In spring 2025, while working for a company in Tokyo, I paused to consider the months ahead.
Where should I live? If I became independent, what work would I do? Deciding on the work alone would not settle the day. Practical arrangements and working hours would eventually return to the same calendar.
There is a considerable distance between “one day” and deciding to move now.
Putting down what I was carrying
I began talking with AI to make that distance easier to think about. What was difficult? What had I not decided? Which conditions mattered? Even when the thoughts were scattered, I could begin putting them into words.
But a long conversation eventually led to a new session, and to explaining the assumptions again. I wanted to continue; instead, we had to meet all over again.
I tried collecting the necessary context in YAML and passing it into the next conversation. I did not want to seal myself into a page of settings. I wanted to see what changed when I reconsidered something.
I was also exploring this form of conversation through the metaphor of a Stand. A separate retrospective on those ideas and designs follows that history.
There is work after the choice
Finding a home in Takamatsu, forming a company, arranging the move, building a website. Written down, the tasks made clear that enthusiasm alone would not move them along.
Conversation with AI helped me consider the order and understand explanations of procedures. Then I had to check, contact people and complete the steps. There was also a distance between organizing the work and making a day happen.
In October 2025, I became independent, based in Takamatsu. Tokyo to Takamatsu. The time zone remained Asia/Tokyo.
The clock setting had not changed. Where I placed my time could. I began arranging the work with that thought in mind.
Making something, sustaining it
That year, I also developed a product called fragment.place. Could a document carry its intention, assumptions, relationships and responsibility alongside its prose? Thinking about how information connects in professional work had become a question I could explore by building.
Beyond the screen, there is explanation, repair and operation. Being able to build something does not settle what business I want to sustain. Today, fragment.place publishes specifications and technical notes. The earlier product ambition is not a description of a continuing current service.
Leaving room in the day
At the end of 2025, I wrote about a day as a “small theater.” There was quiet, there were plans that did not hold, and there were changes of role. Increasing the density of work was not going to make everything fit.
In 2026, Fragment Practice is an independent advisory practice in AI, security and technology risk. Within my own operations, I work with AI to revise how records and execution are arranged. Those experiments and client work have their own information and responsibility boundaries. The scope for client work is determined by purpose, information classification, contractual conditions and working environment.
Even with a system in place, choices remain. What should I delegate? Where must I judge for myself? How far should I go today?
When AI can do more, what should count as progress? The size of the business cannot answer that alone. There is specialist work I want to keep doing, and a life I want to sustain alongside it.
Moving to Takamatsu did not complete the answer. It did give me a place from which to ask again where and how I wanted to work.