Why CEOs Are Coding Again?
·
3 min read
3 min read
P.S. I wrote this thought earlier, but it feels sharper now after seeing leaders return to code with AI tools. The point is not that every leader should become a full time engineer again. The point is that building is becoming easier to enter, and harder to ignore.
Recently, I read something that made me pause.
Mark Zuckerberg was reportedly back in Meta’s codebase after years, using Claude Code.
Garry Tan, the CEO of Y Combinator, has been openly using AI coding workflows and even shared his Claude Code setup.
Tobi Lütke at Shopify explored agent style research workflows on internal data.
These are not junior engineers trying to learn syntax.
These are CEOs.
That is what makes the signal interesting.
They are not coding because they need to prove they can type.
They are coding because AI has reduced the distance between an idea and a working system.
And when that distance becomes smaller, leaders can no longer stay too far from the act of building.
For a long time, many people in tech slowly moved away from code as they grew.
Meetings increased.
Slides increased.
Strategy increased.
The codebase became something other people touched.
But there is a problem with that.
In technology, coding is often where thinking becomes honest.
A product idea may sound perfect in a meeting.
A roadmap may look clean in a slide.
A strategy may feel sharp in a document.
Then someone tries to build it.
Suddenly, the truth appears.
The unclear requirement.
The missing edge case.
The broken data flow.
The dependency nobody mentioned.
The user journey that looked simple but is not simple at all.
Code has a way of exposing imagination.
That is why coding matters.
Not because everyone must become a software engineer.
But because building reveals what talking can hide.
AI makes this even more important.
If AI can generate the first version of code, then the scarce human skill is no longer typing every line.
It is reviewing.
Questioning.
Simplifying.
Connecting.
Deciding whether the thing should exist at all.
That means product managers, architects, analysts, consultants, delivery leads, and managers cannot stay completely outside the workshop.
They may not need to code every day.
But they need to understand how systems are built.
How ideas become logic.
How logic becomes risk.
How risk becomes user experience.
Otherwise, thinking becomes second hand.
We discuss systems we have not felt.
We approve complexity we have not opened.
We judge effort without touching the constraints.
Maybe the new builder era is not about everyone becoming a coder.
Maybe it is about everyone in tech staying close enough to building that their thinking remains real.
Sometimes we code to ship.
Sometimes we code to learn.
Sometimes we code simply to test whether our idea can survive contact with reality.
And maybe that is the point.
Coding is not always the final output.
Sometimes, coding is the thinking.
You might also enjoy
Why CEOs Are Coding Again?
·
3 min read
3 min read
P.S. I wrote this thought earlier, but it feels sharper now after seeing leaders return to code with AI tools. The point is not that every leader should become a full time engineer again. The point is that building is becoming easier to enter, and harder to ignore.
Recently, I read something that made me pause.
Mark Zuckerberg was reportedly back in Meta’s codebase after years, using Claude Code.
Garry Tan, the CEO of Y Combinator, has been openly using AI coding workflows and even shared his Claude Code setup.
Tobi Lütke at Shopify explored agent style research workflows on internal data.
These are not junior engineers trying to learn syntax.
These are CEOs.
That is what makes the signal interesting.
They are not coding because they need to prove they can type.
They are coding because AI has reduced the distance between an idea and a working system.
And when that distance becomes smaller, leaders can no longer stay too far from the act of building.
For a long time, many people in tech slowly moved away from code as they grew.
Meetings increased.
Slides increased.
Strategy increased.
The codebase became something other people touched.
But there is a problem with that.
In technology, coding is often where thinking becomes honest.
A product idea may sound perfect in a meeting.
A roadmap may look clean in a slide.
A strategy may feel sharp in a document.
Then someone tries to build it.
Suddenly, the truth appears.
The unclear requirement.
The missing edge case.
The broken data flow.
The dependency nobody mentioned.
The user journey that looked simple but is not simple at all.
Code has a way of exposing imagination.
That is why coding matters.
Not because everyone must become a software engineer.
But because building reveals what talking can hide.
AI makes this even more important.
If AI can generate the first version of code, then the scarce human skill is no longer typing every line.
It is reviewing.
Questioning.
Simplifying.
Connecting.
Deciding whether the thing should exist at all.
That means product managers, architects, analysts, consultants, delivery leads, and managers cannot stay completely outside the workshop.
They may not need to code every day.
But they need to understand how systems are built.
How ideas become logic.
How logic becomes risk.
How risk becomes user experience.
Otherwise, thinking becomes second hand.
We discuss systems we have not felt.
We approve complexity we have not opened.
We judge effort without touching the constraints.
Maybe the new builder era is not about everyone becoming a coder.
Maybe it is about everyone in tech staying close enough to building that their thinking remains real.
Sometimes we code to ship.
Sometimes we code to learn.
Sometimes we code simply to test whether our idea can survive contact with reality.
And maybe that is the point.
Coding is not always the final output.
Sometimes, coding is the thinking.
You might also enjoy