the shape of the internet
protocols, alexander galloway, bluesky
Here is a short post about why I like Bluesky. (I also wrote about antimemetics for Dirt, the only place that would publish such a thing!)
The internet has an original shape, and it’s not the bells and whistles of platforms that we see today. It is an architecture that incited near-spiritual practitioners, one that appeared to be unique from the top-down or bottom-up control of the real world, the control instead running horizontally on protocols that were necessary to adhere to—HTTPS, HTML, and so on—in order to connect various parts of the web together securely, for various webpages to speak with each other, in order for anything online be able to operate at all.
That meant that every website had an entrance and an exit. They still do, today, even if the destinations they lead to lay fallow, ravaged by a lack of attention and capital. Protocols are, as Alexander Galloway writes, “synonymous with possibility.” They are part of what animates the internet, what gives it a distinctive shape. They still exist, some powerful, others like artifacts.
Substack uses IMAP and SMTP protocols to ensure that this email goes out to any address, whether it’s Gmail or Outlook, and so that it can be opened on any server, whether it’s Apple Mail or the Gmail app. If, in the future, a single email provider and server come to completely dominate the market, there might be a chance we won’t ever need to use these protocols again. They can go obsolete. But they will always exist as a possibility in the un-aging but obscured internet.
This happened with the protocol RSS, which, in the early 2010s, used to act as a reader for websites; a person could add different websites to their RSS feed and visit it to get updates on what they had published. Then, Google Reader, the most popular and powerful RSS reader thanks to the company’s search abilities, shut down in 2013. (This is a good reflection from the team who built it).
Over the next decade, outside of a few tech-savvy internet evangelists, the existing online world and the rest who were coming online, transitioned to closed social network feeds, which were more accessible than what had come before, all of the spiky, technical bits hidden by fancy UXes. In these new social media feeds, we could follow people and have the algorithm suggest content for us; we didn’t have to seek out publications and add them manually. Things were easy. Platforms tracked what people clicked on, what their eyes lingered on, and evolved to meet them. In the early 2000s, you might have to manually add publication URLs to your RSS feed. With social media, you got one-click follows, comments and likes, feeds sorted by relevant algorithms.
RSS moved to the background, save for those places that still used it for infrastructural reasons and the few hold-outs. In this way, bit by bit, the internet became easier to use and understand. More people came online, and the previous hope of everyone getting a bunch of Legos that they snapped into place however they wanted lost its power.
Of course, the new internet undid some of the unevenness between coder and person. You no longer needed an interest in the technical to get into it. Protocols were, as Susan Leigh Star writes, “an embedded strangeness, a second-order one, that of the forgotten, the background, the frozen in place.” They demanded static decisions, like IMAP, which ensures that every email is only assigned to a single folder. In order to change that, Gmail creates labels, which don’t export as easily to other places. Protocols become a nuisance, a background thing to work around.
And they acquiesced, because protocols are weak. They are not a law or a transaction; they do not need to be listened to; they are merely a polite ask from the makers of the internet to keep the place working together; or they are a convention that companies adhere to if they are still so small that they still need to cooperate with others. If one provider comes to own email—not likely any time soon, but a possibility—no one needs to listen to these protocols again. The possibility of entrance and exit is shut off.
In 2016, I used to blog and report for Tech in Asia, a Y Combinator-backed startup trying to do a TechCrunch for Asia, with writers in Pakistan, China, India, Indonesia. I was, like many young people, chock-full of opinions, but with little of the wisdom and controls that come with age.
I wrote about India banning Facebook’s Free Basics, an app that was going to provide free internet—to be precise, Free Facebook’s Apps only—to the one billion or so Indians who hadn’t yet come online, who were stymied by various reasons, including the fact that data was still expensive.
The problem with Free Basics was that it would convince those who were new to the internet that all there was to see was Facebook’s feed and algorithm. They wouldn’t be coming on to a distributed internet with a maze of entrances and exits, but instead on to a company’s space.
The ban happened, but it wasn’t enough. Bans, what Akash Kapur refers to as reflexive statism, haven’t always worked well on the hydra of the distributed web. Facebook returned to providing a light version of its app, meant to work well in India. This light app needed very little data because it ripped out all of the structured HTTP formatting—the protocol that lets browsers and servers communicate—and only retained the TLS layer, or the S you see in HTTPS, effectively creating a custom protocol that didn’t interact with the larger web. The light app didn’t have a fancy UX, but it still grew quickly; in nine months, it hit 100 million users active each month, making it the fastest-growing version of Facebook at the time.
Network effects: the more people that sign on to a platform, the more powerful it is. A few months later, along came Mukesh Ambani, India’s wealthiest man (Beyonce and Rihanna have performed at his kids’ weddings), who launched his 4G network, fighting tooth and nail with other companies, eventually crashing data prices until they were dirt cheap. Facebook’s network effects compounded, earning the company its biggest market in the world. In 2024, India had nearly 370 million people using Facebook, or around 25% of its population.
What we had feared had come true; as always, the big guys had won. Many people came online, not to the internet, but to a specific social media site. All they knew was what it felt like for content to be targeted and offered to you. It was surprising that the polite controls of the web were, at the end of the day, so weak.
I was recently catching up with a cousin who lives in a small village off Mangalore; he was lamenting phones. Then, he told me a theory about why he’s grown so addicted to his. Facebook now has content for him, whereas it didn’t before. Personalized, targeted, made exactly for him, understanding him better than anyone else. The more he invests in watching this content, the more he wants to watch season 1, 2, 3. There are people to follow to expand the content. The longer he watches, the less of a reason he has to get off.
A shape is the only way to describe the original internet. “A distributed network is always caught…au milieu, meaning that it is never complete, or integral to itself,” writes Galloway. As long as there are protocols that have to snap into other places, the web’s life goes on.
But what we use today is not the web. It’s more like enhanced television, with lots of interactivity. There are blocks of text and video and images in the same place. We can comment back to people previously out-of-reach, drag or flirt with our favorite influencers. This last ability is getting less and less possible, the “for you” feed dominating the “following,” comments to influencers becoming more like calling into a radio show and getting our voice on the air.
X/Twitter is the new Fox News. When AI makes it cheap enough to have an around-the-clock feed, when we no longer need humans to produce original content, when we no longer need to spend money moderating these humans, I imagine the New York Times will create their own feed, along with various others—CNN, Kai Cenat, Cenk Uygur and Hasan Abi, maybe, until we have a series of hyper-engaging channels to flip through and comment on and act as public-access celebrities on.
Underneath all of this, protocols still exist on the internet, like some buried ultramarine structure that has sunk down next to the internet’s pipes. Most might not know they are there, but they feel its presence. It is what older millennials refer to when they say that the internet used to be more fun. When protocols ran the show, when companies had to interact with each other because no one place had a ton of power, things used to be roomier. It is what they say when the internet used to seem like much more than an entertainment channel, before it was a place to watch large-scale spectacles from the sidelines, fires and genocides and entire governments being redone and placed into the hands of the unelected.
These protocols might not retain even their weak power for long. Various parts of the world have been advocating for more nation-owned internets, clamping down on protocols that allow access to the world; domain names—everything that comes after HTTPS:// (like substack.com), which are vertically controlled—are entering increasingly heated fights between ICANN and various governments; maps are another grounds of dispute, a place to uselessly rename bodies of water in the thin Arial font of Google Maps and act as if these decisions lead to any material impact.
Which brings me to why I like Bluesky, the X alternative that over 30 million people use, including some who have left old-Twitter. Bluesky—”Bluecry”—has many haters. Some jeer at its culture, its conservative moderation policies, conjecture at the snowflakes who will come back to Twitter/X if its’ culture changes. Still others think it’s foolish to invest in another social media platform that has raised venture capital and might have platform ambitions. It is neither owned by no one, like Mastodon, nor is it collectively owned by everyone. It runs on the AT Protocol, which theoretically lets people build out who and what they follow. But the protocol hasn’t yet evolved to allow this to happen, which, even now, has created a Twitter/X-inverse, and not much of its own original culture.
But the old internet presented plenty of alternatives to big platforms, and this choice still wasn’t enough to drive people outwards. The only thing that seems to win is a culture and a story, like Bluesky being a way to get off of Twitter/X. Bluesky is a good way to remember the internet’s original shape while it still exists, a resting place to stop at before we move on to whatever is next—something that is stronger, more robust, long-lasting, impossible to control by the power-hungry.
I miss RSS feed so much. And as millennial the idea of feeling the presence of protocols still is funny but resonates
this is super interesting, thank you