Zhang Pinpin · Hands-On AI Coach
← Back to home

You Don't Need to Learn AI — But the People Teaching It Deserve a Pop Quiz

Long-Form Video · EP0092 September 4, 2026 03:09
What this episode covers

Somebody recently bragged about pulling in thirty million yuan doing FDE consulting — over a hundred projects, supposedly. In Chinese, the same syllables that mean "ten million" also mean "be sure to," so "thirty million" doubles as a triple blessing: be sure to stay healthy, be sure to stay happy, be sure to stay safe. That's the only thirty million I buy. People don't even bother rehearsing their lies anymore.

AI implementation in the enterprise is a heavy-service business, not a matter of spinning up some SaaS and walking away. Even with AI doing the heavy lifting, seeing a single project through from start to finish takes a month at minimum. One project, one team — run more than three in parallel and your people collapse. Do the math: those self-proclaimed FDE instructors probably haven't even heard of a hundred real projects between them, let alone delivered any. It's all talk.

Somebody recently bragged about pulling in thirty million yuan doing FDE consulting — over a hundred projects, supposedly. In Chinese, the same syllables that mean “ten million” also mean “be sure to,” so “thirty million” doubles as a triple blessing: be sure to stay healthy, be sure to stay happy, be sure to stay safe. That’s the only thirty million I buy. People don’t even bother rehearsing their lies anymore.

AI implementation in the enterprise is a heavy-service business, not a matter of spinning up some SaaS and walking away. Even with AI doing the heavy lifting, seeing a single project through from start to finish takes a month at minimum. One project, one team — run more than three in parallel and your people collapse. Do the math: those self-proclaimed FDE instructors probably haven’t even heard of a hundred real projects between them, let alone delivered any. It’s all talk.

So how do you tell whether an AI instructor is worth your time? The simplest filter: the ones who train full-time are the least credible — when would they have time for actual projects? Zero hands-on experience, pure information arbitrage. They absorb material just long enough to regurgitate it, half-understanding it themselves before they start teaching. The test is dead simple — ask what they’ve built and have them pull it up and show you. Anyone with even one shipped project, or one that’s turning a profit, is worth a conversation. The vast majority of instructors today are querying AI and copying prompts on the fly, armed with a handful of skills that look dazzling in a demo but fall apart the moment you try to apply them to real work.

Here’s the thing: AI genuinely does not need to be “learned.” If you manage to get your hands on the very best tools — Claude Code, Codex — you’ve already surpassed 99.9 percent of the people on this planet. Everything after that is simply treating the tool as a mentor, asking it real questions, and thinking critically about what it tells you.

The sooner a CEO or founder personally gets their hands on a top-tier AI tool, the sooner they realize this is nothing like PowerPoint or Photoshop. These tools are not built for the people who execute — they’re built for the people who decide. Every piece of execution can be handed to AI entirely. I already have clients who’ve rolled out subscriptions across their entire companies, spending one person’s salary to lift productivity for every department lead by at least three times — and I’m being conservative. People who truly internalize AI see twenty times or more. Of course there are counterexamples: bosses in the group who send their employees over with instructions to “go learn some AI from Teacher Pinpin.” Sometimes I honestly don’t know what to say.

How powerful is AI right now? Take my own situation. An account like mine — heavy, original content going out every single day — used to need at least a small team of one or two people. Now it takes less than twenty minutes a day. No filming. No editing. I just say everything I want to say in one go, and the system turns my words into a finished video and publishes it — fully automated.

I suppose I’m a strange one. Clients show up ready to pay for monthly consulting, and the first thing I tell them is to take half their money back. Once I’ve walked them through Claude Code hands-on, we’re done. The rest is exploration, and nobody on earth understands their business better than they do. Anyone without a real business, or who hasn’t figured out what their business actually is, I turn away outright. It’s a hands-on course — without a business to practice on, what’s the point? You won’t make money; you’ll just waste it. This leads to an absurd dynamic in my group: reverse admissions. Over a hundred people have signed up, and I’ve turned away dozens. I only agree to take someone on when I’m confident the value delivered will far exceed what they paid. Don’t I want to make money? Of course I do. But more than making money, I want to invest through cognition — to spend my limited hours answering the questions that are worth the most.

Teach a hundred casually curious people and ninety of them will never put it to use. That’s a worse bet than teaching ten founders who’ve spent years deep inside their industries. If those ten start wielding AI — forget creating wealth; the sheer social value they’d generate would be enough to make me proud for a long time. That’s just who I am: achievement-driven, through and through.

Teach a hundred casually curious people and ninety of them will never put it to use — that's a worse bet than teaching ten founders who've spent years deep inside their industries.