are vector databases actually useful for anything or can I just mess around with pgvec and call it a day

BDFL of Quilibrium, Eng @ Farcaster, ex-Coinbase, always opinionated, never hydrated

/quilibrium

thanks, got it. primary use case right now is just storing messages, embedding them, and clustering those that are semantically similar

pgvector’s featureset is smaller than other vector databases, but this gap will shrink in time. The big question is what works best for your data pipeline needs, and that’s ultimately something only you can answer. 🫡