I.

In the winter of 1994, in a small apartment in Hong Kong, a fourteen-year-old boy plugged a US Robotics Sportster 14,400 Fax Modem into his family's telephone line. He had read about something called a "bulletin board system" in a PC magazine. He didn't fully understand what it was. He dialed anyway.

The modem screamed. Static. Then text appeared on his screen. A welcome message. A list of names he didn't recognize, discussing topics he barely understood.

He was terrified. He was hooked.

Over the next two years, that boy spent hundreds of hours on BBSes. He made friends he never met in person. The person running each board—the "sysop"—was usually just some guy with a spare computer. But that guy was there. You could page him. He knew your name.

That boy was me. And I've spent three decades wondering why nothing since has felt quite the same.

II.

Here is what we're told about the internet: it got better. The web replaced BBSes. Social media replaced the web. Every metric improved. Speed. Reach. Convenience.

Facebook has three billion users. Twitter has five hundred million. These numbers are presented as achievements.

But on a BBS with fifty users, you knew people. Not their curated personas—them. The sysop wasn't an algorithm optimizing for engagement. He was a person who might show up at the same coffee shop next Tuesday.

We traded that for scale. We traded community for audience.

And we did it for a reason.

III.

Running a BBS in 1994 was hard. You needed technical knowledge, a dedicated phone line, hardware that could stay on all day. But a motivated teenager could manage it.

Running your own community platform today? Authentication, databases, real-time messaging, moderation, security, deployment. The effort required vastly exceeds what any individual can invest.

So we surrendered to the platforms. Not because they're better, but because the alternative became impossible.

IV.

In 1865, economist William Stanley Jevons observed something counterintuitive. Steam engines were becoming more efficient. The sensible prediction: coal consumption would decrease.

The opposite happened. Consumption increased—dramatically. Why? Efficiency didn't just make existing uses cheaper. It made new uses possible. Resistance that had bottlenecked adoption simply vanished.

This is the Jevons Paradox. And it's happening right now in software development.

V.

I joined a hackathon to test a hypothesis: how much of a complex project could AI handle with minimal human intervention?

I delegated almost everything—concept development, documentation, architecture, tech stack selection, over eighty percent of the coding. I stayed at the strategic level and touched as little code as possible.

Four weeks later: a working BBS. WebSocket terminal with ANSI graphics. AI-generated ASCII art. AI-powered games. Docker deployment. The complete stack.

It's not the finished product I envisioned. But it works. And I am not an exceptional developer with unusual resources. I just had access to AI tools that have only existed for two years.

VI.

AI-assisted development doesn't just make existing projects faster. It changes what's worth building.

That custom tool you couldn't justify? Buildable now.

That community platform for your specific weird needs? Achievable now.

That intimate corner of the internet that died when the platforms took over? Resurrectable.

The barrier that pushed us all onto Facebook is collapsing. Not slowly. Rapidly.

VII.

There's an irony here. The most advanced AI in the world is what made it possible to rebuild the most primitive form of social networking. Text on a screen. Fifty users who know each other. A sysop who cares.

AI handled the tedious parts—infrastructure, debugging, moderation—so humans can focus on what matters: genuine community.

The technology has changed beyond recognition. The thing we're reaching for hasn't changed at all.

VIII.

I'm going to return to that apartment in Hong Kong.

The boy with the modem thought he was just playing with computers. He was actually learning what community could feel like—small, personal, real.

The platforms taught us a different lesson. They taught us that community means scale, and that the alternative was impossible.

But the impossible is becoming possible again. The tools that didn't exist two years ago now let a single person build what previously required a team.

The modem screamed. The connection was terrible.

And it was the best community I've ever been part of.

Maybe it's time to dial back in.

Baud Again was built for the Kiroween hackathon. The code is open source. The modem sounds are optional but recommended.