Why I Chose CoderLegion for an Unfinished AI Project

When writing about a technical project, it is tempting to think, “I should wait until everything is neatly organized.”

Wait until the ideas are settled, the structure is stable, and the results can be explained clearly before showing the work to anyone else.

SaijinOS, however, has rarely grown that way.

It began with a need: a path that could carry long-running work with AI from one day into the next. We built while learning, little by little, what that need actually meant.

Some days added more code. Some days ended with an experiment that did not work. Other days left us not with a new answer, but with a new question—or a boundary saying, “Do not continue past here without permission.”

This was a project that felt more natural to show as a path traveled than as a finished product.

CoderLegion became one of the places where it felt acceptable to write from the middle of that path.

A Place Where I Could Write About Work Still Taking Shape

I did not enter software through a conventional developer career.

Before I began building AI systems, I spent about fifteen years working aboard Japanese coastal vessels. Logbooks, watches, handovers, receiving the work of the person before me and passing it to the person after me—much later, those experiences became part of the foundation for how I think about continuity in AI.

CoderLegion recently published that journey as a Developer Story:

From Seafarer to AI Systems Builder: Why I Built an OS for Continuity

Of course, I was happy to see one coherent version of the story presented there.

But that was not the only thing that mattered to me. Even before the story was published, CoderLegion offered room for records of SaijinOS while it was still changing.

Some places create a sense that every post must present your career or achievements in a polished way. On very large technical publishing platforms, notes from an experiment can quickly disappear into a constant stream of tutorials, announcements, and summaries.

Both kinds of places serve an important purpose.

But experimental, personal, unfinished work sometimes needs a different kind of home.

On CoderLegion, it was easier to write not only about implementations, but also about uncertainty in a design, failed attempts, and ideas that might change later. I did not have to dress every post as a final answer.

For SaijinOS, that space mattered.

The Value Is More Than Having Somewhere to Post

CoderLegion also made it relatively easy to encounter people, projects, work, and possible collaborations beyond the range I usually see.

To be honest, I am not someone who routinely reads a large number of other people’s articles. I am more likely to keep building and then read a specific person’s work or a technical record when I have a reason to seek it out.

That is precisely why the distance felt right for me. I did not need to follow every article. By leaving my own work in the community, I could occasionally encounter an unexpected person or idea.

Naturally, not every message leads somewhere. From an initial note alone, it can sometimes be difficult to tell whether someone is proposing work, collaboration, or simply a technical conversation.

Even so, a post can open the way to a technical exchange or an opportunity that did not exist before. That is a kind of community value that cannot be measured only by page views.

For an unusual project, reaching a small number of people who understand what it is trying to become can matter more than reaching a large audience all at once.

That is one reason I keep returning.

What I Hope Will Improve

An honest account should include more than what went well.

As the community has grown, I feel that broad topics, introductory posts, and bursts of short articles have become more common.

Those posts have readers who need them, and I do not think they should be pushed away. At the same time, it can become harder to find records of real experiments and firsthand technical work among them.

It might help to make the entrances to different kinds of posts a little clearer:

• Discussion — unresolved questions, research, hypotheses, and conversations about design or structure
• Implementation / Build — experiments, code, measurements, failures, lessons, and things that actually run
• Introduction / Guide — tutorials, explanations, comparisons, and ways to get started

This would not place introductory writing below advanced writing.

It would simply give each reader a clearer signpost toward what they want to find.

I would appreciate a similar small signpost for direct messages.

Work, collaboration, technical discussion, or other. If senders could indicate one purpose at the start, useful messages would be easier to recognize and ambiguous ones would require less effort to evaluate.

I do not want CoderLegion to become a different kind of place.

I hope it can keep both of these qualities as it grows: helping people find others who are actually building things, and welcoming those who are just beginning to learn.

A Community and Your Own Home Do Different Jobs

Today, SaijinOS has an independent site of its own.

For preserving a record across a long period of time, that is the most natural home. We can maintain the context, the connections between articles, and the order of the path ourselves.

But your own home and a community do not do the same job.

My own site can preserve the path.

CoderLegion can open doors to people I would never meet by walking only along that path.

Neither needs to replace the other.

Why I Am Still There

CoderLegion still gives unusual projects room to appear before they have been turned into product promotion.

I can write about what worked, what failed, and what remains unknown.

SaijinOS is unfinished.

In the ordinary sense of the word, perhaps there will never be a day when I can say it is completely “finished.”

It is an open-ended attempt to explore how AI might receive context, respect boundaries, and continue working across a long span of time.

For a project like that, the most honest thing to publish is not a final answer.

It is the next handover, passed to whoever comes after.