18 Feb 2026
Network effects. They are the main reason that most social media startups, both earlier in the smartphone era (~2013-2016) and the secondary push to chase new social primitives (~2020-2022) left us with relatively few distinct social experiences - mainly deriving from the likes of Meta, X, Snapchat, or TikTok.
Having users on a social media network is the only way for a social media network to work. Without sufficient network density, there isn’t an incentive to keep opening the app, which stalls engagement and stalls growth. It’s for this reason that apps like Meta have chased engagement above all else - resulting in a hyper-polarized, centrally-controlled social empire dominated by just a few parties.
When Meta launched Threads, they did something interesting. They cross-advertised that your Instagram account was already a Threads account - all you had to do was grab the app and sign in. For this reason, Threads quickly eclipsed 100 million daily active users in December of 20241. They took an existing network, with network effect in full force, and simply reused it.
This brings me back to an experience I had this morning. I opened a new app called Popfeed; it feels like a mashup of Letterboxd and Goodreads, but built AT Protocol. I signed in with my AT Protocol account. Lo and behold, all the people I follow on Bluesky were just… there! My feed was already filled with people that I recognized, interacting and rating movies. And unlike Threads, Bluesky doesn’t own Popfeed - Popfeed was just built on AT Protocol.
I believe AT Protocol will win. If for any reason, it’s because of the positive feedback loop of network effects, and that any developer can contribute a new app to this existing network.
Source: TechCrunch ↩