Nitya Tarakad

[Nitya Tarakad @nitarakad]

TIL on @neeva gist that this beautiful underwater waterfall in Mauritius is an illusion 🤩😮 try the beta today at https://t.co/dk9xCo0HcY #neevaai

[Dwarak Rajagopal @dwarak]

Exciting insights from @rajhans_samdani @nitarakad @lukemerrick_ and @pxyumass. The latest article delves into how retrieval and chunking strategies remain crucial for accurate AI-generated financial insights, even with long-context LLMs.

https://t.co/aaBV4v1v9V

[Rajhans Samdani @rajhans_samdani]

🧵 1/ Recent hype suggests long-context LLMs remove the need for retrieval in RAG pipelines—"just put all your docs in context." We tested this theory rigorously in finance, focusing on SEC filings.

Spoiler: Retrieval & chunking strategies still dominate.

[Daniel Campos @spacemanidol]

Having built search systems for a while now I firsthand know how hard it is to fiddle with all the possible knobs to maximize performance. Luckly now @SnowflakeDB's cortex search does it for you. 🌑https://t.co/sKig7CCmyJ

[Rajhans Samdani @rajhans_samdani]

Work led by @nitarakad and @spacemanidol
https://t.co/JL5BWBRJcB

[Vivek Raghunathan @vivek7ue]

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

We promised to build #SnowflakeArctic in the open, and here we are, with the third edition of our cookbook series, this time on data ...

Data ablations are the lifeblood of any LLM training run, whether processing, filtering, deduping, data composition, or schedule.

To read about our data approach to Arctic, see the blog below

[Snowflake @Snowflake]

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

[Daniel Campos @spacemanidol]

Have you ever been up late looking at the stars and the vastness of space and wondered how computers understand similar words with different queries? If you are like me and have spent sleepless nights, @SnowflakeDB, @lukemerrick_ , and I got you. https://t.co/ohaAiI5xs6

[sridhar @RamaswmySridhar]

Tomorrow, the future of AI ⚡️⚡️@SnowflakeDB unfolds.

Are you ready?!

Register now. https://t.co/hsckIpkw9m

[sridhar @RamaswmySridhar]

Exciting stats from the Snowflake team about the SQL generation performance of Meta's Code Llama. And this performance is oh so close to SOTA gpt4.

[Vivek Raghunathan @vivek7ue]

Excited to be joining forces w/ Frank, Benoit, Thierry, Christian, Greg, and everyone at @SnowflakeDB !

[sridhar @RamaswmySridhar]

Excited for the @Neeva team to be joining @SnowflakeDB to bring the power of search, AI and LLMs to people!

[Neeva @Neeva]

Big news! 💥

@Neeva is on this year’s @forbes AI 50!

This list recognizes the most promising privately-held companies building businesses out of artificial intelligence. 💡

[sridhar @RamaswmySridhar]

Thrilled to announce @Neeva ‘s transparency and accountability center!

(Um we do design doc reviews in that conf room!)

[Neeva @Neeva]

Break free from the echo chamber.

With #NeevaAI and Bias Buster, you can access multiple perspectives & viewpoints on news and issues.

Here's an example from a trending story today ⤵️

[Neeva @Neeva]

Watching our competitor chatbots hallucinate, fail at basic math, & fall in love — we felt like something was missing at @Neeva.

That ends today with: #NeevaAI personalities 🪄

🚜 Feeling country?
🌐 On Bard's waitlist?
🫡 GenZ? Bet.
🏴‍☠️ Captured by pirates?

Follow along🧵

[Vivek Raghunathan @vivek7ue]

Got access to Bard.

Tried my fizzbuzz question: [why is the sky blue?]

v0: abstained
v1: hallucinated a citation. The first citation is a 404.

[Neeva @Neeva]

Couldn't agree more, Bard!

At @Neeva, we put our users first.

That means:
❌No ads
❌No advertisers

Your privacy is always protected and we never profit off of your data.

[Nitya Tarakad @nitarakad]

use #neeva :)

[Neeva @Neeva]

Bard may not have citations but it’s right on this one!

*eats ice cream*

👀🍦

[sridhar @RamaswmySridhar]

1/ And so it begins…

Just as @vivek7ue and I predicted this year that @OpenAI would soon be "Closed AI"...

(via @verge ⬇️)

https://t.co/nl71v9n00M