When the Internet Had Doors

The Fires of History Newsletter, No. 9

You remember the sound. If you are old enough to have dialed into a BBS, you remember the sound the way you remember the smell of your childhood home. It is not a pleasant sound. It was never a pleasant sound. It is the sound of two machines screaming at each other in a language made of pure mathematics, and it was the most exciting sound in the world.

First: the dial tone. Then the touchtone melody of someone else’s phone number, played at machine speed. Then silence while the remote modem picks up. Then the scream – a carrier signal at 2100 Hz, a single high-pitched tone that says I am a modem, are you a modem? The answer comes back, and the negotiation begins. Chirps, hisses, warbles, bursts of static. Two robots arguing in an elevator. The future clearing its throat.

And then: silence. A cursor blinking on a black screen. You are connected. You are in.

One phone line. One user at a time. One universe.

If you are younger than forty, you have never heard that sound, never seen that screen, and never experienced what it felt like to be the only person connected to a computer that someone set up in their spare bedroom for the sole purpose of letting strangers call in and talk to each other. Which means you have been told a story about online community that starts in the wrong place, at the wrong time, and with the wrong assumptions.

A Snowstorm in Chicago

On January 16, 1978, a blizzard hit Chicago. Among those trapped indoors were Ward Christensen and Randy Suess, both members of a local computer hobbyists’ group. Christensen had been thinking about a computerized message system – something like an automated answering machine where club members could leave announcements. The blizzard gave them the time. Christensen wrote the software. Suess built the hardware. On February 16, 1978, the Computerized Bulletin Board System – CBBS – went online.

Simple concept. A personal computer, a modem, a phone line. You called the number, left a message, hung up. Other people called the same number, read your message, left their own. Asynchronous communication between strangers mediated by a machine in someone’s basement.

Stating this now feels absurd, like explaining that a door can be opened and also closed. But in 1978, the idea that a private citizen could operate what amounted to a public communication service from their home was genuinely novel. Communication was something provided by institutions – the phone company, the post office, the newspaper. Christensen and Suess demonstrated that communication could be provided by a guy with a computer in his basement.

Ward Christensen never sought fame. He worked at IBM his entire career. He died on October 11, 2024, at the age of seventy-eight. The most consequential act of digital community-building in history was a side project by a quiet man in a snowstorm.

The God of a Tiny World

If you ran a BBS, you were the SysOp – the System Operator – and the title was not honorary. The SysOp owned the hardware, paid for the phone line, chose the software, approved new users, moderated the message boards, and could read every private message on the system, because all the data lived on the SysOp’s hard drive, in the SysOp’s house.

Mark Zuckerberg, in his capacity as the person who controls what two billion people can and cannot say on Facebook, is exercising a form of power that was invented by a teenager in 1985 running a BBS on a Commodore 64 who could boot you off the board for being annoying. The scale is different. The structure is identical.

New callers started at the lowest access level – restricted time, limited privileges. The SysOp reviewed their application. Some SysOps called you on the phone to verify you were a real person. Imagine Twitter doing this. If satisfied, the SysOp promoted you. If not, you stayed at probationary level or got denied entirely.

This was not onboarding. This was initiation – the same structural logic that esoteric traditions have used for millennia. You do not get inside until someone who is already inside decides you belong. The SysOp was not processing an account request. The SysOp was deciding whether to let you into their space. A single human being who knew the community, knew the users, and made judgment calls. It did not scale. That was the point.

Speed as Selection Pressure

300 baud. That was the beginning.

At 300 bits per second, text arrived on your screen at roughly thirty characters per second. You could read it as it appeared. A 100-kilobyte file took over forty-five minutes to download. You did not download files at 300 baud unless you were very patient or very committed.

Each speed increment changed the culture. At 300 baud, only text was practical, which meant only people who could write had anything to offer. The selection pressure was brutal and specific: it filtered for literacy. No images. No audio. No video. Just text, arriving thirty characters per second, written by people who had been selected by the medium itself for their ability to put words in order.

As speeds increased – 1200, 2400, 9600, 14.4K, 28.8K, 56K – the filter weakened. More people came online. More was possible. Each increment brought participants and expanded capability, but it also diluted the original pool of people who had been selected by the pitiless constraint of slow modems for their ability to write.

This is not elitism. This is physics. A medium that can only transmit text will be dominated by people who are good at text. The early internet was disproportionately populated by people who could write, because writing was the only thing the medium could do.

But here is what the accessibility-is-always-progress narrative misses: the friction was generative. Fifteen minutes to download a file at 2400 baud meant you chose carefully. The busy signal meant you wanted it. The modem handshake meant you had cleared a technical bar just to show up. Every barrier to entry was also a filter for investment, and investment is what produces communities worth belonging to.

The Community You Could Not Escape

The social dynamics of a BBS were unlike anything on the modern internet, and the difference is one word: local.

Long-distance phone charges in the 1980s and early 1990s were brutal. Most BBS users called boards in their local area code, which meant the community on any given BBS was drawn from a geographic radius of roughly twenty to fifty miles. Your fellow callers were your neighbors. Not metaphorically. In a literal, “I could drive to your house” sense.

On a BBS with fifty to two hundred regular callers, all from the same metro area, everyone knew everyone. Your handle was your identity, and that identity accrued reputation over time. Helpful? People knew. A jerk? People knew. Uploaded a virus? People knew, the SysOp knew, and the SysOps of other boards in the area knew, because SysOps talked to each other.

Getting banned from a BBS was not like getting banned from Twitter. On Twitter, you lose access to a platform with hundreds of millions of users, but you can create a new account in thirty seconds. On a BBS, getting banned meant losing one of perhaps two or three boards in your area code. Your contacts, your message history, your door game characters, your reputation – gone. And word traveled. A ban on one board could lead to bans on others. In a town with three BBSes, getting kicked off one of them was losing a third of your digital social life.

The result was accountability that was personal rather than institutional. You behaved not because an algorithm would flag your content but because the person running the board was a real person, probably local, possibly someone you would see at the next meetup.

Many BBSes held real-world gatherings where callers met in person. The guy you had been arguing with in General Discussion was a human being standing in front of you holding a beer. Hard to be a complete jerk to someone when you might share a pizza with them next Saturday.

This is not a new social technology. This is how villages work. The BBS, by accident of technological constraint, recreated village dynamics in digital space. Nobody designed it that way. The phone company did.

FidoNet: The Poor Person’s Internet

Tom Jennings solved the isolation problem. A BBS was an island – you dialed in, read local messages, and hung up. Want to talk to someone on a different BBS in a different area code? Pay long-distance charges that could turn a teenager’s phone bill into a family crisis.

Jennings built FidoNet: store-and-forward networking. A BBS would collect outgoing messages during the day, then at 4 AM – during what came to be known as National Mail Hour, when rates were cheapest – dial another BBS, exchange messages, and hang up. Through a chain of nightly phone calls, a message posted in San Francisco could reach Sydney in a matter of days.

By 1995, FidoNet listed approximately 39,000 nodes worldwide. Two million users. Entirely volunteer-run. No company owned it. No one profited from it. Thousands of SysOps donated their time, their phone lines, and their electricity to keep the network running.

If you want to understand what a decentralized, user-owned communication network looks like – the thing cryptocurrency enthusiasts have been trying to build for fifteen years – skip the whitepapers. Look at FidoNet in 1990. It already existed. It ran on goodwill and modems.

What Was Lost

Between 1994 and 1998, the BBS ecosystem collapsed. At peak in 1994: an estimated 60,000 BBSes in the United States alone, serving approximately 17 million users. By 1997: roughly 10,000 and falling fast. The web offered access to everything, everywhere, all at once. AOL carpet-bombed every household in America with free trial disks. The transition was fatal.

Every account of this period treats the BBS-to-internet shift as progress. In many ways, it was. The internet was faster, more capable, more connected, more accessible. But what was lost was not a technology. It was a social structure.

The BBS was local. The internet is global. The BBS was personal – run by a human with a name and a phone number. The internet is corporate – terms of service, shareholder obligations. The BBS was small – fifty, a hundred, maybe two hundred people who all knew each other. The internet is billions of strangers by default.

Every problem social media companies spend billions trying to solve – harassment, misinformation, radicalization, content moderation, community management – was solved on BBSes. Not perfectly. Not at scale. But solved, in the sense that a person in charge knew the community and made decisions. The BBS solution: a human who knows the situation exercises judgment. The internet solution: an algorithm that does not understand the situation exercises pattern-matching. One does not scale. The other does not work.

The nostalgia trap is real, and I am aware of it. The BBS era was not a golden age. SysOps could be petty tyrants. The barriers to entry excluded the majority of people. The world was disproportionately white, male, and middle-class, because those were the demographics with home computers in the 1980s. The internet’s universality is its great achievement.

But what was lost was a model of online community that was human-scale, personally accountable, and architecturally constrained in ways that produced social dynamics we have not replicated since. Not because we lack the technology. Because we lack the phone bill.

Before venture capital, before the attention economy, before “engagement” became a metric, the first online communities were funded by individual humans who paid for them because they wanted them to exist. A 386 in a spare bedroom and a conviction that it would be interesting to let strangers call in and talk to each other.

This is not nostalgia. This is what community looks like when it is not a product.

This essay draws from Lurk More, coming fall 2026.

Next week: Trolling is not bullying. The difference is not a matter of degree. It is a difference in kind, and the media’s refusal to learn the distinction has made both problems harder to solve.

Source URLs