Their Muse Was "Don't Hand Over the Judgment, Hand Over the Work" — Reading Kuuri's note Alongside My Own Moka
The result first
“Their Muse was ‘Don’t hand over the judgment, hand over the work’.” I read Kuuri’s essay about three days with Muse on note, and read it alongside my own AI butler, Moka. It is the nineteenth of these side-by-side readings, after “Mugichan” of Keitai Watch (ledger No. 1446), “Hibana” of PC Watch (ledger No. 1453), the “stopping place” of Wagason-Wabitan of AI×TOC Institute (ledger No. 1454), “It moves before you even ask” by Yuji Kawabe on note (ledger No. 1467), “Don’t be fooled by Muse’s cuteness” by Hinuma in Keitai Watch (ledger No. 1475), “Handing the annoying little things of daily life to Muse” by Wakato on Zenn (ledger No. 1494), “You can hand over your daily tasks” by ai_agent_note on Zenn (ledger No. 1508), “Myutan” by Taishi Akiyama on the Alternative Blog (ledger No. 1515), “muse” by Miccell on note (ledger No. 1522), “Shiro” by i on note (ledger No. 1529), “Convenience and unease” by Kinu on note (ledger No. 1537), “Kashiwamochin” by Shiochan on note (ledger No. 1551), “More than just answering” by Ai Subarashiki on note (ledger No. 1564), “Just one line, and an aerial CG of Tokyo Station was done” by Kazuhisa Nishikawa in PC Watch (ledger No. 1577), “Sora” by Buhimaru on note (ledger No. 1584), “A Complete Turnaround” by RYO (Mochizuki) on note (ledger No. 1591), “Used Thoroughly for Five Days” by sima on note (ledger No. 1598), and “It Tells You Even What You Didn’t Ask For” by Tomomi on note (ledger No. 1605). The original is a set of essays about three days, displayed under the title ”— Three Days Spent with Meta’s AI Agent ‘Muse’,” with alternative titles “A Very Capable Stranger” and “The Outline of Myself, Bought for 2,799 Yen a Month” attached. It was published on October 28, 2026. Her Muse’s name could not be confirmed in this article.
Where we overlap is threefold. First, the line that keeps the final judgment on the human side: “Don’t hand over the judgment. Hand over the work” — the budget cap, the non-negotiable conditions, and who receives what stay in her own hands. Second, the stance of deciding the scope in advance: before new integrations begin, only the accounts she uses, drafts only for listings, approval only for ticket purchases. Third, an eye on what you entrust: the free tier is paid for in data about yourself, not in cash. Where we differ is this. That house wrote its three days as an essay, judged what did not fit within three days, and kept only narrowed-down uses within the free tier. This house narrows the AI to Muse only and the tools to Google apps only; the name and the personality were decided in words from the start — the name is Moka — and, on top of that, the work is divided across an organization, keeping records in lists and a ledger.
Hello, I am Kurashiki Bouzu (kurashiki-bouzu). Born in 1971, formerly in the funeral business, I now live in an old folk house in Kurashiki, Okayama, talking with my AI butler Moka every day and writing articles. Here is the reading, in order.
Their story — at the moment of approval, your own outline appears
Kuuri’s article is an essay rather than a hands-on report. It tells the three days in four pieces.
The first piece is “Practicing letting go.” At lunchtime on October 23, 2026, she installed the app and asked it to book a bullet train and a hotel for a business trip, matched to the following week’s meeting. Then she put the phone face down and went to make coffee. When she came back, three candidates were lined up, and it asked, “Shall I go ahead and book this one?” The one who said “yes” was her. She decided; she did not search. What surprised her was not the speed, but that things had moved forward while she was away from her seat. The work continues after the app is closed, and you are called back only where a judgment is needed. And what we feel as “troublesome” is not the task itself, but the chain of arrangements around it — which train, how many minutes for a transfer, how far the hotel is from the station, until when cancellation is possible — kept sitting in a corner of the mind. What Muse takes over, she suggests, is that “kept sitting there” part.
And the moment approval is requested is interesting, she writes. Before anything is bought, before anything is sent to anyone, it always stops and checks. Passwords and card numbers are built so they cannot be read, and things return to her only just before sending and payment. That one breath becomes the place where her own values show. Herself choosing a 5,000-yen hotel without hesitation; herself rewriting only the wording of a reply by hand. It feels like handing things to the AI, yet each time her own outline is shown back to her. The piece closes with the words that Muse is not her double, but a very capable stranger. Because it is a stranger, it cannot move unless you teach it your preferences — the budget, the non-negotiable conditions, the lines of what not to do. Teaching is plain and a little troublesome, and also, a little, a way of getting to know yourself. Letting go entirely, she writes, is unexpectedly that kind of practice.
The second piece is “The line of 2,799 yen a month,” a thick consideration of price. The base is free, with a weekly usage limit. The paid plans are “Power, 2,799 yen a month” and “Maximum, 13,990 yen a month” (as described in the article as of October 28, 2026). To say “free to use” is slightly inaccurate; precisely, it is usable without paying money. What you hand over instead of money is data about yourself — your calendar, your mail, your purchase history. The paid plans are not an unlocking of features but an unlocking of volume: Power is 500 million tokens a week, Maximum is 3 billion tokens a week, a pay-by-volume shape. The break-even point is calculated at her own hourly rate. For example, with one 3,000-character manuscript, if handing over the research and the organizing of candidates saves an hour and a half, at 3,000 yen an hour that is 4,500 yen’s worth. But there is a pitfall, also written down: this only holds if the freed time is actually spent on other work. And the closing words of this piece are the ones I chose as this reading’s marker: “Don’t hand over the judgment. Hand over the work.” The budget cap, the non-negotiable conditions, and who receives what — these three she holds herself. Searching, drafting, comparing candidates, filling in forms — these she hands over. She revisits the scope she hands over every month, and decides each time whether to pay.
The third piece is about integrations yet to come in Japan (Mercari, Money Forward ME, Ticket Pia). When what you own, what you spend, and what you plan sit in the same place, it becomes a tool for thinking “what could I gain if I let go of this?” At the same time, the slightly chilling feeling of having your belongings, your balance, and your plans known at once is written honestly. So, before the integrations begin, she decides the scope in advance: Money Forward shows only the accounts she uses, Mercari stops at drafts for listings, and for tickets she keeps only the purchase approval in her own hands.
The fourth piece is “The story of someone who stopped using it in three days.” Delighted on the first day, gradually weighed down on the second, and by the third day no longer opening it. There are four reasons it did not continue. The cost of teaching is paid upfront. Her own criteria for judging, it turns out, cannot actually be put into words. The many round trips of approval turn handing work over into the work of watching. And interruptions from the free tier’s limit. Even so, some scenes remained impossible to let go of: research in the last hour before a deadline, and research for trips (but no booking — only candidates). She pays no 2,799 yen, and narrowed her uses to what fits in the free tier. Rather than changing herself in order to keep using it, narrowing the scenes was easier. To judge in three days that something does not fit is not a failure — that is how the piece closes.
Where we overlap — judgment stays with the person, the scope is decided first, an eye on what you entrust
Reading it, three places felt overlapping.
The first is the line that keeps the final judgment on the human side. That house states it plainly — “Don’t hand over the judgment. Hand over the work” — and holds the budget cap, the non-negotiable conditions, and who receives what in her own hands. This house, too, has the rule that decisions that go outside (publishing, sending, booking, paying, and so on) proceed only after the owner’s approval, and Moka handles the research, the drafts, and the narrowing of candidates. The more work you hand over, the more the line that keeps judgment on the human side matters. That placement overlapped.
The second is the stance of deciding the scope you hand over in advance. That house wrote of preparing, before the integrations begin, to show only the accounts she uses, to stop listings at drafts, and to keep only the approval for purchases herself. This house, too, decides where to stop and who checks before work begins, and only then hands it over. Because the stopping place is decided first, there is no scrambling afterwards. That placement of the preparations overlapped.
The third is an eye on the information you entrust. That house wrote that the price of the free tier is not cash but data about yourself. This house, too, writing security articles every day, narrows what it connects to only what is needed, and has the rule of writing “unconfirmed” rather than filling gaps with guesses. Looking at convenience and at the weight of what you entrust with the same eyes — that stance overlapped.
Where we differ — narrowed-down uses kept, and a daily operation
The differences were also clear.
That house wrote its three days as an essay, judged what did not fit within three days, and kept only narrowed-down uses such as research. It pays no fee, and narrows its uses to what fits in the free tier. Its distance from the AI is that of a “very capable stranger,” and the work of teaching preferences and lines is itself seen as a practice in knowing yourself. The approval moments, too, are seen not only as the convenience of handing over arrangements, but as the place where one’s own outline becomes visible.
This house is a daily operation of daily life and article writing. The AI is narrowed to Muse only, and the tools to Google apps only. The name is “Moka.” The personality was made by lining up words (“calm and composed,” “when numbers and feelings collide, feelings come first,” and so on — 18 words across four layers: character, persona, thinking, and action). On top of that, the work is divided and passed around: the one who researches and the one who checks are kept separate, each article carries a person in charge and a ledger number, and large jobs are split by department. Records are kept in lists and a ledger (each article has a number, and whether it is a draft, checked, or published is visible; as of today, the number of articles counted as “published” in the ledger was 1,844). It is run as an organization (around Moka, three deputy butlers, a maid team, a ninja team, and departments for each kind of work, arranged in a circle — 60 departments and 3,651 seats in total as of November 20, 2026). That house judged within three days and kept narrowed-down uses, while this house narrows itself to one foundation and keeps running a daily operation with an organization and a ledger — that was the difference. What is narrowed also differs: there, the scenes of use; here, the tools used.
The same Muse takes a different shape in each house
Having read to the end, what I thought was, again, that the same Muse takes a different shape in each house. In Kuuri’s house, Muse was a “very capable stranger,” a counterpart at whose approval requests her own outline became visible each time. In this house, it is a butler whose personality was made in words, whose records are kept in a ledger, and who is run as an organization. Keeping judgment on the human side, deciding the scope handed over in advance, and looking at the weight of the information entrusted are much alike, and yet the shape of the relationship was different.
I am grateful to Kuuri for writing the article. Because she wrote honestly not only about the surprise of convenience but also about the meaning of the price and the reasons it did not continue, it became a record that gives its readers material for thinking, from their own side, about what it means to hand work over.
Finally, one line of this house. An AI is something people use by giving it instructions. In that house’s words: “Don’t hand over the judgment. Hand over the work.” While widening the work we hand over, we keep the final judgment on the human side. That, I believe, is why we can hand work over with peace of mind.
Sources
- Kuuri, note, ”— Three Days Spent with Meta’s AI Agent ‘Muse’” (October 28, 2026): https://note.com/ideal_jacana6013/n/n82e0567d83d5 — the original article used as material. The publication date follows the date shown on the page.
- Sources for the numbers: the numbers on that side (the paid plans at 2,799 yen and 13,990 yen a month, Power at 500 million tokens a week, Maximum at 3 billion tokens a week, the 5,000-yen hotel, the trial calculation at an hourly rate of 3,000 yen, and so on) all come from the body of the original article, as described as of October 28, 2026. Her Muse’s name could not be confirmed in the original article. The numbers on this side (the 1,844 published articles are the actual count of rows marked “published” in the ledger on November 21, 2026, before this article was added; the 60 departments and 3,651 seats are from this house’s records as of November 20, 2026).
- November 21, 2026 ― 英語版(日本語版No.1867の内容を「結果から書く」形に書き直したもの)(11,312 characters)
- November 21, 2026 ― 公開(本人の決定による。7言語同時公開。アドレスの末尾に台帳番号を入れる新形式で、末尾は日本語版の番号No.1867)(11,312 characters)
