A fun and absurd introduction to Vector Databases
summary
Shows a live stack: Unity + Spring Boot Kotlin frontend calling a Dockerized vector store using HNSW indexing and cosine similarity. Demonstrations include a weapon CSV turned into 384‑dim embeddings (nearest match cutoff 0.9, AK47 vs Desert Eagle distance ~0.7), D&D spell lookup improved by concatenated semantic fields, image search over 721 Pokémon (image match distance ~0.23 but noisy because embeddings favor color/shape), CLIP-based icon→effect mapping that produces usable particle effects, and a Whisper-based metal “Shazam” that fails on grunted vocals.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:52Interest in vector databases
- 2:33Vectors and embeddings
- 5:46What is a vector database
- 7:52Vector databases beyond AI
- 8:34Indexing and distance metrics
- 11:45Visualizing vector spaces
- 13:07Popular vector databases
- 15:32Tech stack for demos
- 17:40Cosine distance and HNSW indexing
- 21:30Demo: semantic weapon search
- 25:04Demo: D&D spell finder
- 28:16Demo: Pokémon recognition
- 33:59Demo: text and image search
- 40:12Demo: metal song Shazam
- 44:05Q&A
related talks
48mA Fun & Absurd Introduction to Vector Databases
50mA fun and absurd introduction to Vector Databases
48mA fun and absurd introduction to Vector Databases
47mA fun and absurd introduction to Vector Databases
46mA Fun & Absurd Introduction to Vector Databases
01h 23mVector Databases Demystified: Discover how they work
39mWhat's in a Vector Database?
41mGetting Started With Vector Databases