PrimeTek libraries on NPM have reached 100 Million downloads. Thank you all for trusting in our work and stay tuned for the exciting roadmap featuring Headless Mode, Accessibility, New Showcase, Theme Generator from Figma, Prime Web Components and more... #alwaysbetonprime

Chat agents often face information overload; #NLP solutions can alleviate operational hurdles by incorporating semantic search, question answering and snippet extraction. Learn how @apexchat uses Vectara to deliver exceptional customer service: hubs.ly/Q01xXG8h0

Announcing the Vectara Galactic Headquarters at 395 Page Mill Rd, Suite #275, Palo Alto, CA94306. (We just subleased the executive briefing center of the Cloudera bldg in Palo Alto).

1/8: Lots of discussion recently about #ChatGPT and how amazing the recent #LLMs are. Underlying the their success is the application of Reinforcement Learning with Human Feedback (RLHF), which has been recently applied to #LLMs (including #GPT4, #Alpaca and #Dolly).

As AI evolves, how will the industry solve issues like hallucination? Our CEO Amr Awadallah 🤖 shares Vectara’s goals for ChatGPT-like business functionality and how to prevent hallucination at the source with Richard MacManus in @TheNewStack: hubs.ly/Q01JB2f80

What is your biggest hurdle to adopting LLMs in your business?

We open sourced this today: vectara-answer is a React project that provides a configurable conversational search user interface. With it end users can ask questions of your data and get back accurate, dependable answers. github.com/vectara/vectar…

We Race Capital are proud of lead @Vectara $28.5M seed round to bring Generative AI to the Enterprise along with our friends at Databricks Matei Zaharia, angels Stewart Butterfield & many others. techcrunch.com/2023/06/13/vec… by @kyle_l_wiggers TechCrunch

Tonight in SF, let's talk about LLMs & hallucination. Looking forward to meeting you all.

New meme: "there is an AI model for that".

Jokes for AI people: "Which movie made a big impact in AI?" Transformers. "What would be good name for a baby Megatron?" Perceptron. "What is that all you and Perceptron need?" Attention.

“Move from AI models to AI systems to combine multiple tools such as RAG to avoid hallucination and improve quality” - Matei Zaharia at #vectara event.

“LLM fine-tuning is not the right way of dealing with the hallucination issue” - Amr Awadallah 🤖 at #vectara event.

Only if you think I deserve it, I would appreciate if you put in a vote for me in the form linked below. Further, and only if you think I deserve it, please reshare this post. Thanks 🙏

Seriously though — what did Ilya see?

I like the bluesky brand, new home of tweeting birds. bsky.app/profile/oztali…. Who else is there?

We are building a new cloud—one that is open source, can run anywhere, and costs 3-10x less. Join our Rebel Alliance, as we launch part of Y Combinator W24! ycombinator.com/launches/KdH-u…

Vinod Koshla "90% of investors add no value, 70% add negative value" youtube.com/watch?v=TYt5yu…

Hope you invested in Redpanda Data in 2020. Now it is a $1b company and it is too late for Confluent to acquire. Congrats 🕹️ Alexander Gallego ⚡️, well-done! businesswire.com/news/home/2025…

A great founding team setup is Walter White and Jesse Pinkman. One founder zeroes in on building the best product in the world and the other handles early distribution.