Personal Blog

2025 Recap

This year I read over 30 books, went from being an FRC season finalist to moving to a new school in a new city, and founded OpenEd. Life feels like it's thriving.

Then I discovered that every Columbia student has their own brilliance—either academically diligent and intelligent, or achieving top-tier excellence in a particular field. Essentially, college admissions officers are VCs.

However, I feel demystified about the elite school halo. The world is a giant improvised stage, and HYPSM/Ivy League schools are no exception.

Although Stanford remains my dream school—it's free, innovative, and brave; it's the utopia for all tech entrepreneurs.

But the students who are truly best suited for Stanford have all dropped out to start companies. I think I'm one of them.

Because in this era, going to school is actually a high opportunity-cost decision. The chance to change the world is right in front of us.

This year I also learned the skill of coffee chats, which allows for rapid networking. I really like a quote I learned in Stanford's MS&E class: "Your network is your net-worth."

It was actually during that summer session at Stanford when I started asking myself: "Do you want to be a coder for life, or seize the opportunity to change the world?"

Without a doubt, I chose the latter.

Without a doubt, I will always choose the latter.

Because I either live brilliantly or die peacefully.

You should bet on me, like I am Sam Altman at 21.

2026 Preview

OpenEd will become the Perplexity of education:

All knowledge from K-12 to college can be learned through OpenEd's video generation/interactive animations/quizzes. We've invented a new learning paradigm.

In fact, it's already more useful than ChatGPT. Now when I want to learn a new concept, I use our own product instead of ChatGPT.

The reason I say it's the next Perplexity is because Perplexity is essentially a wrapper optimized for research, and OpenEd is essentially a wrapper optimized for education.

Including Manus recently being acquired by Meta for $4 billion. I remember when it first came out in March, the internet was flooded with criticism, saying it was just a wrapper based on Claude and MCP. But that didn't stop it from reaching $125M ARR. Technical difficulty doesn't equal market size.

I think the key to judging whether an AI product will explode is: Is it better than ChatGPT in a certain field? Manus and Perplexity are such products, and so is OpenEd.

I think entrepreneurs need to have extreme confidence in their products and extremely strong psychological resilience. Just as Sam Altman and Jensen Huang said, resilience and determination are more important than anything in entrepreneurship.

So I think by the time I graduate, my Columbia degree will become the most inconspicuous line on my resume: when your product changes humanity's learning paradigm, no one cares which school you went to.

Finally, I believe Fei-Fei Li's World Labs will become the next OpenAI, and spatial intelligence will be the last piece of the puzzle on humanity's path to AGI.

From then on, humanity will move toward fully automated production.

"Dream big, work hard."

Dedicated to all entrepreneurs who dream of changing the world.