With AI This Powerful, What Are Humans Still Good For? — An Opening Word, and the Manifesto for a Free Course

Epigraph: This is not an article built to sell you anxiety. Quite the opposite — I want to take the anxiety out of your hands first, and only then talk about learning.

1. With AI This Powerful, What Are Humans Still Good For?

Let me start with a concept: RSI — recursive self-improvement, where AI iterates and improves on itself.

The idea is that AI is no longer just helping people do things; it has started training itself, improving itself. When that day truly arrives, even “working on algorithms” will be something AI does on its own. It may sound far off, but the trend is already sitting there in plain sight.

So here's my judgment: Ordinary people don't need to master much of the deep theory. Get good at the application layer, and you can already solve 90% of the problems in daily life. As for the algorithm layer, odds are AI will take care of that itself.

So what is our value, as humans?

My answer: living well.

Everything that follows — this course, these articles, every video — boils down to that one sentence: how to live well in the AI era.

2. What This Course Looks Like

“Living well” is not a platitude. It takes tools, it takes craft, it takes a system. The whole course is built around that sentence, in three parts:

Part One: Mindset. The underlying logic of the AI era: why this time really is different, where ordinary people fit in, and how humans and AI divide the work. I move through it fairly fast, but every point is a judgment I've verified myself — nothing copied out of a book.

Part Two: the Live Part. The complete process of building a product: how to use AI to design a good front end, how to make games and music, plus my thinking on video editing — including how that 800,000-plus-view video was cut together, step by step. I'll open the project files and lay them out for you to see.

Part Three: Life & Systems. How to understand the state of a person's life, then turn those ideas into tools that run on their own, one after another. You won't have to feel your way in the dark — follow along and I'll build them with you, piece by piece.

Put together, these three parts are simply the opening sentence, expanded: how to live well in the AI era.

On money, I want to be sincere. Back in high school, online course materials were expensive — too expensive for me — so all I could do was hunt down pirated copies. I've never forgotten that feeling. That's why this recorded course is free, permanently. I want it to be there for everyone. If anything paid shows up later — say, a bootcamp where I work alongside you — the price will stay under ¥1,000. If you think it's worth something, vote with your wallet; if not, freeload away — you're just as welcome. There's a line in The Almanack of Naval Ravikant: the best work is your creative expression in the free market. The market will keep giving me the incentive to make better things.

Update schedule: starting with this one, at least two updates a week. I'll be online every Saturday and Sunday without fail, and this round will be done before Chinese New Year. In between, I'll run the occasional livestream showing the full Vibe Coding process — so you can watch, with your own eyes, an idea become something real, line by line.

3. What Gives Me the Right to Teach This

You must have a question mark hanging in your head: you're a student — what gives you the right?

First, my credentials: I'm a sophomore studying software engineering at a 211 university in Nanjing (the “211” label means top-tier in China's system). No decade of industry experience, no big-tech title.

But by nature I'm something of a rebel. I never wanted to keep grinding inside the exam-driven system, and those two years of 11th and 12th grade beat me up pretty badly. So in my very first semester of college, I started chewing on one question: besides chasing GPA, is there anything else worth doing?

As it happened, AI had just taken off. Other people used these tools when an assignment demanded it; I went to play with them on my own — one tool at a time, one idea at a time. Through the second half of freshman year, nearly all my energy went into this single thing: experimenting, hitting pitfalls, summarizing over and over. The course system you're looking at now was ground out exactly that way, bit by bit.

So when I say “my edge is speed and room to fail,” it isn't talent — it's a choice I made: I took the time other people spend grinding GPA and grinding competitions and spent it on trial and error. In the time it takes someone else to test one approach, I can test ten — then take every pit I stumbled into and turn it into experience to hand to you.

At this point, you may expect me to start preaching — “figure out what you really want,” “don't just copy what everyone else does.”

No. Say that out loud and you instantly sound like a lecturing dad — even I find it tiresome. I'll only speak for myself: back then I asked myself exactly one question — what do I actually want? I couldn't answer it, and that's why I began to try. If you already have your answer, you're ahead of me. If you don't, then reading this article right now already counts as trying.

This next part is for a small slice of you: if pure studying genuinely holds no interest — if you don't want to grind GPA, grind competitions, grind all that clutter — I want to tell you there's another road in college: take that energy and spend it learning something real, making things you genuinely believe are valuable. Not the mock-exam-style “projects” assigned in school, but actually solving problems out in the world — find a need, find a pain point, and use AI to solve it. Get out into society more. Practice more.

Of course, I'm not telling you to stop studying; I haven't skipped a single one of my own classes. Schoolwork is the floor, not the ceiling. What I mean is that the leftover energy you're burning yourself out on could go somewhere better.

Here are a few things I've picked up these past two years: champion of a university admissions promo-video competition, Top 350 in the semifinals of the TRAE AI Creativity Competition (invited to the in-person grand finals in Shanghai), grand champion of an in-person hackathon roadshow, Gold Award at a national creativity competition, Shenzhen Sci-Tech Innovation Star... I'm not putting these here to show off. I just want you to know: the things I've built with AI along this road have been tested for real.

There are also my conversation logs on TRAE and ZCode, and the sheer volume of tokens I've burned. I want to make one thing clear: I'm someone who uses AI at high frequency to build real things — not someone who read a few articles and then came out selling courses. Every piece of experience I teach was paid for out of my own pocket, then sorted and distilled. And one more piece of evidence — that video with 800,000-plus views: in the course, I'll open its editing project files right in front of you.

Finally, so we're clear about where I stand: I'm not a teacher walking ahead of you; I'm just a classmate who happens to be a little further along, for now. Toward this rat-race system, I'm a rebel; toward every individual person, I'm humble. What I want to do is mark the pitfalls on this road in advance, so those moving a little slower can take fewer detours and plug into the AI era sooner — speeding up my own growth, and sparking more people's creativity along the way.

4. What I Don't Know Either

I'm a sophomore. Cutting-edge theory, model algorithms — I understand only a little.

But remember that judgment from the beginning — get good at the application layer, and you can solve 90% of life's problems. I teach the part I'm actually good at, and I don't pretend to understand the part I'm not. That's why I dare to stand here.

5. One Thing I Didn't Get Right

Lastly, I owe you an apology.

My English-learning game missed its launch. I didn't make the date I promised you. After the competition ended, my life was a mess for a while, schoolwork was genuinely heavy, and it took a long stretch to get my routine back in order. There's nothing to argue here — I made the promise first and failed to deliver on time, and that's on me. I hope you can understand.

You surely want a firm answer. Honestly, I don't quite dare to slap my chest and promise anymore, but here's a high-probability one: after the National Day holiday and before October 17, the WeChat Mini Program goes live. I want to take it into the upcoming competitions, so it will be ready before then.

While I'm at it, here's the current setup: I did the game's planning and design, but the engineering has been handed over to professional developers — it involves servers and distributed systems, which come with performance pitfalls; my stress tests crashed more than once during the competition. I'm catching up on that material myself, but I'd rather get it into your hands soon than stubbornly shoulder the parts I'm not good at all by myself.

A Closing Word

Compressed into one judgment card:

What this course is — a somewhat restless college student who is, for now, a little further down the road, laying out all of his experience making things with AI, teaching how to live well in the AI era. What it is not — not a masterclass, no selling anxiety, no telling you to quit studying; and in the theory sections, I share exactly as much as I understand.

If this course suits your taste, go ahead and bookmark this site. The course videos will be updated in sync on Bilibili, and all the written scripts will live here on this site — you can browse the full course tree on the course page right now. The English game will most likely go live before October 17, and I'll tell you here the moment it does.

Next time, we officially begin.