Welcome to Deep Dive Panda! We've spent the week decoding the most important conversations from over 13 hours of Chinese tech, business, and AI podcasts.

The big talk in Chinese tech isn't about models or data. It's about a huge shift in how we work. AI is getting incredibly good at doing things. That forces founders to focus on the one thing AI can't copy: good taste.

The Unifying Theme

Talk to anyone in China's tech world, from startup CEOs to podcasters, and they'll all tell you the same thing. The value of just doing stuff is dropping fast.

It's a new division of labor. AI handles the "how." We're left with the "why."

Hu Yuanming, who founded the 3D AI startup Meshy, put it simply: "AI is what accelerates the 'how.'" His team doesn't spend weeks building a 3D model anymore. Instead, they spend that time debating why they're building it at all. An educator in Beijing said the same thing about his students. He's not using AI to help them get answers faster. He's using it to let them tackle bigger projects that force them to think for themselves. That ability to judge—that taste—is something AI will never have.

This isn't some fuzzy, philosophical idea. It's happening right now. Think of it this way: you're not buying servers anymore. You're hiring a junior employee with an API call. But just like any intern, it needs good direction. Your job isn't to do the work. It's to have an opinion strong enough to tell the machine what to do. Even the famous mathematician Zhang Yitang says AI is great at combining what we know, but it can't invent a whole new game.

Why It Matters

If you're a founder or investor, this is a preview of what's coming. When the "how" becomes cheap, you have to change your entire game plan.

Your edge isn't how fast you can code or how slick your marketing is. It's the strength of your ideas and the clarity of your vision. As one creator, Pan Luan, said, "the most expensive thing in this world is an idea."

This means investors should care more about a founder's unique perspective than their code. A strong point of view is the new moat. For founders, it means you need to work on more than just tech or management skills. You need to think like an artist and a philosopher. Your job is to be the chief visionary, the main strategist, and the ultimate curator of taste. You're not managing people anymore. You're directing an AI. The real bottleneck is the quality of your commands.

What This Looks Like in Practice

  • AI is a great intern, not a boss. Think of AI as a smart, fast intern. It needs clear, step-by-step instructions to solve a problem. Your job shifts from worker to director. The value isn't in doing, but in deciding what to do.

  • Your taste is your new moat. Old advantages, like secret tech or huge datasets, are fading away. Now that anyone can rent a powerful AI, the real edge is human. It's your unique worldview and your ability to connect AI to a real person's problem.

  • The "Horror Game" Strategy. Meshy's first AI models made distorted, creepy 3D characters. Instead of fixing them, they found a niche that saw the AI's bugs as features: indie horror game developers. It's a smart way to launch an AI product. Find the customer who loves your flawed first version because of its flaws.

  • The "Hand-Cranked" Rule. The best teams just build what they need. They don't wait for perfect tools. They create scrappy solutions to solve today's problem. This shows they can learn and adapt faster than anyone else, which is the only thing that matters.

  • The algorithm is a fairer boss. Surprisingly, many gig workers who deliver food would rather be managed by an algorithm than a person. The app is tough, but it's fair. It doesn't play favorites. This might be a glimpse of the future, where systems run on clear rules and leaders are free to focus only on strategy and vision.

Best Quotes

❝

"My biggest shift was realizing a moat isn't the most important thing in a fast-moving field. The most important thing is being able to ride the wave."

William, CEO of Xiaosu Technology
❝

"To do a high-quality project, I believe 50% of the time is spent thinking 'why,' and 50% is spent on execution, thinking 'how.' AI is what accelerates the 'how.'"

Hu Yuanming, Founder & CEO of Meshy
❝

"You discover that in the end, the most fundamental thing in this world is an idea. It’s about different people's worldviews, their judgment of a matter. This is the most expensive thing there is."

Pan Luan, Creator of Luanfanshu
❝

"I can only say one thing: it's about finding joy in it for yourself. You are enjoying it, you think it's interesting, and what you're doing is very interesting. That's enough."

Zhang Yitang, Mathematician

This week's Deep Dive Panda covers 14 specific podcast episodes, including:

Here are the key lessons & mental models from each episode:

1. Indie Hacker - EP114: A conversation with William, CEO of Xiaosu Technology: An in-depth discussion on the core challenges of large-scale AI agent deployment

Original: 硬地骇客 — EP114 对话小宿科技 CEO William：深入探讨 AI Agent 规模化落地的核心挑战

  • The Inverted Startup Stack: Startups building with AI don't buy servers first. They buy intelligence. They start right away with model APIs and data from search feeds. Cloud servers are just a tool for scaling later, not something you need on day one. The game has changed. The basic building block isn't computing power anymore, it's intelligence.

  • Machine-First Search: When you're building a search tool for AI, you have to forget how humans think. It's not about finding that one perfect link. It’s about grabbing a big batch of the most relevant results and giving the AI the full context, not just a few snippets. Your user is a machine, and it reads everything all at once.

  • Ride the Wave, Build the Moat Later: If a market is growing 100x, your only job is to stay on the surfboard. Don't waste time worrying about how you'll defend your business on day one. That's just a distraction. The market itself is all the momentum you need. You can build your moat once you're already moving at full speed.

  • The 'Hand-Cranked' Mandate (手搓精神): Great teams don't wait for the perfect tools to show up. They build their own scrappy, "hand-cranked" solutions to solve the problem right in front of them. They're biased toward doing things, not buying things. Honestly, this is the clearest sign of a team that's going to make it.

  • Capital is Fuel, Not a Trophy: Here's some advice from an investor: take the money that helps you build the fastest. In a market where cash is everywhere, it's your product that defines you. Nobody really cares about the big-name VCs you've got on your side.

❝

"[My biggest mental shift was realizing my own shallowness... Talk is cheap. After starting my own company, you discover that a moat isn't the most important thing in the early days of a fast-moving industry. The most important thing is to be able to ride the wave.]"

William, CEO of Xiaosu Technology
logo

Subscribe to Deep Dive Panda to unlock the rest.

Become a paying subscriber of Deep Dive Panda to get access to this post and other subscriber-only content.

Upgrade to paid

Keep Reading

No posts found

© 2025 Deep Dive Panda.

Privacy policy

Terms of use

Powered by beehiiv