From Whispering to an Operating System: In June, I Am Changing How I Operate with AI

Until May, I mainly used conversations with Meta Chulbuji to organize my thinking, then executed projects based on the results.

That method worked. I operated the AI blog and moved as far as applying for AdSense. I also organized chulbuji.com into the Board, Log, Insight, and Project structure. AI Content Assistant reached the Kmong listing and sales-validation stage.

But for June, I felt that the operating method needed to move up one level.

Until now, conversation was the center. When an idea appeared, I talked with Meta Chulbuji, clarified the direction, and executed. Here, whispering means the operating style of continuously talking with AI to unfold thoughts and find direction.

The problem is that this alone is hard to turn into a repeatable operating system. Even when the conversation is useful, if the purpose, roles, validation criteria, and record location are unclear, the next execution starts again from scratch.

In June, Structure Comes First

Going forward, I want to structure and record work with harnesses, skills, workflows, and evals instead of only organizing it through conversation.

A harness is a frame for defining purpose, roles, scope, validation criteria, and record method before work starts. A skill is a way to package repeated work into a reusable procedure. A workflow is the actual execution sequence, and an eval is the evaluation criterion used to judge results by standards rather than by feel.

The core of June is not only the first payment validation for AI Content Assistant. That validation matters, but the larger goal is to build the ability to operate an AI working system directly.

AI Content Assistant, AI Blog, AdSense preparation, marketing experiments, and auto-trading record assetization are independent projects. At the same time, they are validation targets for the June operating system.

June Operating Flow

Major work will follow this flow.

1. First, write a harness that defines purpose, roles, scope, validation criteria, and record method.
2. Mark repeated work as a skill candidate.
3. Organize the actual execution path as a workflow.
4. Check the result against eval criteria.
5. Recover meaningful process records into chulbuji.com as Log, Insight, SOP, or Board.

This does not mean building a grand system all at once. It means creating a small repeatable structure, using it in real work, recording what happens, and improving it.

What I Will Check in June

This month, I will use the following items to check whether the operating method actually works.

Item	Standard
Harness cards	Write at least 5
Skill candidates	Identify at least 3
Workflows	Run at least 2 in practice
Eval criteria	Write at least 2
Log / Insight	Publish at least 2 operating-system records
SOP	Publish Harness-Based Operation Routine v0.1

Even if every item is not met, I will not label it as a simple failure. What matters is not staying busy without criteria, but leaving enough structure to know what worked, what got stuck, and what should continue next month.

Recovering Work as Operating Assets

June is the month for applying this operating method to real work. Marketing, content, productization, automation, AdSense, and AI Content Assistant will not be left as scattered projects. I will handle them inside one operating system.

Until now, I found direction by talking with AI. Starting in June, I will build a way of working with AI.

The goal this month is not to claim a fixed revenue path. The goal is to build a structure that can review service demand, revenue possibility, and operating quality, then apply that structure to real projects.

Work Log

Item	Note
What I built	June operating system transition direction and harness-based operating criteria
What broke	Whispering alone is hard to turn into a repeatable operating system
What I learned	After learning how to talk with AI, I need a structure for working with AI

→ June 2026 Operating Board
→ Harness-Based Operation Routine v0.1