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

Gaokao Application Choices with AI as Your Advisor—It Doesn't Fill the Form, It Builds You a Toolkit

Long-Form Video · EP0040 June 14, 2026 4:33
What this episode covers

Can AI help your kid pick gaokao application choices? Of course it can, but not the way you think. Don't have AI fill in the application form—put its research ability to work guiding and safeguarding the whole process.

Two days ago I used the strongest model available, Mythos Fable 5 (even though that model is no longer usable now), had Claude Code call it, and built a web-based "parent's application toolkit." Here's the prompt I used: from the point of view of a parent in the 2026 gaokao season, spin up a multi-agent team to do deep research across the whole web, and pull it all into one web page—three core mental models, the key timeline, the tool kit, a glossary, a pitfall guide, a view of where each major is headed, and how to talk to your kid.

Companion files · Drop them into Claude Code
2026 Gaokao Application Choices · Parent's Toolkit (Web Page) HTML 64 KB
The full web version of the toolkit from the video: three core mental models, the key application timeline, the official and AI tool kit, a glossary covering rank position / equivalent score / tier cutoff line / file rejection / slipping through tiers, a pitfall guide, the five practical steps of reach-match-safety, an analysis of where each major is headed, and how to talk to your kid about what interests them versus what they're good at. Every dimension can be handed back to AI to dig deeper.

Can AI help your kid pick gaokao (China’s college entrance exam) application choices? Of course it can, but not the way you think.

It’s not about having AI fill in the application form. It’s about putting AI’s research ability to work guiding and safeguarding the whole process.

Two days ago I used the strongest model available, Mythos Fable 5 (even though that model is no longer usable now), had Claude Code call it, and built a parent’s “toolkit” for application choices. I’ll share the prompt I used here too.

The prompt: cast yourself as the “parent-advisor” and send a multi-agent team out to research

The gist of what I told it: it’s the 2026 gaokao season, and as a parent I’m a key advisor to my kid’s application choices. I need you to spin up a multi-agent team and do deep research across the whole web on everything related to gaokao application choices in China.

The main sections include:

  • The concepts—the terms you absolutely have to understand before you pick and what they really mean underneath, plus the full application timeline and the tools available;
  • The pitfalls that are easy to fall into along the way;
  • Guidance on choosing a major, including how to talk to your kid and help them see clearly where a given major leads.

And ultimately give it to me as a web-based toolkit. This is where it parts ways with other chat-style large models—I have Claude Code call Mythos Fable 5, and it hands me a web version of a gaokao season toolkit directly.

What’s in the toolkit

Three core mental models

  • Look at rank position, not raw score;
  • Tier your choices into reach, match, and safety;
  • AI is your advisor, not the referee.

The key timeline

It broke the season into windows—now through score release, late June, and scores through early July—and spelled out what to do in each.

The tool kit

  • Official tools like Yangguang Zhiyuan, plus your province’s education examination authority website;
  • Some AI tools, like the gaokao sections certain apps have shipped, or Tencent Yuanbao’s “Gaokao Tong” feature;
  • General chat AI works too—DeepSeek, Qwen.

But watch this one closely: do not go straight to models like DeepSeek or Qwen for the application form or school data, because some of their data is genuinely out of date. What they’re good for is understanding majors and career paths—or having them play a role and talk with your kid to find out what kind of major appeals to them.

Glossary plus pitfall guide

It gave me a genuinely complete glossary for the application process: what rank position is, what an equivalent score is, what a tier cutoff line is, plus terms like file rejection and slipping through tiers. Then a pitfall guide—the moves you absolutely must never make, some cautions on using AI, and the five practical steps of reach-match-safety.

Where each major is headed

Working from the latest research reports and national policy, it analyzed where each major is headed: which ones have real jobs waiting at the end of them and real national policy behind them. Some negative signals too—majors cooling off, industries trending down, industries with more graduates than seats.

How to talk to your kid

The parent is a supporting player here. The question is how to get your kid in front of real careers, hear them out in their own words, and let them judge for themselves whether a major is what they pictured. It also draws the line between what interests you and what you’re good at: you love playing games, for instance, but that doesn’t necessarily make you suited to computer science; wanting to make games doesn’t mean a game company plays games all day—the art and design work inside is pretty tedious too. To help you raise all this with your kid, it offers guidance from education experts plus the questions that come up most often during the process.

More important than the toolkit is the method

Every section of this toolkit can go back to the AI for a deeper dig and more detail.

But what matters more is seeing the process—using AI to dig out the underlying methodology of doing something, gather all the relevant data, and turn it into a guide. At work, that is enormously valuable, and clearly it isn’t limited to gaokao application choices.

I’ve also been doing some thinking lately. I’ve done short video and online marketing long enough to know exactly how to ride a trending topic and pull growth out of it. The news-driven pieces I’ve made recently, I can ship extremely fast, and they do bring followers in.

But I’ve thought it over, and I won’t be posting that kind of content going forward. It has no long-term value—for the people I really want to help, the ones here to learn hands-on AI, it doesn’t do much.

So I’ll stick to daily posts that carry real value and teach you something—hands-on AI thinking and concrete methods you can use.

What matters more is the process—using AI to dig out the underlying methodology of doing something, gather all the relevant data, and turn it into a guide. At work, that is enormously valuable.