Lately, a lot of friends — both in the U.S. and back in China — have reached out after seeing the recent K-Scale update. I thought I should write down my side of the story.

When I joined K-Scale, people were understandably puzzled.

I was in my forties, a father of two, and had already been a GM at Lenovo U.S. — that was eight years ago, back when I thought my path in tech was already set. Not exactly the profile of a startup risk-taker. A Confucian by temperament, I believed in balance and duty, not chaos and burn rate.

So why didn’t I just take a safe job at Google or Meta like everyone else my age?

Last fall, my son was born, and I was on the longest break of my career — six months of parental leave. After my in-laws came to help, life slowed down. I spent my days between diaper duty and long coffee chats with friends about the new wave of AI. One of them happened to be an angel investor in K-Scale.

He introduced me to Ben:

“I know his manager at Tesla — he’s talented, but he’ll need someone with real hardware experience on the team.”

My first meeting with Ben was, honestly, a bit disappointing.

He seemed like a good engineer — earnest, a little awkward — but when he talked, his thoughts jumped all over the place. They wanted to build a hardware business but didn’t know the basics: FOB, payment terms, final balances.

“He’s adorable,” I told my friend afterwards, half joking.

“I don’t think he could survive five minutes in the swamp of China’s hardware ecosystem. Still… maybe I can help him.”

That night, I noticed Ben had given me access to their internal wiki.

Out of curiosity, I started reading it — half browsing, half feeding my baby in the dim light of the living room.

Then I came across a letter he’d written to investors. I was stunned.

(I still revisit that post sometimes — Why K-Scale Wins. His writing there was a hundred times clearer than our first conversation.)

I turned to my wife and said, “I kind of want to join them.”

She asked, “Does it sound reliable?”

“Not really,” I said. “But I like them. The kid has talent — and vision. They’re hosting a hackathon this weekend.”

That hackathon is still hard to describe.

It had a kind of raw energy I hadn’t seen in years — maybe only the early days of Xiaomi’s MIUI community came close.

There were so many young people, so many wild ideas flying around. The air itself felt charged. Many of today’s well-known founders — the ones who have since raised millions — were there too, back when none of us had much more than dreams and 3D prints.

JX, Kalsey, and Dynes even built a tiny robot to battle the one Ben made.

That scrappy little prototype would later evolve into what became Z-Bot.

Who could’ve guessed something that started as a weekend hackathon would shape everything that came after?

When I got home, I told my wife, “I’ve decided to join them.”

She asked again, “Is it reliable?”

I shook my head. “No. But I really like them.”

A few months later, at another K-Scale hackathon, I invited Pete Lau, the founder of OnePlus.

After spending a day there, he laughed and said,

“This energy — the last time I felt something like this was when we were building OPPO’s first Blu-ray player. lol.”

K-Scale was always a little crazy — in the best way.

Back then we ran fourteen-day sprints, each with some huge, impossible goal. We never actually hit one, but that didn’t matter. Everyone kept pushing.

I was usually in around 10 a.m. — the only one who didn’t live in the office, because I had two kids. Everyone else looked half-awake, clutching coffee.

“What time did you sleep last night?”

Almost no one ever said before 3 a.m.

If Chinese startups are 996, K-Scale was 10-3-7.

It was wild.

JX was always arguing with Ben — loud, passionate debates that somehow ended in laughter.

Afterwards Ben would quietly tell me,

“You know, JX is exactly the kind of person I like, and he represents the culture we want to build here.”

Pawel, meanwhile, sat in his corner, headphones on, mumbling to himself as he coded — clearly about to pull off something big.

For Ben, the pressure was even heavier.

Because of family reasons, he was flying back and forth between California and New York — red-eye flights, endless context switching.

Sometimes I told him to stop coding while he was in New York, but it never worked. He just couldn’t stop.

We kept running — fast, messy, full of hope.

And of course, we made plenty of mistakes along the way.

But it’s hard to be mad at young people for making mistakes.

If anyone’s to blame, it’s me — the old guy in the room who didn’t know enough about robotics, and wasn’t firm enough in correcting what went wrong.

Still, we hit our first real milestone.

On the eve of AI Day, when most of us had already accepted we probably wouldn’t have a walking demo, the robot suddenly started moving.

JX shouted, “Wesley, I love you!” and ran into the garage.

Pawel and I couldn’t stop grinning — we ended up walking laps around Atherton at 11 p.m., laughing like kids.

Then Pawel looked at me and said,

“You know what this means? If it’s walking in the garage now, our sim-to-real is finally working. Soon it’ll walk on grass.”

We were too young to realize how far we still had to go.

The real sim2real breakthroughs came much later, after Scott and a few other brilliant engineers joined the team.

When I think back, I’m still not sure if we were chasing the future or just trying to get away from the present.

Either way, those nights in the garage stick with me — the tired faces, the noise, and that crazy moment when something finally worked.

It wasn’t easy. Half the time we didn’t really know what we were doing. But it felt real, and alive in a way that’s hard to describe.