Web3's Eternal September

Jun 10, 2022

Usenet was a networked discussion system built in 1980 that predated WWW and lived alongside the Internet.

Usenet used UUCP – Unix-to-Unix Copy – a (roughly)1 peer-to-peer (p2p) protocol that lets users transfer files, email, and text between computers.

Users could subscribe to threaded articles posted to Usenet, organized into different newsgroups, e.g., comp.software, misc.education, talk.politics. Articles would be pushed and propagated to all subscribers. Think of Reddit's subreddit concept in Google Groups form.

Moderators curtailed spam in the most popular newsgroups. Otherwise, in the first decade of Usenet, the small community of early computer enthusiasts established and followed online norms quickly.

Then Eternal September came.

It's all moot now. September 1993 will go down in net.history as the September that never ended. – alt.folklore.computer

There was one exception to the close community of Usenet. Every September, an influx of new college students were given access to Usenet through their University. Experienced users complained of low-quality posts and users that didn't follow the rules (the acronym FAQ was first used on Usenet).

Around 1993, ISPs started to offer Usenet to their subscribers. AOL provided a Usenet service in 1994. After that, the influx of users never stopped.

A similar phenomenon would happen when Facebook opened its network to high schoolers in September 2005 (I joined the following September, on my first day of high school). College students wanted nothing to do with the high schoolers on the platform (their version of Eternal September). Now everyone, including our alt-right third-cousins, is on Facebook, and we want nothing to do with them.

But Usenet couldn't handle the influx of users like Facebook did. In addition, spam was a significant issue since authentication was easily bypassed, and there was no way to track bad actors on the platform (Sybil attacks?).

Usenet increasingly became a destination for pirated software (that's why I joined2). Big binary files stressed the proto-p2p network. Legal issues followed – it was impossible to remove content once uploaded (see: the Blockchain).

Services like Deja News3 and, later, Google started to index and archive all Usenet posts and put them behind a searchable interface. Finally, AOL offered a gateway from the Internet to Usenet.

Usenet was clunky and difficult to use – the Internet was proliferating with better UX (Marc Andreessen launched the Mosaic Web browser on Usenet).

Usenet eventually capitulated to TCP/IP, SMTP, and the Internet. But, unfortunately, by then, it was too late.

Web3 faces many of the same issues as Usenet.

  • Privacy and data issues built into the protocol
  • Adjacent services offer alternative ways to consume and use the content (Dune Analytics, Chainalysis)
  • Bad UX and developer experience
  • A destination for illegal activity (although, like Usenet, I believe that there's much more to it than that)
  • An influx of new users that don't share the same norms as the early adopters

Pundits have compared the adoption of web3 to the early Internet. Maybe that's true, but has web3 had its Eternal September yet?

Is web3 Usenet or TCP/IP?

Are things different this time?

1Usenet's peer-to-peer (p2p) design was very different than p2p today – both uploaders and downloaders were not identifiable from the network once complete. In most p2p schemes, you can identify a peer by its IP (Usenet wasn't TCP/IP!). So, practically, it worked closer to the client/server model.

2I'm too young to have been a Usenet user in its peak form. By the time I had joined as a kid (late 90s), it had already migrated to the Internet and was a ghost town. Interestingly enough, traffic on Usenet continues to grow year over year.

3Deja News was acquired by Google in 2001. It became the basis for Google Groups.