Jaipur — There is a romantic image every engineer carries around in his head. It usually involves a workshop. Not necessarily a fancy one, just a workshop. A desk covered in tools. A soldering iron still warm from the last prototype. A 3D printer humming in the corner. Drawers filled with mysterious screws that somehow never match the one you’re looking for. Whiteboards covered in terrible sketches. Half-finished robots staring back at you like they know something you don’t.
When I was younger, that was the dream. Build bigger things. Smarter things. More impossible things. The metric was simple - If the project has more wires or more lines of code than the previous one, I was moving in the right direction.
I still love that image. Honestly, I hope it never stops.
Every engineer has a memory like this. An LED blinking for the first time. A piece of code compiling for the first time. A robot taking its first wobbly step. These moments feel magical because they represent something deeper than the objects themselves. They’re proof that an idea escaped your head and entered the real world for the first time.
That moment you changed reality, even if only by a few millimetres.
That’s addictive. That makes you chase larger projects.
The tiny robot becomes a bigger robot. The bigger robot becomes an autonomous rover. The rover becomes a vehicle. The software becomes an ecosystem.
Every year the projects become more ambitious.
Or at least that’s what I thought ambition looked like.
In December 1947, three men at Bell Labs got a sliver of germanium to amplify a signal. The transistor is arguably the most consequential object of the twentieth century, and the story is usually told as a story about that object.
But the transistor was, in a sense, inevitable. Solid-state amplification was in the air; several groups were circling it. What was not inevitable was Bell Labs - an institution that put metallurgists next to mathematicians, gave them decade-long horizons, tolerated the failure rate that implies, and made it socially normal for a chemist to wander into a physicist’s problem. That arrangement had to be invented and defended, year after year, against every reasonable person who wanted to know what the quarterly return on it was.
The transistor happened once. The lab produced the transistor, information theory, the laser, Unix, and C. If you had to choose which achievement to be able to reproduce on demand, only a fool would pick the transistor.
This pattern repeats with irritating consistency. We remember the object; the object was downstream of an organization that was the actual invention.
I think every builder eventually runs into the same invisible wall.
At first, the problem looks technical. Learn a new programming language. Design a better battery. Buy a better printer. Optimise the algorithm.
Eventually the robot works. The software scales. The prototype functions.
And yet…
The impact doesn’t grow nearly as fast as your technical ability does.
Why? Because the bottleneck quitely changed…
It stopped being about technology. It became about people…
The robot isn’t useful because there aren’t enough people who know how to use it.
The startup struggles because nobody knows how to work together.
The lab sits empty because nobody believes they belong there.
The project isn’t limited by engineering anymore.
It’s limited by culture…
So, if institution-building is the higher-leverage craft, why does every builder I know avoid it?
Because it has no compile step.
That is the whole answer, really. When you write code, the universe answers you within seconds. It tells you that you are wrong, specifically, with a line number. That tight loop is what makes programming addictive and what makes competence accumulate quickly. You can improve at it alone, at night, without permission.
Organizations offer none of this. The feedback loop is measured in months. The error messages are ambiguous and arrive as moods — someone is slightly less enthusiastic than they were in March, and you will not learn why until August, if ever. There are no unit tests. You cannot roll back. Credit is diffuse and often goes to the loudest participant rather than the person who designed the thing. And the work itself looks, from outside and sometimes from inside, like not working: a conversation, a reshuffled responsibility, an argument about who decides.
There’s also a status problem. In maker culture, the artifact is the proof. You can photograph a circuit board. You cannot photograph a norm.
An organization is one of the most complicated systems you will ever interact with. Think about it. It has memory. Feedback loops. Distributed processing. Failure modes. Version control through traditions. Information transfer through stories. Self-repair through mentorship. Emergent behaviour that nobody explicitly programmed.
Except unlike software…
every component has free will.
Debugging Python suddenly feels relaxing.
That’s why I’ve become fascinated by institutions.
A robot can build one thing.
A talented engineer can build a hundred things.
An institution can produce thousands of engineers.
The scale changes completely.
Suddenly your project isn’t the robot.
Your project is the environment where robots keep getting built long after you’ve gone home.
That’s a very different kind of engineering.
Maybe that’s what maturity looks like for a builder.
Not abandoning machines. Expanding the definition of what a machine can be.
A workshop full of tools is still exciting, But perhaps the greatest workshop is one full of curious people.
The raw materials become ambition instead of aluminium.
The components become trust instead of transistors.
The output isn’t a prototype.
It’s possibility…
I still want to build absurd things.
I still want the basement workshop.
I still want 3D printers, oscilloscopes, robots, AI assistants and electric motorcycles that look like they escaped from a science fiction film.
Those dreams haven’t gone anywhere.
But I hope they aren’t the biggest things I ever build.
I hope that one day, years from now, I walk into a workshop and find it full of people I’ve never met before.
They’re arguing over designs.
Testing prototypes.
Helping each other.
Failing loudly.
Building things I could never have imagined.
And somewhere in the corner, there’s an old workbench with a soldering iron on it.
Not because I still need to use it every day.
But because every institution, no matter how large it grows, begins the same way.
With one person.
One idea.
And a desk covered in wires.