Lantern

We did a deep dive into how pgvector stores HNSW indexes in Postgres. Read more from Varik below ↓

Did you know there are over 300 parameters used to configure a Postgres service? Narek Galstyan covers some of the differences in default settings across Cloud providers. https://lnkd.in/g8ZtTvaj

Hybrid vector search = dense vectors + sparse vectors, combining semantic + keyword search to improve search results ?? In our latest blog post, we evaluate hybrid vector search with embeddings from OpenAI, Jina AI, and SPLADE using the BEIR framework. https://lnkd.in/gb4bXAfN

We respond to Pinecone's recent blog post comparing Postgres and Pinecone. We show that Postgres can outperform Pinecone in the same benchmarks Pinecone covered in their article. https://lnkd.in/gxgaxtfk

At a recent AGI House hackathon, we built support for dynamically loading extensions in Postgres in the browser! ?? We built on top of ElectricSQL pglite, which uses a modified version of Postgres to enable running Postgres inside the browser. In the video below, we can use Postgres and perform vector search with lantern or pgvector, and encryption with pgcrypto, all inside the browser. ?? To try it out: https://pglite.lantern.dev ?? To learn more about how we built it: https://lnkd.in/dX7BF5K5

Narek Galstyan walks through some of the basics around reading, interpreting, and editing permissions in Postgres https://lnkd.in/gkC2q_hK

Do you stream vector embeddings into a Postgres instance? This might be relevant to you. TLDR; a small schema change may make your pipeline 30-50% more efficient. How? - Generate less WAL, use fewer CPU cycles on WAL, and fewer CPU cycles applying table / index changes. https://lnkd.in/gDEzf3YG

My cofounder Narek Galstyan wrote a blog post about how different operations like indexing affect write-ahead logging (WAL) sizes in Postgres ?? https://lnkd.in/g-WEEjzY

Engineering updates for April! We've added asynchronous tasks, weighted vector search, Ubicloud control plane, and dashboard improvements. Learn more below ?? https://lnkd.in/gJV_6v22

Thanks for the shoutout TechCrunch :) We're building Postgres for AI companies at Lantern. if you're using a specialized vector DB or want to improve your Postgres performance, please reach out! We'd love to help.