Ask HN: What are you using Vector Databases for?

To search the best movie for me to watch

As well as heavy LLM use, we're also using vectors (with pgvector) for searching text content which is displayed in a table. Works very well.

What is unique about your data or use case that makes a vector database a good choice compared to more traditional storage methods?

I want to search based on semantic meaning, rather than partial or exact keyword search.

I'm using pgvector within postgres, so I get all of the postgres benefits.

A reverse image search where I look for similar images by transforming images to a dense vector set by a finetuned triplet loss model.

Building mixture of experts datasets of multimodal embeddings.

Specifically on the Australian power grid at the moment, but more generally on a diverse array of topics.

Anywhere you have data with high dimensionality. E.g. gene expression datasets.

Do you have some resources to learn more about this?

Face recognition.

No using any vector databases at the moment

Stylometry.