souravroy78

Got an update. Some startups are using it. It comes with Zscaler/Netskope/etc compability as well!

This is amazing!

Not sure. But thanks for the attention though have loads of features in main - to be released soon!

8 million rows in. 600,000 out. 5.7 seconds. On a 16GB RAM laptop.

Duckle joined 4 sources at 2M rows each - an ADBC (Arrow) source, a CSV file, a MySQL table, and a second ADBC source - through one visual mapper: a 3-way join, 9 expressions, and a filter, straight to Parquet.

No cloud. No servers. Just Duckle on your laptop/desktop. This is what local-first data engineering looks like now.

Repository: https://github.com/SouravRoy-ETL/duckle

It pretty much holds good as of now. We are currently in a very early stage.

We are working hard to create an alternative to DuckDB and keep it completely open Source. No paid support like DuckDB, ClickHouse, etc.

Thanks bruh!

We are going to publish results real soon. Stay tuned!

Hehe! Irony leads the way!

Haha! Sloth's are actually faster than duck in Water!

I am working on Publishing Results soon in a day or two stay tuned and thanks for the kind words

It is becoming impossible to work with Opus 4.7. The calculations Opus 4.7 is making is not adding up and hence we feel like it is hallucinating. May be they tuned down RAG, context secretively.

I'm not clear on the actual use case how can this be leveraged?

what monitoring Token usage along with the framework or does it already have one?

Can you explain the use cases we can cover by leveraging this?

Skynet is upon us! But how does the blog run itself?

awesome work though! Much needed for Nvidia Folks!

Lol guys we are all being xhigh nerfed!

This is in development stage Fellow Hackers! Open for collaborators and ideators!

Yeah guys they nerfed it for sure.

Nice idea though! Would like to try it out!

It is normal to feel this way. We are adapting since inception. We should not fear this AI Boom TBH. The same people who opposed Cars during the Horse Era were driving around Ferrari's. We should make the most of AI as long as it brings value to Humanity.

Nice

Are they done with so called state of the art Gemini models

Don’t know if Karpathy even wrote this version. Where are the citations?

Can these support local LLM’s?

Cool!