How a 17-Year-Old Became a Co-First Author on a Kimi Paper: AI Education Advice for Parents

Let me start with a fact that may make a lot of parents sit up.

There is a kid in Shenzhen named Chen Guangyu. He was born in 2009.

In November 2025, he became co-first author on a paper from Kimi (Moonshot AI), listed alongside senior researchers, proposing a technical approach called Block Attention Residuals.

When he did this, he was still in high school.

(To be clear and avoid any confusion: he is not Kimi's founder. The founder is Yang Zhilin. He was a high school intern and a co-first author of the paper. But it is precisely because he is a kid that every parent should pay attention to this story.)

A 16- or 17-year-old, without waiting for university, without waiting for a diploma, without waiting for anyone's permission, put his name on a paper at a top AI company.

How did he grow into that?

1. First, Look at What This Kid Actually Did

Lay out his timeline, and remember: all of this happened while he was in high school.

Earlier: ran an online store and short-video accounts, real business attempts with real stakes;
Earlier: organized a teen open-source community. Not joined. Organized;
February 2025: entered a high school hackathon, where he started learning AI research;
Summer 2025: spent 7 weeks interning at an AI startup in Silicon Valley;
November 2025: was discovered through the open-source community, joined Moonshot AI, and became co-first author on the paper.

When reporters asked him about it, he did not claim the credit. He stressed that "every contributor was essential" and declined to single himself out. The way others described him came down to four words: "dares to dream, dares to try."

"Dares to dream, dares to try" plus "every contributor was essential": these happen to be the two ends of an ancient model.

In my work at JIC, I break human capability into five elements (Aspiration, Time, Space, Love, Anchor) that combine into eight powers. I call it Five Elements, Eight Powers. Chen Guangyu's path draws on almost all eight.

And what parents can really do is this: for every power, give your child a real arena to practice it.

2. Breaking His Growth Down into the Eight Powers

The Power of Aspiration: "Dares to Dream"

He dared to think, "I'm a high school student, and I can do AI research too."

Most kids are not held back by ability. They are held back because they have never allowed themselves to think that way.

One for parents: don't set goals for your child, and don't rush to pour cold water on theirs. When they voice an aspiration that seems beyond them, your job is to catch it with one question: "So what's your first step?" Aspiration isn't something you shout into being. It is something you allow into being.

The Power of Execution: "Dares to Try"

Online store, short videos, hackathons... he didn't just think about them. He actually built them, one after another. Time turned his aspirations into action.

One for parents: attach a delivery date to every "I want to do this." Not "go look into AI," but "by this weekend, build a small thing that runs and show me." If it doesn't work out, that's fine. The act of doing is itself training the power of execution.

The Power of Influence: The Door to Being Seen Has Changed

Moonshot AI didn't find him because of test scores. They found him because he kept building things in an open-source community and got noticed.

Open-source communities and hackathons are the new imperial exams of the AI era. Keep making noise on one specific point, and real echoes will find their way back to you.

One for parents: in the AI era, the door to being seen has moved from exam papers to work and community. Take your child into a real arena: GitHub, open-source communities, competitions, hackathons. Let real people see and use what they make. One star, one citation, is closer to the future than a perfect score.

The Power of Creation: From Nothing to Something

He organized a teen open-source community (building an organization from nothing) and proposed a technical approach (building a thing from nothing).

Creativity is not finishing homework. It is staking out something new on empty ground.

One for parents: know the difference between completing and creating. Finishing 100 practice problems is completing. Making a small thing nobody has made before is creating. Each week, leave your child one task with no standard answer. That is where creativity grows.

The Power of Persistence: He Never Stopped

Online store → short videos → open source → hackathon → internship → paper. It looks scattered, but it is really the same drive, never broken.

One for parents: don't dismiss your child for "switching projects every time the novelty wears off." What matters is not which one they stick with, but whether that drive to always be making something ever breaks. Jumping around isn't scary. Stopping is.

The Power of Attention: From Scattered to Focused

He started out in business (an online store, short videos) and later narrowed down to one direction: AI research.

One for parents: casting a wide net early is right. That is how kids find their direction. Your job is not to force early "focus," but to help them protect it once they find what they truly want to hold on to, so other temptations don't pull them away again.

The Power of Love: The Hardest One

When the paper came out, he stressed that "every contributor was essential" and declined to single himself out.

A 17-year-old already understands that "the credit belongs to the team." Many adults never learn this power in their whole lives.

One for parents: when your child succeeds, asking "Who helped you?" matters a hundred times more than saying "You're amazing."

3. Three Counterintuitive Truths for Anxious Parents

① Don't wait until they're "qualified."

Chen Guangyu didn't wait for university or a diploma. The biggest dividend of the AI era is that the admission ticket is gone. On GitHub, a high school student's open-source project and a PhD's project stand as equals. The earlier you let your child into a real arena, the earlier they get seen.

② Knowing how to use AI matters less than knowing how to direct it.

The kids with a real advantage in the future won't be the ones who are "better at answering questions with AI." They will be the ones who can plug AI into real tasks, design workflows, and direct it to get work done. That is exactly what Chen Guangyu did. He wasn't writing code alone. He co-produced a paper within a team working with AI.

③ AI can grind out scores for your child. It cannot grow the Eight Powers.

Exams, calculation, memorization: AI can do all of these. But aspiration, attention, creativity, and love can only grow inch by inch in the child themselves.

AI flattens efficiency, but it cannot flatten your child. It cannot flatten their aspiration, their love, their place, or their time.

Final Thoughts

Chen Guangyu is not a genius myth.

He is an ordinary kid who was allowed to dream big, given a real arena, and never stopped moving.

In the end, parents can do three things:

Allow them to dream (Aspiration);

Give them a real arena where real people can see them (Execution + Influence + Creation);

Help them hold their direction once they find it, and teach them to see others when they succeed (Attention + Love).

The AI era will take care of the rest by tearing up the admission ticket.

You are not teaching your child "how to beat AI."

You are raising someone AI cannot replace:

a person with their own aspiration, their own love, their own place, and their own time.

"Where Wuxing meets code. Just I Can."