Networks like Farcaster enable us to shift focus from aggregating attention to unifying identity. They give us the capability to embrace the fragmentation of experience across multiple applications, unified by a single social graph. Incidentally, this is exactly why Farcaster and its competitors are uniquely positioned to capture a growing number of users who are seeking genuine, specialized connections on a post-social, ad-driven web.
Note: none of this was written by AI.
The rise of digital slop across social networks is apparent to the extent that it's impossible to ignore. If you're a frequent user of LLMs (or any AI-driven product), you can typically tell within a few words whether the response you're reading has been written by a human or not. On the premier social networks, AI-generated content dominates the responses above the fold before you reach an ad. If you're not already reading something inauthentic, someone will ensure you are by tagging Grok in the first comment. The Dead Internet Theory is playing out before our very eyes.
Content itself is no better. Threads, articles, images, videos, and even news are being generated by the hallucinations of a machine. It creates a nether-realm where you begin to wonder if you're reading anything genuine at all. Are these machines communicating with each other to facilitate a mechanism that keeps our attention, and nothing more? We're consuming the shadow of truth, the idea of authenticity. And we're accepting it knowingly.
Rather than blindly eat what is put in front of us like farmyard hogs, many have begun to flee the feed and seek more relevant sources of entertainment, news, and connection. I recall reading an article several years ago about how more social interaction would occur within insulated group chats, rather than being broadcast across social graphs that have become popular over the past 15 to 20 years.
Both anecdotally and materially, this has occurred. The flight to directed, specific conversation is a response to the immense amount of noise one must deal with on broadcast networks and the very slim and constantly reducing amount of signal one finds.
Hence, news feeds have shifted to Telegram, where one can receive exactly the type of information they want to read, with absolutely no noise blocking the signal. Similarly, group chats facilitate conversation with people you want to hear from, providing the information and connection you actually desire.
As such, niche social networks are on the rise.
When the fight for the dominant social graph was at its peak and users began to fall into their respective subcultures, the common mantra among the zeitgeist was, "Nah, I'm not downloading another social app. I already have too many." However, today that has begun to change. Niche social networks, such as Substack, which caters to readers and writers (wordcels), or Strava, which caters to fitness enthusiasts who like to share their stats, have begun to capture growing attention. Whereas in the past bolting a social feature to your app would have been seen as unfocused or even desperate, now the ability to refine what one consumes is quite attractive.
I believe this new commonality is why networks like Bluesky try to lean into the "curate your feed" concept. Whether this slight iteration is enough to mitigate the fact that they are ultimately still a traditional broadcast network remains to be seen. The bottom line remains clear, though. People will continue to gravitate towards specialized applications that provide them with what they want or need quickly, as long as those applications remain focused on their use case.
Reddit was once a great place to discuss shared interests, but it has become a counter signal for almost everything. Dribbble was once a valuable resource for designers, but it has since become little more than a glorified moodboard. Many such cases.
There is a genuine risk that attention will be fragmented to the extent that no network is populous enough to be interesting. A unified (ideally decentralized) social graph fixes this issue. Hop from one app to the next, but authenticate with one persona. Alternatively, stay in one app but have many specialized mini-apps within it.
The solution does not need to interact with the attention layer at all. We've spent too much time trying to aggregate attention. This misdirected intention has led to the issues stated above. We need to aggregate the identity layer. Unfortunately, solutions like OAuth only exacerbate the problems created by the incumbents who have set aflame what was once a genuine way to connect in pursuit of year-over-year profit.
Where OAuth engrains users into an unescapable, centralized graph tied to a single entity, decentralized solutions are already being well-funded and developed at scale by a variety of brilliant minds. You have Fediverse, Farcaster, World, and Telegram all vying for the top spot in this burgeoning vertical. Where Fediverse and Farcaster offer "sufficiently decentralized" components and infrastructure, World and Telegram take a more centralized approach.
Still, all of these provide tooling that enables devs to build within apps where users already exist, and users to find experiences that reflect the current meta of doing what you actually want online rather than what an algorithm desires.
It's clear that we're aiming to escape the machine and seek authentic connections and community. I'm encouraged by the fact that more and more people are unwilling to accept the egregious, compounding tradeoffs presented by the FAANG cabal. Social is evolving, and that's inspiring.
Nicky Sap
Support dialog
