Clickhouse and vector databases

In the video ClickHouse for AI - Vectors, Embedding, Semantic Search, and more by Alexey Milovidov, i learn a lot about databases and AI.

The video started with a brief introduction of some key concepts such as semantic search, vectors and embeddings. Then proceeded with some optimization technique, such as executable pooling and chunking. the bulk of the video is on how to do query faster and faster.

At the end of it, Alexey Milovidov presented some other vector databases such as Milvus/Zilliz and Weaviate. Both of them are implemented in Go. However, Qdrant is implemented in Rust. Best 15 vector databases for 2024 [Top Picks] gives some more vector database products. Those I picked up for a closer look is because:

• open source
• implemented in the language that I am interested in learning

The other day one of my friends told me that I can spend some time on AI. Probably this is where I started. It seems that there are no good books about vector databases. Two books I can find so far is Database Internals and Designing Data-Intensive Applications.

Database internals gives an under the hood overview of database implementation. It focuses mainly in B-tree which I am not sure if it is related to the vector database. But it has discussed column and row databases.

For Designing Data-Intensive Applications, it is more on the side of application design and architecture that cope with large database sizes.

Probably will take a closer look later.