Excited to host changhiskhan and Haotian Liu this Friday.

Tune into our webinar to learn about:
- LanceDB: github.com/lancedb/lancedb
- LLaVA: github.com/haotian-liu/LL…

Excited to be talking about Lance format!

ChatGPT wrote me a script that finds 'further reading' articles from my blog 😁

using LanceDB (YC W22) for the search

Very excited about these two PRs from LanceDB (YC W22)

- github.com/lancedb/lanced…
- github.com/lancedb/lanced…

⚙️384 adds a Pydantic-powered ORM for querying with and saving pydantic models.

🤏385 makes pandas dependency optional - it's marginal footprint over pyarrow and numpy is tiny

11. Lance (LanceDB (YC W22))

Multi-modal AI is revolutionizing industries, but data management remains a challenge. Lance offers a solution, optimizing storage and handling for this unstructured data, increasing performance, reducing costs.

- saar gur, CRV

First PR of Lance sprint merged EuroPython thanks to Lindt!

Line around the room to ask Ritchie Vink (@[email protected]) questions about polars data

Shaan Khosla (linkedin.com/in/shaan-khosla) has been using #LanceDB a ton over the past few weeks. Interested in discovering what he found? Take a look at his NLP newsletter:

shaankhosla.substack.com/p/easy-free-an…

#vectordatabase #python #opensource #oss

I feel awkward asking for the check in Prague. I keep expecting the server to be like “what do you mean sir there are millions of us”

Anyone going to EuroPython next week? Come hear my talk on Lance format and/or join our sprint!

ep2023.europython.eu

Introducing #LanceDB integration with #Voxel51 , boosting CV researcher productivity tenfold! Discover it now!

Amazing work from Ayush Chaurasia

blog.lancedb.com/scalable-compu…

Fast SSD is a killer feature, and for more than just BI ;)

Rippling was great in the beginning but the support experience is just awful. Are there good alternatives for startups?

It is out.

Pydantic V2.0 is finally live!

After 16 months of work, with help from hundreds of people, we've finally done it.

Why you should never use pgvector (e.g. Supabase Vector Store) for production:

😮 pgvector is 20x slower than a decent vector DB (e.g. Qdrant)
🤯 And it's a full 18% worse in finding relevant docs for you

And this can happen at as little as 10K documents when chunked!

Does Azure gives out credits for open source testing against Azure services (blob store in this case)? A kind soul (alikefia) helped build Azure integration in Lance but we don’t have an account to test against.

Got a bunch questions around how the new delete feature actually works in Lance. Will Jones wrote a great post to explain this in more detail. Big step forward for Lance to support reproducibility in your LLM-apps

blog.lancedb.com/you-can-now-de…

Pretty excited about this release. Deletions that support time travel, 3x faster index creation, and an even lighter-weight install (no openblas needed).