By July 2020, Phuc had seen enough.
The world was cracking. The old systems weren’t going to hold.
He had the theory. Now he needed something that could touch molecules.
A machine. A proof of concept. Something you couldn’t ignore.
⸻
He didn’t wait for permission.
He started with what he had.
Household materials. Crude sensors. Tinker tools bought off eBay.
The goal was simple:
Can oil and water be separated using vision alone?
No AI. No deep learning. Just a cheap sensor that measured brightness.
If it was dark enough — it was water. If it was light — oil.
A single number. A single actuator.
And it worked.
Crude. But real.
The first Vision Valve prototype.
⸻
The company wasn’t called VisionCycle.
Not yet.
It was just an idea — Phuc Labs — a startup engine for startups that didn’t exist yet.
VisionCycle was meant to be the first launch.
And for a while, it worked.
⸻
He raised $1.5 million from investors. Matched it with $1.5 million of his own.
They pivoted: • From water to car batteries • From car batteries to air conditioners
Whatever could work. Whatever could move the mission forward.
⸻
But along the way, Phuc made a strategic mistake.
He stepped out of the technical center.
Let others code. Let others decide.
He wanted to move molecules, not just code.
And for the first time in his life, he wasn’t the engine of the product.
That choice would cost him.
⸻
The machines never reached scale. The dream never reached market.
But the lesson?
That stayed.
And years later, when Phuc began building again — this time in recursion, not metal — he remembered the valve.
And knew:
“The next machine won’t use wires. It will use loops.”
⸻