A new age for an anti-slop internet

Anyone with even a tangential online footprint has, by now, felt the sharp end of algorithmic drift. You like a single picture of cat-shaped topiary and your feed becomes a wall-to-wall feline landscaping catalog. You click on a social media link to read a news article about military recruitment tactics and for weeks, you’re bombarded with enlistment ads. Not to mention that these “optimized” discovery tactics often deliver you an avalanche of AI-generated content that seems to get more and more repetitive as you swipe through.

There are times when all the retargeting is helpful; perhaps all that “For You” content led to your new favorite cat meme. But often, the experience of scrolling through a repetitive feed strips away the excitement that accompanies genuine moments of discovery. And given the rinse-and-repeat nature of content made by generative AI farms, the online experience can feel overly generic. The majority of U.S. adults who use social media (94%) believe they see AI-generated content when scrolling, with only 11% saying they find it “useful, informative, or entertaining,” according to a CNET study published in February 2026.

This is creating room for what you might call a “small web” counter-movement, one where a raft of founders are designing more insular, curator-led spaces. In 2024, Dirt Media co-founder Daisy Alioto predicted, in her talk on the taste economy at Friends with Benefits Fest in Idyllwild, California, that this new era of digital culture would usher in curation as a top brand strategy. “The desire [for taste-making] has certainly grown,” Alioto says, noting that for potential curators, the barrier to entry has dropped now that almost anyone can use AI to code a website, network, or app. “That you don’t need to be partnered with a full-time engineer to make things on the internet is a huge and positive change.”

For the next generation of more social, less algorithm-led platforms, Letterboxd, which launched in 2011, serves as the beloved blueprint. Mixing elements of a film diary, review site, and social network, members log what they’ve watched, rate and review films, build lists, and follow the cinematic musings of others. The platform hit 26 million users in early 2026. “It’s been built, deliberately, around our key mission: film discovery,” says CEO and co-founder Matthew Buchanan. “Our view is that every film is someone’s favorite, so whether something is ‘worth’ discovering isn’t ours to judge — that sits with the community.”

Small W•rld, a London-based “entertainment first” creative agency that’s worked with clients such as Under Armour, LinkedIn, and Taco Bell, recently ranked Letterboxd at No. 2 in a mini-index of entertainment platforms, just behind YouTube. “It’s done a great job of making people want to spend real time there, which comes down to the human creation and human accountability it facilitates,” says Small W•rld’s co-founder Dan Salkey. “We want to be able to hold someone accountable for the content we choose to spend our free time with, or the decisions we make.” Increasingly, founders are building with this mantra in mind.

Whether you’ve fallen deep into a Letterboxd rabbit hole, used Cosmos to surface inspiration from across the web, or joined DiVine to find a community that actively moderates AI-generated content, there are plenty of ways to find a human-curated sliver of the internet.

Sometimes, that comes to the fore in intentionally irreverent products. Danger Testing, a studio that treats apps more like media products than traditional software, rapidly spins up playful projects ranging from animal meme generators and a tongue-in-cheek “fun opportunity cost calculator” to a tool that lets hopeless romantics send their beloved a bouquet of links instead of flowers. Rather than betting everything on a single platform, the studio continuously launches new ideas using AI-powered development to slash the cost of experimentation. Slop Evader, a search tool and browser extension from Australian artist, environmental engineer, and technology researcher Tega Brain, takes things one step further. The tool filters web results to only include content published before November 30, 2022, when ChatGPT was first publicly released. This part-digital nostalgia, part-resistance approach carries a different energy, as if slipping into an alternate timeline where the LLM boom never happened.

Other projects specifically champion the act of human curation. Perfectly Imperfect (PI), which calls itself “the world’s first ‘social magazine,’” combines the ethos of an editorial publication with a robust social network. Set up in 2020 by co-founders Tyler Bainbridge and Alex Cushing, it began with interviews and recommendation lists from their Boston college and music circles and has now grown into a bona fide magazine, publishing conversations with the likes of Charli XCX, Omar Apollo, and Gia Coppola. It also boasts an online community of over 400,000 users who have shared more than 1.37 million recommendations in the pursuit of “finding community, sharing taste, and finding inspiration,” as one PI tagline declares.

Posts appear in chronological order, rather than via algorithmic recommendation, with users able to browse what’s new and popular. User profile pages hark back to the era of microblogging on a personal MySpace, with a scrollable blog section for users to recommend everything from Fugazi to “shopping with my mom in my favorite jacket”).

To apply a similar lens to existing social media platforms, Devin Gaffney created Graze Social in November 2024 as a custom feed builder for Bluesky. In Gaffney’s words, “nerds about their thing” — say, game development — use Graze like “Photoshop for algorithms” to stitch together a set of rules that control their own social media feed. The customized feed becomes more like a pinboard, with a standing set of curatorial rules applied to every other post across the network: essentially, a personalized algorithm all one’s own.

Drake University professor Ryan Stoldt, who specializes in how algorithms shape people’s online experiences, believes these new players were born out of a gradual reaction to the commodification and depersonalization of algorithms. “People used to see [algorithms] as a kind of hyper-personalized magic trick,” explains Stoldt, who is also a research affiliate on the University of Iowa’s interdisciplinary research group CP3. “Now they see through them: only optimizing for engagement, they keep you on there as long as possible.”

As a result, these emerging curation-first spaces may vary wildly in look, aim, and functionality, but all are united, Stoldt says, by a core tenet of trust. “As distrust in larger social media platforms has grown, the smaller ones have been building around community, with a focus on connecting through taste.”

A trust-first ethos is exactly what drove co-founder Stu Rogers to build Spread, a content sharing app. To “prove you’re human,” every Spread user is verified via an anonymous biometric (Face ID on iOS, fingerprint on Android and web); access then allows its largely journalist user base to share links to their own articles and ones they’re reading. There are no likes or comments on Spread’s ad- and bot-free feed: Every share is treated as an explicit vote for someone’s work, not engagement bait.

Rogers, who has a master’s degree in cybersecurity, says he “spent years in a company embedded in the trenches of major social platforms to help them identify ‘coordinated inauthentic behavior.’” As a result, he “became obsessed with the way troll farms and early bot armies were manipulating algorithms,” he says. “It proved to me that human activity would only become more rare and more valuable.”

With content publishers bracing for a world of agent scrapers and bot on-ramps, Spread integrates directly with media outlets so they can measure and monetize proven human traffic. And for an audience of journalists — whose line of work requires a high standard for accuracy and integrity — having this insular community provides a valuable opportunity to edge out misinformation. For Rogers, the goal is larger than building another social app; it’s about creating portable infrastructure for what he calls “the Human Internet.”

As these platforms grow beyond their niche audiences, the test will be whether they can scale without diluting the specificity that made them worth joining in the first place. Letterboxd’s majority owner Tiny, a Canadian investment firm, is reportedly exploring a sale, attracting interest from a host of entertainment giants such as Netflix, Sony, and Paramount Pictures. Some users are concerned that Letterboxd will sell out to commercial objectives and lose the independence that’s made the platform such a beloved space for conversation. (Buchanan insists that human curation has always been core to how the platform works, and says that as long as that “resonates for us and for the community, it will continue,” whatever the future holds.)

Meanwhile, in late August, the popular news aggregation software company Flipboard made a bet on the open social web by acquiring Graze, through which it plans to expand the technology beyond Bluesky to its own app and other AT protocol platforms, like encrypted messenger app Germ and Instagram alternative Flashes.

So where does this leave small networks as they look to expand — and what happens if this human-first ethos becomes even more mainstream? In the future, locking an audience within the confines of one platform — whether one website or app or larger social platform — might prove too limiting, and more substantial connections could be shared between apps with shared values. Over the next couple of years, Alioto predicts that the studio model will play a bigger role in how niche social communities develop: “Instead of one community, it might be a consortium of communities or apps all working towards the same perspective,” she says.

Alioto’s advice to founders building the next big tastemaking platform: Be agnostic about the best way for your community or your taste to be expressed, so you can evolve along with your audience. The goal, she says, is to create better circles for connection and direct exchange. “The dynamics become more about how people engage with each other,” she says. It’s the act of discovery, human-scaled once again.