Comparing China’s Youth Employment Problem with Japan’s “There Are Jobs, But People Still Sink” Problem
1. The viral claim: “Don’t marry, don’t buy a car, don’t buy a house. Pay for AI.”
A Chinese-language post on X went viral with a very extreme message.
In short, it said:
Young people in their 20s should not marry, buy a car, buy a house, date, or socialize.
Live alone, pay $200 per month for Codex or Claude, and use up all your token limits every month.
Keep doing it for two years. If you still have not succeeded after two years, come back and insult me.
Read in Japanese or English, this sounds like an intense ideology.
“Do not date.”
“Do not socialize.”
“Go all-in on AI.”
Taken literally, it is dangerous. Cutting off every human relationship and betting only on a screen is not a stable life strategy.
But in the Chinese youth context, the post reads differently.
It is less about worshipping AI and more about this:
If the normal life path is too heavy, build personal leverage with AI before taking on fixed costs.
That is not entirely unreasonable.
2. In China, the normal path of job, marriage, house, and car is heavy
China’s surveyed urban unemployment rate was 5.2% in April 2026. Youth unemployment was much heavier: the jobless rate for people aged 16 to 24, excluding students, was reported at 16.3%, while the rate for those aged 25 to 29 was 7.4%.[^china-youth]
For young people in China, the problem is not merely that a job may be boring.
The entrance itself is heavy.
On top of that, housing, marriage, cars, family expectations, education costs, and urban living costs can become major fixed burdens.
Marriage registrations have also fallen. China recorded 6.106 million marriage registrations in 2024, down 20.5% from the previous year, according to state media citing official data.[^china-marriage]
This does not simply mean young people dislike romance or marriage.
It is more likely that work insecurity, housing costs, education costs, and family expectations have made marriage and family formation feel extremely heavy.
In that context, the message,
Do not buy a house. Do not buy a car. Do not marry. Pay for AI first.
starts to look less like pure extremism and more like a survival poem.
3. Japan does not have the same “employment entrance” problem
Japan is different.
Japan’s unemployment rate was 2.5% in April 2026. The April 2026 Labour Force Survey was released by Japan’s Statistics Bureau on May 29, 2026.[^japan-stat] Market summaries also show Japan’s unemployment rate falling to 2.5% in April, with the jobs-to-applicants ratio at 1.18.[^japan-unemployment]
A rough comparison looks like this:
| Indicator | Japan | China |
|---|---|---|
| Overall unemployment rate | 2.5% | 5.2% |
| Youth unemployment | Around the 3% range for 15–24 | 16.3% for 16–24 |
| Job availability | Jobs-to-applicants ratio: 1.18 | Entry-level market is heavy |
| Main pain point | Jobs exist, but meaning, autonomy, wages, and reward feel weak | Youth unemployment, housing, and marriage costs |
Of course, definitions differ by country. China’s youth unemployment figure refers to urban workers aged 16–24 and excludes students, so it is not a perfect apples-to-apples comparison with Japan.
Still, the structural difference is clear.
Japan is not a society where young people cannot find work in the same way.
Instead, Japan’s problem is this:
There are jobs.
But people still sink inside them.
4. Japan’s pain is less “no jobs” and more “jobs without clear purpose”
In Japan, statistically, jobs exist.
But the lived experience often feels different.
Many workers face environments where:
- the purpose of work is weak
- priorities are unclear
- nobody knows why a task matters
- meetings and reports increase
- decision ownership is vague
- responsibility is pushed downward
- holes in the system are filled by individual goodwill
- hard work does not clearly lead to pay or autonomy
- the people who sustain the workplace are not necessarily the ones rewarded
In other words, Japan’s problem is not always “I cannot work.”
It is:
I can work, but working does not necessarily reward me.
“Being rewarded” does not only mean salary.
It can mean:
- higher pay
- more autonomy
- easier workflows
- accumulated skills
- higher market value
- appreciation
- more personal time
- better future options
- assets or outputs that remain
If at least one of these comes back, work can still feel meaningful.
But often, the reality is:
I worked hard, but only more meetings appeared.
I worked hard, but I became the convenient fixer.
I worked hard, but evaluation stayed vague.
I worked hard, but my pay barely changed.
I worked hard, but nothing became my own asset.
That is Japan’s version of sinking.
5. Bullshit jobs can damage people even when they are paid
This connects to David Graeber’s concept of “bullshit jobs.”
Graeber defined a bullshit job as a form of paid employment that is so pointless, unnecessary, or harmful that even the employee cannot justify its existence, even though the employee is required to pretend otherwise.[^graeber]
This does not mean every job in Japan is a bullshit job.
But the feeling of “the purpose is weak,” “the priority is unclear,” and “I do not know why this matters, but I still have to do it” is very close to the problem Graeber described.
What makes the Japanese version hard is that it is not necessarily idle.
It can be very busy.
- creating documents
- attending meetings
- reporting
- explaining
- adjusting formats
- handling last-minute changes
- coordinating stakeholders
- but not making real decisions
People are busy, yet the work does not feel like it is moving forward.
That drains people.
They are not unemployed.
They are paid.
But they sink inside the work.
That is a different pain from China’s youth unemployment problem.
6. Wage hikes alone do not solve the problem
Japan has recently seen strong wage-hike headlines.
Rengo has published 2026 spring wage negotiation results and updates.[^rengo] Early 2026 shunto figures showed wage hikes above 5%, continuing a high-level trend.[^jil-wage]
Reuters also reported that Japan’s real wages rose 1.0% year-on-year in March 2026, marking the third straight month of growth.[^real-wages]
But from the worker’s perspective, this can still feel like:
A wage hike forced by inflation, labor shortages, political pressure, and annual negotiations.
Wage hikes are not bad. They are necessary.
But they do not automatically fix the deeper issues:
- unclear purpose
- weak priorities
- little autonomy
- vague evaluation
- broken workplace systems
- good workers being used as duct tape
- inflation and workload rising faster than rewards
- more pay without more freedom or meaning
Wage hikes are necessary, but they are not sufficient.
Even if pay rises slightly, people may not feel rewarded if meaning, autonomy, accumulation, and freedom do not improve.
From the ground, it can feel like:
The ship is still sinking, but the seat cushion got a little thicker.
7. Japan’s realistic answer: use the company lightly, build assets with AI
The Chinese post can be summarized as:
The normal route is too heavy, so build another route with AI before taking on fixed costs.
How should that be translated into Japan?
In Japan, saying “quit your job, never date, never socialize, and go all-in on AI” is too risky.
Jobs exist.
Unemployment is low.
Salary from employment can still function as a safety device.
So the realistic answer in Japan is:
Use employment lightly as a safety base, while using AI to build assets on your side.
That means:
- do not invest your entire life into the company
- use salary as a safety device
- treat AI subscriptions as production equipment
- write articles
- build apps
- improve landing pages
- work on SEO and ASO
- write sales messages
- translate
- test
- accumulate outputs outside the company
If your work only gets absorbed inside the company, it may not reward you much.
But if AI helps you build articles, apps, landing pages, code, data, tests, and distribution channels outside the company, those can become assets.
8. Is $200 per month for AI expensive?
Yes, $200 per month is expensive.
But whether it is expensive depends on how it is used.
OpenAI’s ChatGPT Pro $200 tier offers 20x higher usage than Plus and includes advanced tools such as Codex.[^openai-pro]
Claude Max also offers up to 20x higher usage than Pro.[^claude-max] Claude Code can be used with Pro and Max plans.[^claude-code]
If this is merely “paying for a chatbot,” it is expensive.
But if used seriously, it can become a small personal development company, editorial team, and research lab.
It can act as:
- a coding assistant
- an editor
- a landing-page optimizer
- a translator
- a specification assistant
- a bug-checklist writer
- a sales copywriter
- a researcher
- a thinking partner
- an implementation assistant for personal projects
The question is not the subscription price itself.
The question is:
What remains outside your head after using it?
9. The trap: using AI becomes the goal
There is also a major trap.
The most dangerous failure pattern is when using AI becomes the goal.
A common failure looks like this:
- Pay for an expensive AI plan
- Talk to AI every day
- Feel like you made something
- Never publish
- Never sell
- Never test
- Never improve
- Burn tokens and stop there
This is an AI version of a leaking zoo.
You pour water in, but it leaks out from the bottom.
The real goal is not to use AI.
The goal is to turn AI output into published, sold, tested, improved, and accumulated assets.
For example:
- write an article
- save it as Markdown
- pass it to Codex
- reflect it in a landing page
- create app-store pathways
- aim for SEO traffic
- improve ASO
- post on social media
- check clicks and traffic
- return the learning to app improvement
Only when this loop exists does an AI subscription become an asset pipeline.
10. Do not take “never date” or “never socialize” literally
The most dangerous part of the viral post is the “do not date” and “do not socialize” part.
A better translation would be:
Do not spend on status-driven romance.
Do not waste time on draining social obligations.
Do not take on huge fixed costs too early.
Keep time and money for increasing your own productive capacity.
You do not need to cut every human relationship.
Good relationships should remain.
Keep relationships that:
- bring information
- help you recover
- allow you to build together
- do not destroy your freedom
- do not drain money through status or obligation
Reduce relationships that:
- force status consumption
- exist only as empty drinking sessions
- produce meetings without progress
- turn into endless complaints
- drain money and time
- offer neither growth nor recovery
That is the practical interpretation.
11. Conclusion: China has an entry problem; Japan has a sinking problem
The Chinese “go all-in on AI” argument is not just an extreme ideology.
In China, youth employment, housing, marriage, and living costs are heavy. The argument says: before you enter the heavy normal route, build another route with AI.
Japan is different.
Japan has low unemployment, and jobs exist.
But inside those jobs, many workers face:
weak purpose,
weak priorities,
weak autonomy,
vague evaluation,
low reward for effort,
wage hikes that do not necessarily increase meaning or freedom.
That is a different way of sinking.
So the Japanese version of the conclusion is:
Prioritize AI-based productive capacity over status-driven fixed costs.
But do not make “using AI” the goal.
Publish, sell, improve, and accumulate.
Use employment as a safety base, and build your own asset pipeline with AI.
A $200 AI subscription is just an expensive subscription if nothing comes out of it.
But if it produces articles, apps, landing pages, translations, sales messages, code, tests, and improvements, it can become a small personal development company, editorial desk, and research lab.
The important question is not whether you paid for AI.
It is:
Did the AI help you build something that remains outside you?
Sources
[^china-youth]: Trading Economics, “China’s Youth Unemployment Eases in April,” reporting China’s 16–24 youth unemployment rate at 16.3% in April 2026 and broader jobless rate at 5.2%. https://tradingeconomics.com/china/youth-unemployment-rate/news/552296
[^china-marriage]: State Council Information Office of China, “China's marriages down 20.5 pct in 2024,” citing 6.106 million marriage registrations. https://english.scio.gov.cn/pressroom/2025-07/31/content_118005719.html
[^japan-stat]: Statistics Bureau of Japan, Labour Force Survey results page, April 2026 results released on May 29, 2026. https://www.stat.go.jp/english/data/roudou/result.html
[^japan-unemployment]: Trading Economics, “Japan Unemployment Rate,” reporting Japan’s unemployment rate at 2.50% in April 2026, youth unemployment at 3.60%, and jobs/applications ratio at 1.18. https://tradingeconomics.com/japan/unemployment-rate
[^graeber]: David Graeber, “I had to guard an empty room: the rise of the pointless job,” definition of a bullshit job. https://davidgraeber.org/articles/i-had-to-guard-an-empty-room/
[^rengo]: Japanese Trade Union Confederation, 2026 Spring Wage Offensive page with wage negotiation updates and result summaries. https://www.jtuc-rengo.or.jp/activity/roudou/shuntou/index2026.html
[^jil-wage]: Japan Institute for Labour Policy and Training, “2026 spring wage negotiations first tally,” reporting 5.26% wage hikes in Rengo’s first tally. https://www.jil.go.jp/kokunai/topics/mm/20260327a.html
[^real-wages]: Reuters, “Japan real wages climb for third month in March,” reporting real wages rose 1.0% year-on-year in March 2026. https://www.reuters.com/sustainability/sustainable-finance-reporting/japan-real-wages-climb-third-month-march-backing-case-boj-rate-hike-2026-05-07/
[^openai-pro]: OpenAI Help Center, “About ChatGPT Pro tiers,” explaining Pro $200 usage allowance and Codex access. https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers
[^claude-max]: Claude by Anthropic, Max plan page, describing up to 20x higher usage limits. https://claude.com/ja/pricing/max
[^claude-code]: Claude Help Center, “Use Claude Code with your Pro or Max plan,” explaining Claude Code access through Pro and Max plans. https://support.claude.com/en/articles/11145838-use-claude-code-with-your-pro-or-max-plan
