Skip to content
The recordings / Farcaster protocol

The Farcaster protocol, ins and outs.

Cassie (Cassandra Heart) · @cassie · Quilibrium · Builder track - live workshop, recorded

Technical8/10Deep technical
The full transcript is live below. The video is being published in parts - the players land here as each goes up.
Read the transcript Farcaster org (FIPs) Quilibrium

Cassie (Cassandra Heart, founder of Quilibrium and a contributor to Farcaster) gives a deep, candid tour of the Farcaster protocol: what "sufficiently decentralized" really meant, the hubs era and its scaling limits, and the rewrite into Snapchain. She lays out the two centralization problems - the validator set, controlled mostly by Neynar, and the fact that you cannot build a real client straight off Snapchain - then walks through her Hypersnap fork as an additive side chain: threshold ECDSA via MPC, the Snap token and gasless bridging, consensus, the external OP Mainnet contracts, gasless signers, spam and trust scoring, and the roadmap. She closes with the five North Stars for getting Farcaster to a million daily actives, chief among them making it possible to build a client in a weekend.

The talk in four parts

1. Why Farcaster is the way it is 0:48 - 9:11

The backstory. Cassie unpacks "sufficiently decentralized" as a semi-permissioned label, traces the move from a centralized backend to peer-to-peer hubs (written in JavaScript, no incentive mechanism, plenty of chaos), and the scaling wall around 100,000 daily actives that forced the full rewrite into Snapchain.

  • 0:48What ZABAL Gamez is, and the intro
  • 3:27"Sufficiently decentralized" and the hubs era
  • 6:05Enter Snapchain, and the scaling wall

2. The fork, and inside Snapchain 9:11 - 16:55

Why she forked Snapchain into Hypersnap: the two centralization problems - the validator set is mostly Neynar, and you can't build a real client straight off Snapchain - and the goal that anyone should be able to build a client in a weekend. Then Snapchain itself: a real shard-based, Byzantine fault-tolerant blockchain with three shards.

  • 9:11The two centralization problems + why the fork
  • 11:10Inside Snapchain: the three shards
  • 13:28Hypersnap as an additive side chain (threshold ECDSA)

3. Inside Hypersnap 16:55 - 41:39

The deep half. The external OP Mainnet contracts (ID, storage, key, FNAME registries) and the new gasless signers; the Snap token and gasless bridging to any EVM chain via the validator quorum's signature; consensus trade-offs (Malachite/Tendermint, one-second blocks, and the speed-of-light limit on geography); the actor-driven architecture and spam/trust scoring; and the roadmap.

  • 16:55External anchors: the OP Mainnet contracts
  • 20:16Gasless signers (born from a real compromise)
  • 22:18The Snap token and gasless bridging
  • 25:12Consensus: block times and the speed of light
  • 29:47Architecture: actors, DKG, the token state machine
  • 38:47The roadmap

4. Q and A 41:39 - end

Audience questions. The five North Stars to get Farcaster to a million daily actives (a real query layer - done; a template-plus-SDK for clients; channels and curation as first-class; mini app discovery at the protocol level; native monetization), then Snap EVM Compute and the FIP2 / channels question.

  • 41:39The five North Stars (easier clients)
  • 47:39Snap EVM Compute
  • 49:24FIP2 and channels
Key quotes
  • "Control of the validator set is effectively control of the network."
  • "It should be very easy to build a client - you should be able to do it in a weekend."
  • "The speed of light does not care about your ambitions."
  • "Everybody thinks they have a plan till they get punched in the mouth." (quoting Mike Tyson, on migrating the data)
By the numbers
6validators (5 Neynar, 1 Uno) ~100kDAU where hubs hit the wall 1s / 12sSnapchain / Hypersnap blocks 250msmin round trip, 3 global peers 1,000,000daily-actives goal in a year 5North Stars
Topics
Farcaster protocol Snapchain Hypersnap building a client Snap token gasless signers consensus decentralization
Key takeaways
  • Snapchain is a real shard-based BFT blockchain; its validator set is mostly Neynar today, which is the live centralization question.
  • You cannot build a meaningful client straight off Snapchain - you need your own indexer and backend. Making a weekend-buildable client is the top goal.
  • Hypersnap is an additive side chain: threshold ECDSA via MPC, the Snap token, and gasless bridging to any EVM chain, without breaking Snapchain.
  • External anchors matter: the OP Mainnet ID/storage/key/FNAME registries, and the new gasless signers born from a real signer-compromise incident.
  • The five North Stars: a real query layer (done), a template plus SDK for clients, channels and curation as first-class, mini app discovery at the protocol level, and native monetization.

All recordings  -  Session recaps  -  Upcoming schedule

Transcript

Hover any line and click # to copy a link straight to it.