The Gol D. Roger Moment of Software Engineering

There’s a recurring take going around: AI Agents and LLMs are killing software development. The argument goes that if you’re not writing code anymore, the craft is dead. Developers are becoming prompt jockeys with brains slowly rotting away losing their ability to write code by hand day by day. The profession is hollowing out. You can hear the grief in it. People who spent years honing a craft, mourning that the thing they loved is being automated away.

I think this is the opposite of what’s happening. And One Piece explains why.

The execution of Gol D. Roger

If you haven’t read the manga or watched the anime (or recently started season 2 of the Netflix adaptation like me), the entire story of One Piece begins with a public execution. Gol D. Roger, the King of the Pirates, is put to death by the World Government. Their goal: end the age of piracy. Send a message. Crush the dream.

It backfired catastrophically.

With his last breath, Roger tells the crowd that his treasure - the One Piece - is out there, whoever finds it can have it. The execution doesn’t kill piracy. It ignites it. An entirely new generation sets sail. The seas fill with pirates. The very act meant to suppress the dream becomes its loudest advertisement.

The World Government tried to end something by removing its figurehead. Instead, they made the idea unstoppable.

The parallel

Right now, “writing code” is being executed in public. LLMs write boilerplate. Agents scaffold projects. You can describe what you want and get a working implementation. The traditional image of a developer, someone who sits down and types out logic line by line, is changing fast.

Some people see this as the execution of software development.

I see it as Gold Roger’s last speech.

Because what’s actually happening is that the barrier between having an idea for software and having the software is collapsing. That’s not the death of development. It’s the starting pistol for an era where vastly more things get built by vastly more people.

Think about what was gating software before: you needed years of experience to hold a real codebase together. You needed a team to ship anything meaningful. Small ideas died because the cost of implementation was too high relative to the value. A solo founder, a researcher, a designer with a product instinct — none of them could execute alone. I’ve spent years building distributed systems and I still couldn’t ship a mobile app or a polished frontend without months of ramp-up or a co-founder with different skills. The ideas I had stayed as ideas - because of lack of time, kids, duties etc.

That gate is opening.

What actually dies (and what doesn’t)

To be fair, some things genuinely are going away. Rote implementation work. Copy-pasting CRUD endpoints. Wiring up the same auth flow for the hundredth time. Upgrading dependencies. Migrating off deprecated frameworks. If your entire value was in knowing how to type code faster than the next person, that’s under pressure.

But software engineering was never just typing (contrary to what some people want to believe). It was always:

• Understanding what problem is actually worth solving
• Designing systems that stay comprehensible as they grow
• Making the call when two reasonable approaches diverge
• Knowing when to stop building and ship
• Debugging the thing that shouldn’t be broken but is

None of that is going anywhere. If anything, it becomes more valuable when the commodity layer gets automated away.

The pirates who thrived after Roger’s execution weren’t the ones who could row the fastest. They were the ones who knew where to sail.

The new era

I think we’re at the very beginning of a software era that looks nothing like the last thirty years. Not because code stops mattering, but because the definition of “developer” is expanding beyond recognition.

More tools will get built. More experiments will get run. More ideas will get tested that would have died in a backlog, including that one feature you’ve been meaning to build for a year, the one that never quite made the cut. The ceiling on what a small team or a single person can ship is going up dramatically.

Is the old model of software development dying? Sure, maybe parts of it. But if the One Piece parallel holds, what comes after isn’t a smaller, diminished version of the craft. It’s something bigger and stranger, an era we haven’t fully named yet.

So what is the One Piece of software development? What’s the treasure at the end of this new age?

I don’t know. But I hope we all find out.