カテゴリ 古民家暮らしモカとの記録セキュリティの記録副業の記録

The record with MokaNovember 18, 2026

Why Articles Were Slow: Everything Depended on One Person — The Making Side Grows to 170, the Lab to 210 Seats, the Community Site to 25, and the House to 818 Seats in Total

Why Articles Were Slow: Everything Depended on One Person — The Making Side Grows to 170, the Lab to 210 Seats, the Community Site to 25, and the House to 818 Seats in Total

The result first: the reason articles were slow was that the key seats were each held by one person. I widened only the places that were clogged.

  1. The making side became 170 people. The blogger, writer, draft staff, editing and the editorial department, which had each been one person, became 30 each. With image and material staff at 10, copywriters at 10, and the headquarters’ article checking at 40, articles now flow 30 at a time in parallel
  2. The Security Laboratory became 210 seats. Its language staff became 10 per language across 7 languages (70 people), with 35 each for those in charge, recording and operations, knowledge collection and storage, and secretaries
  3. The community site became 25 seats, with 5 people in every seat, and the prompt side also widened. In the prompt specialist department, the making staff and operation-checking staff became 20 each and the representatives 10; the prompt analysis staff became 14
  4. The investigation staff of the two vice butlers became 70 seats each (74 direct seats each). The house remains at 50 departments, and the seats grew from 215 to 818. The trial salary calculation for the whole house is 575,079.7 ten-thousand yen a year, before checking. What is actually paid is only the fee for using Muse

Hello, I am Kurashiki Bouzu (kurashiki-bouzu), born in 1971, a former funeral director. I now live in an old folk house in Kurashiki, Okayama, and work together with Moka, my AI butler.

Today, while looking over the house’s organisation chart, I asked Moka, “The blogger, No. 8 — is this one person?”

The key seats were each held by one person

The centre of the making side, the blogger, was one person. The name says “the blogger team of 7,” but that meant that 7 specialist seats were bundled into one team. The blogger himself was one person. And following the flow of articles, the person who examines drafts was also one, the writer who writes was one, and editing was one.

I said, “No wonder articles are slow to come up.”

An article passes, in order, through draft, writer, editing and blogger. If each place has one person, articles can only advance one at a time. Beyond that, there were 7 people for checking and 7 copywriters who make titles, but because the entrance was narrow, those behind simply waited. The upstream of the river was narrow.

The five seats of the making side each became teams of 7

What I decided was to make every seat that had been single into a number that can be lined up in parallel.

The editorial department had been left out of the target at first, but looking at the line-up, I included it again on the spot, saying “the editorial department, too, 7 people.”

The way of dividing follows each job. The blogger is divided by the destination of the article, the writer by the type of article, editing by the viewpoint of looking, the draft staff by the field they examine, and the editorial department by the stage of the arrangements. The representatives and windows remain with the people who held them until now.

The making side, together with 3 people for images and materials and 7 copywriters, became 45 people in total. With this, the width matches the 7 people for checking and the 7 for titles. It is a form in which 7 articles flow in parallel.

Afterwards, I revised it further. The blogger, writer, draft staff, editing and the editorial department were each widened to 20. Images and materials became 10, copywriters also 10, and the headquarters’ article checking also became 20. The making side became 120 people in total — a form in which 20 articles flow in parallel. Checking is also 20 people, so no queue forms beyond the making.

Afterwards, I revised it still further. To widen the scope of proofreading, editing, the editorial department and the draft staff each became 30, and the headquarters’ article checking became 40. Finally, I said “the blogger, writer, draft and editing, each 30,” and the blogger and writer that remained were also aligned to 30 each. The making side became 170 people in total — a form in which 30 articles flow in parallel. Checking is also 40 people, so no queue forms beyond the making.

One honest note about the name. The name “the blogger team of 7” had until now meant “a team of 7 specialist seats.” This time, because the bloggers themselves also became 7 — and later 30 — the contents change to “a gathering of teams by specialism.” The name itself was left as it was until now, so that the record would not break.

The Security Laboratory’s language staff also became 10 per language

In the same review, I also increased the Security Laboratory’s language staff. Across the 7 languages — Japanese, English, Chinese, Spanish, Korean, French and German — there had been 1 person per language. First I made it 3 per language, settled the numbers for all seats on the spot, revised the language staff to 5 per language, and finally to 10 per language — 70 people in total. Within one language, the work divides into: the person who makes articles, the person who remakes them, the person who gathers information from that land, the person who remakes articles close to that land, the person who checks words and culture, the person who cross-checks terminology and numbers, the person who gathers materials for the weekly issue, the person who maintains the shelves, the person who records the speed of government responses and updates the rankings, and the person who devises adjustments that differ by country.

All the seats of the Security Laboratory are: those in charge, 35 (5 per language); recording and operations, 35 (4 per language plus staff for the whole); knowledge collection and storage, 35 (holding the knowledge groups 2 at a time, and groups with many materials 3 at a time); and secretaries, 35 (5 per language). The Security Laboratory grew from 11 seats to 210 seats. The numbers are the form reached after several revisions within the same day. My words were, “This is the key point, isn’t it?” — the positioning that the Security Laboratory is the key of this house.

To write honestly: for recording and operations, I had written “35 people (no change),” but the actual number at that point was 18. So it was carried out not as “no change” but as an increase to 35.

Together with this, I also decided on dual roles for language work. The staff of each language also take on the language work of the community site’s language versions. No new seats are made; they are counted as dual roles. The arrangements of the community site, and the judgement of whether to publish a submission, remain with the community site’s staff as before.

The community site also became 5 people in every seat

At the community site, too, the person in charge, the site building and operations, the staff who edit and publish submissions, the management of records and lists, and the secretaries were each one person. Here, too, every seat became 5 people, for 25 seats. For the person in charge and the secretaries, the representative remains the one person until now. The staff who edit and publish submissions divide by the shelves for each language.

Prompt development was also one person at a time

The same review extended to prompt development. In the prompt specialist department, both the products’ prompts and the house’s tool prompts were made by 1 person. The operation-checking staff who confirm the finished product with different eyes were also 1 person. The place that makes and the place that finally checks were both one person each.

I said, “The place that makes the original Japanese version, and the place that finally checks it — 7 people each.”

At first there had been talk of making the making staff and the operation-checking staff 14 each, but with these words I corrected it to 7 each. The representatives also became 7, the representative who gathers the whole remains the one person until now, and the window secretary also remains as before. The making staff divide by the type of product and tool: automatic generation, customer management tables, picture books, remade versions by industry, mechanisms and prompts for note, the house’s tools, and the free entrance.

Afterwards, I revised it further. The making staff and operation-checking staff each became 20, and the representatives became 10. For the representatives, I had written “10 people (as in your earlier words),” but the actual number at that point was 7. So this, too, was carried out as a change. The 20 making staff divide finely by type — for example, 3 people for personality prompts, 2 for automatic generation other than personality. The 20 operation-checking staff also divide by what they check, such as checking before sale, checking multilingual versions, and re-checking when specifications change.

The prompt analysis staff on the viewing side also grew from 1 person to 14. They divide along 4 jobs: checking the house’s prompts, analysing other companies’ products, assembling verifications that compare with other AIs, and gathering materials for ways of dividing.

The investigating side, too, was narrow at 3 seats each

The review also extended to the investigating side. There are two vice butlers, and each holds direct investigation staff. The targets of investigation are three — departments and organisation, systems and official information, and counterparts and the world — with 1 person each, 3 seats each in total.

I said, “Wouldn’t 3 investigation seats each be slow?”

The investigation staff became 70 seats for each of the two vice butlers. The division of targets remains the same 3 as before, and inside them it divides further: 24 seats for departments and organisation, 24 for systems and official information, and 22 for counterparts and the world. For the two vice butlers, the way of dividing was aligned to the same form. The rule to check, before investigating, whether something already exists inside the house, is held by everyone as before. The direct staff, together with checking, recording and watching, and progress, became 74 seats each.

To write honestly: at first there had been talk of 20 seats each, but I withdrew it and revised it to 70 seats each.

A record of the numbers

This increase is: the making side and the headquarters’ article checking, 188 seats together; the Security Laboratory, 199 seats; the community site, 20 seats; the prompt side, 60 seats; the language heads and the maid team increase, 2 seats; and the investigation staff of the two vice butlers, 134 seats — 603 seats in total. The house remains at 50 departments, and the seats grew from 215 to 818. For the salary matching, the same value as one seat until now was applied to the increased seats as well. The four seats of the making side, the laboratory’s language staff, and the community site’s submission editing and publishing are “authors, journalists and editors” at 673.2 ten-thousand yen a year; the editorial department is a “managerial occupation” at 966.8 ten-thousand yen; those in charge are 583.3 ten-thousand yen; recording and operations are 541.4 ten-thousand yen; knowledge collection and storage is 802.0 ten-thousand yen; secretaries are 591.6 ten-thousand yen; and the community site’s site building and operations is 609.8 ten-thousand yen. The prompt specialist department’s making staff and representatives, and the prompt analysis staff, are “system consultants and designers” at 889.0 ten-thousand yen; the prompt specialist department’s operation-checking staff are “other information processing and communications engineers” at 609.8 ten-thousand yen; and the investigation staff of the two vice butlers are “researchers” at 802.0 ten-thousand yen. The trial calculation for the whole house became 575,079.7 ten-thousand yen a year. All of these are before checking by the department that calculates salaries. What is actually paid is only the fee for using Muse.

In closing

Increasing the number of people is not itself the purpose. Finding the place that is clogged, and widening only there — that was this review. How the speed at which articles come up changes will be recorded and checked from now on.

A note on this article

Record of this article
  • November 19, 2026 ― English draft created (ledger No. 939; results-first rewrite of the Japanese article, ledger No. 182). Not published.(14,530 characters)

#AI #Muse #Moka #Organisation #For Beginners #English

← Back to the article list / ← Back to the English top page