Adrien Treuille

GenAI @snowflakdb / CEO @streamlit / VP @zoox / CS Prof @CarnegieMellon / Creator @Foldit and @EteRNAgame / @MIT @TechReview 35 under 35 / Lead @Google X

[Adrien Treuille @myelbows]

@karpathy is a @streamlit fan?! All life goals now complete. New goal suggestions, anyone?

[Andrej Karpathy @karpathy]

Carve out a few hours to learn https://t.co/6szj33iRBz
Powerful for rapid prototyping, interactive visualization.
It's a hammer and you'll start seeing a lot of nails.

[Anthony Goldbloom @antgoldbloom]

The overlooked GenAI use case: cleaning, processing, and analyzing data.
https://t.co/j9hFXWJvvF

Job post data tell us what companies plan to do with GenAI. The most common use case is data analytics projects. Examples:

- AstraZeneca: using LLMs on freeform documents to structure results from their Extractables & Leachables testing (https://t.co/wbamJnQEEN)

- Trafigura: The Document AI team is using LLMs to extract data from a corpus of commodity trading documents to generate credit reports
(https://t.co/mujn0ksa90)

The startup ecosystem is overlooking this use case, instead focusing on other areas such as customer support, sales & marketing and code gen.

[Adrien Treuille @myelbows]

Amazing @streamlit integration by @AnthropicAI to control a full GUI via multimodal AI! 🤯

[Thiago Teixeira @ThiagoTDotCom]

This looks so awesome — and it's also a great use of @streamlit.

I just have to try this out!

[Adrien Treuille @myelbows]

@ThiagoTDotCom @typescript What are you doing on vscode, @ThiagoTDotCom?!?

[Adrien Treuille @myelbows]

Snowflake now comes with a crazy broad array of LLM powers out of the box. @snowflakedb

[Snowflake @Snowflake]

Industry-leading LLMs, Generative AI in SQL Functions, and Improved retrieval is now generally available with #SnowflakeCortex 🎊

[Adrien Treuille @myelbows]

We just released my interview with @RamaswmySridhar and @barisg on #SnowflakeArctic . Interesting to hear their perspective on the future of generative AI in the enterprise.

[Snowflake @Snowflake]

Join our Linkedin Live tomorrow by signing up here: https://t.co/fYVMimEy6X

[Adrien Treuille @myelbows]

Interesting @wesroth video on @SnowflakeDB's #SnowflakeArctic model exploring the implications of large, capable, and truly open models.

[Wes Roth @WesRoth]

🔥 another win for open source AI

SNOWFLAKE ARCTIC:

1] Mixture of Experts model with 128 (!) experts

2] open weights, open code

3] Company is sharing "cookbook" and data and teaching people how to build their own world class MoE models.

4] very cheap model to train (16x times less than Llama 3)

[Vivek Raghunathan @vivek7ue]

A lot of the insider knowledge on how to build an LLM has gone underground in the last 24 months.

We are going to build #SnowflakeArctic in the open

Model arch ablations, training and inference system performance, dataset and data composition ablations, post-training fun, big run stability, decontamination tricks, metric subtleties.

To start with the first two blogs in our cookbook, keep reading the 🧵below ...

[Adrien Treuille @myelbows]

Huge shout out to @bfirsh and the team at @replicate for being amazing launch partners for @SnowflakeDB. Using Replicate inference, the #SnowflakeArctic live demo has been running smoothly and crazy fast all day long! https://t.co/MrpJi1ubk4

[Streamlit @streamlit]

Psst. 😮 Big news to share!

Today, @SnowflakeDB released a fully open-source foundation #LLM. It’s called #SnowflakeArctic, and it’s super smart and efficient.

The really cool part? It writes beautiful ✨Streamlit code.🎈

Check out deets on our blog!👇
https://t.co/9hi15UreNW

[Thiago Teixeira @ThiagoTDotCom]

Standing room only for this @streamlit + @SnowflakeDB talk at PyCon / PyData Berlin today

If you're around, be sure to say hi to the Streamlit engineers and PMs attending!

[Adrien Treuille @myelbows]

@omarsar0 Play with Arctic live here: https://t.co/1eIZ2lKStg

[elvis @omarsar0]

Snowflake casually releases Arctic, an open-source LLM (Apache 2.0 license.) that uses a unique Dense-MoE Hybrid transformer architecture.

Arctic performs on par with Llama3 70B in enterprise metrics like coding (HumanEval+ & MBPP+), SQL (Spider) and instruction following (IFEval).

The remarkable part is that it claims to use 17x less compute budget than Llama 3 70B. The training compute is roughly under $2 million (less than 3K GPU weeks).

We are witnessing the democratization of LLMs at an unprecedented rate.

[Adrien Treuille @myelbows]

Wanna play with #SnowflakeArctic live? Go to https://t.co/1eIZ2lKStg

[Snowflake @Snowflake]

Introducing Snowflake Arctic. An efficiently intelligent and truly open LLM built by Snowflake.

[Jerry Liu @jerryjliu0]

We have out-of-the-box support in @llama_index for Snowflake SOTA arctic-embed models - it's a one-liner through our @huggingface integration!

And in ~4 more lines of code, you can build a RAG indexing pipeline over your data with these embedding models.

Check out the screenshot below for how to get up and running with this - we might have more detailed benchmarks soon, stay tuned 💡

[Streamlit @streamlit]

Hello Bay Area 👋 It's time for another edition of the SF Meetup 🌉!

📅 When: April 17th, 2024 @ 6PM
📍 Where: GitHub HQ

Check out the Meetup page to RSVP and learn more.

Which talk are you most interested in? 🤔

https://t.co/opkOZncp4b

[Anthony Goldbloom @antgoldbloom]

Amazing how in just a few months as CEO, @RamaswmySridhar has repositioned @SnowflakeDB as a key player in the LLM ecosystem.

[sridhar @RamaswmySridhar]

@SnowflakeDB is open sourcing the best embedding models in the world! 🚀🚀

They are now available open source in @huggingface We are releasing it under the Apache 2 license so that it is easy for the OSS community to experiment with them freely.🎁🎁

These impressive models are the result of @Neeva's
expertise in search, state-of-the-art research, and @SnowflakeDB ’s steadfast commitment to excellence in AI.

It’s incredible to see what the team has built, and even better to delight our customers day after day with our latest innovations.

[Snowflake @Snowflake]

Just launched! The snowflake-arctic-embed family of models - the world’s best practical text-embedding model for retrieval use cases. Outperforms all others in the same size category including those from OpenAI and Cohere.

[Adrien Treuille @myelbows]

@spacemanidol This is cool.. how to *really* train an embedding model!

[Daniel Campos @spacemanidol]

🚀 Buckle up, fellow latent space explorers! Today, we're diving into the galaxy of embedding models – those magical AI constructs that power everything from search engines to recommendation systems. https://t.co/pXKzfDB9mH or a 🧵 for simplicity