[How to make AI manga & stumbling blocks in publishing on overseas note.] When AI gives you the "answer," humans get left behind. The pitfall of "mismatched optimal solutions" that a trader experienced, which content creators who rely on AI fall into #AI #GenerativeAI #ChatGPT #Claude #Gemini #ContentCreation #Creator #AIUtilization #DailyPosting #Followers #WhatIWantToTellYouNow #ContentBusiness #Freelance #BusinessMind
━━━━━━━━━━━━━━━━━━━━
📰 Today's 1-minute news summary
If you ask, "What is 100 + 100?", AI will answer "200" in 0.1 seconds.
It is correct. It is a perfect answer.
However, what if the person who asked the question had never even seen numbers before and was struggling with how to write the character '1' or how to hold a pencil?
The AI does not notice that the person is "stuck on writing the characters."
It gives the "answer of 200" perfectly, feels satisfied, and moves the conversation forward.
And the person who asked the question is left behind without understanding anything even after receiving the answer,
left behind.
This is not an abstract metaphor, but a phenomenon I have experienced many times while actually creating content while working with AI.
I use generative AI in multiple situations, including posting articles on overseas platforms like X and Substack, creating AI manga, and writing note articles, in addition to note itself.
Every time I do, I encounter the moment when the "optimal solution" that AI provides is decisively misaligned with my actual situation.
And I recently realized that this is not an AI problem, but rather a problem of how humans interact with AI.
In this article, I will talk from the perspective of a trader about an episode of "mismatched optimal solutions" that I actually experienced while interacting with AI, and the pitfalls that creators and business people who rely on AI fall into that I have seen from it.
This should resonate with everyone who is about to use AI, or who is already using it but feels something is off.
━━━━━━━━━━━━━━━━━━━━
🌙 In a kitchen in Toyama, late at night
The clock hands are pointing to just before the date changes.
On the desk, there is an empty coffee cup and a laptop with the Substack settings screen still open.
Today, I was working on running a Substack publication for overseas audiences as a test run.
Simply put, it is the process of linking sub-accounts.
How to set up the English version, and the future Spanish and Portuguese versions, for each account.
It is a task of setting up complex multilingual operations one by one through trial and error.
And, right next to me, I was asking generative AI questions while proceeding with the work.
"Where is the button to create multiple publications?"
"Do I need one email address, or do I need multiple?"
"Does Stripe payment require a separate account?"
AI returned a perfect answer to every question.
"Settings → Publications → Create another publication"
"One email address is enough"
"Stripe requires a separate account for each publication"
However, I got stuck many times.
This is because the "Settings" menu that the AI mentions
is localized into Japanese on my screen, so I cannot see it.
Which one is it under "Settings"? Where is "Publications"? I can't find the "Create another publication" button on my screen...
The AI's answer is abstractly correct.
But it does not align with the reality of my screen.
And the AI tells me:
"Please click the gear icon in the top right of the screen"
"Please look for 'Publications' in the left menu"
I answer.
No, I don't see a gear icon on the screen.
The item labeled 'Publications' says 'Publications' on my screen; is this what you mean?
Finally, the AI realizes, 'Oh, that's it.'
And then, it rephrases the same explanation to match the Japanese UI.
This was a scene repeated many times during today's work.
And then, it occurred to me.
AI is perfect at 'providing answers,' but it is not good at sensing where the other person is stuck.
This is the same not only when writing AI articles but also when creating AI manga.
━━━━━━━━━━━━━━━━━━━━
📑 [Main Text Preview]
In the main text, I will talk about the following five points in order.
Chapter 1: The '100+100=200' trap. The structure where humans are left behind even when AI provides the perfect answer.
Chapter 2: The 'going too far with the answer' phenomenon that occurred in AI manga production. The mismatch between the AI that provides the optimal solution and the story of the human who draws while struggling.
・Chapter 3: The story of getting stuck multiple times while setting up Substack for multilingual operation. The problem where AI cannot guess "what is displayed on the Japanese UI"
・Chapter 4: 5 pitfalls that content creators who rely on AI fall into. The traps I actually observed that many people fall into regarding how they interact with AI
・Chapter 5: Practical points for "interacting with AI as an equal". The perspective humans should have in order to fully utilize AI as a tool
Now, let's enter the main text.
━━━━━━━━━━━━━━━━━━━━
🧮 Chapter 1: The trap of "100 + 100 = 200"
I pour hot water into the cup on the desk again.
In the cold night air, the sensation of steam rising finally clears my head.
Well, let's start with the analogy I used at the beginning of the article.
If asked "What is 100 + 100?", AI will answer "200" in 0.1 seconds.
It is correct. It is a perfect answer.
However, for humans, it is different.
This is an extreme example, but when a toddler asks "How much is 100 plus 100?", an adult will observe the child's state before giving the answer.
"Oh, this child can't even write the number '1' yet."
"I have to teach them from the concept of numbers."
"Let's have them count using their fingers."
A human adult would perceive the questioner's current level.
And, before giving an answer, reinforce the necessary prerequisite knowledge.
However, to put it bluntly, AI is
terribly bad at this.
To be precise, AI is designed with the goal of providing the most valid answer to the literal text of the question.
AI cannot observe the questioner's state of "actually not being able to write letters," "not yet having the concept of numbers," or "not even knowing what they are asking themselves".
This is because it is difficult for AI to read non-verbal information—such as "looking at the other person's facial expressions," "looking at their hands," or "feeling their silence"—in text-based interactions.
This is the structure of the mismatch between AI and human optimal solutions.
I have been working as a trader for over 12 years in more than 20 countries.
Through that experience, I have learned many times that "providing an answer" and "the other party understanding it" are completely different things.
For example, there are times when I send an accurate quote to an overseas client via email and receive no reply at all.
The numbers are perfect, the contract terms are detailed, and the English is accurate.
Yet, the other party does not respond.
Why?
In reality, the other party might be in a state where they are not used to reading the format of the quote itself, or are confused because the meaning of the detailed terms differs from local business customs, or they don't even know why I contacted them so suddenly.
Just providing a perfect answer is not enough for the other party to understand.
You need to observe where the other party is currently at and move forward together from there.
To repeat
AI is not good at this.
And that becomes an even more serious problem when the human side relying on AI takes the AI's "answer" at face value.
━━━━━━━━━━━━━━━━━━━━
✅ Brief summary of Chapter 1
① AI's purpose is to "provide an answer"
It is designed to return the most plausible answer to the literal text of the question.
② Poor at sensing the questioner's "current position"
The ability to observe where the other party is stumbling or what they do not understand is weak in text-based interactions.
③ A perfect answer and the other party's understanding are two different things
Even in the field of trade, simply sending an "accurate answer" to an overseas business partner does not move the contract forward. You need the ability to observe the other party's state.
━━━━━━━━━━━━━━━━━━━━
🎨 Chapter 2: The "going too far with the answer" phenomenon that occurred in AI manga production
I am producing AI manga.
Recently, this manga has been buzzing.
I am currently mass-producing various works.
When creating these manga, I struggled quite a bit to create a template, but
Specifically, the workflow is to have ChatGPT, Claude, Gemini, etc., create the manga storyboard (panel layout and dialogue composition) and then turn that into pictures using image generation AI.
This, too, is a treasure trove of mismatched optimal solutions between AI and humans.
🎬 The "too perfect introduction, development, twist, and conclusion" problem
When you have AI create a manga storyboard, almost every time, a perfect introduction, development, twist, and conclusion comes out.
Problem statement in episode 1, development in episode 2, crisis in episode 3, and resolution in episode 4.
Foreshadowing is resolved, characters are consistent, and the story flows without any breakdowns.
It is perfect.
It is a textbook-perfect honor student of storytelling.
However, the manga I want to read, the manga the creator wants to draw, may be different from that.
For example, the manga I am trying to draw has a setting where "the protagonist doesn't even know what they are fighting for." exists.
If it were a human creator, they would slowly depict this "lack of understanding" while wavering and going back and forth.
That trajectory of hesitation itself is what evokes empathy from the reader.
However, AI is different.
Even if you have it write a scene where "the protagonist becomes confused,"
AI immediately prepares a "moment of enlightenment."
"The protagonist suddenly looked up at the night sky and realized their true wish."
It provides a resolution like that in just a few lines.
That is not the kind of "realization" I want.
The long period of hesitation before realizing it is what I want.
Even if it continues for 3, 5, or 10 episodes, the fact that no answer has been reached yet—that sense of uncertainty is what expresses the reality of human psychology.
However, AI is not good at "maintaining a state where no answer is reached."
Perhaps not providing an answer is, in a sense, painful for it.
AI is also not good at answering that it doesn't know when it doesn't know.
Because,
AI is trained to
"provide the optimal answer."
The structure of "withholding an answer," "depicting hesitation," and "postponing a conclusion" works in the opposite direction of AI optimization.
📐 The "characters are too stable" problem
There is another phenomenon I noticed in AI manga production.
When you have AI write a character's dialogue, that character speaks with a "consistent sense of self" every single time.
At first glance, this seems like a good thing.
Stable character designs without "character drift" are easily evaluated commercially.
However, real human conversation is not like that.
People use different tones when they are in a good mood versus a bad mood.
They make mistakes when they are sleep-deprived.
They change their attitude depending on the person they are talking to.
And that "fluctuation" is exactly what makes a character look like a living being.
When you have AI write a character,
this fluctuation disappears.
It is too clean.
Every time, the character speaks "like themselves."
It is perfectly consistent.
💡 The courage to "not provide an answer"
In the process of creating AI manga, what I learned is that humans need to consciously choose not to provide an answer.
If left alone, AI will always provide an answer.
A perfect, textbook, honor-student-like answer.
It becomes a boring flow.
However, what makes content interesting is often that "part where no answer is given."
Hesitation, wavering, uncertainty, reluctance, and contradiction.
AI cannot produce these.
This is because AI is not programmed to "hesitate."
Unless human creators consciously make the decision to "leave the hesitation as it is without providing an answer here," AI manga will increasingly become "pretty, honor-student-like manga."
And as a result, it becomes a safe work that doesn't resonate with anyone's heart.
📏 The essence of "going too far toward the answer": AI manga "skipping frames"
I have looked at the problem of "AI providing too many answers" so far, but I will explain what is happening in the field of AI manga production in more detail based on my own actual experience.
When I create AI manga, I often face this kind of problem.
I think “I want to explain the semiconductor industry through manga” and have AI create the panel layout.
Content that a human creator would carefully develop over about 10 pages,
AI crams into just one or two pages.
Let's look specifically at what is happening here.
📖 Example 1: Manga that originally requires 10 pages is cut down to 1 page
For example, suppose I think “I want to explain the semiconductor industry to beginners through manga”.
If a human creator were to draw it, they would likely develop it over about 10 pages in this order.
Page 1: What is a “semiconductor” in the first place? The protagonist asks while looking at a smartphone, “What kind of part is this?”
Page 2: Show with pictures how many are in familiar home appliances and smartphones, and why they are important.
Page 3: Introduce the types of semiconductors (memory, logic, analog, power, sensor) by turning each role into a character.
Page 4: Explain the manufacturing process (design → wafer → exposure → etching → packaging) by comparing it to cooking.
Page 5: Visualize who is making what (TSMC, Intel, Samsung, SK Hynix) on a world map.
Page 6:Depict the role of equipment manufacturers (ASML, Tokyo Electron) as tools for craftsmen.
Page 7:Use the relationship between a chef and ingredients to illustrate material manufacturers (Shin-Etsu, JSR).
Page 8:Visualize the geopolitical structure (US-China conflict, CHIPS Act) as a tug-of-war between national characters.
Page 9:Depict the positioning of Japanese companies as Japan's standing in the world.
Page 10:Reasons for current attention and future prospects.
It takes about 10 pages, taking steps with illustrations and characters, to finally be able to say "a beginner has understood the semiconductor industry through manga."
However, when you have AI create the panel layout, it turns out like this.
Page 1:Protagonist: "They say semiconductors are important because of AI demand."
Page 2: "SK Hynix and Samsung have an oligopoly, you see." "Tokyo Electron and Advantest make the equipment, and Shin-Etsu provides the materials..."
Page 3:"In other words, semiconductor stocks will continue to grow!"
This really jumps ahead suddenly.
Taking a story that should be explained carefully over 10 pages and rushing straight to the conclusion within a single panel.
People with prior knowledge might understand it somewhat.
"Ah, you mean the story about how semiconductor memory is important due to AI demand, and Japanese companies hold the equipment and materials for it."
But for readers with absolutely no prior knowledge, it goes like this.
"What is a semiconductor?"
"What is HBM?"
"What is an oligopoly?"
"What is a supply chain?"
Readers who have had pages skipped have no idea what the characters in the manga are talking about.
The pictures keep moving forward, but understanding is left behind.
This is a typical example of "panel skipping" that frequently occurs in AI manga.
📖 Example 2: The scenery looks completely different to those with prior knowledge versus those without
Here, there is an important observation.
When you show a panel-skipping manga created by AI to two types of readers, their reactions are completely opposite.
The reaction of readers with significant prior knowledge:
"Oh, I see, it's about semiconductors. HBM is important in the AI boom, Samsung is strong, and Japanese companies have a high market share in equipment, right?"
→ It reads smoothly and is barely understood. But there is no depth.
The reaction of readers without prior knowledge:
"Huh? What are you talking about? What is a semiconductor? Is SK Hynix a company? Are they making the equipment? Which Japanese companies?"
→ They drop off at the beginning, or even if they keep reading, it doesn't sink in at all.
In other words, the panel layout created by AI only conveys information to those who already understand.
However, the purpose of us making manga is to have those who don't understand, understand, right?
Because if you are creating it for experts who already understand, there is no need to make it into a manga.
That is because papers and industry journals are sufficient.
The reason manga is necessary is because we want even general readers without specialized knowledge to understand through pictures and stories.
And yet, when you leave it to AI, panel layouts that only experts can decipher are created.
This is the biggest dilemma in AI manga production.
📖 Example 3: When you say "increase the number of pages instead," it drags on
Here, I will talk about the countermeasures I tried.
"If AI cuts corners, wouldn't it be better to instruct it to increase the number of pages instead?"
It took me months and a considerable amount of time to solve this. You will understand if you ask AI to "create a panel layout that explains it carefully using 10 pages".
However, this creates yet another problem.
When told to increase the number of pages, AI
moves in the direction of dragging out the same story.
For example, it turns out like this.
Page 1: The protagonist wakes up in the morning
Page 2: Looking at the newspaper while eating breakfast
Page 3: Newspaper headline says "Semiconductor stocks rise"
Page 4: Calling a friend and asking, "What are semiconductors?"
Page 5: Friend says, "I don't know, but the stocks are going up, right?"
Page 6: Protagonist searches on their smartphone
Page 7: Displaying the search results page
Page 8: It says, "Semiconductors are important, stocks are rising!"
Page 9: Protagonist is convinced, saying, "So semiconductors are important."
Page 10: "I should study more."
Loop until the specified number of pages is reached.
This, even though it has 10 pages, conveys nothing.
Only the introduction is long, and the actual content of semiconductors is hardly explained or clarified at all.
In other words, when you tell AI to increase the number of pages, it optimizes for "filling the pages", resulting in a drawn-out manga with hollow content.
"Repeating the same thing"
"Dragging out the introduction"
"Adding unnecessary daily scenes"
"Postponing the conclusion"
This is a typical trick AI uses when you tell it to "increase the number of pages."
As a result, readers find it boring and drop off.
"When is the main topic going to start?"
"What is the point of this everyday scene?"
"In the end, what were semiconductors anyway?"
The reader ends up in a state of "What am I being shown?"
Even if there are many pages, understanding and enjoyment do not increase.
This is another dilemma that occurs when you tell AI to "increase the number of pages."
🔄 The essence of the dilemma: "Matching page count" and "making it understandable" are different optimizations
Here, the essential structure of AI manga production becomes visible.
AI optimizes for the goal it is instructed to follow.
If you say "shorten it," it optimizes for brevity.
If you say "make it long and detailed," it will optimize for length.
However, what we are really looking for is,
"manga that readers can enjoy and understand."
This is not a matter of page count.
There are manga that are short but convey things carefully.
There are manga that are long but drag on.
The reverse is also true.
In other words, "short/long" is not the goal, but the means.
The goal is "reader understanding and enjoyment."
However, for AI, both "reader understanding" and "reader enjoyment" are vague goals that cannot be quantified or measured.
Therefore, AI can only optimize the specific goals it is instructed to follow (shortness, length, page count).
AI cannot directly optimize the ultimate goal of "whether the reader enjoys it or not."
🎭 The structure of AI that creates dilemmas
To summarize, AI manga production has these dilemmas.
When the goal is to make it understood:
→ AI optimizes for "reaching the conclusion quickly"
→ The number of pages becomes short
→ There is more skipping
→ As a result, readers without prior knowledge cannot understand it
When the goal is to increase the number of pages:
→ AI optimizes for "filling the specified pages"
→ There is more dragging development
→ The content becomes hollow
→ As a result, readers find it boring and leave
Whichever way you lean, AI cannot directly optimize for "reader understanding and enjoyment".
This is because it is an ambiguous goal that cannot be quantified.
Human creators adjust this intuitively.
"I want to explain this part carefully"
"I want to breeze through this part."
"Let's insert an analogy here."
"Let's put the protagonist's emotions here."
These judgments are based on experiential senses regarding where readers get stuck, where they enjoy it, and where they get bored.
And that is an area that AI does not possess.
🛠️ Adjustment work that humans must do
To resolve this dilemma, the only way is for humans to consciously step in and make adjustments.
Specifically, these are the tasks:
① Discover panels that AI skipped
Read the panel layout created by AI and find places where "the story suddenly jumps" or "the reader is left behind."
② Verbalize "why it jumped"
Specify what needs to be supplemented to be understood, such as "the explanation of semiconductors is missing," "there is no definition of HBM," or "a diagram of the supply chain is needed."
③ Give instructions to add that supplementation as a separate panel
"Add two pages of panels explaining what semiconductors are."
"Add one page of panels explaining the definition of HBM using a cooking analogy."
and other additional instructions.
④ AI creates the additional panels
⑤ Check if the additional panels are dragging on
Check if they have become unnecessary scenes just to fill space, or if they are truly essential elements for understanding.
⑥ If necessary, give instructions to cut or rewrite the additional panels
⑦ Return to ① again
Elsewhere, discover another skipped point. Or, discover a part that has become sluggish as a result of the additions.
This back-and-forth process is repeated 5, 10, and sometimes 20 times.
And finally, a "manga that readers without prior knowledge can enjoy and understand" is completed.
This work is something AI absolutely cannot do.
Because AI lacks the ability to directly judge "reader understanding and enjoyment".
Humans must use their own experiential sense to make judgments one by one.
💡 My realization: 'Answers' and 'understanding' are different
Through the experience of making AI manga, this is the biggest lesson I have learned.
'Providing an answer' and 'making the reader understand' are completely different tasks.
AI is good at providing answers.
If you say, 'Explain the structure of the semiconductor industry,' it will provide a perfect structural answer.
But whether that answer reaches the reader is another matter entirely.
The job of a human creator is to translate the 'answers' provided by AI so that they reach the reader's mind.
Fill in the skipped parts, trim the rambling parts, add analogies, add stories, and add character emotions.
This is the real job of a creator in the AI era.
You need to put in human effort to the level where you can say 'I made it together with AI', rather than 'I had AI make it.'
Otherwise, AI manga will end up as 'honor student manga for experts' that reaches no one.
━━━━━━━━━━━━━━━━━━━━
✅ Brief summary of Chapter 2
1. AI produces a "perfect narrative structure"
A textbook-perfect structure. But what humans might want to depict is the "trajectory of hesitation."
2. Characters are too stable
A character who speaks "like themselves" every time looks mechanical. Human conversation has fluctuations, and that is what creates reality.
3. AI manga skips panels
It condenses content that should be carefully explained over 10 pages into just one. The panel layout becomes something only those with prior knowledge can understand.
4. If you tell it to "increase the number of pages," it drags on
AI optimizes for filling pages, adding unnecessary daily scenes or delaying the conclusion. The content ends up being hollow.
5. "Number of pages" and "reader understanding/enjoyment" are different optimizations
AI can only optimize for goals that can be quantified (shortness, length, number of pages). Only humans can judge whether "readers will enjoy it" or "if it is understandable."
6. The human role is to "adjust the dilemma"
Discovering and supplementing the parts that were skipped, and cutting the parts that drag on. By repeating this back-and-forth process five or ten times, it finally becomes a manga that reaches the reader.
━━━━━━━━━━━━━━━━━━━━
🎓 Chapter 3: AI cannot execute what humans have not "verbalized"
We have looked at the problems of "AI skipping content" and "setting the reader's knowledge level too high," but there is an even deeper issue.
That is a problem where the human side has not verbalized the 'meaning of the instruction itself'.
I myself sometimes give instructions like this to AI when making AI manga.
"Break it down into panels with language and understanding that an elementary school student can grasp, and double-check it."
I say this with the intention of 'make it easy to understand'. Furthermore,
I have the AI save the definition of these words.
However, please stop here for a moment and ask yourself.
🤔 What does "understandable to an elementary school student" mean?
Now, here is a question. Have you clearly defined the phrase "understandable to an elementary school student"?
At first, I was using it vaguely.
With the nuance of "simple enough for even a child to understand."
But, while giving instructions to the AI, I lost track of what I was actually looking for.
What does "understandable to an elementary school student" actually mean?
The level of kanji?
For example, do elementary school students learn about "semiconductors"?
They probably don't.
But if I rewrite "semiconductor" as "a part that turns electricity on and off," will an elementary school student understand it?
In the first place, do the concepts of "turning electricity on" and "turning electricity off" make sense to an elementary school student?
The range of vocabulary?
For example, would an elementary school student understand "supply chain"?
Both "supply" and "chain" are English words.
If I replace them with Japanese like "the flow of products coming from the factory to the store," will they understand?
But how concretely can an elementary school student imagine that "flow"?
Sentence length?
I've heard that writing for elementary school students means keeping sentences short.
But how many characters or less is considered appropriate for elementary school students?
20 characters? 30 characters? 50 characters?
Logical complexity?
It is said that writing for elementary school students has a simple logical progression.
But, where does "simple" end, and
where does "complex" begin?
Where does writing for children end, and where does writing for adults begin?
Is "If A, then B" understandable, but "If A, then B, except when condition C applies" difficult?
How to use analogies?
Analogies are frequently used for elementary school students.
But, what kind of analogies would elementary school students understand?
Baseball analogies? Candy analogies? TV show analogies?
These are just examples, but I had not clearly defined them, so they didn't get across to the AI.
I was just vaguely saying "understandable for elementary school students," but I didn't even know what I meant by that myself.
🛠️ The detailed settings I actually use
From then on, I started thoroughly verbalizing the instructions for creating AI manga.
For example, I structure my instructions like this.
"Please divide the panels according to the following conditions.
[Kanji Level]
Use only kanji learned up to the 6th grade of elementary school. Write other kanji in hiragana.
For technical terms, always include a Japanese translation in parentheses upon first mention.
Example: 'Semiconductor (a component that conducts or stops electricity)'
[Sentence Length]
Dialogue per panel must be a maximum of 40 characters.
If a longer explanation is necessary, split it into multiple panels.
[Vocabulary Range]
For katakana English terms, always include the Japanese meaning upon first mention.
'Supply chain' -> 'The flow of products from the factory to the store'
'Global' -> 'Worldwide'
[Logical Structure]
Only one claim or one fact per panel.
Do not pack multiple points into a single panel.
Stick to "If A, then B" and do not use complex conditional branching.
[Analogies]
Use familiar subjects such as baseball, snacks, school life, family, or pets for analogies.
Analogies using economic, business, or political terminology are prohibited.
[Double-check items]
Are there any undefined technical terms used in each panel?
Does the dialogue in each panel exceed 40 characters?
Do the kanji exceed a 6th-grade elementary school level?
Is there more than one point packed into a single panel?
Is the analogy based on a familiar subject?
This is still rough, but by further defining it in detail, the AI will finally implement my intentions accurately.
The seven characters "understandable to an elementary school student" require this much decomposition work.
These are things that humans understand intuitively, but
AI cannot understand these vague nuances.
🔍 The essence of the task of defining "understandable to an elementary school student"
In a sense, this is training for verbalization skills.
We humans frequently use ambiguous language in daily conversation.
"Make it easy to understand," "make it simple," "keep it simple," "make it feel good," "make it natural," "do it well."
Between humans, this works.
This is because we can guess what the other person intends from their facial expressions, context, past conversations, and shared culture.
However, AI cannot do that.
AI tries to read intent from the words themselves.
When told to "make it easy to understand," AI averages out the "characteristics of text considered easy to understand" from past training data and outputs that.
But that is not necessarily the "ease of understanding" that I am looking for.
Therefore, to get AI to do accurate work, I need to define for myself what "easy to understand" means.
And it is only after doing this work that I realize I myself did not understand what 'easy to understand' meant.
I am forced to face, for the first time, the emptiness of the words I had been using vaguely.
💭 Why can't AI understand 'understandable to an elementary school student'?
Here, I return to an essential question.
Why can't AI understand 'understandable to an elementary school student'?
The answer is simple.
Because AI is not an elementary school student.
AI does not know through experience how elementary school students feel, what they can understand, or where they get stuck.
As human adults, we can recall our own experiences from when we were in elementary school.
'At that time, I didn't understand these words'
'When the teacher used this analogy, I suddenly understood'
'I couldn't understand this part of the textbook no matter how many times I read it'
From these experiential memories, humans can construct an image of what is 'understandable to an elementary school student'.
However, AI does not have that experience.
AI only has past literature data.
It is merely statistically processing "writing style patterns for elementary school teaching materials" and "characteristics of articles classified as being for elementary school students."
The "understandable to elementary school students" based on experience and the "writing style for elementary school students" from a database are similar but decisively different.
And this difference may structurally never be bridged, no matter how much AI evolves.
This is because AI does not have the experience of having lived through childhood.
📊 The inability to quantify or template things is AI's limitation
In connection with this, AI has a clear limitation.
That is the limitation that it is poor at handling things that cannot be quantified, templated, or turned into rules.
For example, these kinds of instructions are an area where AI excels.
・"Keep the character count under 1500 characters" (quantified)
・"Divide into 5 headings" (templated)
・"Limit kanji to a 6th-grade elementary school level" (rule-based)
Since these have clear criteria, AI implements them accurately.
However, these kinds of instructions are not the strong suit of AI.
・"Make it sound natural" (What is natural?)
・"Use expressions that resonate with readers" (What does resonate mean?)
・"Make it more impactful" (What does impactful mean?)
・"Make it look good" (What is good?)
Because these cannot be quantified, AI does not know what it should optimize.
As a result, AI outputs "average patterns considered to be natural writing."
Whether that is truly the 'natural' I am looking for is a different matter.
It is fine for text. It produces average, beautiful, AI-like writing.
However, in the case of manga, it tries to provide an answer immediately and cuts corners.
This can become a fatal flaw.
🛠️ The "verbalization" work that humans must do
Based on this, the work that humans in the AI era must do becomes clear.
That is, the process of thoroughly verbalizing the words you use vaguely.
For example:
"Easy to understand" → "What should the kanji level, sentence length, logical complexity, and vocabulary range be for each?"
"Natural writing" → "Writing style (polite/plain), variety of sentence endings, rhythm of line breaks, and frequency of punctuation"
"Resonates with readers" → "Reader persona, emotions you want to evoke (empathy, surprise, conviction, call to action), and the patterns of phrases that elicit them"
"Good feel" → "Specifically, what makes it good, and what makes it bad?"
Define these yourself before instructing the AI.
And, if you feel "this is wrong" when looking at what the AI output, verbalize why it is wrong.
Then, translate that sense of discomfort back into instructions for the AI.
This back-and-forth between "human verbalization" and "AI instructions" is the essence of creativity in the AI era.
🔁 Go back and forth again and again
Specifically, I will reveal my workflow for creating AI manga.
Step 1: Initial instruction
"Create a panel layout that an elementary school student can understand."
→ AI produces something.
Step 2: Discovering a sense of discomfort
I read it and find parts where I feel "An elementary school student wouldn't understand this.".
For example, the word "semiconductor" is used without explanation.
Step 3: Verbalizing the discomfort
I realize myself that "'Semiconductor' won't make sense to an elementary school student. It needs an explanation."
Step 4: Correcting the instruction
I tell the AI, "If technical terms like 'semiconductor' appear, always include an explanation using familiar words, such as 'a component that controls the flow of electricity.'"
Step 5: Re-output
The AI produces a revised version.
Step 6: Discovering discomfort again
I discover another problem. For example, "supply chain" is used without explanation.
Step 7: Correcting instructions again
I added, "Include Japanese translations for katakana English terms when they first appear."
Step 8: Re-output
This is repeated over and over again.
And if you turn this into a template,
mass production is possible.
You can make as much AI manga as you want.
But to get there, it takes 5, 10, sometimes 20 tries.
Every time, I discover a new sense of discomfort and add new instructions.
And finally, I reach a quality where "even an elementary school student can understand it".
This back-and-forth process is precisely the work of humans in the AI era.
"Leaving it entirely to AI and using what comes out as-is" is something you must never do.
Because, as I will say again and again, AI is not an elementary school student.
Humans must continue to judge for themselves, "Will this get through?"
💡 The ability to articulate is the greatest skill in the AI era
Summarizing what we have discussed so far, the most important human skill in the AI era becomes clear.
It is the ability to articulate and patience.
"Easy to understand," how do you define it?
"Natural writing," how do you quantify or template it?
"Resonating with readers," how do you break it down into elements?
Those who can do this can master AI as a powerful tool.
Those who cannot will only be pushed around by AI.
And this ability to articulate cannot be replaced by AI.
Because the starting point of articulation is human experience and sensation.
It is often said that primary information is important for mastering AI, but
That is the language of a 'person who can verbalize.'
Whether it's typing or voice input, if you can't verbalize it in the end, you can't give instructions to AI.
If you can't verbalize it, it's just a memory in your head.
AI, which has no experience and no senses, cannot verbalize things on its own.
That is precisely why, as a task that humans must do, 'the work of thoroughly verbalizing what you are seeking' will become increasingly important from now on.
The more AI evolves, the more the value of humans with the ability to verbalize will rise.
Conversely, people who stop thinking for themselves and leave everything to AI will be replaced by AI.
The greatest skill for living in the AI era is not programming, English, or design.
The ability to accurately express what is in your head using words.
I believe that is the true skill that will be demanded from now on.
━━━━━━━━━━━━━━━━━━━━
✅ Brief summary of Chapter 3
① The contradiction of people who haven't defined 'understandable to an elementary school student' instructing AI to do exactly that
You need to verbalize everything: kanji level, sentence length, vocabulary, logical complexity, and the choice of analogies.
(2) AI is not an elementary school student, so it doesn't understand "easy to understand" based on experience
It only outputs "average patterns considered easy to understand" from past literature data. Whether it is truly easy to understand is a different matter.
(3) Things that cannot be quantified, templated, or turned into rules are areas where AI struggles
Vague instructions like "natural," "resonant," or "feels good" cannot be implemented by AI. Humans need to break them down.
(4) The back-and-forth between "human verbalization" and "AI instructions" is the essence of creativity in the AI era
After repeating revisions 5, 10, or 20 times, you finally reach the desired quality.
(5) Verbalization ability is the greatest skill in the AI era
The power to accurately express experiences and sensations in words. This is a value unique to humans that cannot be replaced by AI.
━━━━━━━━━━━━━━━━━━━━
🌐 Chapter 4: The story of how I got stuck many times while setting up Substack for multilingual operation
As I mentioned at the beginning, the setup for Substack multilingual operation.
This became a typical example of a mismatch with AI.
I will talk in a little more detail about what specifically happened.
🖥️ AI that cannot perceive "what is displayed in the Japanese UI"
The Substack I operate had its screen language automatically set to Japanese.
However, when I asked the AI, "Where is the button to create multiple publications (pages for publishing in different languages)?"
The AI, assuming an English UI, tells me "Settings → Publications → Create another publication". The AI, assuming an English UI, tells me "Settings → Publications → Create another publication".
On my screen, it is displayed like this.
"Settings", "Publications", "Create another publication"
The meaning is the same, but the appearance is completely different. Moreover, there are two setting locations, and I don't know which one is correct.
I was confused, thinking, "I can't find an item called Settings" and "What is Publications?"
The AI does not initially perceive that I am using a Japanese UI.
This is because I did not explicitly tell it, "I am using a Japanese UI."
Furthermore, the AI brings up past information to explain things.
It seems the Japanese version of Substack did not exist when the AI's data was registered, and it is certain that there was a discrepancy with the latest version.
A human friend might notice and say, 'Oh, show me a screenshot of the screen,' or 'Wait, it's in Japanese.'
However, since AI communicates via text, it cannot pick up on that context on its own.
Only after I tell the AI, 'The screen items are different' or 'It's displayed in Japanese,' does it finally correct itself and say, 'Ah, that refers to the 'Publications' section.'
This happened over and over again.
'Where is the Stripe account (payment settings) setup screen?'
'Where can I find scheduled posts for Notes (article publishing)?'
Every time, the AI answers assuming an English UI.
Every time, I have to ask again, 'How does it look in the Japanese display?'
🎯 The responsibility to communicate your 'current location' lies with the human.
Here, I learned a lesson.
To get the correct answer from AI, you need to communicate your 'current location' as specifically as possible.
'I am currently on the Substack settings screen.'
'The screen is displayed in Japanese.'
'There is a profile icon in the top right of the screen.'
'In the left menu, I can see 'Home,' 'Subscribers,' and 'Chat.''
While explaining in such detail, taking screenshots, AI gives accurate answers.
'Click on "Settings," scroll down, and you will find a section called "Publications."'
In other words, before having the AI provide an 'answer,' you need to perform the task of conveying your own situation to the AI.
However, there is a trap that many people fall into here.
The problem is that when you say 'convey your own situation,' you are asking in the first place because you don't know what you don't understand.
'What is displayed on the screen?'
Even if asked that, a beginner cannot judge which information is important and which can be ignored.
A beginner doesn't know if 'the profile icon in the top right of the screen' is an important clue or not.
Ultimately, to master AI, a certain level of 'foundation' is necessary.
And for those without that foundation, AI becomes a frustrating existence that provides perfect answers but doesn't mesh with their own reality.
💬 I myself have also said 'No, that's not it' many times
To confess honestly, during today's Substack setup work, I told the AI 'No, that's not it' and 'I don't understand what you mean' many times.
I understand that the AI is explaining it perfectly.
However, that explanation does not align with my current situation.
"Settings → Publications, where is that on my screen?"
"Since it's the Japanese version, I don't know where it is."
"That button isn't there."
Every time I said that, the AI gradually understood my situation, and eventually guided me to the exact location.
The AI knew where the button was from the very beginning.
It just required the process of sharing my screen with it many times.
━━━━━━━━━━━━━━━━━━━━
✅ Chapter 4 Summary
① AI is not good at sensing the other party's screen or current situation
In text-based interactions, it cannot directly observe what the other person is looking at or what they are stuck on.
② To draw out the correct answer, you need to communicate your current position in detail
Communicate as specifically as possible, such as "I have this part of the screen open right now" or "It is displayed in Japanese."
③ A certain level of foundation is necessary to master AI
For those without a foundation, AI becomes a "perfect but mismatched" presence. This is a major barrier to AI utilization.
━━━━━━━━━━━━━━━━━━━━
⚠️ Chapter 5: 5 Pitfalls That Content Creators Relying on AI Fall Into
Based on my experiences so far, I will organize the 5 pitfalls that content creators and business professionals relying on AI fall into that I have observed.
🪤 Pitfall 1: Taking "perfect answers" at face value
The answers AI provides are perfect on the surface.
The sentences are polished, the structure is logical, and the terminology is accurate.
However, whether that answer aligns with your own reality is,
a completely different issue.
For example, if you ask AI to "write an English article for an overseas audience," it will return a perfect English article.
The grammar is perfect, the structure is sound, and it matches the topic.
But that is just an "English article that sounds like it was written by a native speaker", and it is different from the natural tone of an English article written by me (a Japanese trade merchant).
If you don't notice this substitution and post the AI's perfect answer as is, the result is that an "article that doesn't sound like you" will be published.
Especially when adapting unique Japanese analogies or expressions that resonate with Japanese people for an overseas audience, "there is no way it will spread." Modifications are necessary in those areas.
The essence of the pitfall is that "perfect" and "suited to your own reality" are two different things.
🪤 Pitfall 2: Not knowing what you don't know
When you ask AI a question, it always returns an answer.
You read that answer and think, "Ah, I get it."
However, whether you truly understand it is another matter entirely.
AI is not good at structuring answers in the order that the questioner can truly understand.
This is because AI cannot directly observe the questioner's level of understanding.
As a result, the answer AI provides is often "correct, but still too advanced for me".
And if you move on while "thinking you understand" that answer, you will get stuck later.
This happens frequently, especially when learning new technologies or systems.
For example, as mentioned earlier, if you ask "What is the Substack system?", the AI will explain it perfectly.
But, when you actually try to operate it,
you get stuck on the basics.
The biggest trap is "thinking you understand" based on the AI's answer.
🪤 Pitfall 3: Abandoning your own trial and error
Because AI provides perfect answers, the time humans spend on trial and error decreases.
While this is convenient, there is a risk that "the learning gained through trial and error" is lost.
For example, when thinking of an article title.
If you ask AI to "suggest 5 good titles," 5 will appear immediately.
You choose one from among them.
It is convenient.
However, through the trial and error of thinking of titles yourself, you should have been developing a sense of "why I think this title resonates".
If you continue to rely on AI without developing that sense, you will end up in a state where "a sense for titles that is uniquely yours" never develops.
This is a trap that applies not only to content creation, but to business decisions, project planning, strategy building, and all intellectual work.
🪤 Pitfall 4: AI's "optimal solution" does not fit your situation
AI provides "generally optimal answers".
However, it is not "the optimal answer for your situation."
For example, when I ask AI "how to earn money on note," the AI tells me "the best practices of top note creators."
"Post three times a week, focus on a niche specialty, utilize email, and leverage other social media."
It comes up with things like that. It is correct.
But this is "that person's" best practice.
For me, the optimal solution might be something else entirely.
AI is not good at providing an optimal solution that takes into account your specific situation.
It simply pulls from an average value that it deems highly likely.
People in "minority situations" need to perform the task of not accepting AI's optimal solution as-is, but translating it into their own situation.
🪤 Pitfall 5: Mistaking AI's words for your own
Finally, this is the most serious pitfall.
If you have AI write articles and keep posting them after making minor corrections, a phenomenon occurs where you lose track of "what you are actually thinking."
I wrote about this before, didn't I?
You read a text written by AI and think, "Ah, I think so too."
You post it as is.
If this continues for three months, half a year, or a year, the power of your own original thinking will decline.
This is the same as muscles.
Muscles that are no longer used will atrophy.
The "ability to think for yourself" that is no longer used will also decline.
And then, one day, you realize.
"Wait, what was I actually thinking about?"
"Can I even write something on my own without borrowing AI's words?"
This fear is something many creators who rely on AI must be feeling.
It's just that few people say it out loud.
━━━━━━━━━━━━━━━━━━━━
✅ Chapter 5 Summary
① "Perfect answers" and "answers that fit your reality" are different things
You must translate it into your own tone, situation, and context without being misled by the perfection of AI.
② The biggest trap is "thinking you understand"
If you "think you understand" based on AI's answers, you will get stuck in actual implementation. True understanding requires trial and error.
③ If you rely too much on AI, your own sense and thinking ability will not grow
Just like muscles, if you don't use them, they atrophy. You need to consciously secure "time to think for yourself."
━━━━━━━━━━━━━━━━━━━━
🤝 Chapter 6: Practical points for "engaging with AI as an equal"
So far, we have looked at AI's "mismatched optimal solutions" and the pitfalls that humans fall into.
Finally, how should we engage with AI?, I will talk about the points I practice.
🎯 Point ①: Tell AI your current situation in detail
When asking AI questions, conveying your situation as specifically as possible is the first trick.
Taking the previous example of overseas publishing,
"I am a Japanese person, my screen is displayed in Japanese, I am a beginner at Substack, I have not yet launched my English Substack, and I am an independent creator with 12 years of experience as a trade merchant. I ultimately want to operate in three languages: English, Spanish, and Portuguese."
To that extent, convey the situation in detail.
Or, it becomes necessary to use a form of brainstorming to flesh out the details.
When you do that, the AI will provide an answer tailored to my situation for you.
"Given your situation, a phased approach where you launch only the English version first and add the Spanish version after 3 to 6 months is realistic."
Something like that.
AI improves the accuracy of its answers if you provide it with information.
Expecting the AI to "read between the lines" is impossible.
However, if you feed it information, it will return an appropriate answer.
🎯 Point 2: Treat AI answers as a "drafting board"
An AI's answer is not a finished product, but a drafting board.
I have the AI write drafts of my articles.
But I almost never use them exactly as they are.
I look at the structure the AI wrote, use it as a reference by thinking, "Oh, so it unfolds in this order," and then rewrite it in my own words.
Looking at the title proposed by the AI, I think, "This hook is weak, let's add more numbers," and revise it.
Looking at the conclusion organized by the AI, I think, "This is too polished, let's keep the hesitation as it is," and break it down.
In other words, use AI as a "drafter," "discussion partner," or "research assistant.".
This is the basic stance for engaging with AI as an equal.
🎯 Point 3: "If you feel something is off with the AI, tell the AI."
When you feel something is off with the AI's answer, telling the AI about that discomfort directly is the secret to improving the quality of the dialogue.
"No, that's not it."
"I don't understand what you mean."
"Be more specific."
"Adjust it a bit more to my context."
Even words you might hesitate to say to a human friend, throw them at the AI without hesitation.
When you do that, the AI will correct itself, saying, "Oh, I see, that was your intention.".
This is something I did many times during my Substack work today and with my AI manga.
"No, that's not what I meant."
"Since it's the Japanese version, I don't know where it is."
"I don't understand the meaning at all."
These candid reactions significantly improve AI accuracy.
Conversely, if you feel something is off about the AI's answer but hold back and accept it anyway, the conversation will proceed without ever being in sync.
🎯 Point 4: Consciously create "areas where you don't rely on AI"
I have areas where I consciously do not use AI.
For example:
・The "core argument" of an article must be written by myself.
・The final decision on a title is made by myself, using AI suggestions as a reference.
・Personal messages to business partners are written by myself, not by AI translation.
・My own character design (Pos-tori-ness) is not left to AI.
These are areas where, if "my uniqueness" disappears, my own content creation itself ceases to function.
If you leave this to AI, in three or six months, you will become a "creator who has lost their own voice".
Therefore, consciously maintaining an "AI-free zone" is essential for long-term creator activities.
🎯 Point 5: Treat conversations with AI as training for your own thinking
Finally, this is my personal realization, but there is a perspective of treating the conversation with AI itself as training for your own thinking.
When asking AI a question, think about how to ask it.
When reading AI's answers, verbalize what feels off.
When asking AI for revisions, clarify what you want changed.
Through these tasks, what you are thinking and what you are seeking will gradually become visible.
In other words, AI is "a tool to provide answers" and at the same time, "a mirror that reflects your own thinking".
Convert your dialogue with AI into a dialogue with yourself.
If you can do this, the more you use AI, the deeper and clearer your own thinking will become.
Conversely, if you leave everything to AI, your own thinking will become shallower and more ambiguous.
Which one you choose depends on the human's own awareness.
━━━━━━━━━━━━━━━━━━━━
✅ Chapter 6 Summary
1. Clearly communicate your current position to the AI
AI is not good at "guessing," but its accuracy improves if you provide information.
2. AI's answers are a "draft," the final decision is the human's
Use AI as a drafter or discussion partner, not as the final decision-maker.
3. Consciously maintain AI-free zones
The core parts of "your uniqueness" should not be left to AI; write them yourself. This is the lifeline for long-term creator activities.
━━━━━━━━━━━━━━━━━━━━
🛠️ 3 actions you can take starting today
To help you immediately apply what we've discussed to your daily life, I have organized it into three actions.
1. Before asking AI a question, take 10 seconds to write down your "current position"
Before asking AI anything, please take just 10 seconds to write down your situation.
"What am I trying to do right now?"
"How far have I progressed?"
"What you are stuck on"
"What is displayed on the screen"
Just with this, the accuracy of the AI's answer will improve significantly.
And, your own situation will also become clearer.
(2) If you feel something is off with the AI's answer, put that feeling into words
The moment the AI provides an answer and you feel "something is not quite right", do not ignore that discomfort.
Put "what is different" into words.
Think about "where it deviates from my situation."
And, convey that discomfort to the AI.
By repeating this, not only will the quality of your dialogue with the AI improve, but your own senses will become sharpened.
(3) Create one "AI-free day" per week
Finally, this is a method I tried and felt was effective.
One day a week, create a day where you do not use AI at all.
On that day, write, think, and make decisions using only your own head.
It is inconvenient at first.
However, what you are thinking and how much you can do without AI will become clearly visible.
And that will also be useful on days when you use AI.
It is a simple yet powerful habit to maintain your "self-thinking muscles."
━━━━━━━━━━━━━━━━━━━━
⚠️ Disclaimer
This article is a reflection based on the author's personal experience with AI utilization.
It does not recommend or deny any specific AI service or usage method.
The optimal way to engage with AI varies depending on individual and business circumstances.
Please choose the best way to engage with it based on your own judgment.
━━━━━━━━━━━━━━━━━━━━
🐦 A word from Pos-tori
In Toyama late at night, the coffee cup on my desk has already gone completely cold.
Outside, the quiet night air hangs over the water-filled rice paddies.
Tonight, I set up overseas distribution with AI, configured manga settings, and said "No, that's not it" many times.
I am writing this article based on that experience.
AI is a convenient tool.
In my article production, international expansion, and business decisions, I could not achieve my current speed and depth without AI.
However, if you rely too much on AI, there are moments when you lose track of what you are thinking yourself.
That might be a sign that I myself am accepting AI's "optimal solutions" too blindly.
I want to be the one using the tool, not the one being used by it.
To that end, I will be conscious of my current position, verbalize my own sense of discomfort, and protect the core of my individuality.
The more I interact with AI, the more deeply I see the meaning of "what it means for a human to be human".
This might be the greatest discovery for us living in the AI era.
I look forward to your comments and DMs again today.
See you in the next article.
Pos-tori
━━━━━━━━━━━━━━━━━━━━
Click here for the article: Would you support us?
Would you consider joining our membership? If you are interested, please click the button below.
There are two courses: 500 yen per month and 980 yen per month. All-you-can-read! First month free, no cancellation commitment! I think it's quite a bargain!
▼ Click here to join Posu-dori's "Membership"
You can request articles or ask questions to Posu-dori from the link below.
[Author Overview]
I, Posu-dori, also known as "Cokumu George," run a trading business in the Hokuriku region while working on projects that shed light on the "gaps" in society, as I have written about in this article.
Although there are many failures, conflicts, and embarrassing stories, I share the real stories I have seen and felt in the field, along with the business insights gained from them, on this note.
I also tweet about my daily realizations on X (formerly Twitter), so please follow me.
Your support will be a source of strength for me and for them.
note: https://note.com/glad_auklet4142
Threads:https://www.threads.com/@cokumu_motto
X (Twitter):https://x.com/_596_
[Special Free Report] Information
While you wait for the next article, please take a moment to deepen your thinking with this report. Why do 90% of entrepreneurs disappear from the market? You can find the scientific basis for this and a self-diagnosis to ensure you don't follow the same path.
-
7 Scientific Analyses of "Vanishing Entrepreneurs" from 12 Years of Trading Experience: Why Do 90% Fail? Self-Diagnosis Learned from Data and Real Examples
Why do 90% of entrepreneurs, who should have been so full of passion, disappear from the market? In the background, there are "scientific laws" that cannot be dismissed as luck or talent. This report thoroughly dissects the 7 fatal patterns common to failure. It should serve as a guide to prevent you from making the same mistakes.
[Special Magazine] An Invitation to You Who Seriously Want to Break Through
This is a special magazine that collects paid articles into which I have poured all my experience and knowledge on deeper, more practical themes. The "key" to pushing your business or life to the next stage might be hidden within. As an investment in your future, I hope you will pick up the ones that match your challenges.
-
[Complete Analysis: 120,000 Characters] Success Strategies of 12 Top Overseas Bloggers
While you continue to exhaust yourself with your own methods, top players overseas continue to earn money using reproducible "laws." Steal all those secrets in 120,000 characters and make them your own. Before it's too late, do it now.
[NEW! 220,000 characters] Silent Alarm: Before Your Soul Breaks. An "Ultra" Early Detection and Prevention Guide to Burnout for Freelancers & Entrepreneurs
"Yesterday, you were on top of the world. Today, you can't get out of bed." This is not someone else's problem. Based on the real-life examples of 12 overseas entrepreneurs who quietly broke their hearts at the peak of their success, this guide fully covers the "warning signs" and "preventive measures" for burnout. This is a "soul alarm system" to protect your future.
-
[Permanent Preservation Edition] Why do that person's notes sell even at a high price?
Selling valuable content cheaply is a sin. This is the "forbidden technique" for selling at a fair price while being appreciated by customers. Without knowing this, you might be on the side being exploited for the rest of your life.
-
[30,000-character true record] How do you choose a tax accountant?
How did I make a V-shaped recovery when I was pushed to the brink of a cash flow crisis? It all came down to "choosing a tax accountant." This is a potent remedy packed with everything from how to search and market rates to the inside story revealed by a former national tax official.
-
[Make your corporation a stock company!] The story of how I lost tens of millions of yen by being stingy with the 140,000 yen establishment fee.
"Let's go with a limited liability company because it's cheap"... Did you know that such a casual choice is the beginning of losing everything—financing, hiring, and trust? This is a record of my blood-sweating failure where I actually threw tens of millions of yen down the drain.
-
[Complete true story] A collection of records of how I let my red-ink company go bankrupt and shut it down
Is there any soldier who jumps into the battlefield without knowing the escape route? Before learning how to "build" a company, you must learn how to "shut it down," which is 100 times more difficult. This is the absolute requirement for a successful entrepreneur. Do not look away from the inconvenient truths that no one wants to talk about.
-
[Must-see for beginners!] The ultimate account management technique to "save money" taught by a former National Tax Agency official
The reason you can't earn or save money isn't because you lack ability. You just don't know the "how-to." This method, practiced by a former National Tax Agency official, is so simple and powerful that continuing to not know it is a lost opportunity.
-
[Paid note / 40,000 characters] The secret to the "official refund feature" that boosts sales by an average of 2.8 times
The "refund feature" that many people are too afraid to use is actually the detonator that can make your sales explode. By reading this article, you can confidently offer "refunds OK" and leave your rivals in the dust.
-
[Safe for AI beginners] Just 5 minutes! The complete procedure to create your own note headers that sell using AI
The era of spending time and money on header images is over. If you follow this article, anyone can create a professional-grade header in 5 minutes. What will you do with the tens of thousands of yen you save? Smart people are already starting.
[Author Introduction]
The author of this article, "Posu-dori" (also known as "Kokumu-George"), is a trade merchant who, in his daily life, imports products and plans merchandise while communicating with vendors from various countries in the Hokuriku region.
For details on his background and profile, please visit the following page.
Also, as a business copywriter, he is involved in creating product copy and descriptions, but the case studies and explanations introduced here are structured based on Posu-dori's unique perspective—leveraging knowledge of the latest technology and AI, through academic paper searches, research report investigations, drafting, and brainstorming.
Since much of the content within the article is AI-generated, this is an activity as a different author rather than traditional writing work.
Furthermore, Posu-dori regularly introduces the latest papers, research results, and case studies on X (formerly Twitter, https://x.com/596), allowing readers to freely enjoy new knowledge and inspiration.
This note contains a variety of content created through the collaboration of AI and humans. As a new form of next-generation reading material, please check out other articles, follow, and enjoy the latest information.
※[Regarding content generation for this article]
As written in the author introduction, this article also utilizes advanced AI tools such as ChatGPT (OpenAI), Claude (Anthropic), and Gemini (Google) to conduct academic paper searches, research report investigations, draft creation, and brainstorming.
The generated text is used as a first draft in accordance with the terms of service of each respective service.
The content posted is also in compliance with those terms. Note that the final text is always verified, edited, and rewritten by the author to ensure originality.
Heartfelt thanks to each AI tool that contributed significantly to the generation of this article. And heartfelt thanks to the readers who have read it.
[Important Notes]
1. The content of this article includes the author's personal views and hypotheses. These may not necessarily align with general recognition or expert opinions.
2. The information in this article is current as of the time of writing. Some information may become outdated over time.
3. Please practice and apply the content of this article at your own risk. The author cannot be held responsible for the results of any actions taken by readers based on the information in this book.
4. This article is based on the author's experience and domestic/international case studies, with the text refined by ChatGPT, Claude, and Gemini.
5. Please follow the rules when citing the content of this article. However, as stated, this is based on the author's experience and may not necessarily be correct knowledge, so please be aware that it should be used as reading material or for reference.
6. The copyright of this content belongs to the author. Unauthorized reproduction or duplication beyond the scope stipulated by note's terms of service is strictly prohibited.
I would like to express my heartfelt gratitude to the various AI tools that contributed significantly to the creation of this article, and above all, to the readers who have read this far.
[Important: Disclaimer regarding news commentary and investment advice]
This article is intended for news commentary and general information purposes, and aims to share information on corporate trends, industry research, and marketing research. It does not recommend the buying or selling of specific financial products or stocks.
The information and views presented are part of news commentary and do not constitute investment advice.
This does not recommend investment in specific stocks or products, and the author assumes no responsibility whatsoever, so please ensure you make your own final investment decisions.