016
Fact: Facebook didn't conquer the world because it was well-designed, but because it captured your friends. Twitter became a global conversation not because the app was beautiful (it never was), but because it let you follow anyone. TikTok's "For You" page is basically a hyper-optimized interest graph that knows you better than your friends do.
The biggest companies on earth have been built on these invisible maps of connections: social graphs. Social graphs aren't a side feature; they are the product. They decide who sees what, which apps survive the cold start problem, which influencers blow up, and which memes go viral.
But here's the issue: the social graphs we have are broken. They're optimized for advertising, not belonging. They flatten dynamic relationships into likes and follows, and in the process make us feel more like spectators than participants.
This essay isn't a eulogy, but an autopsy. Under the hood, the geometry of social graphs explains why some people dance but most just scroll. Maybe, if we look at them differently, we can figure out how to build social graphs that actually make us want to show up, and dance, again.
A social graph is the map of who's connected to whom. Networks of user identities that relate to one another via standardized social data such as followers, likes, comments, and favorites. For example, Facebook's graph revolves around mutual friends, with strong network effects which grow rapidly as each new user's circle of friends generates more engagement, personal data and content. Reddit's graph centers on communities formed around certain topics, where content visibility is determined by self-curation. And Twitter's graph is heavily driven by an asymmetrical follower/following model: users can follow someone else's post without reciprocity. Each network has a different variation of social graph, but all are the same in that they are made up of nodes (people) and edges (relationships).
Simple enough. But this simple math ends up shaping almost everything online. Social graphs matter because they're invisible infrastructure. You don't notice them until you do. They decide if you feel seen or lost, if your community feels alive or hollow. Every party, group chat, and application has a graph humming beneath it.
And here's the kicker: the geometry of the graph matters. Flat, wide graphs lead to shallow connections end endless scrolling. Dense, clustered graphs lead to intimacy and culture. Some people end up with audiences; others end up with dance floors.
Web2 social graphs are walled gardens. Facebook, Instagram, TikTok - they each own your connections, and they monetize your attention. You don't own your edges; you rent them back in the form of reach. That's why if you are an Instagram influencer and decide you want to move to a different platform, you can't take your following with you. These types of graphs reward scale: the more nodes, the better. Which is why they end up looking like Costco: huge, efficient and soulless.
Web3 promises something different: portable, composable social graphs. Identities and connections that aren't locked in a single app. In theory, you could take your network with you, like luggage. But in practice, most web3 social experiments have either over-financialized (Friend Tech) or failed to reach threshold scale at all (Lens, Farcaster) because they offer nothing more than web2 counterparts besides onchain action.
The idea is right though: social graphs should belong to people, not platforms, But the execution so far feels like trying to rebuild Facebook with tokens or other blockchain primitives stapled on. A follower graph with a blockchain attached is still a follower graph.
This is the part the internet refuses to understand. Online, “cool” means follower counts, impressions, virality. But virality is fragile. It burns hot and then disappears, leaving everyone refreshing their notifications for scraps of serotonin.
Scene "cool" is different. It’s offline, embodied, and slow. The person who throws a party that 150 people actually show up to is cooler than the influencer with 150k disengaged followers. The geometry of their graphs is denser, and more triangular. Clusters of trust, repetition and rapport. These are the social graphs that compound.
Online "cool" is a performance. Scene "cool" is a presence. Followers spectate; scenes dance. And honestly, most people would rather spectate. It’s safer, easier, and people are sheep (on average). But the few who actually dance end up carrying culture forward.
We’re stuck between two models: the infinite, flattening graphs of Web2, and the half-baked experiments of Web3. One optimizes for clout; the other optimizes for speculation. Neither has figured out how to encode belonging.
Maybe the next step isn’t building bigger graphs but better ones. Smaller, weirder, more intentional. Graphs that preserve intimacy without smothering it in metrics. Graphs that don’t just measure attention but create space for presence. That's what we're attempting to build, through a variety of products, tools and experiences at ninetynine labs. Relationships and human-to-human interactions built by rapport instead of followers. Social graphs powered by access, cultural capital and shared rituals instead of likes.
The jury is still out on whether or not any technology or new social networking concept can fully capture that. The geometry of dancing, of being in a room with people you care about. But I do know this: the networks that matter most are the ones you feel, not the ones you count.
And maybe that’s sentimental. Maybe it’s naive. But the truth is: followers don’t dance. And I still believe some people, relationships, and interactions are worth dancing for.
US Government to take ownership in more companies.
Share Dialog
knownquantity.eth
Support dialog
293
