If you remove Threads (Instagram drives distribution, so derivative of that graph) and TikTok (billions of dollars of paid acquisition + ability to plug into a mature ML model from Bytedance), the last sustained broadcast network created by a startup was Telegram (and it's mostly messaging).

11 year drought.

So no meaningful new sustained public social graphs since early 2010s.

1. Public social network growth playbook is likely a mirage and mostly due to mobile adoption S-curve. cf. @pmarca's the only thing that matters

2. If you want to build a new public social network graph in 2024, start from first princples.

🐵 ← 🙈🙉🙊 → phil.cockfield.net, db.team →  ƒ(trust / time) →  ƒ(open-commons) → ƒ( /sys )  ← 🧫 <cell primitives>

We're pretty far off from "at-scale". But making progress!

there's an old @lessin screenshot essay about vintages of social graphs, like sets of connections based on the era the network came out. 

having trouble finding it, but think it would be additive to thinking here