June 2026

June 2026 Operating Board

From Whispering to an Operating System — Harness-Based Execution

A monthly execution Board summarizing what chulbuji.com is focused on operating this month.

June is not only about validating the first payment for AI Content Assistant. The larger direction is to move from conversation-centered work with Meta Chulbuji to a repeatable operating system based on harnesses, skills, workflows, and evals.

June 2026 chulbuji.com Operating Board dashboard image

June Strategy Update

Until May, I used whispering with Meta Chulbuji to clarify direction and execute projects. In June, those conversations are recovered into operating assets through pre-work harnesses, reusable skill candidates, execution workflows, and eval criteria.

  • 01
    Move from conversation-centered work to system-centered operation

    Instead of stopping at AI conversations, each major task starts by defining purpose, roles, scope, validation criteria, and where the record will live.

  • 02
    Apply harnesses, skills, workflows, and evals to real work

    Rather than designing a large system at once, apply the concepts to small repeated routines: AI Content Assistant, blog publishing, AdSense checks, marketing experiments, and auto-trading records.

  • 03
    Treat each project as a validation target for the operating system

    AI Content Assistant, AI Blog, AdSense, marketing experiments, and auto-trading record assetization are independent projects, but they also test the June operating system.

Operating System Validation Targets

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

Core Direction This Month

  • 01
    Write a harness before starting major work

    Before execution, define the purpose, roles, scope, validation criteria, and record location. June is for turning this into the default operating habit.

  • 02
    Validate market response for AI Content Assistant

    After the Kmong listing, observe industry samples, test deliveries, inquiries, and payment possibility. Treat the results as demand-validation data, not guaranteed outcomes.

  • 03
    Recover results into Log, Insight, SOP, and Board

    One-off work ends as a Log. Repeated work becomes a skill candidate and eventually an SOP. Judgments are checked through eval criteria, not only by feel.

Shared Foundation Capabilities

AI Content Assistant, AI Blog, and chulbuji.com look like separate experiments, but all three run on the same six foundational capabilities. June is about demonstrating these inside a harness-based operating system.

Planning ability Knowing which experiments to run and why
Content production Turning ideas into actual outputs
Product-building Completing what's built into something that works
AI tool utilization Knowing which AI to deploy, when, and how
Marketing understanding Explaining why what's built is valuable and to whom
Monetization experiment design Building and validating a path to revenue

Operating Status

Focused Operation · Operating System Validation 3
Service demand validation · harness applied

AI Content Assistant

After the Kmong listing, observe industry-specific samples, test deliveries, inquiries, and payment possibility. Validate real demand and operating procedure through a content-format service before building an app.

AdSense preparation · publishing routine eval

AI Blog / AdSense

Operate around 2–3 posts per week and 25+ cumulative posts as a working target. Record AI writing, SEO review, and AdSense preparation as routine and quality validation, not as an approval promise.

Home base · Log / Insight / SOP / Board recovery

chulbuji.com

Recover execution records from AI Content Assistant, AI Blog, marketing, and auto-trading into the home base. Meaningful processes become Log, Insight, SOP, or Board entries.

June-Only Pilot 1
Maximum 10 videos / under 2 hours/week / stop immediately if conditions violated

Shopping Shorts

June-only pilot experiment. Validating whether a repeatable production routine for product-information-format shorts is achievable. Stop immediately if the home base schedule is displaced or 2 hours/week is exceeded.

June Experiment — Completed 1
Shorts Generation Experiment / AI Collaboration / Local MVP · Complete

Shorts / Reels Generator — Local MVP Phase 1 Complete

Built a local MVP with Claude Code that generates 9:16 vertical MP4s from photos and short clips, reaching v0.1.4. Confirmed BGM, Ken Burns motion, transitions, intro/outro cards, per-scene captions, workout record input, text style controls, card preview, and custom download filename — then judged Phase 1 complete. Phase 2 will separately review drag editing, AI scene analysis, OCR, deployment, and user-saved sessions.

Supporting & Independent 3

Auto-Trading

Strategy validation and operation report asset-building continues. Operated independently, separate from content monetization projects.

Chulbuji Official

Supporting channel for driving traffic to the home base and services. Used as a secondary distribution path for key content.

Supporting channel

Music Channel

Steady accumulation of AI music content outputs. Anonymous operation maintained.

Anonymous operation

June Operating Principles

  • 1
    Write the five harness questions before starting new work

    Define purpose, roles, scope, validation, and record location. Work that starts without criteria is hard to recover as an operating asset.

  • 2
    Mark repeated work as a skill candidate

    If a task may repeat, do not leave it as one-off execution. Record it as a candidate skill that can reduce future work.

  • 3
    One-off work ends as a Log

    Do not over-document everything. For one-time work, leave the judgment and result as a Log entry.

  • 4
    Work repeated more than once becomes an SOP

    Once a repeated pattern appears, promote it to an operating document. Start with a v0.1 standard that helps the next run.

  • 5
    Judge results by eval criteria, not only by feel

    Check whether the predefined criteria were met. If there were no criteria, record the result as pending instead of forcing success or failure.

  • 6
    Reflect progress on the Board once a week

    Do not let project execution scatter. Recover public-safe summaries into the monthly Board.

Operator Growth Track

Manager (chulbuji) is moving from someone learning AI to someone operating with AI.

AI Learning Tools and methods
AI Execution Shipping real outputs
AI Project Operation Structure · record · improve
AI Collaboration Operator — Year 1 ← Current stage

Disclosure Scope

This Board covers only what can be made public about the operating direction.

The following are not disclosed: actual account balances, specific investment performance figures, internal development structure, non-public service URLs, and the names or handles of anonymously operated channels.

Auto-trading content is not intended to guarantee investment returns. It is a record of strategy structure and risk management processes.

Explore the full operating record