我有个朋友，工作八年，公司从 300 人减到 80 人

Let me be clear: this isn't my own story, it's a friend's.

His name is Lao Zhou (pseudonym), a backend developer with eight years of experience, age 31 this year.

I used to think that 35 was the cutoff for programmers. After hearing what he went through over the past six months, I'm more inclined to believe that this cutoff just arrived early.

1. Eight years ago, he also believed technology was his foundation

He entered the industry in 2018.

Back then, the internet wasn't as cold as it is now, and people who could write code were genuinely in short supply. After training, his first job was at an outsourcing company. The pay wasn't high, but there was plenty of work. In a year, he saw everything from CRUD operations to production incidents.

Later, he jumped to his current company, starting from writing APIs and eventually taking charge of a business line.

For eight years, he kept chasing: Spring Boot, microservices, Docker, K8s, message queues, database partitioning, and then cloud native. Every time he switched tech stacks, he felt he'd gotten a little thicker.

He truly believed the phrase "experience is a moat."

He even made a checklist for himself: how many production incidents he'd resolved, how many modules he'd refactored, how many new hires he'd mentored. He calculated that as long as he learned a little more each year than the last, by age 30 he'd be worth something.

The calculation didn't pan out.

2. Technology didn't depreciate; its scarcity did

When ChatGPT first got popular, he was still stubborn, saying AI couldn't write complex business logic and could at best help with a regex.

Later he used Cursor and then Claude Code. An API endpoint, including unit tests, comments, and API docs, came out in an afternoon. Previously it would've taken him two days, plus he'd have to align on fields with the frontend three times.

He said the hardest part wasn't that AI wrote faster than him, but that he found the things he was proud of — the pitfalls he'd stepped in, the instinct for locating problems, his familiarity with a certain framework — AI and recent grads could do seven out of ten as well.

The act of writing code itself is becoming non-scarce.

What's becoming scarce is something else: the ability to think through requirements clearly, to judge whether what AI produces is usable, and to take an idea from zero to something people actually want to use.

But the problem is that this shift can't happen overnight. He spent eight years hammering out code line by line. Now he has to relearn how to work with AI, and he's somewhat resistant.

He also asked me a question: if what he learned over all those years is depreciating fast, what were those eight years worth?

I had no answer.

3. That day he walked into the meeting room, it was a Friday afternoon

The company went from 300+ people to 80+ in less than two years.

His team went from 12 people down to 4.

At 3 PM on Friday, HR sent a message on WeChat Work: "If possible, come to the meeting room." When he walked in, the Leader was already sitting there, with a printed document on the table.

He'd seen others get laid off before. Back then he'd instinctively think: probably their ability wasn't good enough, or their attitude was a problem, or the company was really failing.

When it was his turn, he realized none of those were the real reasons. It had to do with which business line you were on, which slot you occupied on the cost sheet, and less to do with whether your performance rating that year was A or B.

n+1, two rounds of negotiation. His hand was steady when he signed.

After leaving the meeting room, he didn't go straight back to his desk. He slipped into the stairwell and sat on the steps for over ten minutes. He didn't dare go straight home that day either. He bought a rice ball at the store downstairs and sat for an hour before leaving.

4. What truly panicked him wasn't the unemployment itself

He had some psychological preparation for losing his job. He was watching the industry situation every day.

What kept him up at night was a few figures.

Mortgage: over 8,000 RMB a month, with 22 years remaining.

His child just turned two, starting kindergarten next year.

And there was a more hidden ledger: a large part of what he'd accumulated over the past eight years was internal to this company. Why this permission system was designed this way, why this business chain took this route, why this legacy code can't be touched. Outside this building, how much of that could he take with him?

He used to think "knowing the business" was valuable. Now he understands there are two kinds of business knowledge: one that becomes worthless as soon as you leave the company, and one that's still usable across industries. He'd hoarded too much of the former.

5. He started using AI to build something of his own

He was at home for two months. He sent out over 30 resumes, interviewed at four companies: two blocked on education, one on age, one went well but the position was frozen.

In his spare time, he used AI to make a small tool.

His wife works in cross-border e-commerce and gets headaches calculating first-leg shipping costs and tariffs every day. So he wrote a mini-program that takes weight, volume, and destination and automatically outputs a quote.

He finished it in two weeks, launched it, and posted it in several seller groups.

Now it has over 200 users, generating less than 300 RMB a month.

He said the meaning isn't the money. It's the first time he went through the full loop from requirements, development, launch, being criticized, to iteration — whereas before, that required a team.

But he also clearly said: AI doesn't equal making money. Writing something nobody uses is the same as not writing it.

6. How he thinks now

He has no comeback story, and no plan to turn a side project into a full-time income. He's taken on two freelance gigs, waiting for interview results while continuing to maintain that mini-program.

After several conversations, there are a few things he repeatedly mentions.

First, he no longer chases a checklist of new technologies. He used to have hundreds of saved articles, but never built anything from them. Now he only learns what solves his current problem, especially AI — he throws it straight into a project and sees how much time it saves.

Second, he no longer defines himself only as someone who writes code. He now looks at costs, traffic, and why users install and then uninstall. These used to be product manager concerns, but now he feels they're his business too.

Third, keep a path that you can take with you. Not necessarily starting a business, nor necessarily becoming a content creator. It can be a transferable skill, a small thing that people actually use, or a reputation in a certain field.

Fourth, don't be pushed around by anxiety. Seeing others succeed in career changes makes him anxious. Seeing others make money with AI products makes him anxious. Seeing people who haven't found a job in six months also makes him anxious. But everyone's mortgage, family, and technical background are different. Other people's paths are for reference, not necessarily for copying. He now sets a self-check every three months: compared to three months ago, has he actually made a change?

7. Eight years, and still under thirty-five

He's 31 this year. By conventional wisdom, he still has four years before the so-called cutoff.

But his feeling is that this cutoff isn't about age. It comes with the company's cash flow, with the industry's position, and with whether the capabilities in your hands can be carried away and whether anyone will pick them up once you do.

One day he said something to me that I still remember:

"In the past I thought that as long as I kept learning, I'd never be left behind. Now I know that whether I learn is up to me, but whether I can stay is not."

I'm not sure if that counts as clarity.

Finally, I want to ask you readers: how many years have you been working? Was there a moment when you suddenly realized you should prepare a second path?

I'm not selling anxiety, nor forcing optimism. We're all having a hard time. Being able to speak more truthfully is probably more useful than anything.