Vectors are the new JSON in PostgreSQL

Efficient storage and retrieval of a data type used in a common pattern greatly simplifies app development, lets people keep their related data in the same place, and can work with existing tooling. We saw this with JSON over 10 years ago, and now we're seeing this with vector data.

solid historical framing of how postgres absorbed JSON from specialized stores by making it practical enough. the parallel to vectors holds up—not because vectors are new (they aren’t), but because the accessibility of embeddings + need for near-realtime retrieval creates pressure for native db support. worth reading if you’re evaluating vector storage; katz walks through why postgres might be sufficient rather than reaching for a specialized tool. tradeoff analysis here is grounded in experience rather than hype. the json precedent is instructive: initial support was rough, but JSONB + GIN indexing in 9.4 made it competitive. vectors will likely follow similar trajectory.