Those Without This One Thing Will Lose Their Jobs to AI: The Reality Presented by DeNA's Tomoko Namba

"I should learn skills so AI doesn't take my job."

If you think that, you are half right and half wrong.

I can say this with certainty.

What determines whether AI takes your job isn't the amount of skills you've acquired.

It is something else entirely, represented by a specific concept.

This isn't just my opinion. Tomoko Namba, who founded DeNA and returned as President in June 2026 after 15 years, specifically named this concept in a lecture.

I will give the answer in Chapter 3. You can skip ahead if you like.

However, the answer won't be effective unless you read Chapters 1 and 2. If you read it without knowing the extent to which the person who said it has pushed boundaries, it will just end up being another "nice quote."

This is a long post, so I recommend saving it if you want to look back later.

Chapter 1: Who is Saying This?

First, understand who is behind these words. If you don't, it will just be a nice story.

Tomoko Namba. Born in 1962. After graduating from Tsuda University, she joined McKinsey, earned an MBA from Harvard, and became a McKinsey partner (executive) in 1996.

She founded DeNA in 1999. She stepped down as President in 2011 and has since served as Chairwoman, owner of the Yokohama DeNA BayStars (since 2015), and Vice Chair of Keidanren (since 2021).

Then, on June 27, 2026, Ms. Namba returned as Representative Director and President of DeNA after 15 years, also serving as CEO.

The reason given by the company was:

"There is an urgent need to significantly increase management speed and accelerate the transformation of organizational operations and business models based on the future business environment."

A founder returning to the top of the field after 15 years—at age 64. You don't have to be a business owner to realize how extraordinary this is.

DeNA is a listed company with 147.7 billion yen in revenue and 18.694 billion yen in operating profit for the fiscal year ending March 2026. The top of that company came back to take the helm personally.

Why did she return? AI.

Chapter 2: What Happened One Year After Saying "We are going all-in on AI"

Let's rewind a bit.

"DeNA is going all-in on AI."

The content of that declaration was quite aggressive. She planned to maintain and grow the existing business, which was then run by about 3,000 people, with half the staff using AI. The other half would be shifted entirely to new businesses.

Many companies say they will do things but don't. So, look at the numbers one year later.

On March 6, 2026, at "DeNA x AI Day 2026" held at Shibuya Hikarie, Ms. Namba reported these results:

• In some development projects, 5% human / 95% AI, with 20x productivity.
• Legal contract checks saw a 90% efficiency improvement.
• QA (Quality Assurance) man-hours were halved.
• Screening costs for the live streaming service "Pococha" were reduced by 60%.

They even created a unique evaluation metric called DARS (DeNA AI Readiness Score). It's a 5-level self-reporting system, where Level 1 is "tried recommended conversational AI once" and Level 5 is "can improve others' productivity and formulate new strategies centered on AI."

This means AI utilization isn't just about enthusiasm or training; it has become a yardstick for personnel evaluation.

What did a leader who has gone this far say about "what happens to the people?"

That is today's main topic.

Chapter 3: The Concept Tomoko Namba Named

Here is the answer.

"Kiten-ryoku" (Originating Power).

You've probably never heard this word. It's not in the dictionary. Ms. Namba herself wrote it in quotation marks on her lecture slides.

It was a statement during the "About Human Resources" slide at the end of the February 2025 keynote. It's a bit long, but I will quote it exactly because shortening it changes the meaning.

"Even in a future where AI moves exactly as instructed, I believe the

Origin is human.

Having the will to make things happen, the power to be absorbed in something, having desires, having cravings—these things will become increasingly important.

Work will become much 'denser.' Since 'thin' work will be left entirely to computers,

I believe this kind of will, this 'Originating Power,' will be extremely important."

This was the line written on the slide:

Will, "Originating Power," the power to be absorbed, desires... Work becomes denser.

On the same slide, Ms. Namba said two more things:

"Will engineers become unnecessary? → No!"

Regarding the "no more engineers" debate raging on the West Coast, Ms. Namba's answer was a clear no. Rather, she said the demand for "engineers with the desire to create something using AI" will only increase.

"All products will be made of AI x Industry Knowledge."

Therefore, the demand for people with specialized industry knowledge, business knowledge, data, and access to customer bases will also rise.

And finally, Originating Power.

Translating "Thin Work" and "Dense Work" into simple terms.

This is the backbone of the article, so let's go carefully.

What Ms. Namba calls "thin work" is this:

Tasks where you move your hands to finish something someone else decided, using a predetermined method.

Formatting documents. Taking minutes. Transcribing numbers. Rewriting the same email draft. Pouring information into a fixed format.

On the other hand, "dense work" is this:

Work where you decide what to do in the first place.

What theme to post about. Which customers to target. Deciding what NOT to make. Who to partner with. When to quit.

AI only takes the former. Most people probably agree with this so far.

The problem is what comes next.

When all thin work moves to AI, only "dense work" remains in your hands.

This isn't a story about things getting easier. It's a story about having nowhere to hide.

Think about when you hire an outsourcer. If you just say "make it look good," you'll never get something great. What to make, why to make it, and how far to go—the person ordering always decides that.

AI is the most capable and cheapest outsourcer in history.

But it will do nothing for someone who can't write the order form.

That is what "The origin is human" means.

Chapter 4: This is Not Spiritualism. Data Says the Same Thing

For those who thought, "Ms. Namba can say that because she's amazing," stopping your thinking there will cost you.

Actual salary data shows the same thing.

There is a paper published in 2025 by Erik Brynjolfsson and others at the Stanford Institute for the Digital Economy. The title is "Canaries in the Coal Mine?". It analyzed individual-level salary data from ADP, the largest payroll service in the US.

The results were:

• In occupations most affected by AI, employment for 22-25 year olds decreased by about 13%.
• For software developers in the same age group, it decreased by about 20% from the peak at the end of 2022.
• Meanwhile, employment for experienced layers in the same occupations did not decrease.

This is the part often quoted.

But the most important part of this study is not that.

The decrease in employment was concentrated in "tasks where AI replaces humans," and did not occur in "tasks where AI augments human capabilities."

This says exactly the same thing as Ms. Namba. Professions aren't disappearing. Tasks are. Work that is just "executing a fixed method" is disappearing, while work that is "using AI while making decisions" is not.

The World Economic Forum's "Future of Jobs Report" also predicts that by 2030, 92 million jobs will be displaced and 170 million new jobs will be created globally. That's a net plus of 78 million.

In short, the total amount of work will not decrease. Only the content will be swapped.

When that swap happens, will you be on the side that "decides what to do"? That's all it is.

And here is the tough part for Japan.

According to the Information and Communications White Paper published in July 2026, the percentage of individuals in Japan who have used generative AI is 58.8%. It doubled from 26.7% the previous year. Looking at the numbers alone, it's a rapid surge.

However, the US exceeds 70%. Moreover, Japanese usage is centered on text, such as document creation, and the usage rate for images, videos, and audio is lower than in the West.

Even after doubling, we still haven't caught up. This is where we are.

Chapter 5: Translating "Originating Power" into Today's Actions

"I understand Originating Power is important. So, what should I do?"

From here is the main point. Based on Ms. Namba's statements and what DeNA is actually doing, I've broken it down into five actions.

Originating Power ①: Sort your work into "Thin" and "Dense"

Many people skip this.

They suddenly start touching tools to "improve efficiency with AI" and get bored in three days. The cause isn't ability. It's because they haven't decided what to hand over.

What Ms. Namba calls "thin work" is different for everyone. So, you have to take inventory of your own hands first.

If you're doing it today, paste this directly into an AI:

I will paste the tasks I did from yesterday to today below. Please sort them into two categories.

A: Tasks where the method is fixed and I just moved my hands as decided. B: Tasks where I decided what to do and how to do it myself.

After sorting, for each item in A, write one line on "What do I need to tell the AI first if I were to hand this over?"

[Paste your tasks from yesterday here in bullet points]

The "A" that comes out is your "thin work." And the more A you have, the more dangerous it is.

Work that disappears: Time spent vaguely sorting in your head and ultimately handing nothing over.

Originating Power ②: Move first, even if not asked

At AI Day 2026 in March 2026, there was a phrase Ms. Namba repeated:

"Move it first."

Instead of moving after a perfect design is done, move while fixing it. In the same lecture, she also said:

"Product superiority without velocity is meaningless."

Velocity. Even if the UI or UX at this moment is superior, it's meaningless if you don't have the speed to fix it.

This sounds like company talk, but it applies directly to individuals.

People waiting for someone to say "try this with AI" will never get their turn.

Waiting for instructions is the most dangerous posture in the AI era. This is because the job of giving instructions is "dense work," and that's where humans remain. The moment you stand on the side waiting for instructions, you are lining yourself up on the "thin" side.

If you're doing it today: Pick one task no one asked you to do and throw it to an AI. It could be meeting minutes or organizing next week's schedule. One thing, within the range you can do without asking permission.

Work that disappears: Time spent thinking of excuses like "My company can't use AI yet."

Originating Power ③: Discard mediocre expertise and multiply it

At AI Day 2026, Ms. Namba used quite a bold expression:

"Foundation model players and LLM players are more ruthless than expected."

"I know that mediocre expertise will be struck down in one blow."

OpenAI, Google, and Anthropic, who are investing tens of trillions of yen, will take every area they can. So, expertise like "being a bit familiar with AI" or "being able to write a bit of code" will disappear the day the next model comes out.

So what remains? Ms. Namba's answer is that one line from earlier.

"All products will be made of AI x Industry Knowledge."

AI alone cannot win. The industry knowledge you've built for 10 years, your on-site intuition, and seeing the faces of your customers. AI can never write this.

Make it a multiplication. Don't be an "AI expert," be someone who "can solve problems in [Industry X] with AI." There are millions of the former, but almost none of the latter.

Work that disappears: Time spent touching every new AI tool that comes out and ending up mediocre at all of them.

Originating Power ④: Stop polishing prompts and design the environment

Ms. Namba organizes the transition of AI utilization technology into three stages:

1. Prompt Engineering: The era when how to write instructions to AI was important.
2. Context Engineering: The era when how to teach background information was important.
3. Environment Engineering: Now.

The third stage means that since AI agents now go to get information themselves, we are in a phase of designing the environment itself, such as "how far the AI can look" and "what it is allowed to do."

This is exactly what I've been saying.

AI doesn't fail because its ability is low. It fails because it's lost.

Instead of rewriting prompt phrasing 100 times, preparing one file for the AI to refer to is more effective. Deciding what to let it read, how far to let it touch, and what to let it output. That is the human's job, and it is exactly the "origin."

If you're doing it today: Prepare one text file with your premises (who you are, who you're targeting, what tone) for the AI you use often, and start from there every time. Three lines is enough.

Work that disappears: Redoing work because you introduce yourself from scratch every time and get a different answer every time.

Originating Power ⑤: Go get information yourself

Finally, the most plain but most effective one.

AI won't teach you if you just wait.

If you don't go get information yourself, it will progress before you know it, and you'll be left behind. What's worse is that by the time you realize it and try to catch up, everyone else has moved too far ahead and it's too late.

I've been hit by this many times. A tool I thought "I'll look into later" became a prerequisite three months later, and it took many times longer to catch up from there.

In the AI world, it's normal for a topic that trended three days ago to already be old news.

So, you have to be in a situation where you are always moving yourself. This isn't a matter of ability, but the nature of this field.

If you're doing it today: Decide on just one place where information flows and make it a habit to look there for 5 minutes every morning. You don't need to follow everything. Just have one place you chose of your own will.

Work that disappears: Time spent frantically trying to catch up six months after something becomes a hot topic.

Chapter 6: To be Honest, I Tripped Up Here Once Too

I've written all this quite grandly, but let me talk about myself.

I was an elementary school teacher for 7 years and then quit. No one asked me to. No one told me to quit, and no one told me I should. I decided myself, and I quit myself.

From there, I started posting about financial education, and my Instagram followers grew to 59,000.

Then, at the end of May 2026, that account vanished overnight due to a Meta account BAN. 59,000 people became zero.

What I started after that was AI content. My X account reached 8,000 people in 2 months. I now run two of my own products (a money school and an AI course).

Why did I write this?

Because neither quitting teaching nor rebuilding with AI after the account ban happened because someone instructed me to.

It's not a pretty story. When the account was banned, it was truly dark. But at that time, the option of "waiting for someone to show me the next path" didn't exist from the start. Because I knew no one would come if I waited.

Originating Power is probably something like that. It's not talent, but whether you stand on the premise that "no one will give me instructions."

And the time I suffered the most in AI utilization was exactly the "postponing" in ⑤. I pushed a tool I was interested in to "next month because I'm busy this week," and by the time I actually touched it, everyone else was already talking about the next thing based on that tool. It took many times longer to catch up than if I had done it from the start.

That wasn't a failure of efficiency; it was a failure of Originating Power.

Summary: If You Do Just One Thing Today

To recap:

1. What determines whether AI takes your job isn't the amount of skills, but Originating Power.
2. Thin work (doing as decided) all goes to AI, and only dense work (deciding yourself) remains.
3. Stanford data also shows employment only decreased in "tasks AI replaces." Tasks are disappearing, not professions.
4. Originating Power consists of five things: ① Sorting ② Moving first ③ Multiplying ④ Designing the environment ⑤ Getting info yourself.
5. "Will engineers become unnecessary? → No!". What disappears isn't the job title, but the posture of waiting for instructions.

Let me place Ms. Namba's words once more.

"The origin is human."

"Work will become much 'denser.' Since 'thin' work will be left entirely to computers."

If you do just one thing today, paste the prompt from Chapter 5 ①. It takes 3 minutes.

Your work from yesterday will come out divided into A and B. The more A you had, the more room you have to change starting today.

Finally

What did you think?

One last thing.

The conclusion of today's article was "Let go of thin work and spend time on dense work."

However, the most common response I get after people read this is, "I understand what you mean, but I don't know how to let go."

That's natural.

So, I will name one thin work you should let go of first.

Creating presentation materials.

The reasons I name this first are three:

1. The time consumption is abnormal. Someone who spent 3 hours on one slide deck and makes one a week spends 156 hours a year on it.
2. Judgment is almost zero. "What to convey" is dense work, but "how to arrange it and what colors to use" is almost entirely thin work.
3. The area where AI has become the most proficient. Rather than text generation, image generation is what's growing now.

I have turned this mechanism of taking "judgment" entirely out of slide creation into a tool.

It's the Brain: "Automatic Slide Generation Skill."

Here it is in numbers:

• 50 Layout Templates (extracted from actual seminar materials: bullets, comparisons, graphs, Venn diagrams, timelines, flowcharts, etc.)
• 50 Design Bases (all with color codes specified. Includes industry-specific ones like finance, medical, beauty, parenting, education, travel, etc.)
• 100+ Atmosphere Adjustment Words (specify "flashier," "luxurious," or "spring-like" in one word)
• When multiplied, over 250,000 combinations
• 3 Generation Routes: Just paste into ChatGPT / Leave it all to Codex / Batch parallel generation via API

There are only three things to do: Place one folder. Tell it the theme or script. Give the OK to the proposed structure.

While you leave it alone, the structure is built, the design is decided, all pages are generated, inspected, and come out as a PDF. There is no limit on the number of pages.

I will also honestly write who it's NOT for.

The output is images and PDFs. It's not for people who want to retype text in PowerPoint.

Also, AI is fickle. It's a tool based on the premise that 10-20% will need to be remade. That's why I've included 37 recovery prompts that fix things just by pasting them according to the symptoms.

It's faster to use the words of those who have used it than for me to say it:

200 copies sold in one week since launch.

The price is 1,480 yen.

For reference, the market rate for outsourcing slide design is 1,500 to 3,000 yen per slide. If you ask for 20 slides, it's 30,000 to 60,000 yen. The price is two orders of magnitude different.

▼ Click here for "Automatic Slide Generation Skill"

https://brain-market.com/u/masakiiii/a/b5UTO3UjMgoTZsNWa0JXY

The sales page also lists all the conditions for people it's not suited for. Please read it before you decide. You don't have to force yourself to buy it. Just the prompt from Chapter 5 ① will change how you use your time tomorrow.

References

• [DeNA x AI Day || DeNA TechCon 2025] Opening Keynote (Held Feb 5, 2025 / DeNATech Official YouTube) — Source of "The origin is human," "Originating Power," and "Will engineers become unnecessary? → No!"
• Fullswing by DeNA "DeNA's Tomoko Namba Talks 'Company Management and Growth Strategy in the AI Era' Full Transcript" — Official full transcript of the above lecture.
• Fullswing by DeNA "DeNA's Tomoko Namba 'Move it First, Expand the Business' Speech Full Transcript" (DeNA x AI Day 2026 / March 6, 2026, Shibuya Hikarie) — Source of 20x productivity/95%, Legal 90%, QA man-hours halved, Pococha 60% reduction, "LLM players are more ruthless than expected," "Mediocre expertise will be struck down in one blow," "Product superiority without velocity is meaningless," and Environment Engineering.
• DeNA News Release "Notice Regarding Change of Representative Director" (Announced May 12, 2026 / Tomoko Namba assumed office as Representative Director, President & CEO on June 27).
• DeNA Fiscal Year Ending March 2026 Financial Results (Revenue 147.7 billion yen, Operating Profit 18.694 billion yen).
• Nikkei XTech "DeNA, where all employees become AI talent under Chairwoman Namba's command, evaluates skills with unique metrics" — Definition of DARS (DeNA AI Readiness Score) 5 levels, full operation from August 2025.
• Erik Brynjolfsson, Bharat Chandar, Ruyu Chen "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence" Stanford Digital Economy Lab (2025) — 13% decrease in employment for 22-25 year olds in high AI exposure jobs, ~20% decrease for young software developers vs peak, decrease concentrated in replacement-type tasks.
• World Economic Forum "Future of Jobs Report" — 92 million jobs displaced and 170 million created by 2030 (net increase of 78 million).
• Ministry of Internal Affairs and Communications "Information and Communications White Paper" (Published July 2026) — Japan's generative AI individual usage rate 58.8% (26.7% previous year), usage centered on text like document creation.

Note: All statements by Ms. Tomoko Namba in this article are quoted from publicly available lecture videos and official transcripts. This article has no affiliation with her or DeNA Co., Ltd.