2000: The Bubble Burst, the Net Stayed
· 43 min read · Views --

2000: The Bubble Burst, the Net Stayed

Author: Alex Xiang

Essays

In 1999, I wrote about bank front-end systems and sockets.

That was a hard kind of experience: SCO UNIX, production systems, transaction links, timeouts, logs, and failure boundaries. It taught caution. It also taught that systems are not toys.

In 2000, the scene changed abruptly.

I came to Beijing.

The city changed, and so did the air around the industry. The word “internet” was no longer only a novelty inside technical circles. It arrived together with capital, media, recruiting, stock options, portals, IPOs, layoffs, and all kinds of rumors. That year, I talked with CSDN about an opportunity, but the salary did not meet my expectation, so I eventually joined a foreign company.

That sounds plain now.

But in 2000, it carried a very typical smell of the time: the internet seemed to make anything possible, and everyone was judging where they should stand. Some people went to portals, some joined foreign companies, some chose startups, and some stayed in traditional software and industry systems. Everyone knew the network mattered. Very few people really knew what it would become.

The bubble had already inflated.

It was also about to burst.

Beijing Felt Like a Larger Machine

When you move from one city to another, technical work can change its flavor.

Shanghai Jiao Tong University, the early 386, MIS systems in offices, and bank front-end machines were all concrete experiences. In Beijing, many things suddenly became “the industry.” You could see more companies, more opportunities, more technical media, more arguments about the future of the internet. Offices, job advertisements, technical forums, portal sites, and cubicles in foreign companies turned like gears in the same machine.

At that time, Beijing did not feel merely like a workplace to me. It felt like a larger machine.

It absorbed programmers, product managers, editors, salespeople, marketers, investors, and every kind of imagination about the future. Some believed portals would become the entrance to everything. Some believed e-commerce would change retail. Some believed online communities would change content. Some believed enterprise software was still the steadier direction.

Looking back today, each of them was only partly right.

The internet did change everything, but not in the exact way that seemed most exciting then. Many companies disappeared, many concepts changed names, and many business models were reshuffled. But the network itself did not retreat. Browsers, email, search, IM, forums, and portals continued to grow, and eventually became a more complicated internet life.

Bubbles burst. Infrastructure remains.

That may have been the most important lesson of 2000.

The CSDN Episode

In 2000, I had contact with CSDN.

For programmers of that generation, CSDN was a familiar name. It was not just a website. It was an entrance into a technical community. Articles, forums, downloads, discussions, recruiting, and technical news made many people feel that “programmers” were becoming a visible group.

I did not join CSDN in the end because we did not reach agreement on compensation.

But my relationship with CSDN did not stop there.

Later, I applied to start the Linux board in the CSDN forum and became its first moderator. A forum moderator was not a formal title in any corporate sense, but it had real meaning inside a technical community: organizing discussions, maintaining order, answering questions, and helping a direction gather people. Linux was hot then. Many people installed it for the first time, configured drivers, fixed Chinese-language environments, compiled kernels, or built servers, and they went to forums for experience.

I also received a CSDN MVP once.

Today, these experiences are harder to put on a formal resume than full-time jobs, but they represent the relationship between programmers and technical communities in that era. Much knowledge did not come from courses or company training. It accumulated through forums, articles, replies, and repeated troubleshooting. To me, CSDN was not only a company where salary negotiation failed. It later became one of my entrances into the Linux community.

That thread continues in the next essay, because around 2001 the heat around Linux in China truly rose.

The Light Around Portal Websites

In 2000, portal websites were bright.

Sina, Sohu, and NetEase were the entrances of that era. You opened a browser, entered a portal homepage, and read news, finance, sports, entertainment, email, and forums. Search had not yet completely become the daily entrance, and mobile internet was still far away. A portal was like a large square: it put a little bit of everything there and wanted to carry a little bit of everything.

The heat around portals also gave internet work a fresh glow.

People doing content, editing, technology, and marketing suddenly entered the same industry story. Site traffic, advertising, listings, financing, page redesigns, and channel operations became high-frequency words.

I remember an editor from Sina saying that the compensation he received could cover an entire bed.

I remembered that sentence for many years.

It may not represent the full truth of the industry, but it captures the absurdity of that moment. The internet bubble was not an abstract curve. It landed on people: someone received compensation, someone lost a job, someone missed an opportunity, someone changed industries, and someone stayed to build the next round.

The easiest mistake in a bubble is to confuse capital-market heat with technical and commercial maturity.

In reality, a website that opens does not prove a business model. Fast user growth does not prove reliable revenue. An IPO does not prove that an organization can operate for the long term. A full office building does not prove that a product has a moat.

All of this would be proven again and again later.

But in 2000, when you were inside it, it was not easy to see clearly.

The dot-com bubble and internet infrastructure around 2000

Bubbles burst. Servers, networks, browsers, and user habits remain. This is a period-inspired illustration.

Beyond the NASDAQ Curve

When people talk about the 2000 dot-com bubble today, the NASDAQ curve is usually the first image.

It peaked in March 2000 and then fell sharply. Many internet companies lost market value, financing tightened, and layoffs increased. That curve matters, of course. It records the capital market’s sudden cooling toward the internet dream.

But if we only look at the curve, we miss the other half of the truth.

The internet did not disappear after the bubble burst.

On the contrary, the crash was a brutal filter. Many invalid business models were eliminated, and many companies supported only by stories collapsed. But the truly useful foundations continued to move forward: broadband adoption, server growth, Web development, better search engines, email and IM as daily tools, and online communities gathering users.

The valuation burst, not the network.

That sentence is easy to understand today. It was much harder then. When a company lays people off, an industry cools down, and a concept becomes a joke, many people assume the entire direction was wrong.

History often works differently.

A direction may be right while the first companies are too early, too rushed, too expensive, and too bad at making money.

This would happen repeatedly later: mobile internet, the sharing economy, blockchain, the metaverse, and AI all followed similar rhythms. The technology is real, imagination runs too far ahead, capital punishes a group of people through cooling, and finally the part that can enter daily life remains.

The year 2000 was one of the earliest collective training sessions.

The Stability of a Foreign-company Office

I eventually joined a foreign company.

A foreign-company office felt different from an internet startup. Processes were clearer, the organization was more mature, and the work style was steadier. It might not have been the most exciting place, but for a programmer, stability itself had value.

Software work at that time did not have as many remote collaboration tools or cloud services as today. The office was the work center. Computers, intranets, email, meeting rooms, printers, English documentation, process tables, and project plans made up daily life.

Moving from bank front-end systems into this environment let me feel another kind of engineering order.

Production systems teach you to respect failures. Corporate processes teach you to understand collaboration. The former cares about machines and links. The latter cares about roles and responsibility: who raises requirements, who confirms scope, who tests, who signs off, who releases, and who maintains.

Young programmers often underestimate process.

I later felt more and more that process does not replace ability. It makes ability repeatable. A strong individual can put out a fire, but a team cannot run on firefighting forever. Some foreign-company processes were cumbersome, but they at least showed that software projects could be organized more systematically.

This later influenced how I thought about platforms, larger teams, and cross-department projects.

The Software World Did Not Stop in 2000

In the same year that the bubble burst, technology continued.

Windows 2000 was released and pushed the Windows NT line to an important point for enterprise desktops and servers. Mac OS X Public Beta appeared, bringing NeXT’s technical lineage into the future of the Mac. Java, the Web, and server-side development continued to evolve. The open-source world was also growing, and Linux, Apache, PHP, and MySQL appeared more often in developers’ fields of view.

In other words, capital markets cooled down. Technology itself did not.

If programmers only read the news, it is easy to be dragged around by industry emotion. But if you look at tools, systems, communities, and real users, you see another slower and steadier line. Many important changes are not completed at the loudest moment. They keep accumulating after the bubble withdraws.

That is why I do not like writing 2000 only as “the year the bubble burst.”

It was certainly that year.

But it was also a watershed. The internet moved from “everything can be narrated as a story” toward “everything eventually has to prove itself.” Traffic had to prove value. Products had to prove retention. Business models had to prove revenue. Technical teams had to prove stable delivery.

This requirement was harsher than the bubble years, and healthier than the bubble years.

From Production Responsibility to Industry Judgment

The front-end system work in 1999 taught me that systems carry production responsibility.

Beijing and the internet bubble in 2000 taught me that careers also carry judgment responsibility.

When writing sockets, errors appear in logs. When making career choices, errors often reveal themselves only years later. Technical problems can be reviewed. Industry choices can also be reviewed. Neither can eliminate uncertainty completely.

Putting these years together, my path is interesting.

In 1996, I saw the split between low-level development and the Web.

In 1997, I made a webpage move.

In 1998, I put forms into databases.

In 1999, I connected sockets to production systems.

In 2000, I began placing myself inside a larger industry current.

Programmers do not only relate to code. They also relate to cities, companies, industry cycles, capital markets, and organizational structures. Coding ability is important, of course, but it is not the whole story. The stage of the industry you stand in, the kind of organization you join, the people you work with, and whether you choose stability or a bet all change the path afterward.

The year 2000 made me realize this more strongly for the first time.

Next: The Golden Age of Linux

After 2000 came 2001.

The internet bubble cooled, but another thread became clearer: Linux and open source. The golden age of Linux in China was unfolding. Distributions, communities, companies, and technical media were all active. Linux then was not only a server operating system. It also carried a kind of idealism and an imagination around domestic software.

That thread had actually been planted back in 1995, with the Slackware disc on my 386.

In 2001, it would appear again.

The next essay is about the golden age of Linux in China.

IT Milestones of 2000

  • The dot-com bubble burst. In March 2000, the NASDAQ Composite reached a peak and then fell sharply. Internet-company valuations dropped, financing tightened, layoffs and failures increased. But the crash did not end the internet. It pushed the industry back from stories toward real business and infrastructure.
  • Windows 2000 was released. Microsoft released Windows 2000 and pushed the Windows NT technical line further into enterprise desktops and servers. Its stability, networking, and directory capabilities influenced many enterprise environments.
  • Mac OS X Public Beta appeared. Apple released Mac OS X Public Beta in 2000, bringing the NeXT lineage more clearly onto the Mac. Later macOS and iOS ecosystems can trace part of their roots to this line.
  • Apache Ant emerged. Around 2000, Ant became an important build tool for Java projects. It represented the engineering side of Java enterprise development and foreshadowed the importance of build, dependency, and automation tools.
  • Tomcat became an important Java Web foundation. Around 2000, Tomcat became a key open-source container for Servlet/JSP development. Java Web and enterprise application development entered a more active stage.
  • Nupedia, the predecessor to Wikipedia, went online. Nupedia launched in 2000, and Wikipedia would appear in 2001. The direction of collaborative knowledge, open editing, and online encyclopedias was already forming.
  • Baidu was founded. Baidu was founded in Beijing in 2000. It later became one of the most important entrances to Chinese search and trained many engineers in search, advertising, distributed systems, data, and algorithms.
  • Internet companies shifted from expansion to survival. After 2000, many companies could no longer talk only about growth. They had to face revenue, cost, cash flow, and organizational efficiency. This shift would repeat in later technology waves.

References

More in this column

View all posts in "Thirty Years in IT and Me"