Off the Feed: Meet the People Building the Internet Without an Algorithm

Somewhere in a Discord server that doesn't show up in any recommendation engine, about 400 people are having a genuinely good conversation about urban gardening. Nobody got there because an algorithm suggested it. Nobody is being served ads based on what they said. There's no feed. There's no scroll. There's a list of channels, a set of community norms, and a pinned message that says, simply: "We found each other on purpose."

This is what the counter-internet looks like right now — and it's more organized, more intentional, and more populated than most people scrolling through their main feeds would ever know.

The Allergy to Algorithmic Discovery

To understand why people are building algorithm-free spaces, it helps to understand what they're reacting against. Platforms like Instagram, YouTube, Facebook, and TikTok are fundamentally discovery machines. Their primary job is to serve you content you didn't ask for, based on behavioral signals you didn't consciously provide. The scroll is the interface. Engagement is the currency.

For a significant and growing slice of US internet users, that model has started to feel less like a feature and more like a trap. The content gets more extreme because extreme content drives engagement. The communities feel hollow because they were assembled by an engine rather than by human affinity. And the experience of spending time online increasingly feels like being guided through a mall designed by someone who knows exactly how to separate you from your attention — and your money.

"I realized I hadn't actually chosen anything I was looking at in weeks," said Marcus, a 34-year-old teacher in Portland who now runs a newsletter-only community around independent film. "The algorithm had just been choosing for me, and I'd been calling that 'discovery.'"

Marcus's newsletter, which he runs through Substack, has about 1,200 subscribers. There's no recommendation engine pointing people toward it. New readers mostly come from word of mouth — people forwarding it to friends who they think will like it. Growth is slow. Retention is high.

The Formats the Algorithm Can't Touch

The spaces being built outside algorithmic discovery tend to cluster around a few formats, each with its own logic.

Newsletters are perhaps the most mature of these alternatives. The newsletter renaissance of the last five years — driven by platforms like Substack, Beehiiv, and Ghost — is partly a content trend and partly a deliberate architectural choice. Email is push, not pull. You subscribed, you get it. The platform isn't deciding what you see or in what order. The relationship is direct, and the discovery mechanism is human recommendation, not behavioral data.

Discord servers occupy a different niche. They're community-first rather than content-first, and their structure — channels, roles, pinned threads — rewards people who show up regularly and participate rather than people who post content that performs well in an algorithm. Many of the most active Discord communities in the US around topics like personal finance, tabletop gaming, homesteading, and local politics have explicitly rejected any kind of recommendation integration.

Forum-style platforms like Tildes, Marginalia, and a revived interest in old-school message board formats represent a more ideologically committed strain of this movement. These are spaces built on the premise that good content should surface through human curation and community voting, not behavioral optimization.

And then there are the reading groups, listening circles, and IRL-adjacent digital communities — groups that use technology to coordinate but deliberately keep their membership small enough that algorithmic scale is never really on the table.

The Monetization Question

Here's the practical challenge that every algorithm-free community eventually has to answer: how do you make money without engagement metrics?

Algorithmic platforms have a straightforward business model. Engagement data is sold to advertisers. The more time you spend on the platform, the more valuable you are. Remove the algorithm, and you remove the core of that model.

The communities thriving outside that model tend to use one of a few alternatives. Paid subscriptions are the most common — Substack's whole value proposition is built around writers charging readers directly, cutting out the advertising intermediary entirely. Some Discord communities charge membership fees for access to premium channels. A smaller number run on patronage models, where members voluntarily support the community financially because they value it.

"We have 600 members and we cover our costs with a $5-a-month optional contribution," said Deja, who moderates a Discord community focused on Black women in tech. "Nobody is getting rich. But nobody is selling our data either, and that feels like a fair trade."

The sustainability question is real, though. Paid subscription models work well for creators with established audiences and genuine expertise. They're harder for new communities trying to build from scratch, and they create access barriers that algorithmic platforms — whatever their other flaws — don't have.

Could This Actually Reshape Discovery?

The honest answer is: probably not at scale, and probably not soon — but it might not need to.

The movement away from algorithmic discovery isn't trying to replace TikTok or Instagram. It's trying to carve out territory that those platforms can't colonize. And in that more modest ambition, it's having real success. Newsletter open rates for independent creators consistently outperform social media reach. Discord retention metrics — when communities are measured on how long members stay active — often look dramatically better than equivalent Facebook Groups or subreddits.

What's emerging is something like a two-tier internet. The algorithmic tier is loud, fast, and dominant. It's where trends break, where viral moments happen, where the mass of attention lives. The intentional tier is quieter, slower, and more durable. It's where the conversations that actually change people's minds tend to happen — the ones you remember six months later.

At ScrollerBase, we're built around the idea that the scroll reveals something about what people value. And what the growth of algorithm-free communities reveals is that a meaningful number of people value something the scroll can't give them: the sense that they chose where they are, and that where they are was worth choosing.