports that their agent produced without actually even testing it. I would take a stronger stance of explaining what are the parts that we guarantee and what are the parts that will not be fixed because that's not our security boundary. But honestly, this was my first time I was exposed to this world, so I just didn't know how to handle that, and I lost a few months and got a few gray hairs to learn.

当今智能体式基础设施中最大的瓶颈是什么？

What's the single biggest bottleneck in today's agentic infrastructure?

可靠性工具、记忆、演进、管理算力，如果你明白的话。比如如果我在本地机器上运行一个测试，TSGO 会启动 16 个线程，并成为我机器的瓶颈。如果 10 个会话这样做，其中两个可能会超时，我们必须重做。而且没有真正好的系统来管理所有这些。现在，如果我做网络相关的事情，创建云会话很容易。如果我做需要 macOS 的事情，那已经是 99% 的工具让我失望的地方。如果需要我电脑上的其他东西，也一样。而且我们还没有真正构建出能让东西毫无问题地在这里和那里移动的东西。至少我还没看到好的东西在运作。或者我能以可靠的方式驱动一支舰队。就像现在我用太多系统，我通过屏幕共享到一些电脑来分配负载。我不应该那样做。

Reliability tooling, memory, evolving, managing compute, if that makes sense. Like if I run one test locally on my machine, TSGO will spin up 16 threads and will bottleneck my machine. If 10 sessions do that, two of them will probably time out and we'll have to do it again. And there's no really good system of managing all of that. Now, if I do web stuff, it's easy to create a cloud session. If I do something that requires macOS, that's already where 99% of the tools fail me. If it's something that needs other stuff that's on my computer, same. And we haven't really built something yet where things could easily move from here to here without issues. At least I didn't see good stuff working yet. Or where I could drive a fleet in a reliable way. Like right now I use too many systems, I screen share into some computers to distribute the load. I shouldn't be doing that.

你如何保持开源项目有主见并指向一个方向，而不是合并第 N+1 个添加随机功能的 PR？你如何向社区传达这个方向？你有没有因为 PR 偏离项目方向而拒绝过受欢迎的 PR？

How do you keep an open source project opinionated and pointed in one direction rather than just merging the N plus 1 PR that adds a random feature? And how do you communicate that direction to the community? Have you ever said no to a popular PR because it pulled the project off course?

是的。是的。

Yeah. Yeah.

我会说，我拒绝得不够多。当我做新的开源项目时，我会写一个 vision.md 文件，说明它现在是什么以及我期望它走向何方。但当然，这并不准确，它不是完美的科学，而且我需要更好地遵守它，因为添加一个看起来很酷的功能总是很有诱惑力。但你很少想到的是，当你合并这个功能时，它实际上意味着一堆代码，那个人可能并不真正理解，我也不完全理解，而且我还要为此承担责任。

I would say I didn't say no enough. When I do new open source, I write a vision.md file where I explain what it is now and where I see it going. But of course, it's incorrect. It's not perfect science, and it's something I need to adhere to better, because it's always tempting to add this one feature that looks really cool. But what you rarely think about is that when you merge this feature, it really means here's this pile of code that the person probably doesn't really understand, that I don't fully understand, and that I also take on responsibility for.

你预计什么时候能有持续运行、主动工作的智能体？

When do you expect to have agents that are perpetually running and proactive in their work?

说实话，这与其说是技术问题，不如说是 token 问题。我们今天就能做到，但你的订阅额度撑不了多久，而且不是每个人都愿意花那么多 token。还有一个不容易的部分：设计一个不会白白烧掉 token 的系统。即使是我早期的“心跳”系统也太静态，不够主动。尤其糟糕的是，如果你有一个很大的会话，一小时后你调用一次心跳来检查所有东西，那意味着在 KV 缓存清除后，你要把 60 万 token 发回服务器，为大量无用工作支付一笔愚蠢的钱。所以有很多事情可以优化。这也真的非常难。

Honestly, that's not so much a tech problem. It's more a token problem. We could do that today, but you wouldn't get very far with your subscription, and not everyone is willing to spend so many tokens. It's also a part that's not easy: designing a system that doesn't just burn empty tokens. Even my early system of heartbeats was too static and not proactive enough. And especially bad if you have a large session and then one hour later you call a heartbeat to check up on everything, and that means you send 600,000 tokens back to the server after the KV cache cleared, and you pay a stupidly amount of money for a lot of not useful work. So there's a lot of things that can be done to optimize that. It's also really, really hard.

你能多讲讲你自己的个人设置吗？你把智能体托管在什么上面？你运行什么模型？你用什么 harness？你还读自己的代码吗？

Can you tell us more about your own personal setup? What do you host your agents on? What model are you running? What harness do you use? Do you still read your code?

我用我的 MacBook，但我通常用 Jump Desktop 远程到我的工作室，然后使用那里的电脑，因为那台电脑一直开着。在那台机器上，我可以快速运行任何我想运行的东西，而且不会耗尽电池，我可以合上笔记本电脑，事情继续运行。我还有几台其他远程机器，有时我会用 VNC 连进去。这样做的好处是，因为我做很多 Mac 软件，智能体喜欢接管我的屏幕并点击，如果你给它们一台自己的机器，它们就不会打扰你。否则，你会和智能体争抢鼠标光标。

I use my MacBook, but I usually use Jump Desktop to screen share into my studio and then just use the computer there, because that one's always running. That one I can run whatever I want fast and not drain my battery, and I can close my laptop and things keep working. I have a few other remote machines where I sometimes VNC in. The part that's really nice about it is because I do a lot of Mac software, agents love to take over my screen and click around, and if you give them their own machine, it will not bother you. Otherwise, you'll fight with the agent for the mouse cursor.

你还读自己的代码吗？

Do you still read your code?

我想我有点用“风险管理”这个词回答了这个问题。而且，如果我自己做开源，那和在 OpenAI 做软件的风险管理是不同的，在 OpenAI 我们仍然会阅读所有代码。

I think I answered that a little bit with the term risk management. And also, if I do open source by myself, it's a different risk management than if I do software at OpenAI, where we do still read all of the code.

Peter，你会如何着手你的下一个创业想法？

Peter, how would you approach your next startup idea?

我不知道。我应该把这个告诉所有人吗？我觉得这会被大量复制。我的意思是，最重要的是你需要构建你自己想用的东西。否则，它就不会好。其次，我希望你经营好自己的个人品牌和可见度，因为在这个时代，噪音太多了，你最困难的问题不是技术，不是软件，甚至不是人才。现在最困难的问题是获得关注。也许我会再次选择“困难且无聊”这个类别，因为这类问题通常更容易找到真正欣赏你解决问题的人。如果你选择有趣但困难的事情，你会非常艰难，尤其是在人们可以仅凭提示词就生成东西的时代。

I don't know. Should I tell this to everyone? I feel like it gets a lot of copies. I mean, the most important thing is you need to build something that you want to use. Otherwise, it'll just not be good. Second of all, I hope you worked on your personal brand and on visibility, because in this day and age there's so much noise out there that your hardest problem is not the tech, it's not the software, not even the people. The hardest problem now is getting eyeballs. Maybe I would pick something again that's in the category hard and boring, because that's usually a category that is a little bit easier to actually find people that will appreciate when you solved something. If you pick something that is fun, even if it's hard, you're gonna have a very tough time, especially in a time where people can just prompt things into existence.

好了，今天最后一个问题。你希望有人构建什么产品？

All right, last one for today. What's a product you want someone to build?

你知道，我觉得获得一个 Linux 测试机很容易，但获得一个真正好用的 Mac 测试机却难得离谱。而且我得到的那些 Windows 的东西也很烦人。我还没有找到一个真正好的提供商，能快速又便宜地做到这一切。我不确定这是否是你想做的生意，因为开发工具本身就很难，但这是我非常想要的东西。

You know, I feel it's very easy to get a test box for Linux. It is unreasonably hard to get one for Mac that works really well. And all the stuff that I got for Windows is also quite annoying. I haven't found a really good provider yet that does all of that fast and cheap. I'm not sure if this is the business you want to be in, because dev tools are inherently hard, but that's something I would love to have.

谢谢大家。

Thanks everyone.