Probabl Livestream: KNN on a LanceDB Vector DB

We're experimenting with live streams!
In this live-stream Vincent will dive build a K-nearest-neighbor scikit-learn estimator that uses LanceDB as a backend. We'll use embetter to generate the required embeddings and LanceDB to keep things fast if we ever need to scale up. We may also learn that for some problems ... you may not need it!
So tune in if you'd like to see a demonstration and feel free to ask questions

[@probabl_ai]

In case folks are interested, we posted an issue on Github on the LanceDB side and got a nice conversation going that helps explain what is happening. If you're eager to index these kinds of datasets there's a bunch of remedies listed here:
github.com/lancedb/lancedb/issues/1222