从梦想到现实：一位 AI 原生设计师的旅程

我们经常听到设计领袖来到这个节目，谈论好奇心的价值，以及希望看到人们突破这项新技术的极限。但真正成为一个 AI 原生设计师是什么样子？那个闪烁的光标就像是一个邀请，让你对它说：我不知道。我不清楚你怎么工作。这是我的工作方式。我们一起来做点什么吧？然后转眼间，你就有了一个完整的 iOS 应用，里面集成了很多大语言模型，还能做金属着色器。你怎么创造那些属于你自己的体验，并反映出你作为设计师能带来的独特价值？我们有过质量，有过工艺，但我认为人们接下来会极度渴望的是更人性化的东西，对吧？比如灵魂，它将成为今年和明年的年度关键词。欢迎来到 Dive Club。我是 Rid，这里是设计师永不停止学习的地方。本周的嘉宾是我的朋友 Chris Pucket，他曾在 Dropbox、Mercury 领导设计，现在在 Stripe。他的故事完美诠释了如何真正拥抱 AI 这个时代：保持好奇，构建自己的想法，并帮助团队中的其他人提升。Chris 会分享一些他在 Stripe 的实践，也会深入探讨他如何利用这项新技术改善个人生活，首先从他的新移动应用 Epilogue 开始。

We hear a lot of design leaders come on this show and talk about the importance of curiosity and wanting to see people pushing the limits of what this new technology can do. But what does it look like to really be an AI native designer? That blinking carrot is this invitation for you to just tell it like I don't know. I don't have any idea how you work. Here's how I work. Like what can we do together? And then the next thing you know you've got like this full iOS app that has like a bunch of LLMs integrated into it and it's doing like metal shaders. How do you create experiences that are personal to you and reflect what you can bring to the table as a designer? We've had quality, we've had craft, but I think what people are going to be desperate for is more that human side of things, right? Like soul is going to be one of those things that is like that it'll be the word of the year like this year and next year is like soul. Welcome to Dive Club. My name is Rid and this is where designers never stop learning. This week's episode is with my friend Chris Pucket who's led design at Dropbox, Mercury, and now at Stripe. And his story is a perfect example of what it looks like to really lean into this moment in time with AI. To be curious, to build your ideas, and also to find ways to level up other people on your team. So Chris is going to share a little bit about his practice at Stripe, but also goes really deep into how he's using this new technology to improve his personal life as well, starting with his new mobile app Epilogue.

目标一直是设计和构建自己的软件。说实话，这个梦想我做了差不多 20 年。2005 年我买了第一台 iBook G4，当时觉得能自己写软件的人太酷了。但你必须会编程，这对我来说是个难题。大学一年级时，我决定去学如何编写 Mac 应用。那时候我们管它叫 OS 10。你得学一堆不同的编程语言，比如面向对象编程语言，我当时根本不懂那是什么意思。于是，我像所有傻瓜一样去了 Barnes & Noble，走到图书区，那里全是 PC 书籍，在最右下角有大概 10 本关于 Mac 软件开发的书。我完全看不懂他们在说什么。所以我想，那就去上个课吧，比如 CS 101。但那课程全是工程师思维、数学很重的东西，不是我自然倾向的。我几乎说服自己：我不擅长这个，这不适合我，我还是专注其他事情吧。但那个梦想一直还在：要是能自己做出那些东西该多酷啊？后来 iPhone 和 App Store 出来了，我就想成为能做出这些东西的人。所以我不断尝试，无论是 Treehouse 还是学习 Swift 或其他编程语言的课程，但始终没有真正开窍。我也不太确定为什么。但今年夏天，我在用我最喜欢的工具之一 Perplexity。我已经用了好几年，而且用的方式是他们从未预料到的——我会为每本在读的书创建空间。我喜欢实体书，一直是个爱读书的人，也喜欢就书提问。我是个喜欢纸笔和实体东西的人，但这样会错过数字化的力量。结果就是，我填满了笔记本，却忘了自己提过什么问题，或者根本没有去跟进查找。有一次在飞机上，我读着一本书，冒出很多问题，就用 Perplexity 空间为这些问题创建了独立的小图书馆。但你不能添加引用，不能加笔记，它们之间也没有关联。我当时想，要是我读的书能以某种方式相互关联，我能追踪它们就好了。Goodreads 从来没有真正打动我。它的设计不太合我口味，而且也不是为我想做的事情打造的——我想要一个私密的、只属于我自己的提问空间。所以我想要一个实现这个目标的方法。恰巧 Claude Code 刚发布，我就想，为什么不试试呢？于是我真的开始在 Cursor 里构建一个读书应用。但遇到了很多障碍，因为当时它并不适合开发 iOS 应用。后来 Claude Code 发布了，我在 Twitter 上看到很多人称赞它特别适合这种创意流程。所以我试了一下。结果发现，通过它我能比以往任何时候都更有效地构建自己的想法，并开始了这段旅程。

The goal had been to design my own and build my own software. And I had this dream for like 20 years to be honest. I got my first iBook G4 in 2005 and I thought it was so cool that people made their own software. But you have to know how to code and that was one thing like back in college I was a freshman and I was like I'm going to go learn how to code Mac apps. OS 10 was the thing that we called it then. You had to learn a bunch of different programming languages. Like it's an object-oriented programming language, which I did not understand what that meant when I was a freshman in college. So, I went to Barnes & Noble like any goober does and you're like going to go to the books book section. They have all these like PC books and on the very bottom right corner with like 10 books on developing Mac software. And I didn't understand a single thing that they were talking about. And so I was like, "Oh, I'll take a course. I'll take like a CS 101 course." And it was all very engineer, math-heavy, things that I don't naturally gravitate towards. And I kind of convinced myself like, "I'm not good at doing this. Like I'm I This is not for me. I'll just focus on other things." But the dream had kind of always been there of like, "Oh, how cool would it be?" And then the iPhone came out and and the App Store came out and I was like, "I want to be someone who makes these things." And so I constantly would try like whether it was Treehouse or the courses to learn Swift or other programming languages to kind of help get there and it just never really clicked. Uh and I'm not entirely sure why, but what ended up happening this summer, I was using one of my favorite tools called Perplexity. I've been using it for a few years and I was using it in ways that they had never really intended it to be used, which was like I would create spaces for each book that I was reading. I love physical books, always been a big reader, and I love being able to ask questions about the books. And I big pen and paper person and physical things, but then you miss out on the power of like the digital. What ends up happening is I fill out my notebooks and then I forget what questions I had or I wouldn't actually follow up and look them up. So, I was actually on an airplane flying and I was reading a book and I had so many questions about it and I'm using Perplexity spaces to create individual little libraries for the questions I'd have. But you can't add quotes and you can't add notes and it doesn't link together. And I thought it'd be so cool if like the books that I was reading, they'd all link in some way and I could kind of keep track of it. Goodreads never really like hit it for me. The design is unfortunately like not my jam and then it's not really built for what I was trying to do either which was like a private version of me just asking questions. And so I wanted a a way to do that and it just kind of clicked like Claude Code had just come out. Why wouldn't I try to do this? So I actually started building an a book reading app in Cursor. I was running into so many brick walls. Like it really wasn't at the time built for developing iOS apps. And so Claude Code had come out and I saw on Twitter a bunch of people like raving about how much better it was for this specific type of like creative flow. So I gave it a shot. And what I ended up finding was I could build my ideas through that so much more effectively than I ever could before and I was able to get started on this journey.

快速插播一条消息，然后我们继续。好了，我刚刚在 Paper 上有了顿悟时刻。我现在几乎把它看作一个共享白板，可以和 Claude Code 一起使用。所以如果我在构思一个想法，我可以让 Claude 探索六种不同的变体，它会立即在 Paper 的画布上开始设计，然后我就可以跳进去完善我最喜欢的概念。等我准备好了，我可以无缝地将设计传回 Claude Code 或 Codex 来实现它。Paper 是第一个完全互联的画布，它彻底改变了我对设计流程的思考方式。如果你还没试过，请访问 dive.club/paper 开始使用。另外，Jitter 刚刚发布了图片转视频功能，这相当厉害。你只需上传一个素材，Jitter 就会借助 AI 立即生成一个短视频片段。这是一种为静态背景添加动感和深度的简单方法，可以为品牌视觉创建小片段，或者通常用微妙的动作让任何场景更生动。玩起来非常有趣，今天就可以使用。请访问 dive.club/jitter 试试看。现在我们回到节目。

Real quick message and then we can jump back into it. All right, I just had my aha moment with paper. I'm now thinking of it almost like a shared whiteboard that I can use alongside Claude Code. So if I'm building an idea, I can tell Claude to explore six different variations and it'll immediately start designing on Paper's canvas where I can then hop in and refine my favorite concept. And then whenever I'm ready, I can seamlessly pass my design back to Claude Code or Codex to bring it to life. Paper is the first fully connected canvas and it is totally changing the way that I think about my design process. So if you haven't tried it already, head to dive.club/paper to get started. So Jitter just released image to video and it's a pretty big deal. All you have to do is upload an asset and Jitter will instantly generate a short video clip from it with the help of AI. It's an easy way to add motion and depth to static backgrounds, maybe create little clips for brand visuals, or just in general make any scene feel more alive with subtle movement. It's super fun to play with and available today. Just head to dive.club/jitter to try it out. Now on to the episode.

跟我们聊聊你是怎么获得动力的吧。从“我想做这个东西”开始，有那么多不同的路径可以走，你是从哪里开始的？你是倾向于提前把所有事情都计划好，还是只迈出一小步？从之前你无法实现的想法，到现在我们看到屏幕上这个漂亮的东西，你在这个过程中学到了什么？

Talk to us a little bit about how you got momentum. Like going from I want to build this thing and there's so many different paths that you can take in terms of where do you start? How much do you try to like plan it all out up front or just take little baby steps? Like what were some of the learnings that you encountered on this journey of going from an idea that previously you were incapable of bringing to life to, you know, we're looking at this thing on the screen here and it's beautiful.

啊，是的，谢谢。这实际上导致了我创作过程的根本性转变。我意识到自己一直在用“我不知道”这句话来限制自己的创意工作。我总是在等待。

Ah, yeah, thanks. It actually led to a pretty fundamental shift in my creative process. And I realized how much I was like self-limiting my own creative work by the phrase like I don't know. And I was constantly like waiting.

我不太确定自己到底在等什么，但总有一种需要许可的感觉。而我在使用 Claude Code 时发现，我根本不需要有完善的计划。我不需要完美的规格文档或 PRD 文档就能让 Claude Code 开始工作。我想我在这里更激进的观点是，我发现 Claude Code 在这个过程中是一个非常有效的共谋者，像是和我一起的共创者。在 Epilogue（你看到的这个应用）中，有时它会生成一些 UI，我自己在画布上可能不会那样设计，但我真的很喜欢那个初始想法。然后我就可以开始重塑它的输出，直到它变成我喜欢、我自豪、并且我认为更能代表愿景的东西。所以我最终意识到，我只需要问问题。就起步或获得动力而言，当你开始很快看到进展时，获得动力是相当容易的。

I wasn't sure what I was waiting for exactly, but there was this like need for permission. And what I found just working with Claude Code was mostly that I don't need to have fleshed out plans. I don't need to actually have like the perfect spec docs or PRD docs to get Claude Code to start working. I think probably the more hot take I would have here is that I ended up finding that Claude Code was a really effective like co-conspirator in this process. Like co-creator with me. That there are times that in Epilogue, the app you're seeing here, where it would spit out a piece of UI that I probably would not have designed on a canvas that way myself, but I actually really liked the initial thought. And so then I could just start like reshaping what it was doing until it was something that I loved and I was proud of and I thought like this was more representative of the vision. So, what I ended up doing was realizing I could just ask questions. So, in terms of like getting started or getting the momentum, I think it's pretty easy to get momentum going when you start seeing progress happen pretty quickly.

但举个例子，我稍后会回到这个欢迎页面。但在这里看到的图书馆网格，设置起来真的非常有趣，因为我以前从未做过类似的事情。我从未使用过 API。老实说，我几乎不知道 API 是什么。所以，要让一个 iOS 应用（或任何应用）从你搜索或阅读的书籍中拉取封面，你必须连接到一个最新的信息源。在我们的案例中，就是 Google Books API。例如，如果你搜索《奥德赛》，你会得到大量不同的结果，但我想要这个特定的封面。所以我能够过滤结果，找到我想要的。连接服务，就像获得动力，当我能告诉 Claude Code“我不知道怎么做。你只需要以某种方式连接这个服务”时，它会一步步引导我。好的，你需要去 Google Books，去 API，创建一个密钥，复制密钥，做这个，确保安全。它带我走过了所有这些步骤。这就像给动力油箱加满了燃料，让我觉得“哦，好的。哦，我在做事了。我在做事了。”然后你看到它出现在屏幕上。这太疯狂了。它工作了。我 literally 告诉它做什么。如果我卡住了，我可以 literally 说“我不知道。我不知道发生了什么。它坏了。你能修好吗？”我不知道为什么这不起作用。我截取我所看到的任何内容的屏幕截图，然后说“现在怎么办？”我的截图文件夹简直是疯了。如果有人翻看，他们会说“你有这么多没有意义的小 UI 片段。”所以，是的，很容易获得动力，然后“我不知道”变成了一个非常疯狂的邀请，进行更深入的对话，比如“如果你看这是 Xcode。如果你以前从未看过 Xcode，或者即使你习惯了像 Framer 或 Figma 这样的基于画布的工具，这并不容易弄清楚如何构建你的应用？比如，如何把它放进模拟器？这里有一个播放按钮。这很好，但你必须实际知道你的目标是什么？然后，更别提如何实际上传应用到 App Store 了。在后台，这很疯狂，如果我没有 Claude Code 第一次带我走这些步骤，我永远无法做到这一点。

But, for example, this is a I'll come back to this welcome sheet in a little bit. But, on the library grid that we see here, this was really, really like fun to get set up because I had never done anything like this before. I never worked with APIs before. I barely knew what an API was, to be honest. And so, in order to get an iOS app, or any app, to pull book covers from books that you would like to search for or read in your library, you have to connect it to some source of information that is up-to-date. And in our case, it's Google Books API. And so, for example, if you were to search for The Odyssey, you would get a ton of different results, but I wanted this specific cover. And so, I'd be able to filter through the results to find what I'm looking for. Connecting the service, like getting momentum, the minute that I was able to tell Claude Code, like, "I don't know how to do this. You literally just need you to connect this service somehow." It would walk through the steps. Okay, you need to go to Google Books, you need to go to the API, you need to create a key, copy the key, do the thing, let's make it secure. It would walk me through all those pieces. That is a huge, like, fuel in the tank for momentum to be like, "Oh, okay. Oh, I'm doing stuff. I'm doing stuff." And then you see it show up on the screen. They're like, "This is wild. Like, this is it's working. I'm literally telling it what to do. And if I get stuck, I can literally just say, 'I don't know. I don't know what happened. It's broke. Can you fix it?' Like, I don't I don't know why this is not working." I take a screenshot of whatever I'm looking at and I just say, "Now what?" My screenshot folder is unhinged, for sure. If anybody was ever go through it, they'd be like, "You have so many small little pieces of UI that don't make sense." So, yeah, it was easy to get the momentum filled, and then the I don't know became this really wild, like, invitation for a deeper conversation of like, "If you look at this is Xcode. If you've never looked at Xcode before, uh or even if you're used to tools like Framer or Figma, canvas-based tools, this does not make it easy to figure out how would you build your app? Like, how would you get it into simulator? Uh there's a play button up here. That's great, but you have to actually know, like, where are you targeting it to? And then, don't even get me started on like how you actually upload an app to the App Store. On the back end, it's wild, and if I didn't have Claude Code walking me through the steps the first time to do it, I would have just never been able to do this.

我欣赏你的坦诚。我想这就是为什么我如此喜欢这个故事——AI 在很多方面让我们能够触及更远，承担更雄心勃勃的想法。这确实很有野心。而且这不仅仅是让 Google Books 工作。我注意到一些设计细节，比如你甚至使用了金属着色器。跟我谈谈这个，因为对我来说，AI 突然让这些有趣的运动和动画效果变得更容易实现了。对于像“我想做类似的东西，但完全不知道从哪里开始”的人来说，这部分过程是如何运作的？

I appreciate the honesty. And I think that's why I like this story so much is AI in so many ways just allows us to reach further and take on more ambitious ideas. And like this is pretty ambitious. And it's not even just about getting like Google Books working. I mean even some of the design details that I noticed like you have these uh I think you're using like metal shaders even. Like talk to me a little bit about that because that's something that for me AI has made very interesting all of a sudden where it feels much more attainable to have these interesting motion and animation effects. Like for somebody who's like I want to do something like that, literally no idea what to start. How did that part of this process work?

我真正想要的是，这个应用的整个意图是成为一个环境阅读伴侣。环境这个词可能有点模糊，但它意味着它在那里，但又不在那里。我希望很多 UI 都有这种感觉。所以书籍封面自然会让它感觉有活力，与你同在。直到最近，我才让金属着色器工作起来。我对金属着色器一无所知。所以我使用了参数化设计技术来引入它，首先我必须教 Claude Code 关于金属着色器的知识。我找到一个关于金属着色器的很棒的 YouTube 视频，比如金属着色器入门。我创建了一个视频研究代理，它总结所有文字，提取出来，并帮助我创建一个技能。还有《着色器之书》，里面全是文字内容。我通读了所有内容，大部分都不懂。但我把选中的部分复制粘贴给研究代理来开发技能。所以我教 Claude Code 说，这是互联网上关于着色器的所有内容。让我们把它们全部编译成一个技能。然后采用我的想法，使用 Color Cube 提取书籍封面颜色，创建一个网格渐变，并在其上应用金属着色器。这就是我上周与 Claude 开始对话的方式。它最终构建了这个。然后我希望能够调整所有参数。所以我可以调整色温，使其更暗或更突出。我可以调整原点，即着色器从何处进入。所有这些部分帮助我微调着色器中我想要看到的内容。然后我可以复制参数值，回到 Claude Code 中说：“好的，我找到了我想要的调整。现在更新它。”它会更新，然后我可以预览，然后我们可以将其设为硬编码值。我也有不同的预设来测试哪些看起来真正好。所以我们可以浏览一下，现在我不太喜欢这些与调色板搭配的效果，但大致思路就是这样。所以我的过程就是，我不知道怎么做某事。

What I really wanted and like the entire intent of this app is to have this ambient reading companion. And ambient can be like cuz like a wishy-washy word, but it's like it's there, but it's not there. And I kind of wanted a lot of the UI to feel that way. And so the book covers really felt like something that would be a natural way to make it feel like kind of alive and with you. Uh And it wasn't until recently that I was able to get the metal shaders to work. I don't know anything about metal shaders. So what I did was I used this parametric design technique to be able to pull in one, I had to teach Claude Code about metal shaders. So what I did, there's a great YouTube video out around metal shaders, like intro to metal shader. And I've got this video research agent that I created that summarizes all the copy, pulls it out, and helps me create a skill. And then there's the the Book of Shaders, which is all the copy uh content in there. And so I would I read through all that, didn't understand most of it. But I would copy-paste the like selections to it to the research agent to develop the skill. So I taught Claude Code like here's everything about shaders that is available on the internet. Like let's compile it all into one skill. And then let's take this idea that I have of being able to extract the book cover colors using Color Cube to create a mesh gradient that we apply a metal shader on top of. That's literally like how I started this conversation this past week with Claude. It ended up building this. And then I wanted to be able to adjust all of the parameters. So, I can adjust like color temperature, so make it darker, make it more prominent. I can adjust like the origin, so like where is the shader kind of coming in from. And all of these pieces just help me fine-tune you know, what exactly I want to see in the shader itself. Then I can copy the parameter values and go back into Claude code and say, "Okay, I've found like the tweak that I want now. Go ahead and update it." And it will update it, then I can preview it, and then we can make that like the hard-coded value. And I've also got like different presets to test out of like what would actually look really good. So, we can kind of go through and see now I don't really like any of these with the color palette, but that is the general idea. And so, my process was to like I don't know how to do something.

「我不知道」变成了一种超能力，因为我假设自己可以用 Claude Code 来解决它。LLM 会疯狂地产生幻觉，会编造东西。所以我一直在反复核查它的信息。比如我不断让它给自己的计划打分：你产生幻觉了吗？你引用的来源准确吗？就因为我不信任它们。总之，我不是完全信任 AI，我们需要自我检查，但输出结果我觉得相当扎实。我对此非常兴奋。另外一件事是，我想制作一张欢迎卡片。同样，目前这主要是为了我自己。所以当我有时阅读——爸爸的生活、工作、各种事——我读书进度落后了。所以我想要一种方式，有一个小小的欢迎回来卡片。欢迎卡片最初——我把它调出来，肯定在某个地方。这是我设计的欢迎卡片的初始版本。这就是我经历的迭代过程：告诉 Claude Code，‘好的，我想创建一个欢迎卡片。我希望它包含我最近在读的书。如果我有任何与这本书相关的保存引文，我希望它们能显示出来，然后还有行动号召，也许还有一些统计数据。’这就是我们最初想出来的，它用了我这里的琥珀色主题。所以 Epilog 的理念就像你坐在炉火旁，一杯茶，一本书。它相当温暖，带有橙色的琥珀光晕。看起来不错，但我有点想，‘哇，我刚刚做了一个很酷的金属着色器系统，看起来真的很棒，我为此感到自豪。我真的很想把它用在这些欢迎卡片上。’所以我们不断改进，我不得不为卡片专门调整一些参数，这就是我们最终得到的，我认为它实际上看起来好多了。

I don't know becomes kind of a superpower because I'm assuming I could work with Claude Code to figure it out. LLMs are going to hallucinate like crazy, going to make stuff up. So, I'm constantly also rechecking its information. Like I'm asking it all the time to grade your own plan. Did you hallucinate? Did you pull from accurate sources? Just because I don't trust them. And anyway, I'm not trusting the AI fully; we're going to need to check ourselves on that, but the output I think is pretty solid. I'm really pumped about this. And another piece of this was that I wanted to make a welcome card. Again, this is mostly for me at this point. So, when I read sometimes—dad life, work, whatever—I'm behind on my reading. And so, I wanted a way to have a little welcome back card. The welcome back card was initially—I'll pull it up, sure I've got it here somewhere. This was the initial version of the welcome back card that I designed. So, this is like the iteration process that I would go through, which is tell Claude Code, 'Okay, I want to create a welcome back card. I want it to have the book that I was most recently reading. I would love it to have a quote that if I have any saved quotes related to the book, those surface, and then call to action, maybe some stats.' So, this is what we came up with initially, and it's using my ambient amber theme here. So, the idea of Epilog is like you're by a fire, a cup of tea, a book. It's fairly warm with the orange amber glow to it. It looks good, but I kind of thought, 'Wow, I just made this cool metal shader system that looks really good. I'm really proud of. I would love to see that actually on these welcome back cards themselves.' So, we kept refining, and I had to change some of those parameters specifically for the card, and this is what we ended up with, and I think it actually looks a lot better.

也许再放大一点看，自从开始做这个项目，你现在已经在 Stripe 工作了。那么，经历了这些反复，拥抱了‘我不知道’，这个构建 Epilog 的过程作为设计师为你解锁了什么，你现在把它带到了日常工作的实践中？

Maybe even zooming out, since starting to work on this, you're now at Stripe. So, having put in these reps, having leaned into the 'I don't know,' like what did this process of building Epilog unlock for you as a designer that you're now bringing into your practice in your day job?

在构建这个应用和做这件事的过程中，我有点发现，‘哦，我其实不需要许可。’就像我可以直接去实现我的想法。我直到看到‘我在 App Store 里有了一个应用’才意识到这有多强大。现在，它按任何标准都不算超级成功，但对我来说已经很棒了。所以，成功，我想我们可以用非常不同的方式来定义，但它教会了我很多：我不需要许可就能构建我的想法。我不需要担心被原谅。那是一种奇怪的权力动态。就工作以及它如何体现在我的角色中而言，我认为尤其是在 Stripe，它帮助我更多地主动出击、承担风险和抓住机会。所以我把自己推出去。我在创造这些东西，无论是为团队制作 Claude 技能，还是任何我能填补空白的方式。但主要是内在的，它建立了很多信心：如果我在某件事上卡住了，‘我不知道’不必是一扇关闭的门。它可以是一个把手，我只需转动它，然后走过去解决问题。而且我知道我有像 Claude 这样的其他资源来帮助我。

In the process of building this app and doing this, I kind of found that, 'Oh, I actually don't need permission.' Like I can just go build my ideas. I didn't realize how powerful that was until I started seeing, 'I have an app in the App Store.' Now, it's not like super successful by any metrics, but it's been great for me. And so, success, I guess we could define it very differently, but it has taught me a lot that I don't need permission to build my ideas. I don't need to worry about forgiveness. Like that's a weird power dynamic. And in terms of the work and how it's showed up in my role, I think especially at Stripe, it's helped me just kind of take a little bit more initiative and risks and chances. So, I'm putting myself out there. I'm creating these things, and whether it's skills for Claude for the team, or any way I can fill the gap. But mostly internally, it's just built a lot of confidence that if I'm stuck on something, 'I don't know' doesn't have to be a door that's shut. It can be easily a handle that I just turn and walk through to figure it out. And I know that I have these other resources like Claude to help me.

你提到了为团队构建技能，我觉得这很有意思，因为在节目中，用 AI 构建通常看起来像是构建一个一次性的设计工具或某种创意模式生成器。这些都令人惊叹，但我认为有趣的是，你指出无许可构建几乎更多地体现在内部流程、工作方式以及在这些方面推动进展。

You mentioned building skills for the team, which I think is kind of interesting because so often in the show, what building with AI looks like is building a one-off design tool or some kind of a creative pattern generator. And those are all amazing, but I thought it was interesting that you pointed permissionless building almost more on internal processes, ways of working, moving the needle in those ways.

现在用 AI 做的最简单的事情就是生成想法来构建产品、应用、网站等等。而且感觉它在 Twitter 上也表现很好。里面有很多乐趣。内部也很有趣。我认为现在每个做大量设计工作的公司，都在构建自己的内部工具，比如‘我们构建了自己的原型工具，团队正在使用。’这都很酷。对我来说，AI 在工作流方面还有其他的巨大解锁，无论是内容设计、你看不到的东西、组织、分析。所以这些 Claude Code 技能，你可以创建任何可重复的模式。你可以为此创建一个 Claude Code 技能。所以我试图找到填补空白的方法，如果有人时间管理困难，你可以为此创建一个技能。它有很多不同的部分，不需要非常技术性就能利用 Claude Code。我认为这是我试图帮助任何我与之交谈或合作的团队或设计师的一件事，因为我认为我们还处于非常早期的阶段。当然，你的很多嘉宾，我看过很多。我的意思是，我觉得每一集。我认为我们都处于 AI 阶段的非常早期。我们都觉得这一切都很正常。我认为还有大量的人，他们对 AI 的使用是，‘嗯，我和 ChatGPT 聊过几次，或者问了它一些问题。’这完全连可能性的表面都没有触及。无论是构建一个 iOS 应用，还是构建一种组织你一天的方式，Claude Code 可以为我们做更多。所以，一个很好的例子是，我们创建了一个流程来提取所有必须完成的任务，无论是 Google Docs 还是 MCP 的各种部分，因为我们的上下文现在太大了。比如，如果你在一个大型组织中，你有 Google Docs、Linear 工单，如果你使用 Linear 或 Jira 或任何管理系统，所有这些东西都分散在各处，所有沟通也分散在各处。但是，你可以把它连接到 Claude Code，让它提取并编译所有这些内容到一个统一的来源。甚至对我个人来说，比如我开始使用 Linear 让 Claude Code 更新自己的状态。所以我让 Claude Code A，创建所有项目，B，接受我的大脑倾泻，里面是我希望 Epilog 做的所有事情。让我们去构建一堆工单。它为我写了所有工单。现在我可以每天早上在开始一天之前让它过一遍工单。比如我下楼，让它启动几件事。然后我回来，看看完成了什么，审查一下，从待办事项中划掉。Claude Code 可以做很多这样的事情，减轻我们大量的精神负担，这样我们就可以专注于我们真正想做的工作。

The easy thing to use AI for right now is to generate ideas to build something like a product, an app, website, whatever it is. And it feels like it also just does really well on Twitter. Like there's a lot of fun in it. It's fun internally. I think every company that's doing a lot of design stuff right now, they all have their internal tools they're building that are like, 'We built our own prototyping tool that teams are using.' And that's all really cool. There are other massive unlocks to me for AI around workflows, around how—whether it's content design, things that you wouldn't see, organization, analysis. So these Claude Code skills, you can create anything that's a repeatable pattern. You can create a Claude Code skill for that. And so I try to find ways to fill in the gaps of if there are people that are kind of struggling with time management, you can create a skill for that. Like there's so many different pieces of it that it doesn't have to be super technical in order to take advantage of Claude Code. And I think that's the one thing that I am trying to help any team or designer that I've talked with or work on this because I think we're still so early. Sure a lot of your guests that you have on and I've watched a lot of it. I mean, I feel like every episode. I think we are all very early in the AI phase. And we all feel like this is all very normal. I think there's a very large amount of people out there that are still like their usage of AI is, 'Well, I talked to ChatGPT a couple of times or asked it some questions.' And it has completely not even scratched the surface of what's possible. Whether it's building an iOS app or building a way that you would organize your day, Claude Code can do so much more for us. So, a great example, too, was like we created this process for pulling out all of the tasks that have to be done, whether it's the Google Docs or all the different pieces of the MCPs because our context is so large now. Like, if you're in a large organization, you have Google Docs, Linear tickets, if you use Linear or Jira or whatever management system, all that stuff is so spread out and all the communication is so spread out. But, you can connect it to Claude Code to have it pull and compile all those things into a single unified source. Or even personally for me, like I started using Linear to have Claude Code update its own statuses. So, I had Claude Code A, create all the projects, B, take my brain dump that I had of here's everything I want Epilog to do. Let's go and build a bunch of tickets. It wrote all the tickets for me. And now I'm able to just have it go through the tickets every morning before I start my day. Like I come down, I can get it fired up on a couple of things. Then I can come back, see what's done, review it, knock it off the backlog. Like Claude Code can do a ton of those pieces that take a ton off of our mental load so we can focus on the work that we really want to be doing.

好，那我们来角色扮演一下。对，我来找你。我是一名设计师，可能在一家初创公司领导设计团队，团队大概四个人，公司总共四十人左右，规模较小。我对 AI 的使用就是刚开始在 Lovable 里做了一个模式生成器，然后用 ChatGPT 聊天来提问。仅此而已。我没有实现任何可重复或系统级的东西。老实说，我甚至不知道从哪里开始。你会怎么帮我迈出最初几步？

All right, so let's play a little role playing, then. Yeah, I'm coming to you. I'm a designer. I'm maybe leading design at a startup. There's like four of us. And, you know, maybe team of 40 total, you know, smaller company. Basically, my usage of AI is I've just started making like a pattern generator in Lovable, and I use ChatGPT chats to ask questions. And that's about it. I don't have anything repeatable or system-level implemented. And honestly, I don't really even know where to start. How would you help me take those first few steps?

这也是我最喜欢的，因为“我甚至不知道从哪里开始”就是最好的起点。而且非常简单。你只需要告诉它“我不知道从哪里开始”，它就会帮你。比如 Claude Code 有一个提问工具，内置于系统中。如果你让它“我想让你问我一堆问题。我会告诉你我是谁，我做什么。我想让你问我一些问题，帮我弄清楚你能在哪些方面帮我。”它就会照做，然后给出一些可能比较通用的想法，但很可能会有两三个让你灵光一现——你会想“我完全不知道还能这样。我不知道有这种可能性。我们来探索一下。”然后你开始深入挖掘，突然之间，你为一个五秒钟的问题投入了两个小时，并构建出这些新系统。所以，“我不知道”这个基础部分其实是一个很好的起点。我是一家初创公司的设计主管，手下有四个我支持的人，这是我的工作内容，这是我的一般日常。帮我弄清楚你能在哪些方面帮我。这会比使用 Lovable 更快地改变你与这个工具的互动方式。不是说 Lovable 或其他工具不好——它们在创建视觉制品和实际输出方面确实很有趣，但就理解并解锁新的工作方式而言，从那个基础部分开始才是关键。

That's my favorite, too, because the 'I don't even know where to start' is the greatest starting point. And it is so simple. You just have to tell it 'I don't know where to start' and it will help you. So Claude Code, for example, has an ask tool — a question ask tool — built into the system. If you ask it 'I want you to ask me a bunch of questions. I'm going to tell you who I am, tell you what I do. I want you to ask me a bunch of questions, help me figure out where you can help me.' It will do that and it will spit out a handful of maybe a couple generic ideas, but it will probably spit out two or three that will switch something on your brain — you're like 'I had no idea I could do that. I didn't know that was a possibility. Let's explore that.' And then you start digging into it and all of a sudden you're like 2 hours into a 5-second question that you typed in and you're building out these new systems. So the foundational piece is 'I don't know' is actually a great place to start. I'm a design lead at a startup, got four people that I support, and here's what I do, here's my general day. Help me figure out where you can help me. It will change how you interact with this tool faster than using Lovable will. Not a knock on Lovable or any of those other tools — they're really fun for creating visual artifacts and actual output, but in terms of understanding and unlocking new ways of working for you, it's starting with that foundational piece.

如果我们想更战术性地谈谈它还能做什么。我认为很多时候我们觉得自己没有可重复的模式。其实我们有，只是不擅长识别它们。所以这是 Claude Code 可以帮忙的另一个方面：你可以说“这是我的日常，这是我的日历，这是我大部分时间花在的地方。我错过了什么？”这个“我错过了什么”是一个很好的问题，因为 LLM 的核心就是模式匹配，它们非常擅长快速识别模式——尽管人脑是不可思议的、历史上最伟大的模式匹配器，但我们仍然有这些盲点。所以你可以用 AI 作为一面镜子，反射出你可能忽略的关于自己、工作和现状的一些东西。

And if we wanted to get more tactical of like other things it can do. I think a lot of times we don't think we've got repeatable patterns. We do, we're just not great at identifying them. And so that's another piece that I think Claude Code can help with where you're like 'here's my day, here's my calendar, here's like where I spend most of my time. What am I missing?' The 'what am I missing' piece is like a great question to ask because at their core LLMs are just pattern matching and so they are fantastic at very quickly identifying patterns that even though the human brain is an unbelievable — like the greatest pattern matcher in history — we still have these gaps. And so you can use AI to become this mirror that's reflecting back at you some things that you might be missing about yourself and your work and where you're at.

有一件事我认为 LLM 并不擅长，那就是创意探索，但你可以教它快速进行发散性思维。这对团队来说也是一个巨大的解锁。比如，现在设计师的标准做法是“疯狂八格”。这很有趣——也许现在不流行了，可能暴露了我的年龄，但说真的，它仍然很酷。我们别贬低它，我觉得它还是很酷。它的玩法是，你应该在白板或纸上画四个格子。如果你不知道“疯狂八格”是什么，就是画四个方框，拿一个想法，然后快速创建变体。LLM 其实很擅长这个。你只需要沟通清楚你想做什么。所以你可以教它做这些事。但我觉得从根本上说，人们不喜欢这个建议，尽管它对他们来说是最好的建议：如果你不知道从哪里开始，那就直接开始打字。那个闪烁的光标就是在邀请你告诉它：“我不知道。我不知道你怎么工作。这是我的工作方式。我们能一起做什么？”我保证你会得到一些有趣的想法。如果你接着问：“这很酷，但太表面了，我们深入一点。”它就会越来越深入。你给它的关于工作和目标的上下文越多，输出就越好。你只需要一点启动能量。然后不知不觉，你就有了一个完整的 iOS 应用，里面集成了很多 LLM，还能做金属着色器。这真是一个疯狂的起点。

One thing that LLMs I don't think do really well at is creative exploration, but you can teach it to do divergent thinking very quickly. And that is a huge unlock for teams as well. So like, you know, the standard for designers right now is crazy eights. Like it's such a fun — maybe not anymore. Maybe that's showing my age, but you — It's still cool, honestly. We knock it. I think it's still cool. I think it's cool. Like it plays. I mean I don't know about you. Like the idea is you should draw like on a whiteboard or a piece of paper. Like if you don't know what crazy eights are, you draw like four boxes. You take one idea and then you try to like just quickly create variations through that. LLMs would actually be really good at that. You just have to kind of communicate that's what you want to do. So there's things you can do to teach that. But I think foundationally people don't love this advice even though it's the best advice for them, which is if you don't know where to start, you just start typing. Like that blinking cursor is this invitation for you to just tell it like, 'I don't know. I don't have any idea how you work. Here's how I work. Like what can we do together?' I guarantee you will get some interesting ideas. And if you were to follow up and say, 'That's cool. It's super surface-y. Let's go deeper.' Then it really starts digging in more and more. The more context you give it about your work and what you're trying to accomplish, the better the output will be. And really you just need a little bit of gas in the tank to get started. And then the next thing you know, you've got like this full iOS app that has like a bunch of LLMs integrated into it and it's doing like metal shaders. Like that's just — it's a wild entry point.

过去一年，Dive Club 让我清楚地认识到，设计实践正在发生变化。旧的反馈流程在当今世界已经行不通了。这就是为什么我很高兴地宣布 Inflight 正式进入公开测试阶段。它是我一直想要的反馈工具，专为 AI 速度的世界而构建。我可以分享原型，通过视频导览提供上下文，Inflight 让我轻松获得前进所需的确切反馈，无论是投票选择方向，还是为发布新想法开绿灯。所有这些都可以通过一个链接实现，我可以把它丢进 Slack，或者与高级用户分享来测试新原型。我每天都在用 Inflight，它彻底改变了我分享工作的方式。所以我很期待你试用这个产品，如果你想聊聊它，请发邮件到 rid@inflight.co。

One thing that Dive Club has made abundantly clear to me over the last year is that the practice of design is changing. And the old process of getting feedback just doesn't quite cut it in today's world. That's why I'm excited to announce that Inflight is officially in open beta. It's the feedback tool that I've always wanted and it's built for a world that moves at the speed of AI. So I can share my prototypes, give context in video walkthroughs, and Inflight makes it easy to get the exact feedback that I need to move forward, whether it's voting on directions or maybe even getting the green light to ship a new idea. And all of this is available in a single link that I can drop into Slack or maybe even share with power users to test out a new prototype. I use InFlight every day and it's totally transformed the way that I share work. So, I'm excited for you to try the product and if you ever want to jam about it, just email me at rid@inflight.co.

每当我让你戴上顾问帽子时，你都会兴奋起来。所以我就继续深入了，再给你一个——不是假设，这完全是我的实际情况，也是我最近在尝试的一个工作流程：让 Claude 在某个区域构建东西，但我也想直观地探索某个组件或画布上的变体，让它在后台工作。我发现自己经常遇到这样的提示：“嘿，我有这个东西。你能不能——我想说疯狂八格。你能帮我用不同的方式看看吗？我就是想换个角度看。”这甚至和你之前提到的有关：“也许你不喜欢 Claude 在设计中创造的第一件东西，但它帮你看到了稍微不同的东西，把你带到了一个你原本不会去的方向。”我反复发现这就是 AI 的价值主张。

You're kind of lighting up whenever I make you put on your consultant hat. So, I'm just going to keep leaning in here and I'm going to give you another — not a hypothetical, this is very much so my situation and a workflow that I've been experimenting with more recently, which is having Claude, maybe I have, you know, an area over here and I'm building something, but maybe I want to explore visually some permutations of a component or something on a canvas and actually have it kind of work on things in the background. And I'm finding myself in this situation where I have this prompt pretty often where it's like, 'Hey, I have this thing. Can you just — I want to say crazy eights. Like, can you just help me see this different ways? Like, I just want to see this differently.' And it even ties back to something that you were mentioning earlier where you were like, 'Maybe you don't, you know, love that first thing that Claude created design-wise, but it helped you, you know, see something slightly differently and took you in a direction you wouldn't want to go.' I am finding that as a value proposition with AI like repeatedly.

是的，完全正确。这是一个很好的用例。当你给 LLM 正确的上下文时，快速生成变体是它们的强项。你可以说：“这是我正在处理的组件。生成 10 个视觉变体，每个有不同的布局或配色方案。”即使你一个都不喜欢，它们也可能激发一个想法，带你走向新的方向。这种发散性思维是这些模型的超能力。

Yeah, absolutely. And that's a great use case. The ability to generate variations quickly is something LLMs excel at when you give them the right context. You can say, 'Here's a component I'm working on. Generate 10 visual variations of it, each with a different layout or color scheme.' And even if you don't love any of them, they might spark an idea that leads you somewhere new. That divergent thinking is a superpower of these models.

现在我的问题是，我该如何投资那种我反复使用的提示词，比如‘嘿，帮我换个角度看问题’。我该如何突破现有的局限，拓宽对可能性范围的认识？这是否是一个培养某种技能的机会，或者我该做些什么，既能提高那个时刻的上限，又能让我更高效地迈出那一步？

Now, my question is, how do I invest in that type of prompt that I keep going back to the well on, which is like, 'Hey, help me see this differently.' How do I push past the thing that I have and broaden my horizons for what this spectrum of possibilities could be? Is that an opportunity to build some kind of a skill or just anything that I should be doing to A raise the ceiling for that moment and then also B allow me to take that step more efficiently?

是的，100%，这个问题提得很好。技能的核心就是指令。如果你在网上看到人们谈论技能文件、skills.md，那是 Anthropic 创建的，现在我认为它已经成为其他前沿模型（如 Gemini 和 ChatGPT 等）的标准。它就是一个 markdown 文件，纯文本，只是一组可重复的特定指令，用来增强 Claude Code 的能力。你可以调用这个技能。比如，如果我现在打开我的技能文件夹，里面有好几个技能。我有一个自己做的 liquid glass 技能。我之所以要创建这个技能，是因为这是一个重复出现的问题，尤其是在早期。我做了一个非常愚蠢的决定，为自己开发一个应用，而且我对 iOS 26 非常兴奋。说实话，当 liquid glass 发布时，我其实持观望态度。我主要兴奋的是 Apple 正在做的一些基础模型相关的事情，也就是设备端的大语言模型，叫做 Apple Foundations 框架。为了测试这个，我的想法是，能不能用它来帮助我实现 Epilog 的一些功能。所以当我开始做这个，iOS 26 发布时，Claude Code 还没有访问权限。所有关于 liquid glass 以及如何正确实现这些模式的信息都还没有。Apple 的开发者库，Apple 的知识库，全部是用 JavaScript 写的。大语言模型不太擅长读 JavaScript。如果一个页面全是 JavaScript，大语言模型很难处理。我不能直接把 Apple 开发者文档的链接贴给它，它就读不了，真的很吃力。所以我不得不一直粘贴。我甚至觉得我的剪贴板里有一些 liquid glass 的原则，我反复粘贴，这是其中之一。关于使用不同的形状，如何使用等等。所以我一直在做这些事情。当技能功能推出时，它让我能够卸载所有那些我和 Claude Code 一起经历了很多令人沮丧的对话才获得的信息——比如让导航栏实现 liquid glass 效果，拥有自定义图标等等——我现在可以教它如何做到这一点。这就是技能的本质。当我提到 liquid glass 时，技能就会被加载，因为你可以在基础文件（Claude MD 文件）中告诉 Claude Code，用自然语言关联事物。这样我就不必总是手动调用技能，不必用斜杠命令来调出技能。我只需要告诉它我正在处理 liquid glass，它就会自动调出技能来理解相关信息。所以对你或者任何其他做这些重复模式的人来说，技能是一个很好的方法。技能最好的地方在于，你实际上不需要知道如何创建技能，因为内置了一个技能创建器。你可以告诉 Claude Code，我想创建一个技能，它会问你，你想让这个技能做什么？然后你可以告诉它你的重复模式，它会制定一个计划，输出技能，然后你甚至可以让它自我审查。比如，根据你听到的意图，这个技能怎样才能达到 A+ 级别，给自己打个分，它通常会回来说，可能是 B。我感觉 Claude Code 在这方面挺谦虚的。所以我会说，好吧，我不想要 B 级的工作，我想要 A 级。Claude，我的团队里没有 B 级选手，然后它会稍微重构一下，给出一些更高级的想法。然后你就有了一个技能，它会热加载，这意味着你不需要退出 Claude 就能使用它。技能安装到文件夹的那一刻，就可以用了。任何人都可以做到。如果你有经常做的重复模式，你可以这样做。如果你有一些代码或提示词想执行，但更偏向于每天固定时间做某个特定提示，你可以创建一个循环，我认为就是 /loop。这些循环会在设定的时间间隔运行提示或斜杠命令，所以你可以告诉它你希望多久做一次特定的事情，它就会为你运行一个特定的提示。这与技能不同，技能是一组指令，技能可以包含提示来完成目标，但循环主要是针对你有一个提示的情况。比如我每天早上都会用的提示：我想让你查看我的健康指标，看看我的 HRV 数据，告诉我我有多不健康，或者告诉我我的能量水平，这就是一个我可以用循环来执行的提示。我也有其他系统来做这件事，但你打开了一个新的兔子洞，更偏向个人方面，我觉得这非常有趣，而且我很享受这次对话的一点是，我们正在拓宽我通常给出的答案范围，比如作为一个 AI 原生设计师应该是什么样子。其实不是的，如果你知道有哪些工具可用，并且知道如何提出正确的问题，就有很多不同的机会可以利用这些东西。

Yeah, 100% that's a great call out. Skills are at their core just instructions. So, if you've seen online people talk about skill files, skills.md, and they're files that Anthropic created and then now I think they're a standard across the other frontier models like Gemini and ChatGPT and the others. And it's just a markdown file. It's just text and all it is is it's a set of repeatable specific instructions that augments what Claude Code can do. You can call the skill up. So, for example, if I were to go into mine right now, I've got a handful of skills in here. So, I've got like this liquid glass skill that I made. The liquid glass skill I had to create because it was a repeatable problem especially early on in this. So, I made the very stupid decision of making an app that was specifically for me and I was really excited about iOS 26. I was like on the fence about liquid glass, to be honest, when it like was announced. I was mostly excited about some of the foundation model things that Apple was doing. So, the on-device LLMs called like the Apple Foundations framework. And in order to test that out, what the idea was, what if I could use that to help some of the things I want to do for Epilog. And so, when I started working on this and iOS 26 came out, Claude Code did not have access. Like all the information around liquid glass and how to do these patterns properly was not available yet. Apple's developer library, Apple's like knowledge base, is all in JavaScript. LLMs don't read JavaScript very well. So, if a page is entirely in JavaScript, LLMs have a really hard time. Like I can't just paste the link to Apple's developer docs and it reads it. It really struggles. And so, what I had to do, like I kept pasting. I even think in my clipboard there's like some liquid glass principles that I was constantly repasting over and over and this is one of them. Around different shapes to use, like how to use it. And so, I was always doing these things. When skills came out, it allowed me to offload all of the information that we had spent a ton of frustrating conversations together, Claude Code and I, to get the nav bar to do the liquid glass thing and to have custom icons and all these pieces, I can now teach it how to do that. And that's what a skill is. And so the skill is loaded when I mention liquid glass cuz you can tell Claude Code in the foundation file, the Claude MD file, to be like natural language associated to things. So I don't have to always invoke the skills, I don't have to do the forward slash command to get my skills up. I can just tell it I'm working on liquid glass and it will pull the skill up to be able to understand the information. So for you or for anyone else that's doing these repeated patterns, a skill is a great way to do that. The best part about skills is that you don't actually have to know how to create a skill because there's a skill creator built in. You can tell Claude Code, I want to make a skill and it will ask you, what do you want to do for the skill? And you will be able to tell it so for you, your repeated patterns that you're trying to do and then it will come up with a plan and then it will output the skill and then you can actually have it review its own work. Like what would take this skill to be like an A+ based on the intent that you heard me say and like give it yourself a grade and then it'll usually come back and it'll say like maybe a B. Almost I feel like Claude Code's like pretty humble about it. So I'm like, okay, cool. Like I don't want B work, I want A work. Like Claude, we have no B players on my team and then it will refactor it a little bit, give some thoughts of like what it could do that'd be more advanced. And then you have a skill and it hot loads, which means you don't have to like log out of Claude to use it. The minute that the skill is installed in the folder, it's good to go. And so anyone can do that. Like if you have repeated patterns that you're doing often, you can do that. If there are like code or prompts that you want to do, so less instructions and more just like every day I'm around this time doing this specific prompt, you can create a loop that I think is just /loop. And these run prompts or slash commands on a recurrent interval, so you can tell it like how often you want to do specific things and that will run a specific prompt for you. This is different than a skill where a skill is like a set of instructions and a skill can include prompts in order to to accomplish the thing, but this is really meant for just like you have a prompt. Like my prompt for every morning is I want you to review my health metrics, look at my HRV data and tell me like how unhealthy I am and it will come back or like tell me my energy levels and that's a prompt that I could use on a loop if I want. And I've got other systems that I use for that, but you've opened up the new rabbit hole now, which is more of the personal side of things, which I think is really interesting and something that I'm enjoying in this in this conversation is that it's you know, we're we're we're broadening outside of maybe what I would typically give an answer of in terms of what it looks like to be an you know, like an AI native designer, you know, it's like, well, no, there's so many different opportunities that you can leverage this stuff if you just know what tools are available and and and know how to ask the right questions.

你提到了循环。我的生活中没有任何循环。我真的不知道这能解锁什么。所以再多讲讲，你是如何识别在个人生活中使用 AI 的机会的？

And so you've mentioned loops. I have no loops in my life. I have really no idea what this could even unlock. So keep going a little bit more into like how are you identifying opportunities to use AI in your personal life?

我就是出去散步，思考我的生活，想着我是一个父亲，有两个孩子。我想成为一个好丈夫，一个好父亲。我想保持最好的身材。我想把脑子里的一切都创造出来。所有这些纷乱的思绪，而 AI 真正帮助解锁的是，我不再需要一遍又一遍地对我妻子重复同样的话，比如‘天哪，我真希望能做个 iOS 应用’。我可以去散步，把脑子里的东西倒出来，然后问 Claude Code：‘好了，这是我过去一小时里想的所有东西。’

I just go on these walks and I'm like thinking about my life and I'm thinking about I'm a dad. I've got two kids. Like I want to be a good husband. I want to be a good father. Like I want to be in the best possible shape I can be. I want to like create everything in my head. All these like swirling pieces and the thing that AI has really helped unlock is that instead of telling my wife the same repeatable things over and over and over, like, 'Oh my gosh, I just wish I could like make an iOS app.' I can walk through, go on a walk, brain dump, and then ask Claude Code like, 'Okay, here's everything I just was I've been thinking of in this last hour.'

我们从哪里开始？关于你问的“AI 原生意味着什么”，我觉得大部分时候我们可以走老套路，说这就是现在的 AI 原生。比如，我真的想要一个很漂亮的模糊效果，一个漂亮的卡片展开来截断，我花了大量时间琢磨怎么让它看起来更精致，甚至包括微动画也要感觉非常 polished。所以，我花了很多时间用 Claude Code 打磨这个。这绝对是 AI 原生设计师的一个方面。不是每个人都会和这个交互，但聊天界面只会越来越突出。不管别人怎么说，我坚信我们正在对话，而这是人类除了文字之外最主要的交流方式。有些非语言交流、文字，但大多数时候，你和朋友家人聚在一起吃饭，大家都在说话。所以我从根本上认为，这是人们与 LLM 交互最容易的入口。所以这不会消失。AI 原生必须超越“我做了个很酷的着色器”或“我做了个别人都在做的抖动效果”。我今天也在做那个，然后我想，“天哪，我为什么要做这个？已经有上百个比我做的更好的了。”所以我会让别人去做，我借用并给他们 credit。回答你的问题，我主要是从个人痛点出发。我并不独特，我遇到的每个问题历史上都有成千上万、数百万、数十亿人经历过。它们并不特别。所以如果我有这个问题，我得解决它，很可能其他人也有。或者可能不是一个大市场，比如想要阅读应用的人，但可能也有其他人想要类似的东西，我可以分享，把它发布出去看看会发生什么，这感觉很酷。归根结底是解决问题，这也是为什么我非常感激能成为当今时代的设计师，因为突然感觉有全新的问题类别值得解决。技术上我们一直都能解决任何问题，但只是不值得，投资太大或不合理。甚至包括你提到的“无许可构建”，我理解它的一种方式是，也许我是那种感觉更有权限去构建的人，但另一方面，我总是担心浪费时间或不做高 ROI 的事情。毕竟有人雇佣我。而现在，突然之间，我可以在 3 小时内走得很远，看看这件事是否值得做。这甚至不是一个问题，而是一种思维模式的转变。

Where can we get started on? And to your point or your question around like, "What does it mean to be AI native in this?" I feel like it's mostly just we could take the cliche route of like this is what AI native is right now. Like I really wanted a really nice blur and I wanted a nice card open for truncation and I spent ridiculous amount of time figuring out how I wanted this to look even like the more less animation to feel ultra polished. So I mean I spent a lot of time with Claude Code polishing this. That is definitely one aspect of being like an AI native designer. Not everyone in their life is going to be interacting with this. Like chat surfaces are only going to be more and more prominent. Despite I think what others might say in the space, I really believe we are talking right now and that is the only way that humans outside of like the written word mostly communicate. Some nonverbal communication, written word, but for the most part like you get together with friends and family, you have a dinner, everyone's talking. And so I just fundamentally think like that is the easiest ingress for people to work with LLMs. So that's not going anywhere. So AI native has to be beyond just like I made a really cool shader or I made this like dither effect that every other person is making. I was doing that today and then I was like oh my gosh this is like why am I doing this? Like there's a hundred of these that are way better than what I'm making right now. So I'll let someone else do that and just borrow and like give them credit. To answer your question like I mostly think through the friction points personally. I am not unique as a person. Every problem that I've had has been experienced by thousands and thousands and millions and billions of people in history. Like they're just not special. So if I'm having this problem, I've got to figure it out and then probably other people are as well. Or maybe it's not like a huge TAM of like people that want a reading app. There's probably other people who would want something like this and so I could share that and like that would feel cool to put this out in the world and see what happens. It comes down to problem solving at the end of the day, you know, which is why I guess again I'm like so grateful to be a designer in today's era because it's just all of a sudden it feels like there's whole new categories of problems that are worth solving. Where it's like, yeah, technically we could have solved anything at some point, you know, but it just wasn't worth it, you know, like it was just too much of an investment or like it didn't make sense. Even down to like your permissionless building thing. I think one way that my brain takes that is like maybe I am somebody that felt more permission to build, but I also the flip side of that is like I was always worried about like wasting time or not you not doing high ROI activities. At the end of the day somebody's employing me. Where now all of a sudden it's like, actually I can go really, really far in 3 hours to figure out if this is even worth doing, you know? And I don't know that I there's not even a question in there yet, but it's like that's that mental model shift.

我喜欢这个。我深有同感。我可以实验和探索。所以我有了 Open Claw 和我的 Mac mini，我做到了，虽然说出来有点傻，但我真的很喜欢。它正是我一直希望 Siri 成为的样子，甚至更多。它存在于我的信息里，虽然 Matt Smith 今天发短信让我用上了 Telegram，所以我现在用 Telegram 和 iMessage，但这不是重点。重点是我想要达到精英身材。过去一年我有一些健康问题，现在在好转，我用 Apple Watch 记录所有指标，我想看到这些数据如何变化、各部分如何运作。但最大的问题是，我不能直接把 Apple Health 数据连接到 Mac mini，因为我的设备用了不同的 iCloud 账户——我不是疯子，我区分了用途。所以我的 Open Claw 设置有自己的 Apple ID，没有很好的方式共享健康数据。不过你可以建立本地网络并在设备间共享，但需要一个 iOS 应用在后台运行发送同步。所以昨天我用 Claude Code 在 2 小时内构建了整个应用来实现这个功能。我会给你看模拟器，希望它能构建成功。总之，你会看到一个很粗糙的应用，它没什么值得骄傲的，但它能用，而且专门为我工作。这才是最重要的。所以无许可构建，我 literally 可以说“开始共享”，它就会在我的手机上运行——这里不会演示，但它会询问我。所以 Apple Health 中所有被追踪的数据现在每天三次通过局域网发送到我的 Mac mini，它工作了。这太棒了。所以关于无许可构建，我能在 2 小时内构建这个：我有一个想法，告诉 Claude Code 我需要实现什么，它为我创建一些 Linear 工单，然后我就可以处理积压任务，分类、标记完成，然后继续。这有点像你的 Open Claw 界面？你能讲讲设置吗？我想分享一下我的情况：我有 50 个不同的 Open Claw 书签，但一步都没迈出去。现在我不知道这是炒作还是我应该投入的东西？你怎么看？也许你可以对大多数听众说，我猜他们也在观望。

I love that. I deeply connect with that. The fact that I can experiment and explore. So I've got Open Claw and I have like my Mac mini and I did that and I feel like a little bit of a tool saying that, but I think I love it. Like it's exactly like what I've always wanted Siri to be and more. It lives in my message, although Matt Smith and I were texting today and he got me all hooked up on Telegram, so now I'm doing that and iMessage, but it's not the point. The point is one of the challenges though is that like I want to get into like elite shape. I've had some health problems over the last year and like things are getting better, but I'm using my Apple Watch for all the metrics and I want to see how these things are resolving and how the pieces are in play and all this stuff. Biggest problem though is that you can't directly connect like my Apple Health data to the Mac mini because I have different iCloud accounts on them cuz I'm not a lunatic and I have like separation of faces. So my Open Claw setup has its own Apple ID and there's not really a great way to share health care data. You can however make a local network and share them between them, but you have to have an iOS app that's running that sends the sync in the background. So, in 2 hours yesterday, with Claude Code, I built this entire app that does it. Uh and I'll show you the simulator for this. I hope this builds. Anyway, the idea is like you'll see a very janky app on the screen. And when you see it, like it's not anything to to be proud of, but it works and it works for me specifically. And that is like all that matters. And so the permissionless building, like I literally can just say start sharing and it will on my phone, it won't do this here. Well, it'll ask me, yeah. So, everything that's in Apple Health that is tracked now gets sent three times a day to my Mac mini uh on a local area network and it works. And that's great. So, in terms of like the permissionless building, like I'm able to in 2 hours build this Okay, I have this idea. I tell Claude Code about it. Like, this is what I need to have happen. Make some linear tickets. It makes linear tickets for me on this piece and then I can go through, work through the backlog. I'll triage everything, mark everything complete, and I'm on my way. This is kind of like your interface for Open Claw? Is this like Talk me through the setup cuz and I guess I'll share like where I'm coming from is I'm the person who has like 50 different Open Claw bookmarks and I haven't taken a single step. And at this point I don't even know like is it hype? Is this something that I should be leaning into? How do you think about it and maybe you could talk to the majority of us listening who I'm assuming are also on the fence.

嗯，我不认为这是炒作。我肯定不会夸大其词，但也不是极度怀疑。我尝试它是因为它确实解决了我一个大问题。我之前已经在本地使用 Claude Code，每天打开一个终端窗口，截取我的每日健康简报——健康指标，发送给它，它会分析，然后查看我的日历和待办事项，告诉我今天的状况。但这一切都是本地的。如果我想在路上问它问题或让它帮我跟踪某事，我做不到。我可以用 Claude 移动应用远程启动一个新的 Claude Code 会话，但它们并没有真正融合。直到大约两周前，才有了一个很好的方式让我们在移动和本地之间搭建桥梁。

Yeah, uh I don't think it's hype. So, I definitely don't go towards hyperbole. But I'm not like ultra skeptical. I tried it because it actually solved a huge problem I was having. So, I would was already using Claude Code locally. Every day I'd have a terminal window open. I would screenshot my daily health briefings, all right, like health metrics, I would send it to it, and then it would analyze it, and then it would look at my calendar, and it would look at my to-do list, and it would say like, here's where we're at today. But, it's all local. And if I wanted to ask it a question about something, if I wanted it to help keep track of something on the go, I couldn't do it. And like, I could use the Claude mobile app to start a new Claude code session remotely and try to do that, but they didn't really mix. Not until like 2 weeks ago where there was really like a great way for us to like bridge between being on the go and there.

所以，当 Open Claude 出来的时候，那真是一个不可思议的时刻——这就是我一直想要的 ambient 助手。我真的很想要那种能力：可以随时抛出研究问题，可以把脑子里的东西卸下来。我到处是笔和纸，到处都是想法。也许你也是这样。我觉得我们身上那种创造性的部分就是有太多想做的事。我可以直接告诉我的 Shadowfax（我给它起的名字，就是那个实例）去搭建我想要的东西，告诉它我想做什么、想解决什么问题。

So, when Open Claude came out, it was like this incredible moment of like, this is everything I've ever wanted in an ambient assistant. And like, I really want that. I want to be able to fire off questions about research. I want to be able to just take things off of my brain. I just have pen and paper, stuff everywhere, ideas. Maybe you're like that, too. I think the creative piece of us is like, so many things I want to do. I can just tell my Shadowfax, that's what I call it, the instance, to go build my like, here's what I want to do, here's what I want to tackle.

所以，这是炒作吗？不，只是你需要对它能做什么、能为你做什么有一个更理性的认识。但现在，我无时无刻不在跟它说话，以至于我老婆会开玩笑说：“嘿，你应该告诉 Shadowfax 把这个加到你的待办清单里。”然后我就说：“好主意，我这就做。”但关键是，它在帮我做我想做的事，过我想过的生活。

So, is it hype? No, it's just you have to have maybe more measured understanding like what it will do, and like what it can do for you. But, right now, like I talk to it all the time to the point where my wife will jokingly like, hey, you should tell Shadowfax to add this to your to-do list, and it's like, oh, that's a good call, I'll do that. But, the point is like, it's helping me do the things I want to do, like live the life.

Linear 对我来说主要是一个界面，用来告诉 Shadowfax 我想搭建各种项目、我正在做的事情。呃，我对 Epilogue 里所有困扰我的东西做了一次深度审计。我会用语音转录给 Shadowfax，然后它就会为我创建所有的工单。接着我让本地的 Claude 实例遍历这些工单，开始程序化地一个一个解决。所以，一周前我们就能把所有那些长期存在的问题都处理掉了。

Linear is an interface mostly for me to tell Shadowfax like, I want to build my various like projects, things I'm working on. Uh, I did like a deep audit of all the things that were bothering me with Epilogue. So, I would do a voice transcription to Shadowfax, and then it would create all the tickets for me. And then I could have my local Claude instance go through these tickets and start programmatically like knocking them off one by one. And so, we were able to, you know, a week ago like power out all of these long-standing issues that I've had.

嗯，但这就是我跟踪所有进行中工作、了解进展的方式，同时也能获得我需要的帮助。Shadowfax 通过 iMessage 来做这些事，帮我处理所有这些部分。就像你之前说的，感觉我们在这类事情上才刚刚开始。但就是那么一小点——你可以在散步时产生想法和执行计划，也许你正在和 Claude 用语音模式来回交流，或者只是倾倒想法，然后它真的就把这些变成任务，并在另一台设备上替你工作。我不知道，虽然有很多关于 10 倍工程师之类的陈词滥调，但我懂那种感觉，你知道吗？

Um, but this is the way that I can keep track of all the work in flight, where it's at, and then also have just like the help I need, and Shadowfax is the one that does that like through iMessage and so it's helping me kind of tackle all those pieces. Like you said earlier, it does feel like we're just in the first inning with all this kind of stuff. And yet that little piece there where you could literally just be on a walk and you have ideas and execution plans. Maybe you were having a back and forth with Claude in you know, voice mode or whatever or just dumping and then it literally is turning it into tasks and then working on your behalf like on a separate device. It's like I don't know there's all the clichés about the the 10x engineer and everything, but like I get it, you know?

就像你们所有设计师之间的差距——因为听众里大部分可能是设计师——一个在组织里没有 Scaling 自己的设计师，与另一个能够真正做、探索并完成 10 倍、100 倍更多事情的人之间的差距。我觉得这个差距正在扩大。尤其是在招聘经理那边，和早期阶段的人或者正在找新工作但没怎么接触这个领域的人聊天时，这确实是一个警示信号。

Like the the delta between all of you designers because most people listening probably are designers, but the delta between a designer in an org who is kind of just not scaling yourself. I'm not even sure what the how to even refer to it versus the person that, you know, literally could be doing and exploring and accomplishing 10, 100 times more. It feels like that gap is widening. I think especially on like the hiring manager side, too, like talking with either like early stage folks or people that are like looking for new roles and they're not engaged in the space. Like it is a little bit of a flag.

柏拉图有句名言：当心忙碌生活的贫瘠。所以我有点担心，那种“做更多、更多、更多”的想法——是的，它确实能实现。但我不喜欢让智能体完全自主运行。我不喜欢在 Claude Code 里那种危险的“跳过权限”模式，你放一个小提示，Claude 就能去做一堆疯狂的事，你跳过了权限检查，它可能会删除你所有文件。它确实可能。它可能会失控。但人们这样做是为了让 Claude Code 能连续运行好几个小时。这确实是一种与 LLM 协作的心智模型：我告诉你去做所有这些事，你去做，然后我接受你的任何输出，满意了，就继续前进。

Plato has a great quote where he says that beware the barrenness of a busy life. And so like I worry a little bit that the idea of just doing more and more and more and more and more like yes, it enables that. I do not like having agents run autonomously. I do not like having if you're in Claude code, the dangerously skip permissions world where you can like there's a little prompt that you put in that Claude can just go and do a bunch of crazy stuff and you're skipping permissions as in like it might delete all your files. It might. It's it can go off the rails. But the idea behind that is people do that so that Claude code can just like run for hours and hours and hours and hours. That is for sure one mental model about working with LLMs is that I can just tell you to go do all these things and you do it and then I just take whatever output you have and I'm like happy with it and I move on.

但我觉得我们在 Twitter 文章里也看到了这一点，比如每篇文章听起来都差不多，因为它们都用 AI 来写，即使想法可能是他们自己的。但我觉得有一种压力，就是单纯地做更多，每个人对此都有些焦虑。我想跳出这个循环。我不想和别人一起骑那场自行车赛，因为这不是我看待这些工具的方式。它们是力量倍增器，但我希望它们有焦点。我希望它们仍然能让我感觉是真实的我。

But I think we're seeing this even in like Twitter articles, for example, where every article starts to sound the same because they're all using AI to write the articles, even if it may be like generically their idea, but there's this like push, I think, just to do more and everyone's kind of like there's like anxiety around that. And I want off that cycle. I do not want to be riding that bike race with anyone else because this is not how like I view these things. Like they are a force multiplier, but I want them to be focused. I want it to be something that I feel is still like authentically me.

所以，在 Epilogue 上我真正自豪的一点是，即使 Claude Code 设计了所有东西——我没有用画布工具来设计布局或像素——我只是告诉它层级结构、布局，然后给出反馈。我会给出大量反馈，直到它感觉：“是的，这确实像是我自己设计的。我喜欢这样。我可以放心地把我的名字放在上面。”我必须深度参与这个循环。所以，我希望设计师们能受到激励：这些工具可以加速他们正在做的事情，帮助加速探索，把他们不擅长的任务从盘子里拿走。

So, that's the one thing I was really proud of with with Epilog is that even though Claude Code designed all like I didn't use a canvas tool to design the layouts or pixels, I just told it here's the hierarchy, here's the layout, let me give feedback. I would give so much feedback until it felt like, "Yes, this is authentically something like I designed. Like I like this. I feel comfortable putting my name on it." I had to be very involved in the loop. And so, I want designers to feel like to motivated that these tools can accelerate the things they're doing to help accelerate explorations, to take tasks off their plate that they're not good at doing.

比如，你可以直接把 Claude Code 连接到 Google Docs，让它为你创建文档结构，然后你自己填充内容。我在这方面就很差劲。我对电子表格过敏。Claude Code 现在可以为我编写和创建电子表格，然后我去填写。比如，我给它一个关于电子表格要实现什么功能的大致想法。我自己坐下来花一个小时试图搞明白怎么做，效率很低。Claude Code 来做就很高效，只要我的内容没问题。所以，在我开始长篇大论之前，我想说的是：我们现在很容易觉得“天哪，我们可以做所有事情了”。但问题是：“这值得做吗？你对输出本身感到自豪吗？”

Like, you could literally connect Claude Code to Google Docs to like create structures of a document for you that you fill in yourself. Like, I'm terrible at that. I'm allergic to spreadsheets. Claude Code can write and create spreadsheets for me now that I can then and go fill out. Like, here's the generic idea of what I want the spreadsheet to accomplish. Me sitting down for an hour trying to like figure out how to do it is not effective. Claude Code doing it is effective as long as my content. So, you know, the point I'm trying to make before I like go on this crazy diatribe is that it's easy for us right now, I think, to have like, "Oh my gosh, we can just do all the things." And it's like, "Yeah, but is it worth doing? And do you feel proud of the output itself?"

这就是为什么像 Linear 这样的工具能帮我保持克制：我可以回顾、审查每一个正在构建的工单，确认“好的，这就是我想要完成的，所以我们做到了。”然后它被标记为完成，我可以进去审查具体的工单，看看它在应用里意味着什么。它是否实现了我的目标？它比其他人慢吗？是的，但我对输出更自豪。

Which is why like having linear or these tools kind of keep it in check where I can review, I can go back and I can review each of the tickets uh were being built of like, "Okay, this is what I wanted to accomplish, so we did." And then it's marked as done and I can go in and I can review the specific ticket. And show me what this means in the app. Did it accomplish what I'm wanting? Is it slower than other people? Yeah, but I feel more proud of the the output.

你在 Stripe 的面试过程中，是否刻意突出了这段经历或这套技能？你有没有有意把自己定位成一个对 AI 能做什么、能带来什么价值感到自如并积极推动的人？

Were you leaning into this experience or set of skills while in the interview process at Stripe? Did you intentionally position yourself as somebody who is comfortable and like really pushing what AI can do and and bring to the table in your role?

我们确实提到了，肯定有过相关对话。我没有回避一个事实：这个领域有很多人比我更擅长、更聪明。我很清楚，我觉得自己领先，但我可能无法很好地向你阐述我自己的技能组合。

I mean, we brought it up, like had conversations for sure. I don't think I shied away from the fact that there are people in this space who are so much better and smarter at this stuff than I am. And like I'm well aware of like, I feel like I'm ahead, but I probably couldn't articulate to you my own skill set very well.

所以我尽量谨慎地说，我热爱这些东西。过去 10 个月使用这些工具，尤其是 Claude Code，为我解锁了一种全新的创意工作方式，它比画布更适合我。所以我肯定会强调这一点。我会突出我自己的想法、方法论和流程。我对 AI 原生体验应该如何构建，以及我们应该朝哪个方向前进有很多看法。目前，MCP 不会消失。任何说 MCP 已死的人都是在炒作，这很疯狂。但它们会改变很多，我们会看到更多公司倾向于更好的语义 API 和 CLI，而不是 MCP——如果你不了解这个领域，这里有很多缩写。本质上，MCP 是一种在不同层之间通信的方式，但它需要大量信息来回传递才能完成一个目标。所以上下文窗口，即 AI 可以持有和思考的窗口，会很快被 MCP 填满。而且有时有更高效的通信方式。所以我认为模型会开始更高效地选择使用哪个工具以及何时使用。这会影响公司如何设计产品。你的产品是否具有智能体式焦点？或者所有内容是否都能被智能体读取，因为会有更多智能体为人们工作。所以作为设计师，你必须考虑这一点。

So, I tried to be very cautious about like, I love this stuff. The last 10 months using these tools, especially Claude Code, have unlocked a whole new creative way of working for me that actually works better for me than a canvas does. And so, I definitely would articulate that. I would highlight my own thoughts and methodology and process. And I have a ton of opinions on how AI-native experiences should be built and where we should be indexing and moving towards. And currently right now, MCPs are not going anywhere. Anyone who says an MCP is dead is like, that's hype. That's crazy. But they're going to change quite a bit and we're going to see a lot more companies leaning more towards better semantic APIs and CLIs than they will an MCP, which is a lot of acronyms thrown out if you don't know what the space is. Essentially, MCP is a way to talk among different layers, but it takes a lot of information back and forth to accomplish a goal. So the context window, the window that the AI can hold and think, gets filled up very quickly with an MCP. And there are more efficient ways of communicating at times. And so, I think the models are going to start to get more efficient at which tool they use and when to do things. So that affects how companies are going to design their products. Do your products have an agentic focus to them? Or is everything readable by an agent because there's going to be a lot more agents doing the work for people. So you have to think through that as a designer.

所以我肯定会把自己定位在那些对话中，无论是与 Stripe 还是其他公司，围绕我的技能水平或我认为行业的发展方向。我是否对错并不重要，因为另一方面，招聘经理们不断说：‘好奇心是最重要的。’ 过去 6 个月我听到至少 20 个人这么说。我现在有时甚至不问这个问题，因为我知道答案就是‘好奇心’。但问题是：‘这实际上是什么样的？你如何展示好奇心？’ 对我来说，你展示的很多东西，就是这种表现，对吧？如果我招聘某人，你带我了解一些事情，这些事在 6 个月前我甚至不会认为属于设计的范畴，甚至与设计毫不相关。但现在突然间，每个人都变成了通才，每个人都身兼多职，公司想要那些能进来就产生影响、发现问题并解决、使用任何工具组合来解决问题的员工。我们不再只是在画布上解决问题，甚至可能不再用代码解决问题。我们在系统层面和市场竞争定位上解决问题。这就是好奇心，这也是我想问的原因，因为我认为越来越难把它放进一个框里，比如‘如何准备你的作品集’。但如果你作为一个设计师能展示这样的东西，在当今市场上意义重大。有一个时间窗口，通过这样思考和行动，你可以获得巨大的优势。

So I definitely would position myself in the conversations, whether it was Stripe or a few others I was talking with, around where I was at with my skill set or where I thought the industry was going. Whether or not I'm right, I don't think it really matters whether you're right honestly, because the flip side is hiring managers keep coming on and they say, 'Well, curiosity is the most important thing.' I mean, I've heard like 20 people say that over the last 6 months. I don't even ask the question sometimes now because it's like, 'Well, I already know the answer is curiosity.' And it's just like, 'What does that actually look like? How do you demonstrate curiosity?' And to me, a lot of what you're showing, I mean this is kind of that, right? Like if I'm hiring somebody and you're walking me through things that maybe 6 months ago I wouldn't even have considered in the box of design, you know? But hardly even adjacent to design. But now all of a sudden it's like everybody's becoming a generalist, everybody's wearing all the hats, and people want employees who can come in and just make an impact and see problems and solve them and reach for any combination of tools to solve problems. Like we're not just solving problems on a canvas anymore. We're maybe not even solving problems in code anymore. We're solving problems at systems levels and positioning in the marketplace. You know, it's like this is curiosity and that's kind of why I was wanting to ask because I think it's getting a little bit more difficult to put this into a box of like, 'Here's how to prepare for your portfolio,' you know? But if you as a designer can show something like this, it's a pretty big deal in today's market, you know? Like there's a window in time to get a serious leg up by being able to think and act like this.

是的，100%。我认为特别是如果你现在正在找设计岗位，这跟你刚才提到的面试有点相关，我觉得有些人的困惑是：‘我还没有用 AI 构建过应用或任何东西。’ 我说：‘你不一定非要这样做。如果你想做也可以，但前提是你对成果感到自豪，但你必须展示你理解这些系统，并且你实际上在用它们构建东西。无论是另一个着色器插件系统，只要用它做点什么，去观察、探索，并且能够谈论它，讲述整个过程。’ 我想说，现在招聘经理中有一个微妙之处，你看过那么多作品集。你真正看一个作品集可能只有 6 到 10 秒。我知道听到这个的人不喜欢，因为你花了那么多时间，但如果你的网站是用 Squarespace 的基本模板，即使内容很好，你也可能会被跳过。因为在当今世界，你完全可以用 Claude Code、ChatGPT、Gemini 快速合作，制作一个更定制化的模板。有人只用基本的 Webflow 模板有很多原因，这绝不是对这些公司的贬低。这更多是说明我们发展得太快了。我讨厌这么说，但门槛太高了。每个人都想要同一类型的设计师，如果你不使用这些工具来让自己有点差异化，你就会被抛在后面，这很糟糕。看到这种情况很奇怪，但如果你抗拒改变，那就非常困难。所以好奇心当然可能有点陈词滥调，他们可能想说的是好话，但如果你没有表现出你投入于这个领域的发展方向，那很可能就是不行。可能会有一些公司适合那样，但现在所有大型科技公司都希望看到参与这个领域的人。再回到无许可构建，你不需要许可就能做。‘我不知道’本身就是一种超能力。比如你不知道如何创建自己的网站。你可以直接问：‘我甚至不了解托管。我应该用哪个托管服务？’ 它会给出很多不同的优秀选项。‘哦，我现在可以帮你在 Vercel 上设置，你可以有一个免费网站，大概花一个小时，你就能有一个看起来很好的静态登陆页。’ 然后你说：‘好，我们试试看。看看会发生什么。’ 这比使用现成模板要好得多。你今天展示的一切都是你个人的表达。你构建的所有东西都是为了你自己。这些系统都围绕着你提升和成长的意愿和兴奋感，作为一个人和职业生涯的发展。所以我不知道。你的网站、你的副项目，就是你个人的广告牌。它不再关乎案例研究或你的工作。

Yeah, 100%. I think especially if you're a designer looking for a role right now, which is not what you brought up tangentially a little bit like if you're interviewing stuff, I think the confusion that some people have, they're like, 'Well, I haven't built an app or anything with AI yet.' I'm like, 'You don't have to necessarily. You can if you want to, but only if you're proud of what you make in there, but you have to demonstrate that you understand these systems and you're actually building something with them. Whether it is another shader plugin system, just do something with it to see and explore and be able to talk to it and talk through the process.' I would say the thing that is subtle among hiring managers right now, you look at so many portfolios. You maybe have 6 to 10 seconds to look at a portfolio truly. I know people who are hearing that don't love that because you spend all this time into it, but if your site is on Squarespace and it's a basic Squarespace template, you're probably going to get passed on even if the content's really good. Because there is just not a reality in this world right now where you could not have Claude Code, ChatGPT, Gemini work with you quickly to have a more custom version of your own template. There's a lot of reasons why someone would just do a basic Webflow template or whatever it is and that's not a knock on these companies at all. It's more of a statement that we're moving so fast. I hate saying this but the bar is so high. Everyone wants the same type of designer, the same shape of designer that if you're not using these tools to differentiate yourself a little bit, you are going to get left behind and that sucks. It's a weird thing to view and see, but if you're resistant to change, it is really tough. So curiosity for sure I could see it being a little bit of a cliche and they're probably trying to say the nice thing, which is if you're not showing that you're invested in where the space is going, it's probably a no. There's probably going to be companies that would be a good fit for that, but all of the larger tech companies right now, they really want to see people that are engaged in the space. And there's just again going back to the permissionless building, you don't need permission to do it. You literally 'I don't know' is a superpower. Like you don't know how to create your own website. You can literally ask it like 'I don't even know about hosting. Who should I use for hosting?' And it will spit out a bunch of different options that are great. 'Oh, I can get you set up on Vercel right now and you can have a free website and it would take us maybe an hour and you have a static landing page that looks really good.' And you're like, 'Okay, let's do that. Let's try that. And just see what happens.' That would be far better than just using the stock templates. Everything that you're showing today is such an expression of who you are. All the things that you're building, it's for you. It's systems that are deeply revolved around your willingness and excitedness to improve and grow as a person in your career, all that kind of stuff. So I don't know. Your website, the side projects, it's the billboard of who you are. It's not about a case study or your work anymore.

这就像是在问“你是谁？你想呈现什么？”对吧？我们确实会以貌取人，这是事实。既然你提到了网站，我分享一下——虽然这听起来有点夸张，但我真的找到了一个我很喜欢的 WebGL 着色器，一个海洋着色器，你可以四处移动。

It's like who are you, you know? What do you want to put forward? And you can get a pretty yeah, we do judge books by their cover, you know? That's the facts. I'll share this because you mentioned the website and this is like not even true, but I literally went through this and I found this WebGL shader that I really like, this ocean shader. So you can like move around.

哦，酷。

Oh, cool.

它还会根据一天中的时间变化——太阳会根据我的 IP 地址，在丹佛的正确位置出现。然后我放了一些简短的自我介绍。我其实并不在找工作，因为我对现在的公司感激不尽。所以这个网站只是如果你真的想了解我，可以问它问题。我还做了一些很傻的事，比如我想“加下雨效果会怎样？如果是夜晚呢？”现在我还想让窗户对雨有反应，但我不知道怎么实现——Claude 会做。

And it responds to the time of day. So the sun is in the correct position for Denver based on the time of day based on my IP address. And then I just have like, you know, little writing, something about me. I'm also not like looking for a role because I really am like stupidly grateful for the company I work for. So this is just you can ask it questions about me if you really wanted. And then I did like really silly things like I had a chance to like, oh, what would it look like to add rain to it? And like what if it was at night? And so but I now want the windows to like respond to the rain, you know? So I don't know how to do that. Claude does it.

我现在做的东西，分享出来不是为了自夸，而是说在之前的平台上，我根本做不到这个效果，除非用大量我不想自己设计的自定义组件。而且我觉得输出质量也不会这么好，但我有这个愿景。另外，我每年还能在网站上省下大约 500 美元，这对我来说很棒。但最重要的是，我拥有了一些更个人化的东西。

What I have now, purpose of sharing this is not like to toot my own horn, it's like I could not have done this on the previous platform without a ton a ton a ton of custom components that I did not want to design myself. And I don't think I would have had the output as good, but I had the vision for it. Also, I'm saving like $500 a year on the site and the thing, which is great for me. I'm pretty stoked about that piece. But most importantly, I have something that feels more personal to me.

因为我一直觉得，“灵魂”会成为今年和明年的年度关键词。我们有过“品质”，有过“工艺”，但我认为人们真正渴望的是更人性化的东西。他们会怀念一个他们可能从未经历过的时代——比如 MySpace 时代，你的 MySpace 页面对你来说是非常个人化的。在 2004、2005 年，我的 MySpace 页面完全是定制化的 Chris Pocket 完美风格。我认为这种趋势会回归，人们希望你的作品集看起来和感觉上都像你自己。这就是我设计时希望达到的效果。所以，没有理由不这么做。

Because I do think like I've been saying this for a little while, like soul is going to be one of those things that is like it'll be the word of the year like this year and next year is like soul. We've had quality, we've had craft, but I think what people are going to be desperate for is more of that human side of things, right? They're going to be longing for this either they an era they've never experienced because they're younger of like that MySpace, you know, our MySpace generation where you're like your MySpace page was deeply personal to you. Like in 2004 and three, five, like my MySpace page was complete custom Chris Pocket perfection at that time. And I think that we're going to want to see that come back. And I think people are going to want more of those like your portfolio looks and feels like you. And that would be the hope, right? With what I'm trying to design. So, there's not a reason not to have that.

所以，差异化就在于：你的作品是否融入了你自己，还是看起来和其他人一样？你是不是因为大家都在用 Instrument Serif 字体，所以你也用？我们每个人都有很多个性，我很希望看到这些在设计中被体现出来。我好像说得太多了，就此打住。

So, differentiator is also like are you in it or does it look like everything else? Are you using Instrument Serif in all your stuff because everyone else is doing that? Like there's so much personality that each of us has that I would love to see that reflected in their design. I'm on like such a stupid soap box. I'll get off.

不，这太完美了。老实说，我想不出更好的结尾了。这非常鼓舞人心，你就是一个完美的例子，展示了如何拥抱所有新的可能性，挖掘你内心深处的创造力，然后突然释放出来，向四面八方迸发。看着真美。Chris，我非常感谢你为我个人解开了很多谜团。我对你谈到的很多事情都很感兴趣。我知道还有很多东西可以学，我可以在技能和个人系统上更进一步。有人能这样分解讲解，对我帮助很大。所以，感谢你今天抽出时间。

No, it's perfect. I honestly like I can't think of a better place to end it, honestly. Like it's super inspirational and you're like the perfect example of what it looks like to just run at all of these new possibilities and tap into like the creativity that's obviously been inside of you and all of a sudden she's been unleashed and you're just firing in every direction. It's beautiful to watch and Chris, I really really appreciate you even just demystifying a lot of things for me personally. I'm interested in a lot of the things that you're talking about. I know there's stuff there. I know there's more I can grow. I know I can go further into skills and all these different personal systems. And I just having somebody kind of break it down has really been super helpful for me even. So, appreciate you taking the time today.

Red，非常感谢。能上这个播客是我的荣幸。这是我人生中第二次上播客。所以，非常感激能和你在这里交流。谢谢你的时间。

And Red, thanks so much. It's been an honor to be on the podcast. This is like my second podcast of all time. So, beyond grateful to finally be here with you. Appreciate the time.

在你走之前，我想花一分钟介绍一下我最喜欢的产品，因为经常有人问我用什么工具。Framer 用来建网站，Gen way 用来做研究，Granola 用来在评审时记笔记，Jitter 用来给设计做动画，Lovable 用来用代码实现创意，Mobbin 用来找设计灵感，Paper 用来像创意人一样设计，Raycast 是我每一步的快捷方式。我精心挑选了这些公司，这样我才能全职做这些节目。目前，支持这个节目最好的方式就是去看看它们。你可以在 dive.club/partners 找到完整列表。

Before I let you go, I want to take just 1 minute to run you through my favorite products because I'm constantly asked what's in my stack. Framer is how I build websites. Gen way is how I do research. Granola is how I take notes during crit. Jitter is how I animate my designs. Lovable is how I build my ideas in code. Mobbin is how I find design inspiration. Paper is how I design like a creative. And Raycast is my shortcut every step of the way. Now, I've hand-selected these companies so that I can do these episodes full-time. So, by far, the number one way to support the show is to check them out. You can find the full list at dive.club/partners.