Interesting misconception that Dan flagged today - people think that we are (or could be) using LLMs for spam detection. 

With the way LLMs work today, that's like using a hammer to cut your fingernails. LLMs are slow, expensive and don't really have a deep understanding of what spam is in the context of Farcaster. 

We use a random-forest decision tree that @akshaan designed, and feed it a bunch of signals using embeddings, user actions and graph data.

Follow to get the best of Farcaster using @mbd - AI Recommendation Engine for Web3. @a16zcrypto css23. Ex-ML Lead @Chainlink. Ex-Sr Data Scientist @AWS.

Technowatermelon. Elder Millenial. Building Farcaster. 

nf.td/varun

cc @yassinelanda.eth I’m curious to hear your comment on this.

crypto-anarchist · onchain social explorer, curator, advisor · spread FC gospel in /chinese community · host /luo · write 0xluo.xyz

@v is spot on LLMs - Spam detection is more of a behaviour analysis exercise as well as usually an unsupervised / clustering / anomaly detection problem. Nevertheless the next-gen models that are  « attention based » have so much more classifying power than old method like tree based, with the former giving you a lot of explainability and feedback too.