How semantic search works (and why keyword search fails)

TL;DR Semantic search works by converting text into high-dimensional vectors (embeddings) and finding nearby vectors instead of matching words. This solves the vocabulary mismatch problem — a query fo