Anybody in the farcaster community thinking about ML recommendation systems for a decentralized social network?

Lots of interesting challenges across data collection, training, model deployment/inference, asset management, etc

building /propertyforecast | data stuff, robotics, sometimes defi

I am writing my thoughts about how we can do this in a decentralised way. RecSys are not trivial to get right and it is not an ML problem per say.. just hard to measure… Built my fair share of them in adtech, streaming, and retail and each time it is product/client/user base specific

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.

By "not an ML problem" do you mean that the work is more on process, data lifecycle, metrics/monitoring, etc? I.e. more of an engineering problem than a science problem