AI21 Labs

So, what does a 256k context window actually look like? (🔉sound on!)

It was a pleasure to present alongside some of my personal favorite people in the #GenAI and #Partneships space at the Microsoft All Founders event Kudos Nadav Tzuker, Meital Shamia, Efrat Garber-Aran, Yair Adato & Uri Lampert for the great sessions delivered💪 And thanks Raz Bachar & Yarden Dor 🇮🇱 for putting such an insightful event AI21 Labs #GenAI #LLMs #founders

FBI-AGI, running on Jamba, gathered incriminating data of many predators harassing children online. Hackers Alex Sima and Apurva Mishra built FBI-AGI at the recent AGI House Jamba Hackathon. Within minutes of launching Emma’s profile, a fictitious 14-year-old girl, the AI agent received tens of messages from online predators with offensive and inappropriate messages. The #agent then captured the chat data and metadata to identify and incriminate these online predators in real-time. The best part? By leveraging Jamba’s massive 256K #contextwindow, ‘Emma’ ran multiple chats simultaneously, gathering proof of multiple predators at a time. And that’s just one of the many impressive projects made at the Jamba #Hackathon. Check out this blogpost with insights from Yaniv Markovski, our Head of Developer Relations, to read more about the projects built on #Jamba. 👉 https://lnkd.in/djCZFSYc 👇 Enjoy this clip capturing just a taste of the energy at the 12-hour hack

Between the *whitepapers*, *ablations*, *graphs*, and *endless acronyms*, the topic of genAI can seem pretty daunting—especially when considering scaling at the enterprise level. It really doesn’t need to be. Check out this easy-to-digest glimpse showcasing our latest advancements in the field. A warm thank you to BNP Paribas CIB for illustrating this Summary Scribe following Ori Goshen's keynote address at The BNP Paribas Exane CEO Conference. #FSI #NeuroSymbolic #LLM #AISystem

Everyone's been asking ❓Does having a long context window mean that the model actually does something useful with it? ❓Why use long context when you can use #RAG? AI21 Labs’ Co-CEO & Co-Founder, Yoav Shoham, tackles these questions and shares how our team used an aggressive new long context benchmark from NVIDIA to optimize Jamba-Instruct. Unlike the traditional needle-in-the-haystack evaluation, RULER measures how models hold up across the entire length of their context window on real-world, complex #longcontext tasks. Turns out, most language models don’t do so well. #Jamba-Instruct is the only model to maintain consistently high performance across the longest context window AND is optimized for enabling enterprise use cases at the most competitive price. Read more about how Jamba-Instruct was built to excel on long context use cases for the enterprise and why long context and RAG are not an either/or. 🔎 https://lnkd.in/dNRmGGdx

NOW LIVE: Jamba-Instruct 🖇️ Amazon Bedrock! Just moments ago, Yoav Shoham, our Co-CEO and Co-Founder, delivered the #AWSTLVSummit keynote address, announcing that Jamba-Instruct, is officially LIVE on Amazon Web Services (AWS) Bedrock. With #Jamba-Instruct’s 256K context window (equivalent to 800 pages!) and best cost per value in its size class, Bedrock customers can tackle massive document analysis and lengthy summarization tasks with ease —all within the secure environment of the AWS cloud infrastructure. The technology that enterprises need on the cloud platform they can trust. Get started building with Jamba-Instruct on Amazon #Bedrock today > ✨ From AI21: https://lnkd.in/ddkcuswJ ✨ From AWS: https://lnkd.in/diCGW4zs

On Saturday, we hosted our first Bay Area hackathon at the AGI House where developers got to build apps using #Jamba, specifically taking advantage of Jamba’s massive 256K context window. This was the AGI House’s first ever Mamba-based hack! 🚀Highlights from Hackathon🚀 After 8 hours of intense coding, each team presented a live demo with Jamba’s long context being the key differentiator in each winning project. And the winners are... 🥇First Place: FBI AGI A real-time incrimination tool against predators on social media. Alex Sima Apurva Mishra 🥈Second Place: Excel Mamba Complex financial Excel formula-generation using AI agents. Aniket Shirke Bhavya Bahl Rohit Jena 🥉Third Place: AI Pals The team built various avatars that could communicate via voice-over. Leveraging Jamba’s long context, entire books were used as context within a prompt, thereby decreasing latency. Anto Patrex Anand Vallamsetla Justin Jung Lars Traaholt Vågnes A huge thank you to the AGI House for having us, to Mike Knoop, Co-Founder of Zapier and Co-Creator of The ARC Prize, Or Dagan and Sebastian Limeres for the insightful speaker-series, and to Yaniv Markovski for organizing!

💡 Microsoft and AI21 Labs are thrilled to bring the Jamba-Instruct model to Azure AI. This collaboration delivers more flexibility and choice, with a powerful model for language processing. Explore Jamba-Instruct, available first on Azure: https://msft.it/6049Y24Vf #AzureAI

My favorite thing is meeting new people. Yesterday I had the pleasure to get acquainted with the warm and cheerful Ximus community at a Google event and join inspiring people like Yaron Inger and Ruben Hassid on stage, sharing AI21 Labs journey and discuss how #GenerativeAI is transforming the day to day for organizations and individuals. Thanks Elit Ben Basat Nuriel and Yael Warman for inviting me to be part of this fantastic event 🤩

Say goodbye to Jira headaches with the AI agent that one of our Jamba hackathon winners built yesterday at the first ever Jamba hackathon at the AGI House! Stay tuned for a full update on all the winners and on what innovative apps were built using Jamba's long context window.