2 Revolutions in semantics, scale, and similarity
This chapter covers
- Word2Vec's breakthrough in semantic understanding
- FAISS's solution to billion-scale similarity search
- Sentence-BERT's practical transformer similarity
- How these technologies have converged for RAG
By 2020, the stage was set for a revolution in how AI systems access and use knowledge. Three seemingly unrelated technological innovations had quietly converged to create the perfect conditions for what would become retrieval-augmented generation.
Word2Vec, a model that proved semantic relationships could be captured mathematically, addressed the vocabulary mismatch that limited traditional search systems. FAISS (Facebook AI Similarity Search), an approach and a library that made billion-scale similarity search practical, turned high-dimensional vector search from a research curiosity into production infrastructure. Sentence-BERT, which adapted transformers for real-time similarity applications, completed the pipeline from user queries to relevant document retrieval.
Each technology solved a critical limitation that had prevented semantic search from becoming practical for enterprise applications. Together, they established the foundational infrastructure that made retrieval-augmented generation possible. Understanding these foundations is essential because modern RAG systems inherit both the strengths and failure modes of the technologies they’re built on.