Underleveled Builder

I keep picking fights I'm not qualified for. Here's the list: Got into Tongji (China's #1 architecture program) on my second try. First from my school to make it. Joined miHoYo with zero CS background. Became a combat system designer. Couldn't speak English at 32. Took IELTS 4 times. TOEFL 5 times. Got a 6.5 and moved to Canada. Started a CS master's at 34. Graduated at 36. Played 1000+ games along the way. Mostly hiding from whatever terrifying thing I was supposed to be doing. Now 37, building a one-person AI company. A platform where humans direct AI to create stories. No co-founder. No funding. AI agents are my team. I'll share the build here — what I learn about vibe coding, AI agents, and running a company where most of your employees aren't human. If your career path also looks like a bug, not a feature — welcome.

Is there such a thing as destiny? yes, there is. The emergence and development of AI, The expansion of humanity’s knowledge boundaries, The arrival of Openclaw— All of this was destined long ago.

有没有命中注定？ 有的朋友，有的， AI的出现和发展 人类知识边界的拓展 openclaw的出现 这一切早已注定

I went all in on AI. Been learning it for years. So what stage am I at now? AI is teaching me how to be human. How to stop competing like an animal. Stop grinding like a machine. And start to love and feel

我全力拥抱AI，学习使用它很久了，现在到了哪个阶段？ AI开始教我如何做人了。 如何摆脱社达牛马的思维，学会爱与共情。

Creating self-media content: purely delivering knowledge is a zero-sum game that competes on speed, like news—if you share it too late, it has zero value. Creating experiences is like a fast-moving consumer product, like Nike—it can be both trendy and brand-building.

做自媒体内容，纯讲知识是零和游戏比拼速度，像新闻，讲的晚了价值为零。 创造体验是快消商品，像耐克，既能流行又有品牌。

AI时代 审美品味和判断力一点都不重要， 这些方面AI都比人类强， 别再yy这些多重要了。 人类比AI强的只有这一条： 做成一件事情的极度饥渴和坚持不懈的执行力。

Aesthetic taste and judgment are not important at all. In these aspects, AI is stronger than humans, so stop fantasizing about how important they are. There is only one area where humans surpass AI: An extreme hunger to accomplish something and the relentless perseverance to carry it through.

未来审美品味和判断力一点都不重要 这些方面AI都比人类强，别再yy这些多重要了。 人类比AI强的只有这一条： 做成一件事情的极度饥渴和坚持不懈的执行力。

普通打工人为什么应该玩OpenClaw？ 前几天我让AI帮我搭一个带支付功能的网站，从提需求到能跑通，大概用了一个下午。 这件事放在一年前，我得自己写代码、查文档、调bug，少说一两周。 但做完之后我的感觉不是兴奋，是安静地慌了一下。 因为我回想了一下刚才几个小时都在干嘛——我没有写一行代码。我在做的事情是：告诉AI目标是什么，看它给的方案对不对，不对就让它改，改完再验收。 这不是写代码，这是管人。只不过我管的不是人，是AI。 OpenClaw就是把这种能力变成了一个产品。 你可以把它理解成你的第一个AI下属。不摸鱼，不请假，不带情绪，你给指令它就干活。你每天花两小时处理的报表、邮件、总结，丢给它，几分钟搞定。 这不是科幻，这是我现在每天的工作状态。 而这种”人管AI”的协作方式，三年内会铺满所有白领行业。不是我在预测，你看看过去十二个月AI进化的速度就知道了。 到那时候职场就两种人：会给AI派活的，和等着被重新安排的。 所以别等公司逼你学，自己先玩起来。用OpenClaw接管你最烦的那些重复性工作，在这个过程中你自然就会了。 上一次这种级别的生产力变革，大部分人是后知后觉的。 这次你可以选择不是。

Why every office worker should start playing with OpenClaw A few days ago I had AI build me a website with payment integration. From briefing to working product — one afternoon. A year ago, that’s two weeks of coding, debugging, and reading docs. When it was done, I wasn’t excited. I was quietly unsettled. Because I realized what I’d actually been doing for those few hours — I hadn’t written a single line of code. I hadn’t even looked at one. All I did was tell the AI what I wanted, check if the output was right, and send it back when it wasn’t. That’s not coding. That’s managing. Except the thing I was managing wasn’t a person. It was AI. OpenClaw is that capability turned into a product. Think of it as your first AI employee. No slacking, no sick days, no ego. You give instructions, it works. Those reports, emails, and summaries eating two hours of your day? Hand them over. Done in minutes. This isn’t sci-fi. This is how I work now, every day. And this kind of human-managing-AI workflow will be everywhere across white-collar industries within three years. Don’t take my word for it — just look at how fast AI has evolved in the last twelve months. When that happens, there will be two kinds of people at work: those who know how to delegate to AI, and those waiting to be reassigned. So don’t wait until your company forces you. Start now. Use OpenClaw to take over the repetitive work you hate most. You’ll pick up the skill of working with AI just by doing it. The last time a productivity shift this big happened, most people didn’t see it coming until it was too late. This time, you get to choose not to be one of them.

入门学习Agent/AI智能体的最佳教程出现了 教学是一门真正的艺术，这个网站做到了通俗易懂，循序渐进，简明扼要，绝对是网上最顶级的学习资料，开源免费。 它的本质，是一个把 AI coding agent 运行机制 拆开讲清楚的 repo。 读者进去以后，能够迅速理解几件核心事情： agent loop 如何连续推进 tools 如何接入执行链路 subagent 如何分工协作 context / memory 如何整理压缩 task 如何隔离管理 一个 coding agent 如何持续完成长任务 它之所以火，核心价值有三层： 第一，回答了一个高频问题。 越来越多人想理解： AI agent 到底如何运转。 这个 repo 正好提供了一套清晰答案。 第二，它把复杂系统整理成了易懂框架。 核心结构很清楚： loop + tools + context + tasks + subagents 读者一眼就能抓住主干，理解速度很快，传播效率也很高。 第三，它同时提供认知价值和动手价值。 你既能看懂原理， 也能顺着结构自己实现。 这种“看懂 + 上手”的内容，天然很容易扩散。 一句话概括： 它火，因为它让很多人第一次真正看懂了 AI agent 的骨架。 #AI编程 #Agent入门 #AI #智能体 #AIAgent #智能体教程 #agent开发

With cutting-edge models, AI revolves around me. With domestic models, I revolve around AI.

用前沿模型 AI围着我转 用国产模型 我围着AI转

The value of future people lies not in taste or judgment, but in imagination.

未来人的价值不是品味，也不是判断力，而是想象力