lantern VS jvector

Traditional DBs already kinda support vector DBs via pg_vector extensions and such.

There is a YC startup, latnern, that also built their own extension for postgres that is open source and is better for vector DB use cases: https://github.com/lanterndata/lantern

But yeah! Traditional DBs already support this, if you consider this extension to be part of Postgres.

This extension is licensed under the Business Source License[0], which makes it incompatible with most DBaaS offerings. The BSL is a closed-source license. Good choice for Lantern, but unusable for everyone else.

Some Postgres offerings allow you to bring your own extensions, for instance Neon[1], where I work. I tried to look at AWS docs for you, but couldn't find anything about that. I did find Trusted Language Extensions[2], but that seems to be more about writing your own extension. Couldn't find a way to upload arbitrary extensions.

[0]: https://github.com/lanterndata/lantern/commit/dda7f064ca80af...

Install and use our extension here` https://github.com/lanterndata/lantern

Features today + Coming soon

JVector builds on the ideas in DiskANN to provide state-of-the-art vector search for Java applications. I’ve used the JVector Bench driver to visualize how recall (search accuracy) degrades when searching for the top 100 neighbors in data sets created by different embedding models against a small sample of chunked Wikipedia articles. (The data sets are built using the open source Neighborhood Watch tool.) Perfect accuracy would be a recall of 1.0.

Nice to see people care about index construction time.

JVector scales linearly to at least 32 cores and may be the only graph-based vector index designed around nonblocking data structures (as opposed to fine-grained locks): https://github.com/jbellis/jvector/

JVector indexes the Sift1M dataset in under 19s on a 32 core aws box (m6i.16xl), compared to 50s for Lantern in the article.