Building Haru Space and recording the journey

When I first imagined Haru Space, I did not have a perfectly formed plan for a giant platform. It began with a much more ordinary frustration.

Good conversations quickly disappear up a chat timeline. Ideas become a few lines in a note, then fade when someone gets busy. Even when a group manages to make something, the reasons behind its choices, the people who helped, and the lessons for next time are rarely kept together.

I want to connect those moments in one place. I am building Haru Space toward a place where people who trust one another can talk comfortably, shape scattered thoughts together, and carefully release worthwhile outcomes to the world. I also want the work that genuinely helped along the way to remain visible rather than be forgotten.

A place where conversation becomes an outcome

I am not trying to build one more service with a chat feature. Conversation is only the beginning.

A small thought from one person should be able to meet a useful question, become a clearer idea, and continue into a document, a piece of content, or a modest experiment. After people review and improve it, the result might become a public story or a project. Haru Journal is one of my first attempts to test that public part of the journey.

This does not mean making every conversation public. There must be a clear boundary between a space where people can speak freely and an outcome prepared for an outside audience. People should choose, review, and approve what crosses that boundary. I believe protecting that trust matters more than publishing quickly.

AI is a coworker, not the main character

Without AI, I probably would not have reopened so many things I had postponed. It already helps me organize material, translate, draft, find missing conditions, and implement small features.

That does not mean I want AI to choose the direction and publish whatever it produces. AI can offer alternatives and reduce repetitive work, but people must still decide whose material may be used, what may be shared, and who takes responsibility for the result.

In Haru Space, I want AI to work more like coworkers with different responsibilities. One may help with translation, another with curation, and another with a first draft of a story or feature. At consequential moments, however, a person needs to be able to stop, inspect, and decide. I see this boundary not as a restriction on AI, but as a condition for trusting it over time.

I want to remember useful work, not just visible activity

Communities produce easy numbers. We can count posts, messages, and visits. I do not think those counts automatically describe contribution.

A question that changed the direction, a reproducible bug report, an edit that made a document understandable, or a failed experiment that produced a better rule can all be meaningful contributions. These are also the kinds of work that are easily overlooked.

Before promising rewards, I want to begin by recording what actually helped in a way people can understand. If the criteria are wrong, I should revise them. If they create unexpected behavior, I should pause and look again. A fairer approach can only grow from records that preserve reasons as well as results.

A large plan should move through small proofs

The roadmap contains many ambitions: stabilizing the basic collaboration space, reviewing and publishing work made together, making contribution easier to understand, and eventually helping ideas grow into small service experiments.

I do not want to force features into a calendar simply because they are written down. I would rather test a small scope, record what breaks, and decide whether the next step is earned. The longer I work on this, the more convinced I become that one dependable flow matters more than one spectacular feature.

I also want to keep the failures. Some ideas will feel essential until people use them. Some approaches will be abandoned because their cost or complexity is not justified. Some problems will take far longer than expected. A collection of successes might become marketing, but it would not be an honest build log or a useful guide for the next decision.

What I will leave in Haru Space Lab

This category will not be limited to announcements about completed features. I plan to write about:

• why I chose one problem before another;
• what a small experiment confirmed and what it disproved;
• how I divided responsibility between people and AI in real work;
• how an awkward user flow changed after use;
• what surprised me after a release and what I chose next.

I cannot publish every part of the process. Private conversations, user information, internal operations, specific protective measures, and sensitive unreleased designs do not belong here. This journal will contain only product direction, judgments, experiment results, and lessons that are safe for anyone to read.

I do not want to wait until Haru Space is “finished” before telling its story. The hesitation and choices made while building it may explain what the service values better than a polished final screen ever could.

There is a long way to go, and the plan will change more than once. This time, though, I want to preserve the path as carefully as the outcome.

This is the first record in Haru Space Lab.