Banking, tech firms adopt Llama AI

Big Meta blog disclosure on Llama. What makes this post most interesting for me are the enterprise examples of LLama delivering measured value. Of course, the download and usage numbers are off the rails. Net-net, Llama looks like an enterprise juggernaut delivering measured value to some big enterprises across different verticals. Of course, I always want more customer data with stats. This is a great start, and I am hopeful Meta will disclose more over the coming months. And, of course, even better models. Stats: -Downloads: 350M lifetime on Hugging Face, 10x YoY, 20M past month -Cloud token volume: 2x May-July CSPs, 10x Jan-July largest CSPs -LEAP partners: 5x increase since Llama 3.1 Llama Enterprise customer examples: -AT&T: 33% improvement in search-related responses in customer service, reduced costs, and faster resolution -Accenture: expecting 70% productivity and 20-30% quality improvement in ESG report creation -Goldman Sachs: document information extraction -DoorDash: automating daily tasks for software engineers, like creating pull request reviews -Nomura: code gen, log analysis, document processing -Shopify: product metadata enrichment, 40-60M inferences/day -Zoom: AI companion service Key Partners: -CSP & AI enterprise platforms: Amazon Web Services (AWS), Microsoft Azure, Google Cloud -AI/Data Platforms: IBM watsonx, Databricks, Snowflake, ScaleAI -Infra: Dell Technologies, NVIDIA BLOG: https://lnkd.in/ed3_wEfK

Today, Meta posted that Llama 3.1 has been downloaded nearly 350 million times, 10x more than this time last year. I'm proud to share that Accenture is leading the way among early Llama 3.1 users. In fact, we are leveraging it as part of our AI Refinery to build a custom LLM for Environment, Social, and Government?(ESG), which is expected to improve productivity by 70% and quality by?20-30% for ESG reporting. We are testing this ESG LLM with our own Accenture reporting.?We're also looking for other opportunities to use its breakthrough capabilities internally and with our clients -- for example, using its multilingual functionality to enable localized, culturally relevant experiences across the globe. The numbers and Accenture's own experience are clear -- open-source AI models like Llama 3.1 are creating a diverse ecosystem where companies have more choices and capabilities than ever before. ? Congratulations to Meta on your impressive achievements with Llama 3.1 and to our team led by Lan Guan, Karthik Narain, Senthil (Sen) Ramani, and Stephanie Jamison on this early ESG use case for Accenture's AI Refinery. I'm excited to continue bringing customized AI to enterprises. ? https://lnkd.in/gphRhE3a

Meta are making AI Llama models largely free. And they are being embraced by some of the biggest banks and companies in the world. The risk is these AI model inputs and assumptions are controlled by Meta…

Stoked to see Meta acknowledging our groundbreaking work demonstrating the fine-tuning of an ESG Domain LLM

?? Llama's 10x growth in just a year makes it the dominant open source AI model and a bellwether for open source AI development. Some impressive stats from Meta's latest update: → Llama is now the leading open-source model family, with nearly 350M downloads (10x growth YoY) and 20M+ downloads in the last month alone. → Major cloud providers are seeing massive Llama adoption. Monthly usage grew 10x from Jan to Jul 2024 for some, and the 405B variant is gaining the most traction. → Meta's commitment to open AI is driving a vibrant ecosystem. As Mark Zuckerberg stated, "Open source promotes a more competitive ecosystem that's good for consumers, good for companies (including Meta), and ultimately good for the world." The industry use cases are promising: → Enterprises like AT&T, Goldman Sachs, and Zoom are leveraging Llama to drive innovation and efficiency gains across customer care, ESG reporting, and productivity. →Startups and developers are building on Llama at an unprecedented pace, with 60,000+ derivative models on Hugging Face. As Meta continues to push the boundaries with Llama 3.1 and beyond, the shift to open models has gained unprecedented momentum. Exciting times ahead! ?? Read Meta's full update here: https://lnkd.in/gCJrkDpM

At KPMG, as we continue to accelerate our efforts exploring “the possible” with AI, we are empowering our clients to leverage Llama, Meta’s open-source large language model (LLM), to address a multitude of business challenges. With comprehensive training on diverse datasets, advanced fine-tuning capabilities, strict adherence to ethical guidelines, continuous learning, and robust multilingual support, Llama is equipped to support a wide array of industry-specific use cases. Discover real-life examples of how KPMG is utilizing Meta Llama to streamline and enhance complex business processes. #Innovation #LLM https://bit.ly/3Mpsa5d