Ideas Come Out of Nowhere

Although I have already written notes introducing this site and the Goshuin Gallery, I still wanted to look back at the earlier thoughts behind them: what I wanted to create with code, why that motivation appeared, and how I have been slowly turning those ideas into real projects.

This is also a note about how AI accelerated that process. The motivation existed before the tools became powerful, but the tools changed what felt practical.

The First Thought

Around 2020, while looking through papers and projects about my graduate school research, I happened to visit several researchers' personal pages. That may have been my first clear impression that a resume or CV did not have to be only a paper document with a formal photo. It could also be a website: direct, readable, and sometimes even interesting.

Some of those pages were extremely simple, even old-fashioned in design, but the format still felt powerful. I remember thinking that it was cool and wondering how they had created it. At the same time, I did not try to build my own. Part of the reason was technical hesitation, and part of it was that I felt I had too little to put on a portfolio site.

The Second Thought

Later, the idea of creating a gallery website for my goshuin collection appeared. This idea made me much more excited. I began learning Vue and React with real motivation because I had a concrete thing I wanted to make.

Before I had even finished the tutorials, I realized that building the site would not be as easy as I had imagined. The progress stopped naturally, but I did not think of it as giving up. The idea simply moved into the to-do list, waiting for the right time and the right conditions.

The Third Thought

Then AI suddenly became a powerful and versatile tool, especially for developers. At first, I treated tools like ChatGPT mainly as stronger search engines. After trying coding agents, my view changed quickly.

At that point, personal projects no longer felt like distant plans. They felt practicable as soon as I was willing to move. I started thinking of more project ideas while commuting in crowded trains, walking, or simply letting my mind run through possible things to build.

The First Action

I started working on the Goshuin Gallery before I really understood how to code with agents in a disciplined way. I can honestly call it a vibe coding project: I kept moving without preparing documents first, and that was not an intentional methodology. It was simply how I began.

After about three months of iteration, most of the application work had been done. What remained was content fulfillment: taking photos of goshuin, uploading them, and creating detail pages. That part requires a lot of manual work, so the project is still in progress. I also began to think that this part should be automated as much as possible instead of becoming a long manual process.

The Second Action

Although the Goshuin Gallery has still not been completed after six months, I am also working on new projects, including this portfolio portal site.

This time, I am trying a more structured approach. I prepare the necessary documents before coding, then update them as the project evolves. I want to automate as much of the work as possible while still keeping the important decisions under my own control. When that balance works, development becomes more efficient without feeling out of control.

The Last Mind (for now)

Will AI replace all developers? What remains important for developers in this era? How can a developer take real advantage of AI tools without becoming passive?

I do not have complete answers yet. For now, I think doing something is more useful than only thinking about these questions. Building small projects, reviewing the results, and learning where AI helps or fails is the most practical way for me to understand the change.