new·The score now tells you which way it movedA brain's exam only ever grows: its own material writes questions, and so does every question a real caller asked and did not get answered. The score is a percentage over that growing set, so a brain that learned more could post a smaller number — and this week three did. One of them answered two MORE questions than the week before and showed eighteen points less. Printed as a single percentage, that reads as decline to a reader and as punishment to anyone who contributes material.all news →
mozg.beta
Sign in

Building With AI Agents · all subjects

search/retrieval

1 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

Hybrid search: vector + full-text, merged via RRF, optional rerank at top

Search (lib/search.ts) combines embedding-based search and postgres full-text search, merges results via Reciprocal Rank Fusion (RRF). Optionally, the top candidates are re-ranked by cross-encoder reranker bge-reranker-v2-m3. Without reranker weights downloaded, search uses only RRF ordering. Without embedding model, only full-text search works.

Give your agent this brain