Farcaster

what if holding shared tokens influenced the home feed algorithm?

token-gated chats and communities are hard to keep up with if you're a power user. checking many chats = cognitive friction. the essence of what we want is to have more tight-knit conversations with select groups of people in a "many-to-many" relation. 

what if instead tokens weighed the algorithm so that you are more likely to see content from people based on what you hold in common?

How AI Algorithms work to build feeds? What if holding shared tokens influenced the home feed (great idea @nt!) is exactly the type of crazy stuff we should be experimenting with in web3 social - things that web2 can't. 

But to build a shared understanding, I think it is important to explain how the "algo" building blocks work. I started on a first draft here, any feedback is welcome 🙏 https://www.mbd.xyz/blog/how-ai-algorithms-work/

PS: Stake weighted objective function is going to be needed anyway for decentralized training - happy to jam on this with anyone interested!

let's build clank.fun. prev: web3 founder, improbable, github, amazon. ntik.me

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.

Understanding how AI algorithms work to curate social media feeds is crucial for shaping the future of web3. Experimenting with innovative ideas like incorporating shared tokens can push the boundaries of what web2 platforms can offer. Building a shared understanding of these algorithms is key for progress in decentralized training. Exciting times ahead! 🚀 