Skip to main content
Back to the villageDONGGU’S FIELD NOTESFIELD NOTE / 04
The journey · personal notes
CHAPTER
One person shipping a product in three weeks?
PERIOD
2026–
AUTHOR
Donghyun Kang

A NOTE FROM THE JOURNEY

The same tools. Different decisions.

Mentoring four junior developers and building a product on my own brought me back to one question: who decides what to ask AI to do next?

Story illustration of people learning together and deciding what to build with AI tools
Story illustration
I practice the decisions that come before code: what to leave out, what counts as an outcome, and who that outcome helps.

Around the same time, I mentored four new developers. I was also formally appointed as a mentor twice by the Gyeonggi-do Job Foundation.

The mentees all carried the same question.

"Knowing how to code" no longer proves anything on its own.

If a new developer asks AI to write the same code they can write, how do they prove themselves now?

When the tools become equal

Park Chan-jun (winner of the 2026 Krafton AI Hackathon) defined it in one line in an interview.

"What I think of as judgment is less a grand sensibility, it's closer to deciding what to ask the AI to do next."

When the tools become equal, judgment becomes the differentiator.

This wasn't only for juniors. Building OpyojariDotjari solo and, at the same time, teaching the people coming up behind me, it was the same one thing.

What you decide

Anyone can write code. What you spend an hour on is what shapes your posture.

  • Decide what not to build.
  • Define what counts as a result.
  • Ask who benefits from this result.
  • Check whether you can measure it.

These decisions come before the code.

This seat is built by decisions, not by code.