My Manifesto for Educating the Next Generation

This is the thing that lights me up for the next decade. So here's a snapshot of what I'm seeing, what I'm learning, and the direction I'm committed to.

If this resonates, please reach out because I'd love to exchange notes and build real stuff. Together we can drive this future.

Immerse in possible futures

I was late to AI. When GPT-4 dropped in March 2023, I didn't jump in right away. As a writer and educator, I wrestled with the "authenticity" question for a long time: if I use AI, do my words still matter? So I stayed away.

Then, in November 2024, I took on a course project with a YouTuber who loves AI. To move fast, I had to learn the tools. Then I quickly realized AI wasn't as scary as I thought. It was my perception of AI that blocked me from seeing the future with AI. My take is:

AI can both flatten your voice and level you up, depending on how you let it. It's a tool, not a verdict.

Was I late? Yes, but it doesn't matter. AI will continue to change our world, so a year or two is insignificant. What matters to me now is deciding to lean in and learn it in public with you.

Fast-forward to summer 2025, I decided to make a pivot within my love for education.

In my career, I've run a kids' coding school, built teaching software, and in the last 5 years, I've scaled an online program teaching entrepreneurs "building in public". Now, knowing that AI will completely revolutionize our next generation's education, I want to be a part of it.

Soak up and expand my worldview

Education is broad. Even if I've worked in this space for years in my career, I know nothing about what it looks like right now and in the near future.

What problems keep educators up? What actually matters in learning? How do parenting and psychology shape it all? I want to first find out about the possibilities.

So I decided to read a lot. The first book I picked up was Brave New Words by Sal Khan. It was the perfect book to kick off with because it laid out precisely what changes AI could potentially make in each area of education. There weren't answers, just questions.

Here are some of my favorite lines:

"Now the world is waking up to the possibilities of large language models and what is in store for education."

"To take advantage of this technology requires some creativity and bravery too. Not blind bravery, but something I have started to call educated bravery."

"This requires rethinking everything from the role of teachers, to how kids use multimedia, to how people get credentialed, to how to help graduates find jobs once they enter the workforce."

Two things stood out.

Who can get a higher-quality education?

Right now, no matter where you are, the best education still costs you a premium. It buys you smaller teacher-to-student ratios, brighter peers, and a network of wealthy parents.

But access is the problem.

Most of us won't be able to solve this, but we can help in those areas. Knowledge wasn't accessible in the past, but the Internet gave everyone access to it.

Feedback and iteration weren't accessible in the past, and AI will give everyone access to it.

Every student's learning experience is going to be elevated because AI can act like a patient and intelligent teaching assistant to guide based on one's interests and knowledge gaps.

Learning will be more personalized and fun.

Embracing a new learning model

"Many of today's school practices derive from these monastic traditions: transmitting knowledge rather than challenging or applying it."

This line is from another book I read: The Disengaged Teen, by Jenny Anderson and Rebecca Winthrop.

With knowledge, feedback, and iteration becoming accessible, we have to evolve education to the next important step: "challenge and apply."

This realization hits hard.

I've been a lifelong learner and I've also taught classes myself; I think we can all agree that education is not about classes, textbooks, and one-way lectures. We're way beyond these now.

They made sense in the past when one teacher had to handle 30 kids, and there was no better way. But with AI, we can finally break that bottleneck.

How about learning a bit, getting feedback, iterating, and repeating? Today, maybe a student can get a maximum of three feedback loops a week. Tomorrow, she can get thirty a day.

Let's do simple math. That's 156 feedback loops per year vs. 7800 feedback loops per year. A 50x improvement!

From what I see so far, the biggest misconception is that "AI is doing all the teaching". This is not true. In most cases, AI takes up the role of assessing and recommending, but the majority of the teaching happens with human creations like books, workbooks, well-designed EdTech tools, etc.

AI is not, and should not, directly teach a human being.

Alpha School is a public example that I've been following closely. They promote a learning model:

• 2X your learning
• 2 hours of learning a day
• 4 hours+ of project time
• No teachers. Only human "guides"

Honestly, I see these numbers as marketing slogans only. From a detailed review I read, students spend way more time on learning and a lot of time interacting with the impressive "guides".

Putting this particular business aside, what I care about most is the flipped learning model, where kids take the initiative to learn.

Kids drive their own learning. Adults and technology assist.

I can boldly imagine a 10x learning outcome. Of course, it is hard to measure because a kid's development is not, and should not be, solely about academic results. Life skills, soft skills, character, mentality, all these things are super important yet hard to measure.

Some people say this type of learning model is not for everyone. I agree. However, I believe every kid is born with a high level of curiosity and the ability to learn. It's that many factors play into their development in the first 6 years: parents' expectations, parenting styles and values, peers, etc. Only with the right environment to nurture can a kid adapt to this learning model. I think a lot of kids lose their curiosity and ability to learn because they don't have the right environment for the first 6 years.

Is it realistic to do this for every kid? No. We can only try our best to bring this new learning model to more kids.

Build, don't just learn

I did a talk for university students recently, this is what I told them:

The keyword here is 'building projects'. I can tell you that throughout my 13-year career, I have been putting myself into new areas all the time.

Every time I'm like 'Oh my, it's so new. I don't know where to start.'

And I just come up with a small project, thinking 'Okay, let me just do this project and use it to talk to people and then use it to post online. And this project then starts to create conversations, attract like-minded people, and eventually I can figure out what I really want to do.'

When I build projects, I showcase to the world what I'm capable of doing. I start to gather people. I start to see problems. I get my hands dirty to solve them. I get hired to solve other problems. That's creating value. That's being important in the world.

Entrepreneurs know. We have to build to create value.

In education, the same holds. We have to build to learn. No textbook or assignment can give us better lessons.

Connect interests to learning

For younger kids, I love what Heather Shumaker said in her book, It's OK NOT to Share:

"It often comes when we try too hard to do what's best for our kids."

I've seen this firsthand everywhere. Parents are making all the decisions about what's best for their kids. Teachers are creating all the learning structures for what's best for their students. I get it. It is much easier to teach and evaluate when things are well structured.

But the reality is that students don't have to think.

When I reflected on my childhood: first a local school in Hong Kong, then boarding school and university in the U.S. I realized I enjoyed autonomy and the environment the most.

In Hong Kong, I didn't get to choose. Most of us didn't really know why we were studying this or that. I was one of the worst students in my grade.

In the U.S., I got to choose every class and activity. I followed my curiosity. Even if I made a mistake, I owned it because I made the call. My peers were generally happier and more motivated because most of us got to do what we wanted. Once I got into that environment, I knew what to do and became one of the best students in my grade.

When we don't decide for our kids but genuinely follow their curiosity and interests and inject learning objectives, they learn more, and they're happier doing it.

My wife, Lydia, is a math tutor and her style is "play to learn" too.

Lots of parents teach math by asking their kids to do 1+1 = 2, but Lydia uses real-life examples. "If you have an apple and mommy gives you one more apple, how many apples do you have?"

Our older daughter, Avery, finds it easy to learn this way. More importantly, she understands the concept of addition and can do different combinations soon after.

In It's OK NOT to Share, Heather also said:

Young kids don't need to study the alphabet. They need to fall in love with songs and stories.

Too many parents focus on the technical details: the alphabet, words, and math concepts. They forget two bigger-picture goals. First, we should learn to apply knowledge to real life. It's much easier and better to teach with real-life applications in mind. Second, we can find enjoyable ways to learn something boring.

Avery loves food, so we connected letters, words, and math to food. Instantly, it became fun and easy. Play to learn is the truth. That's also why we built FunSheets to bring this to other parents and educators.

Reflecting on this in my city, Hong Kong

When I look at what parents in Hong Kong pay attention to, I understand why most care more about getting into a top school than grooming a self-driven kid.

It is a product of a system that feeds back into the system.

In Hong Kong, most people are still told to attend school, get good grades, and land a job. That's the default path of life.

Our society values credentials and titles over knowledge and purpose.

With these two ideas ingrained, it's easier to work on the technical details to get into an elite school (tangible) than to shape a mentality (intangible). You feel more in control.

It is a vicious cycle.

This won't work when AI and robots take over many jobs. But right now, there's a lack of education on this topic and a lack of alternatives. Most people still need to play this game.

When I talk to current university students, many already know the default path doesn't work anymore. Some are fighting to create their own paths, and some are stuck because that's the only path they know.

But parents of toddlers and teenagers don't see this firsthand. They are still hoping that a school can help their children thrive in the future.

It has to be a collaborative effort between parents and educators.

How can we get more parents to see the brilliant young people on a different path and what they actually do to develop themselves to become who they are? That's the question waiting to be answered.

A few more things I see as important in the future

Hacking Kids' Motivation

Motivation is more important than ever. For a self-driven kid, the sky is the limit.

Most of the time, kids are just bored. I love this quote from The Disengaged Teen:

The solution to kids' boredom is not to ask kids to work harder or jump higher. It's to give them some choice over how they spend their days in and out of school.

It is the age of agency. Students have more of a say about what they learn, how they learn it, and how they demonstrate mastery of it. This requires a better understanding of what successful learning looks like.

As long as a kid can do this, I don't think parents need to worry about the future. This kid is going to thrive in the AI age.

The role of community learning

Some of the best learning in my life was communal.

In 2013, I learned coding in New York with 20 passionate peers. We learned and built 9-6 every day for 6 days a week. All of us became software engineers right after.

I first ran Public Lab as a community for entrepreneurs keen on learning "building in public". I ran online cohorts bringing together 20 committed students at a time. They learned. They posted. They leveled up in marketing themselves and their work.

This style of learning with a like-minded community will never go away, whether it is online or in-person.

Parents' role is shifting

I experienced the shift from parent-educators firsthand as I use Chatty Bestie with Avery.

Today, we parents are responsible for our children's growth, so we do all the teaching, evaluating, giving feedback, enforcing, comforting, and motivation boosting, and so much more.

Sometimes children just don't want that side of their parents.

This inevitably hurts the parent-child relationship.

But with AI, we can give AI a role to take on, primarily for live interactions and giving feedback. We parents can take on a deeper level of evaluation, support, and emotional support.

This, again, I see firsthand, changes the entire learning dynamic. It makes learning more dimensional and more fun.

For younger children, given that parents are still present, this is not a bad thing at all.

Let's get outdoors

As much as I love technology, nothing replaces dirt, sun, and risk.

Just like "building projects", having outdoor experiences gives kids an indirect way to build up all the life skills they need.

So, time to build!

There is so much to explore, learn, and execute on.

As an educator, I want to play my part. By writing this out, I hope to find other passionate educators to collaborate with. Together, we can build this reimagined future.

As a parent, I feel that there is more urgency than ever to shift our energy toward walking on a new path. But I'm also not 100% ready to abandon the traditional path, so I hope to share this to help every parent find a balance.

— Kevon Cheung