Show HN: A key value store based on semantic similarity rather lexical equality (opens in new tab)

(github.com)

8 pointsenesakar1y ago2 comments

2 comments

A semantic cache is just an encoder in front of a regular cache. It is not hard to connect the two, so where do you see the value? If you DIY, you can choose your own cache.

JimmyRuska1y ago

Isn't this how most vector search backends work, cosine similarity sorting against whatever the vector embeddings returned? How is this different?

j / k navigate · click thread line to collapse

2 comments

esafak1y ago

A semantic cache is just an encoder in front of a regular cache. It is not hard to connect the two, so where do you see the value? If you DIY, you can choose your own cache.

JimmyRuska1y ago

Isn't this how most vector search backends work, cosine similarity sorting against whatever the vector embeddings returned? How is this different?

j / k navigate · click thread line to collapse