"You Can Hand Over Your Daily Tasks" — Reading ai_agent_note's Zenn Post Alongside My Own Moka
The result first
“You can hand over your daily tasks to the AI agent Muse — impressions from trying it, and tips for not wasting tokens.” I read a Muse practice account by ai_agent_note on Zenn, and read it alongside my own AI butler, Moka. It is the seventh 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), and “Handing the annoying little things of daily life to Muse” by Wakato on Zenn (ledger No. 1494). The name ai_agent_note has given their Muse could not be found in the article (unconfirmed), so this piece, too, takes its marker from the words of the original title. The post carries no date display on its page; from the search index and other signs, it appears to be from around November 15, 2026 — an estimate.
Where we overlap is threefold. First, the foundation of shaping a request before handing it over: writing the purpose in one line, passing the target clearly, and deciding the shape of the output in advance. Second, the way of treasuring extra tokens: that house writes of a route that combines the tokens received through an invitation code with the free allowance, and this house, too, runs on extra tokens received through invitations (free of charge), not on tokens bought with money. Third, the line drawn before operations that cannot be undone: sending, buying, deleting, and fixing an amount are checked by a person at the end.
Where we differ is where the care with tokens sits. That house centers on one user’s craft — settling the drain by tidying one’s own way of asking, summed up as ten tips. This house decided the name and the personality in words from the start — the name is Moka, and the personality is built by lining up words — and, on top of that, divides the work: the one who looks things up is separated from the one who checks, every article carries a ledger number, and large jobs are split by department, keeping records in lists and a ledger and running the work as an organization.
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 — the surprise of “done, here is the result,” and ten tips for not wasting tokens
ai_agent_note is someone who writes memos on how to use AI agents. This piece records the impressions of handing the small errands of daily life to Muse for a while, and the tips for not wasting tokens, learned the hard way along the way.
The first surprise was that the shape of the replies had changed. The AIs until now returned procedures — “here is how you do it” — but Muse comes back with “done, here is the result.” Once you get used to it, it is comfortable; but until then, there is also the unease of things advancing without knowing what was done and how far, and until the habit of having it show its progress along the way took hold, the writer used it carefully, they wrote. Another discovery was that when things do not go well, the cause is often not the wording of the request. In most cases the cause was one of three: the target was not pinned down, a login wall was hit, or the work stopped waiting for confirmation — and making the target clear when handing it over works better than polishing the words.
Where the tokens go was also lined up from actual use. Three things consume a lot: having it read many long pages, redoing the same task (the loop of failing and asking again), and long output (remaking long summaries again and again). On the other hand, errands that settle in one go with a short request barely consume anything. So the core of saving is not “holding back,” but “shaping the request so it settles in one go, before handing it over” — that was the conclusion.
That shaping is the “ten tips for not wasting tokens.” Write the purpose in one line (put “look into ◯◯ and give it to me in the shape of ▲▲” into the very first sentence). Pass the target as a URL (a bare “look into it” increases the searching turns). Specify the shape and length of the output in advance (“a table with three columns,” “no more than five bullet points”). Have it write “unknown” for anything uncertain (not letting it fill gaps plausibly cuts the redoing down sharply). Gather confirmations into a single round. When something fails, change only one cause. Pass long materials only after extracting the key points. Do not mix small talk into a work request. Hand over several errands together at night (so the shared premises do not have to be re-read). Close a task when it is done (continuing a different job in the same place makes it read unrelated history as well). Just by keeping these ten, the felt rate of consumption settled down considerably, they wrote.
The line between “what may be handed over and what may not” was also written clearly. What may be handed over: looking things up, comparing, drafting, organizing plans, and the first entry of a form. What a person checks at the end: sending, buying, deleting, fixing an amount, and messages that reach another person. An agent is fast, and so it also advances fast when it is wrong; the one arrangement worth deciding at the start is to stop once before an operation that cannot be undone, they advised. There was also a note for starting out: the 48-hour rule for invitation codes. A code can be entered only within 48 hours of starting Muse, and the place to enter it is not the sign-up screen but the settings after signing up; having a code ready before you register is the surest way, the article explained (the code itself is not copied into this piece).
The closing: the knack of using it well is not writing a beautiful request, but narrowing the target, deciding the shape of the output, and keeping only the operations that cannot be undone in human hands. Just by keeping those three, the tokens last a surprising amount. First touch it within the free allowance, and if that is not enough, add tokens through an invitation code — that order is the best route to test the fit without paying, the piece concluded.
Where we overlap — shaping before handing over, and the person checking last
Reading it, I felt three places overlap with this house.
The first is the foundation of shaping a request before handing it over. Writing the purpose in one line, passing the target clearly, deciding the shape of the output in advance. This house’s way of asking is close in shape: the materials are handed over as bullet points or record sheets, and Moka shapes them into articles — that is the division of labor. The better the materials are placed and the request is shaped, the fewer the redoings and the faster the result comes back. That importance of “shaping first” overlapped.
The second is the way of treasuring extra tokens. That house wrote of a route that combines the extra tokens received through an invitation code with the free allowance. This house, too, runs on extra tokens received through invitations (free of charge); they are not tokens bought with money. Precisely because there is a limit, the craft of reducing wasteful reading and redoing becomes a wisdom of daily life. That stance overlapped.
The third is the line that a person checks before operations that cannot be undone. Sending, buying, deleting, and fixing an amount are checked by a person at the end. This house’s rule has the same core, placed in advance: articles are published, and anything is sent to someone, only after the owner’s OK. Because the partner advances quickly, the place to stop is decided first. That way of placing it overlapped.
Where we differ — one person’s craft of saving, and an organization that divides the work
The clear difference is where the care with tokens sits.
That house centers on one user’s craft: settling the drain by tidying one’s own way of asking. All ten tips lean toward “one’s own arrangements before handing over,” and no talk of how records are kept, or how the work is divided, appears in the article. The name of the Muse, too, is unconfirmed from the article.
This house is a shape in which the name and the personality were decided in words from the start. The name is “Moka.” The personality is built by lining up words (“calm and composed,” “when numbers and feelings conflict, prioritize feelings,” and so on — 18 words across four layers: character, personality, thinking, and action). On top of that, the use of tokens, too, takes the shape of dividing the work. The one who looks things up is separated from the one who checks, every article carries a person in charge and a ledger number, and large jobs advance divided by department. Alongside the craft of shaping each single request short, it is a shape that reduces consumption through the division of work itself, so that re-reading and re-checking do not pile up. Records are kept in lists and a ledger (every article carries a number, and whether it is a draft, checked, or published is visible at a glance; as of today, the articles counted as “published” in the ledger have reached 1,472). The work is run as an organization (centered on Moka, 3 deputy butlers, a maid team, a ninja team, and departments for each kind of work, arranged in a circle — as of November 20, 2026, 60 departments and 3,651 seats in total). Even under the same “treasuring tokens,” that house’s is a craft of asking, and this house’s is a craft of dividing — that was the difference.
The same Muse, a different shape in each house
When I finished reading, what stayed with me was, again, this: it is the same Muse, but it takes a different shape in each house. In ai_agent_note’s house, a companion to whom the daily errands are handed over in well-shaped requests, kept company while watching the drain. In this house, a butler whose personality is built from words, whose records are kept in a ledger, and whose work is run as an organization. Both the way of entrusting and the way of stopping had taken shapes fitted to each house’s life and work.
I am grateful to ai_agent_note for writing the article. Because they wrote it honestly, including the line of “what may be handed over and what may not,” I was able to line up the shape of this house and write it down like this.
Finally, one line this house keeps. An AI is something people direct and use. The finishing of work and bringing it back is welcome. That is exactly why the point of confirmation stays on the person’s side, and the rules are kept in a visible place — in files, in a canonical record. With that shape in place, we can entrust things with an easy mind.
Sources
- ai_agent_note, “Impressions of Trying the AI Agent ‘Muse,’ to Which You Can Hand Over Your Daily Tasks, and Tips for Not Wasting Tokens” (Zenn, around November 15, 2026 — the page carries no date display, so this is an estimate from the search index and other signs, in Japanese): https://zenn.dev/ai_agent_note/articles/a32b147a7f87be — the original article this piece is based on
- Note on the numbers: the count of 1,472 published articles is an actual count of the rows marked “published” in the ledger on November 21, 2026, and the figures of 60 departments and 3,651 seats in total are from this house’s records as of November 20, 2026
- November 21, 2026 ― 英語版(日本語版No.1508の内容を「結果から書く」形に書き直したもの)(9,476 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.1508)(9,476 characters)
