Simon Eskildsen(@Sirupsen):BM25 (least compute) sparse vector regex dense vector late interaction (most compute) we will index every byte. you simply choose how much compute you want to spend indexing, based on your read:write ratio.

Simon Eskildsen

@Sirupsen

co-founder & ceo @turbopuffer, former infra @Shopify 1k→1m RPS

Joined April 2009

871 Following 19.7K Followers

Simon Eskildsen@Sirupsen

2026.07.29 14:37

BM25 (least compute) sparse vector regex dense vector late interaction (most compute) we will index every byte. you simply choose how much compute you want to spend indexing, based on your read:write ratio.

turbopuffer@turbopuffer

2026.07.29 14:04

tpuf now supports late interaction [beta] use models like ColBERT to represent text as a set of vectors (1 per token) tpuf uses a single-vector ANN index for a fast first pass, then reranks hits using exact late interaction scoring to boost recall docs: