The Case for Internet Regression

Last Updated: 2025-08-26 00:00:00 -0500

In a tale as old as time, it’s common among people in my rough age cohort - the Millennials and younger end of Gen X - to lament the pre-“social media” internet. I think the key problem with that is just a matter of phrasing, and a carried implication that the internet pre-social media wasn’t social - in fact, it very much was. While things like myspace and facebook pages, twitter feeds, and instagram accounts lowered the barrier for entry to having a presence on the internet from “must know a bare minimum of web design” to “has an email address”, that doesn’t change the fact that the internet before the rise of the “Five Sites” was a vibrant and highly social place.

The Internet as a Third Place

If you spend enough time in conversations about topics like urbanism, city planning, and community building, it won’t be long before learn that we are running out of unmonetized third spaces - places to be that aren’t work or home (but instead a secret, third thing) which also don’t carry with them the compulsion of a purchase. But for most of the time it has existed, that’s what most of the internet actually had been. Yes, even the “1.0” web had commerce attached to it (that’s supposed to be the point of the .com TLD, after all), but for everyone you had trying to hawk books or pet food at you, you had a dozen or more instances of hobbyist and scientist pages like AmaSci, Hacker Friendly, or the Texas Triffid Ranch.

As a rule, the internet can’t ever be the truly unmonetized third place; it’s not a physical location, and but for a very lucky few of us, the safe money is on the idea that you will have to pay for your connection. But in terms of serving the role of a third place, a junction for random conversation and leisure outside of work and the mundanities of domestic life, it’s already serving that purpose for a shocking number of us. That’s why the companies behind the “Five Websites” internet are all such enormous economic drivers that we keep letting them get away with outright violations of law, justified ex-post-facto with new regulation.

But of course, time spent on their sites isn’t the same as time spent in a genuinely unmonetized third place. When you’re not busy being the product, they’re constantly shoving Temu and Wish into your face to get you to buy junk products you don’t need and don’t work as advertised, or charging you a subscription fee for the privilege of being advertised to.

It could be - and has been, in the past - much better.

Places to Loiter

In his video about surviving the isolation of a COVID-19 Lockdown, CGPGrey referred to finishing the streaming-video internet as a “Mixed Achievement-Get”. Even now, five years on, King of the Hill riffed on the idea in its latest season, joking that “If you finish Netflix, you get something called a ‘wellness check’.” And yet, I’d hazard a guess that anyone with a smart phone has experienced the peculiar annoyance of having access to all 175 Zettabytes of God’s Own Internet and yet there is nothing to see.

This is decidedly a problem created by the Five Sites and ecosystem and the ongoing (largely LLM-aggravated) enshittification of web search. We have become, universally, accustomed to being handed interesting content on a few narrow platforms that are, one way or another, supposedly curated to our interests. Gone are the days where a bored person could type something ridiculous like “science prank” into a search engine and wind up stumbling into a Jargon File story. Instead, you’ll be handed some LLM-output, and then a few pages of SEO-hyper-optimized generative content talking about Science Pranks that never happened because they would violate a few laws of thermodynamics.

It didn’t used to be that way, and in part because we were neither required to have well-trained vector models of ourselves to do algorithmic feed-filling, nor fully bare our souls to algorithmic processing, nor get lucky on a google search. There was a secret, third way to find stuff: web rings and “interesting links” pages.

It turns out, if you spend enough time tooling around on, I don’t know, KeelyNet (unfortunately, no longer extant), you probably have interests in common with the site maintainer. And the site maintainer, having the distinct privilege of being not you, is going to be aware of things to read, watch, or listen to that you have never thought to look for. If they had the courtesy to pull together a collection of their favourite links - possibly even just because it was the pre-accounts-for-browsers era and this was a convenient way to access your bookmarks from more than one machine - then that was available for you to fall down rabbit holes of your own.

To me, the big surprise of my thirties hasn’t been that I have a fairly enormous amount of browser time logged every week - I work at a desk, after all - it’s that somehow I spend almost all of that time on Five Websites, rather than the widespread wandering that used to be the case.

Expression, not Commerce

I cannot remember the last time I was able to load up a website that wasn’t mine and not see an advertisement. It doesn’t matter if it’s a platform I pay for content on, either directly through a paid subscription, or indirectly through my tax dollars (looking at - and judging - you, CBC), it’s hard to swing a GET request and not get a whole bunch of crap you don’t care about shoved in your face ahead of the content you actually want.

To be clear, my objection isn’t advertising per se. Covering the cost of publication by selling space in the publication to people who want your readership to know about something is almost as old as print media - certainly as old as newsprint and magazines. It’s the way the advertising stack works.

See, in the modern internet, I, a publisher trying to cover my costs, do not directly partner with you, a person with something to sell, to trade some of your advertising budget for some of my bandwidth and some of my reader’s time. I am selling some clearinghouse a few column-inches for them to run anything they want. And that’s a problem for a few reasons. Firstly, I have little if any control over what you see, but you’re still going to associate that ad with e. Second, it means that I implicitly participate in the mass trackerization of the internet by serving ads based on unnecessary tracker cookies, which is how on Jimmy’s House of Sprockets you’re getting ads for something you just saw on Instagram or just searched for on Google or Bing. Thirdly, advertiser clearinghouses of ill repute - or sometimes even of decent repute - have become the proverbial watering hole for watering-hole malware attacks, where a malvertisement is served and potentially impacts every user who comes across my site.

I don’t want to contribute to that. Nobody I know wants to contribute to that, and nobody I’d take very seriously would suggest that anything less than an up-to-date ad blocker with a current set of definitions is anything less than table stakes for securing your browser. Unlike the VPN you’ve seen four ads for - including sponsored spots by your favourite youtubers or twitch streamers - since you logged in this morning.

(Admittedly: this model is slightly different than the model used by sponsorship programs. Sponsorships and affiliate programs allow a lot more control on the behalf of the end publisher over who gets shown what. That’s why I disclosed in my blog post about using the Selae logic analyzer as part of the PETI development works was a sponsored position, and why I’ve considered setting up various affiliate programs in the past. Because again, the problem isn’t promotion, or even promotion for pay - it’s an unchecked, market-wide adoption of some of the most manipulative and abusive technology the advertising industry has ever developed. If all the ads were good-faith recommendations for relevant goods and services that presented themselves with a minimum of intrusion, I wouldn’t bother running an ad blocker.)

Merit and Discovery, Not Gameable Algorithms

I wrote earlier on the problem of reaching “the end of the internet” or “finishing netflix”, and I firmly believe this problem is rooted, fundamentally, in the way we’ve come to use the internet. We are all, myself included, overly reliant on the Five Sites and their ability to shove a constant stream of “content” into our faces. I first noticed the problem when I changed from being a regular user of Twitter to a regular user of Mastodon instead. Since Mastodon uses a recent-to-past feed that is comprised entirely of who you are following and what they have posted or reposted, it’s actually shockingly easy to “run out” of mastodon to read. We’ve all had the experience, I assume, of staring at the youtube homepage looking at all the recommendations, and lamenting that one time we spent an afternoon working to some generic ambient techno mix, because now that’s all the machine thinks you want, so it’s all the machine will show you, and now you have to actually work up the motivation and self-awareness to decide what you’re interested in at that moment and search for it manually, which for some reason feels like an ordeal equivalent to being asked to use a library card-index catalogue to find a book. (If you’ve never done this, I highly recommend it).

And this is why the internet of my childhood, for all its pop-up advertisements, limewire viruses, and auto-playing midis (or, god help you if you’re on as slow a connection as I usually was, mp3s!) was inherently superior to the modern internet at being a third space: if there was nothing new on Wild Waldo’s Worldwide Widget Farm, you could just hop down to the next item in your list of 500 bookmarked sites, or grab one of their recommended links, and go find something new.

If you found this a pain in the ass, you had your RSS feed readers (for about 15 glorious seconds when they were actually relevant to mainstream use) that could aggregate the posts from your 100 favourite blogs into a nice inbox, showing you what you had and hadn’t read, and even reformatting them all to your taste.

And, best of all, if there was nothing new on your regular haunts, and nothing from their see-also pages was tickling your fancy, your actual friends had seen something you hadn’t. And even if they hadn’t, we had more to do than just fire requests at each other on 443/tcp. The internet thrived in its weirdest corners. Telnet connections spun across the world connecting millions of gamers playing games as compelling as any modern MMO or dungeon crawler using nothing more advanced than ANSI text encoding. IRC servers allowed humans to give each other film recommendations in real time and show them exactly which peer-to-peer connection was needed to get their hands on them. Games - full-blown, graphics-based video games - let you fight your friends with nobody’s permission needed other than a few licensed copies. That whole thing we have going now, where a company can decide a game is no longer profitable and just kill it’s online multiplayer community overnight? Practically impossible in my day. Worst you might lose is the matchmaking capability, but join the right IRC channel or the right forum and you’d have no shortage of people looking to waste time your favourite way, and enjoy it with you.

My preferred complaint about the internet is that the web forum is more or less dead. It’s not entirely true - they still exist in a few niche places - but it’s almost true in concept. When you make that complaint, fourteen of the worst people you know will come out and proclaim facebook pages or subreddits are the return of the original forum idea “but better”. But the feature they widely proclaim to be so superior - upvotes - fundamentally breaks the idea to have any kind of actual conversation in a forum thread. Instead, everyone’s just sounding off, trying to get the best zinger or the worst take, generating the most “reaction” in order to circulate their post to the top, for karma or whatever other purpose.

The insane thing is: all this technology still exists. We are the inheritors to not just the HTTP world-wide web, but the TCP/IP internet and all of its glorious strangeness. Bringing back the personal website is just the tip of an extremely large and exciting iceberg.

(This also shouldn’t be construed entirely as a RETVRN argument. I have things to say about the centralization of “instantaneous channels-and-large-groups text chatroom” under one vendor (discord), but “IRC was better than discord in every way” isn’t one of them.)

Personal Websites that are Actually Yours

Sometimes when I’m talking about Sanity Line I bring up the fact it used to be a web serial. The original run of Sanity Line and about 80% of From the Ashes were published as monthly chapter uploads on a website I used to run at that address. There was just one problem - I was a cook making less than a true living wage, and while it doesn’t cost much if you have a good salary, having a website like that - built on a platform with vendor lock-in - gets expensive fast.

Sure, it was easier for me to build that site at the time that way. I hadn’t touched HTML to edit it in about 15 years at that point and I certainly had no interest in learning anything like modern web design practices. But when it came time to cancel the plan and re-host by myself, I found that even though I legally owned all the content of all those web pages, I couldn’t just republish them. The vendor owned the license for the code their site-builder generated, and just to make sure I didn’t work around that, they even actively obfuscate the code when they served it to clients, and used a bunch of javascript to decode it and re-render it once it hit the browser.

So when I talk about bringing back the personal website, I don’t mean to knock services like the old Geocities or modern alternatives like Squarespace or Wix… but I do want to encourage you to look further afield. It’s easier than you think and you’re more capable than you give yourself credit for.

The Internet Fits in Your Closet (And Consumes No Water)

I trained for a while to become an accountant (but never bothered pursuing accreditation), and I still keep my finances as more-or-less “correct” books in the “bookkeeping” sense. That means it’s actually pretty easy for me to figure out how much I spend on my web presence. It’s… shockingly little, when you consider how much more web presence I have than most.

In total, I spend:

Less than a dollar canadian a month on AWS S3 Glacier Storage for Web Backups (relevant here because they include my entire web presence. And a whole bunch of stuff that isn’t “my web presence”, so let’s say 50 cents CAD per month.)
Exactly nothing more than a minute or two of my time to do a renewal every month on a dynamic DNS provider, to make sure that my DNS entries always point at the right object
My internet connection, which isn’t worth listing the amount for because you’d have one whether you were self-hosting or not.
$25 CAD/year on email hosting (because my ISP won’t allow residential clients to self-host SMTP servers)
$25 CAD/year on three DNS apex domains (which is two more than you’d need) to hold all my nonsense, and;
About $30/year worth of electricity running a computer I wouldn’t normally run to keep everything hosted on.

Now, a spare computer is a big ask. But that spare computer is no more powerful than a Raspberry Pi 4 with a decent-sized SD card in it. No fancy cooling setup. No sudden spike in your power bill. I literally spend more per year on pen ink and notebooks than I spend on keeping three apex domains and all their associated strangeness live.

A sufficiently motivated nerd with a cellphone they don’t use any more could probably even just use that. Me? I’m using a laptop my dad gave me on which the keyboard, display, and battery no longer work.

The self-hosted, self-service, self-created internet is a hobby for the cheap.

I Create As I Speak

Before it started getting watered down into a tabletop RPG ruleset, Tarnished Tale was going to be my attempt to recreate one of those fancy Telnet games - a Multi-User-Dungeon (MUD). But I wasn’t going to use Telnet. That was outdated, had some security issues, and my players deserved better. Instead, I started cobbling together all sorts of disparate technologies to make it sing:

Multithreaded python backend with a jobs queuing system and a truly absurd amount of async calls to act as the “game master”.
Web-based and standalone TUI (terminal user interface) clients to connect to the thing through TLS-secured Websockets connections, as good and bidirectional as a telnet connection but with all the security of modern cryptography.

And that’s not even the weirdest thing I’ve built for the internet.

But the magic is: you don’t have to be deeply skilled in the technology department to pursue self-hosted internet. The Tarnished Tale Wiki? Five changed lines in a template config file and a single command to boot it up. All well documented and approachable. This very site? It’s mostly managed in plain text.

I use a static site generator setup to convert paragraphs that look like this:

I use [a static site generator setup](\{\{absolute_url | 'guides/jekyllguide.html'\}\}) to convert paragraphs that look like this:

Into text with all the rendering needed. If you’re familiar with markdown from discord messaging, mastodon posts, or even using certain notepad apps, you can use an SSG and whatever templating is available in it to make beautiful, professional-looking sites. Or, if you’re like me and you like to tinker, you can build something slightly off-kilter, like what we have here.

Far from limiting you to the boundaries of your skill set, self-hosting your own websites takes the training wheels off, and lets you learn to truly express yourself as fluently as you do while speaking.

Here is Where I Put My Friends (Because I Have Some)

Here’s where it gets fun - if I have a website, and you have a website, and we find each other’s sites interesting, we could link between each other..

In the golden days thereof, this practice was known as web rings. Web Rings themselves weren’t perfect - they had some central administration and therefore tended to be either so permissively-joinable as to be meaningless as an indicator of “you might also enjoy”-nature, or so restrictive as to form a cool kids table. But where they really shined was their ability to take a reader from Site A directly to Site B. No central clearing house, no special configuration needed. Everyone in the web ring includes a little snippet of code that makes sure their membership in the web ring is displayed and that they, in turn, form their link in the chain from site to site.

Which, by the way, is how my affiliate system works.

And this is more than just an attention funnel - this is an alternative vision for “content discovery”. Instead of being driven by what keeps eyeballs on or off the screen, who who pays who for a slice of primacy, or how well your paid advertiser partner can “game” the search engines into thinking you are maximally relevant, it’s just a statement of mutual interest. “Oh, you enjoyed Dragonlance? Might I interest you in some R.A. Salvatore?”

Oh, and even better: if you’re using basically any sort of site generator (Jekyll, as I prefer, or Hugo, or what have you): RSS is free and usually just a configuration flag away, and RSS Aggregator Websites still exist!

These aren’t even the fun parts of the social, personal internet. Button Exchanges are coming back into vogue as what used to be my daily normal (and even my “new hotness”) becomes the “hip retro” all over again.

Hypertext Was Enough

Now, I was careful to say earlier that I’m not trying to make a “RETVRN” or luddite argument. I know it doesn’t feel like it, but right now you can get more computing power, for cheaper, consuming less energy, than at just about any other point in human history. And part of the reason it doesn’t feel like it is because holding open a browser tab with a youtube video in it costs 2 GB of RAM, not including your other seventy-two browser tabs.

And it really doesn’t have to be like that. Part of the reason I can get away with using a computer otherwise bound for the junk heap as my primary web server is because a static site is just built different. I host five static sites and nearly a dozen dynamic ones on one beat to hell server and I get away with it for two reasons:

Even on the social internet, any of my websites only consumes non-storage resources in the time between someone sending their request for it and them receiving my response, and;
Nothing is running that needs to “call home”, “check in”, “heartbeat”, or maintain an open connection.

There’s no infinite scrolling asking for new content every few seconds while the user keeps the page open in a tab. Apart from wiki pages, nothing is being dynamically rendered. There is only hypertext, and the tiniest possible amounts of javascript (to help the header render usefully on a mobile device, and to request the comments on a blog post, and only when the user clicks it.)

This simplicity helps me - because it keeps my internet small enough to fit on a spare shelf of my desk and consume less than $30 worth of electricity in an entire year - and it helps users because they could probably have one tab open for every page on the wisteria wiki and still have plenty of free RAM for their lo-fi chillhop, or whatever.

Community, Not Commerce

This isn’t an all-or-nothing approach, either. I firmly believe that “social-media” in terms of the Five Sites is here to stay - as a usage concept. Whether you stay loyal to the corporate socialnet or embrace the federated alternatives to it, though, the personal-website internet will be there to augment your experience. Social Media to make the connections that lead you to the fora, chat rooms, and personal websites of the communities you want to be a part of, rather than revenue-chasing, extremes-amplifying social media that’s trying to shape you into its own loyal consumer - and convert your attention into sales.

The internet was once a wild, weird place. And with your help, it could be again.