pgvector v0.5.0: Faster semantic search with HNSW indexes

Supabase Vector is set to become faster with the introduction of pgvector v0.5.0, which includes a new type of index called Hierarchical Navigable Small World (HNSW). HNSW is an algorithm for approximate nearest neighbor search in high-dimensional spaces like those found in embeddings. The update allows users to take advantage of the new HNSW index on their column using CREATE INDEX ON documents USING hnsw (embedding vector_ip_ops);. HNSW combines hierarchical and navigable small world concepts, offering improved performance over inverted file (IVF) indexes.