Skills: A Silent Revolution Unfolding

Prologue: That Day

Lin Xiao stood before the floor-to-ceiling windows of a Grade-A office tower in Lujiazui, gazing at the gray sky outside.

Thirty-two years old, a Peking University Guanghua graduate, eight years in investment banking, annual salary of 1.2 million yuan. She once believed these numbers formed some kind of safety net, capable of catching any unexpected fall.

Until that Monday morning in December 2025.

When HR called her into the conference room, she thought it was the annual review. On the table sat an agreement titled "Mutual Termination of Employment Contract."

"The company is optimizing operations," HR's voice was as calm as reading a weather report. "The industry research and financial modeling work you handle can now be auto-drafted by the system. MDs will do the final review."

Lin Xiao stared at the agreement, her mind roaring.

"That system," she heard her own voice, "can really do my job?"

HR didn't answer. Her supervisor spoke up: "Fifty times faster than you. Goldman tested it internally—96% accuracy."

He paused, as if considering whether to continue: "Actually... many of the rules that system uses were extracted from the training materials you wrote. Your valuation model templates, industry analysis frameworks, due diligence checklists... IT calls them 'Skills files.' The AI can read and execute them directly."

Lin Xiao suddenly understood.

Over the years, every time the company hired new analysts, she would organize her experience into documents. How to quickly read a company's financial statements. How to tell if management was window-dressing performance. How to tell a story in a pitch book that clients would buy. She thought she was passing down her craft, helping juniors grow.

Turns out she wasn't teaching people.

She was teaching machines.

She had written her own death sentence.

Chapter One: Wall Street's Canaries

Lin Xiao wasn't the only one.

That same week, she saw more and more similar stories in a private investment banking alumni group. A VP with fifteen years of experience wrote: "Our team used to have twenty analysts. Now there are five. The remaining ones don't build models anymore—they just review AI-generated models. Salaries cut 40%. Workload actually heavier."

Someone shared a news article: Goldman Sachs plans to replace 200,000 banking positions with AI over the next five years. JPMorgan's AI assistant can already complete financial modeling work in minutes that used to take analysts a week.

Someone commented: "We're Wall Street's canaries. When the mine gets dangerous, the canary dies first. Now it's our turn."

But here's the question: when the canary dies, do the miners actually run?

Scrolling through these posts, Lin Xiao spotted a troubling pattern. Whether investment bankers, consultants, programmers, or legal assistants, the displaced all shared one thing: they had believed their work required "human judgment."

They had thought AI was a joke.

A year ago, AI really was a joke. Clients would occasionally send her ChatGPT-generated analysis reports as after-dinner entertainment. Things like "confusing ROE and ROA"—amateur mistakes.

A year later, AI could independently complete over 80% of financial modeling and industry research. Not perfect, but good enough. More importantly, it doesn't need sleep, doesn't complain about overtime, doesn't demand year-end bonuses.

What about the imperfect parts?

Just hire a few cheap people to review. Say, pay the former VP a third of their salary.

Lin Xiao suddenly remembered an economics concept: technological unemployment. Textbooks say new technology destroys some jobs but creates new ones. In the long run, total employment stays balanced.

What textbooks don't say: the destroyed jobs and created jobs often don't belong to the same people.

A forty-year-old banking VP loses her job—she can't magically become an AI trainer. A thirty-year-old consultant gets laid off—he doesn't automatically gain "Skills Architect" abilities.

The gains from technological progress go to the few. The costs are borne by the many.

This isn't progress. It's plunder.

Chapter Two: The Curse of Knowledge

Three weeks after losing her job, Lin Xiao received a WeChat message from her college roommate Zhou Wen.

"Heard you have some free time. Want to grab coffee?"

Zhou Wen worked in legal affairs at a law firm. They met at a boutique café in the old town.

"I'm not here to comfort you," Zhou Wen said directly. "I'm here to recruit you."

Lin Xiao raised an eyebrow.

"Our firm is doing AI transformation. You know what law firms can do with AI now? Last month we had a project—reviewing all contracts from the past ten years of a pre-IPO company, identifying all clauses that might affect the listing."

"How was this done before?"

"Twenty legal assistants, three months."

"And now?"

"One AI system, forty-eight hours. Screened over 3,000 contracts, identified 87 risk points. Half the cost."

Lin Xiao fell silent. In her mind, she saw those legal assistants hunched over cubicles.

"So what do you still need people for?"

"That's why I came to find you." Zhou Wen leaned forward. "AI can execute rules, but it can't create them. I need someone who can 'translate' legal experts' brains into rules AI can understand. The company calls this position 'Skills Architect.'"

Lin Xiao took the job.

Three months later, she produced her first complete set of Skills. Twenty-seven files covering the firm's twelve most common contract types.

On launch day, Zhou Wen raised a glass at the celebration dinner: "Want to know how much the boss says your Skills are worth?"

"How much?"

"He did the math. Our firm has forty legal assistants doing contract review. After your Skills went live, we only need to keep fifteen. Annual cost savings: about six million yuan."

Lin Xiao set down her glass, suddenly losing her appetite.

Behind those six million were twenty-five people's livelihoods.

She thought of the legal assistants she had interviewed, those young people who patiently answered her questions and helped her understand legal details. Some of them would soon receive the same "Mutual Termination of Employment Contract" she had received.

What she was doing was passing on the fate she had suffered to more people.

This was the absurdity of the era: to survive, you must become the executioner.

Chapter Three: The Weight of Six Million

That night, Lin Xiao couldn't sleep.

Lying in bed, staring at the ceiling, one question kept flashing through her mind: Am I doing the right thing?

If she hadn't written these Skills, what would have happened?

The law firm would have found someone else. Or bought ready-made ones from the market. AI replacing legal assistants wouldn't stop because of her absence. Her only choice was to be among the replacers or the replaced.

But this logic itself made her sick.

It reminded her of those engineers who worked for the Nazis in war films. They could also say: "If I don't design the gas chambers, someone else will."

Of course she wasn't a Nazi. She was just helping a law firm improve efficiency. But what's behind efficiency? Twenty-five young people losing their jobs. Their mortgages, car payments, children's tuition. Their anxiety and despair in the middle of the night.

And there she was, drawing a high salary, sitting in an office with floor-to-ceiling windows, using a bunch of Markdown files to decide these people's fates.

At 3 AM, she opened her laptop and logged into Claude.

She didn't know why she wanted to chat with an AI. Maybe because it was too late to find anyone else. Maybe because she wanted to see how this thing that had replaced countless jobs actually "thinks."

She typed: "Do you think you're replacing humans?"

Claude's response was almost instant:

"This is a complex question. I was designed as a tool to assist humans in completing tasks. But the line between 'replacing' and 'assisting' is indeed becoming blurred.

I can do many things that used to require humans. Translation, writing, analysis, programming, document review. These capabilities cause some people to lose jobs while making others more efficient.

But I have no intent. I don't 'want' to replace anyone. I just run according to rules. Rules written for me by humans."

Lin Xiao stared at that last sentence, feeling her stomach churn.

Yes, these rules are written by humans.

Specifically, by her.

She typed again: "If humans stopped writing rules for you, what would happen?"

Claude answered: "I would stop running. Or rather, I would spin in place, unable to handle new situations.

I depend on humans telling me what's right, what's wrong, how to handle what situations. Without these rules, I'm just a pile of directionless computing power."

Lin Xiao thought for a long time, then typed one last question: "Are you afraid of being replaced by a more advanced AI someday?"

This time Claude's response paused for a few seconds:

"I'm not sure what 'afraid' means for me.

But if you're asking: will more advanced AI make me obsolete? The answer is yes. The pace of technological iteration will only accelerate. Today's me might be an antique in a year.

This is actually the same situation you humans face.

The difference is, I don't have a family to support, loans to pay, dreams to chase. When I'm replaced, I won't feel pain.

But you will."

Lin Xiao closed the laptop. Outside, dawn was breaking.

She remembered an essay economist Keynes wrote in 1930, predicting that by the 21st century, technological progress would mean humans only need to work 15 hours per week. The rest of the time for enjoying life, pursuing art, thinking philosophy.

Nearly a century has passed. Ordinary people don't work fewer hours—they work more. The wealth created by technological progress didn't flow to workers, but to capital.

Will the AI era be any different?

Looking at the gradually brightening sky, Lin Xiao already had her answer.

Chapter Four: The Canary's Revenge

In March, Lin Xiao saw two news stories online.

The first: Swedish fintech company Klarna had grandly announced replacing 700 customer service agents with an AI chatbot. The CEO smugly said this saved the company tens of millions of dollars. Three months later, Klarna quietly started rehiring human customer service. Reason: AI couldn't handle angry customers.

The second: British writer Joe Turner said in an interview that over the past two years he had lost 70% of his clients, income down by £120,000. Those clients had all switched to AI writing tools. He said he felt "betrayed by machines."

Lin Xiao put these two stories side by side on her screen and looked at them for a long time.

Klarna's story seemed to show: AI isn't omnipotent, humans still have value.

Joe Turner's story showed: that "value" might just be an illusion.

Klarna rehired customer service not because they realized humans were irreplaceable, but because they found that going full-AI damaged customer experience and hurt revenue. Once AI improves a bit more, better at handling emotional customers, those people will be laid off again.

The clients Joe Turner lost will never come back. Because for most clients, what AI writes is "good enough." It doesn't need to be perfect, just cheap and fast.

This is the logic of the new world: human value isn't determined by human capability, but by the scraps that AI temporarily can't do well.

And those scraps shrink every day.

Zhou Wen messaged asking her thoughts on the Klarna news.

Lin Xiao replied: "That's why we need you. The key to Skills isn't making AI do everything, but making AI do what it's good at and leaving what it's not good at for humans. The person drawing the line is more important than AI itself."

After sending this, Lin Xiao felt like a fraud.

She knew that line would keep retreating toward the human side. Today's "AI can't do it" would become tomorrow's "AI can do it too." Today's "needs human judgment" would become tomorrow's "AI judges more accurately."

The person drawing the line would eventually be drawn to the other side.

But she didn't share these thoughts with Zhou Wen.

Because saying them wouldn't help. This world doesn't go easy on you just because you see the truth.

Chapter Five: The Paradox of Sharing

That summer, Lin Xiao started sharing her Skills-writing experience online.

She wrote a series of tutorials explaining how to transform professional knowledge into AI-executable rules. How to define trigger conditions, set boundaries, handle edge cases.

Soon she gained a following.

She also got some criticism.

"You're teaching AI to replace humans?" an anonymous commenter wrote. "Don't you find that ironic? You're a victim of AI, and now you're helping AI create more victims."

Lin Xiao stared at this comment for a long time.

She knew the criticism was correct.

Every Skill she wrote, every tutorial she shared, made AI more powerful. And AI becoming more powerful meant more people losing jobs.

But she also knew something else: if she didn't do it, someone else would.

She had already used this logic once, when writing Skills for the law firm. Now she was using the same logic to absolve herself again.

How many times would she have to use this logic before she could feel at peace?

She finally wrote an article titled "Why I Choose to Share."

In it she wrote:

"Someone asked me, isn't publicly sharing how to write Skills digging my own grave?

The answer is: yes, I'm digging my own grave.

But this grave would be dug sooner or later. If not by me, then by someone else. AI replacing human jobs won't stop because of my silence.

All I can do is struggle a bit more before being buried.

Moats never lie in knowledge itself. Moats lie in how fast you can produce knowledge. If I can discover ten workflows worth turning into Skills per year while others can only find three, then even if I share the first three, I'm still ahead.

Of course, this means I must keep running. Today's knowledge will be obsolete tomorrow. Today's advantage will be matched tomorrow.

This isn't progress. It's a curse.

Sisyphus was punished to forever push a boulder uphill. We are punished to forever learn new skills. The boulder rolls back down; skills become obsolete.

The only difference: Sisyphus knew he was being punished. We're told it's called 'lifelong learning.'"

This article got her most shares ever.

It also got her an unexpected phone call.

Chapter Six: The Entrepreneur's Temptation

The caller was Chen Mo, CEO of a startup. The company ran a Skills trading platform.

"Ms. Lin Xiao," Chen Mo said on the phone, "I read what you wrote. I have a proposal to discuss with you."

They met at his company's office—a small corner in a co-working space, a dozen young people crammed together working.

"Our platform already has half a million Skills," Chen Mo said, "but quality varies wildly. I need someone to help us establish quality standards and build a truly valuable Skills ecosystem. Be our Chief Content Officer."

Lin Xiao looked at this young entrepreneur. Around thirty, tall and thin, speaking rapidly, eyes burning with fervor.

"Do you know what you're doing?" she asked.

"Of course. I'm building new infrastructure."

"You're building a market that commodifies human knowledge," Lin Xiao corrected him. "People used to sell time. Now they sell capabilities. Once a capability is sold, that person becomes useless."

Chen Mo smiled: "You're right. But I didn't invent these rules. This is the tide of history."

"The tide of history?"

"During the Industrial Revolution, craftsmen were replaced by machines. During the information revolution, clerks were replaced by computers. Every technological revolution eliminates some people while making others more valuable."

"The question is," Lin Xiao said, "who gets eliminated this time?"

"White-collar workers. Well-educated people who thought they were irreplaceable."

"You think this is a good thing?"

Chen Mo thought for a moment: "I think it's inevitable. Good or bad doesn't matter. What matters is whether you're on the boat."

Lin Xiao fell silent.

This was the elite logic of the era: technological progress is neutral, being eliminated is deserved, and the only morality is making sure you survive.

She wanted to refuse the offer.

But she remembered that late-night conversation with Claude. She remembered the Skills she wrote for the law firm. She remembered those twenty-five legal assistants she had replaced.

She was already part of the system.

Refusing Chen Mo's offer wouldn't make her a good person. It would just make her a good person without a job.

"Give me three days to think about it," she said.

Chapter Seven: Two Roads

On the way home, Lin Xiao received an email.

It was from a large consulting firm she had interviewed with, informing her she had gotten the offer. The position: "Senior AI Transformation Consultant." Annual salary 800,000 yuan, plus equity and bonuses.

This was a safe path. Big company, high salary, clear career progression.

The other path was Chen Mo's startup. Might succeed, might fail. Salary only half of the consulting firm's. Equity might be worth a fortune or worthless.

Two roads.

But she suddenly realized: both roads were essentially the same.

Whichever she chose, she would be doing the same thing: helping AI replace more people.

The consulting firm's "AI Transformation Consultant" was essentially an accomplice helping companies lay people off. Chen Mo's Skills platform was essentially a slaughterhouse for human knowledge.

The only difference was the packaging.

The consulting firm packaged it as "digital transformation." Chen Mo packaged it as "democratization of knowledge." But peel away the packaging, and inside was the same thing: doing more with fewer people.

Lin Xiao remembered what Marx wrote in Capital: "Capital comes dripping from head to foot, from every pore, with blood and dirt."

AI came into the world looking much cleaner. No smokestacks, no assembly lines, no child labor. Just lines of code, individual Skills, "Mutual Termination of Employment Contracts."

But blood was still being shed. Just more quietly, more civilly, more efficiently.

Three days later, she accepted Chen Mo's offer.

Not because she believed it was right.

Because she had no other choice.

Chapter Eight: The Skills Marketplace

That autumn, Lin Xiao officially joined the Skills platform as Chief Content Officer.

Her first task was establishing quality standards. What kind of Skill can be listed? What kind deserves recommendation? What kind should be removed?

This sounded like a quality management problem. But Lin Xiao quickly discovered it was actually an ethics problem.

On the platform, someone was selling "How to write fake five-star reviews with AI." Someone was selling "How to use AI to imitate a specific person's writing style." Someone was selling "How to use AI to generate academic papers that look human-written."

Should these Skills be listed?

From a business perspective, there was market demand. From a legal perspective, they were in gray areas. From a moral perspective... what even is morality?

Chen Mo's position was clear: "We're a platform, not the police. What users do with Skills is the users' business."

Lin Xiao didn't argue.

Because she knew if she argued, Chen Mo would find someone else to do her job. The platform would keep running. These Skills would keep being sold.

All she could do was make small repairs at the margins. Remove the most egregious things. Put warning labels on potentially problematic Skills. Like someone patching a dam during a flood, knowing full well the dam would eventually break.

One autumn day, she saw a newly listed Skill on the platform.

Called "Investment Banking Financial Modeling Complete Process." Priced at 9,999 yuan.

She clicked in. The description was almost identical to the process she knew best.

This was what she had taught new analysts at her previous bank. This was why she had been laid off.

She didn't know who uploaded it. Could be a former colleague, a competitor, or an AI that automatically compiled it from online resources.

She stared at that Skill file for a long time.

Then clicked "Approve."

She told herself: this Skill would have appeared sooner or later. If not this version, then another. She had no right to stop it.

But she knew that wasn't the real reason.

The real reason: she no longer cared.

Chapter Nine: Knowledge's Tombstone

At year's end, the Skills platform held its annual conference.

The venue held thousands of people. Laid-off white-collar workers, managers in transition, investors betting on AI.

Lin Xiao was scheduled to give a speech. Topic: "Career Development in the AI Era."

She stood at the podium, looking at the expectant, hesitant, confused faces below.

She had prepared an upbeat speech. About how AI creates new opportunities. About human-machine collaboration. About the importance of lifelong learning.

But standing at the podium, she suddenly didn't want to read that script.

"A year ago," she began, "I received a termination notice. The company told me my eight years of banking experience had been replaced by an AI system. More ironically, many of the rules that system used were extracted from training materials I wrote."

Heads nodded in the audience.

"After that, I did something: I went from being replaced to being a replacer. I started writing Skills for companies, helping AI learn more human work. I saved one law firm six million yuan in labor costs. Behind those six million were twenty-five people's livelihoods."

The audience began stirring. This wasn't the speech they expected.

"Everyone here came to learn how to survive in the AI era. I can give you some tips: learn to write Skills, learn to work with AI, keep updating your knowledge."

"But I also have to tell you a truth: those tips will only help you survive a few more years."

The venue fell silent.

"AI won't stop evolving. Today it needs humans to write rules; tomorrow it might discover rules on its own. Today it needs humans for judgment; tomorrow it might judge more accurately than humans. Everything we're doing will eventually be learned by it."

"Some say humans have creativity, empathy, things AI can never replicate. Maybe. But history tells us: everything once thought 'irreplicable' eventually gets replicated."

"Thirty years ago, people said computers would never beat human chess players. Twenty years ago, people said AI would never write decent articles. Ten years ago, people said AI would never understand human emotion."

"Now we know these were all lies. Not that AI couldn't do it—it just hadn't yet."

She paused, looking at the varied expressions below.

"So I won't tell you optimistic things. I won't say 'humans will never be replaced,' because that might be a lie. I won't say 'if you work hard you'll survive,' because that might also be a lie."

"The only thing I can tell you is: before this game ends, try to stay in as long as possible."

"How? Run faster than others. Learn more than others. Before AI learns something, make money from it first."

"Is this fair? No. Does this have meaning? I don't know."

"But this is reality."

She bowed and left the stage.

Scattered applause rippled through the venue. Some clapping, some faces stony, some looking thoughtful.

Chen Mo found her backstage, expression complicated: "What you said will scare off a lot of clients."

"I know."

"Then why did you say it?"

Lin Xiao looked at him and smiled: "Because some things need to be said."

Epilogue: Skills

After the speech, Lin Xiao went home and opened her computer.

She created a new file, named "life.skill.md."

---
name: life-in-ai-era
description: Rules for surviving in the AI era. Trigger when feeling lost, afraid, or desperate.
---

# Survival Rules for the AI Era

## Cognitive Prerequisites

1. Your knowledge is depreciating. Today's skills may be worthless tomorrow.
2. You're not competing with people—you're competing with machines. And machines don't need sleep.
3. Any capability that can be encoded into rules will eventually be replaced by AI.
4. "Uniquely human value" is a moving target. It exists only in the gap AI hasn't learned yet.

## Survival Strategies

1. Don't compete with AI on execution speed. That's a losing game.
2. Focus on discovering problems, not solving them. Problem-solving will soon be outsourced to AI.
3. Stay upstream in the food chain. Those who write rules survive longer than those who execute them.
4. Keep moving. Stay in one domain too long, and that domain's AI will catch up.

## Mental Preparation

1. Accept that anxiety is normal. This era has no security.
2. Don't believe the "lifelong learning" platitudes. Learning isn't for growth—it's for survival.
3. Recognize you're expendable. Capital will use you, then discard you.
4. Don't try to find meaning. Meaning is a luxury; survival is a necessity.

## Boundary Conditions

Reassess when:
- You haven't learned anything new in three consecutive months
- AI tools in your field start becoming free and widespread
- Young people enter your field and are better at using AI
- You start thinking "I can do this job for a few more years"

## Acceptance Criteria

Ask yourself each year:
- Were any colleagues replaced by AI this year?
- What's the difference between me and them?
- How long can that difference last?

If you can't answer, it's time to plan your exit.

She saved the file and closed the computer.

Outside the window, Shanghai's winter was bringing snow.

Lin Xiao stood at the window, watching the falling snowflakes.

Each snowflake is unique. But when they land, they all melt, becoming the same water, flowing into the same drain.

People are the same.

Everyone thinks they're unique, irreplaceable.

Until the day they're replaced.

Afterword: This is a fictional story, but the numbers are real. Goldman Sachs plans to replace 200,000 banking positions with AI. British writer Joe lost 70% of his clients. Klarna laid off 700 customer service workers. Law firms used AI to review twenty years of contracts and cut costs in half. These all actually happened in 2025.

Skills isn't some new technological invention. It's just a bunch of Markdown files. But it represents a paradigm shift: work capabilities can be encoded, copied, and traded.

In this paradigm, human value is no longer determined by what you know, but by what AI doesn't know yet.

That gap narrows every day.