Innodata Inc.’s Post

I'm deeply honored to share that I'll be speaking at the AI Accelerator Institute Generative AI Summit in NYC on Jun 12. I'm thrilled about the opportunity to explore the transformative potential of Large Language Models. During my presentation, I'll discuss impactful use cases and showcase adapting transformer-based language models for Text Summarization. I'd be delighted if you could join me and participate in the conversation. #LLMs #AIAINYC #GenerativeAI

Wanting to learn more after the PassiveLogic 2023 Launch Event? 👀 In a new blog series, CEO Troy Harvey goes into more detail about the four key breakthroughs that make generative autonomy possible and enable autonomy for any controlled system. First up: A deep-dive on generative autonomy and why we need it to transform the built world. Autonomy is more than single purpose chatbots and single-purpose operations that require millions of dollars and months to train. True autonomy is flexible, dynamic, and capable of adapting to understand and manage any controlled system in a matter of minutes. True autonomy—by definition—requires generative autonomy 🤖 Stay tuned for the deep dive on the 4 breakthroughs powering generative autonomy:  1️⃣ A Revolutionary Ecosystem of Digital Twin Applications  2️⃣ A Breakthrough Platform for Generative Autonomy  3️⃣ The 1st Digital Twin Language for Composable AI 4️⃣ The World’s Fastest AI Language and Toolchain Read more from Troy on generative autonomy. Join the software waiting list to have first access to the tools of generative autonomy at pssvlgc.com/launch ⬇️ https://lnkd.in/g-CafdKf

The promise of artificial intelligence is immense. Given the scale and scope of the opportunity, business leaders are energized to get started -- but also uncertain as to where and how. If you're feeling this state of frisson yourself when it comes to AI, join my colleagues for a live discussion this Wednesday. They'll cut through the hype by drawing on tested Innosight lenses for understanding the impact of new, disruptive technologies, and offer smart ways to get your footing in this new arena. Save your space. https://lnkd.in/eTV6DQjT

Our latest issue of Member News has landed! 📰 Explore captivating stories, from insights into our Professional Registration process to unlocking the potential of AI in engineering practices. 🚀 Dive in and stay informed! https://spkl.io/60484FPDm

💚 This is what we live for - helping nonprofits have more impact through AI + BI. 👋 If you are at #BBCON, come say hi to the #Dataro team at booth 5 in the Marketplace and chat all things #AI in #fundraising.

Great post and insights by Vibhor Rastogi on state of the #GenAI market. Pay particular attention to point #3 on GPU costs exceeding payroll costs for GenAI companies (#gettingthere) and #4 so we avoid a repeat of the hype and valuation collapse we saw over the last 10 years (#beenthere) >> Exits via M&A will be difficult for companies raising at hundreds of millions to billions of dollar in valuation at the peak of the current hype cycle. At these valuations, companies may not have a choice, but to optimize for liquidity via public markets

I will be flying to London this Wed 12th, to participate in a panel at the AI Summit. I'll be joining representatives from Boston Dynamics and Google. I'll discuss our recent paper from the #Mastercard AI Engineering team, focusing particularly on 'Why Tech Companies Need to Match Innovation with Responsibility.' Feel free to join me if you're around! #mastercard #LLMOps #responsibleAI #ai_engineering_at_mastercard

Successfully completed the "Adopting AI in you Organisation" module, a part of Quantic's CTO Program (https://lnkd.in/edu7jzS9). It included how to navigate essential technology decisions for crafting AI-driven applications, exploring model evaluation, ethical considerations, and current regulatory trends. 

One of the things I enjoy reading on LI are Vibhor's insights on AI/MLs and portfolio companies. One of their points in his summary on the recent Efficient Generative AI Summit got me thinking about a question I've asked AI practitioners over the last year -- what is driving organizations to invest in modifying existing models like Meta's Llama2 for their own use versus picking up pre-trained models. PS... another tidbit that Vibhor shared that deserves repeating: "Domain specific, smaller models may prove to be more effective for applications such as Q&A on corporate data than general purpose or even finetuned foundation models." I know Debojyoti has strong opinions (for) on the same.

Thank you to organizers of Efficient Generative AI Summit (https://lnkd.in/ghNea9eN) for arranging a great panel discussion with Michael Stewart, Rashmi Gopinath, Dr. Yvonne Lutsch and Gayathri Radhakrishnan. Key takeaways 1. Open source models such as Llama2 which allow commercial licensing (https://lnkd.in/ghRGxNem) are game changers and this could be the Android moment for the Gen AI Industry. 2. Domain specific, smaller models may prove to be more effective for applications such as Q&A on corporate data than general purpose or even finetuned foundation models 3. GPU costs are now exceeding payroll cost for GenAI companies, which needs to be factored by VCs as they think about cash burn and runway 4. Companies are exploring multiple tuning and optimization techniques to reduce training and inferencing costs - https://lnkd.in/gtmtycGu and https://lnkd.in/gvXAZzGT 4. Exits via M&A will be difficult for companies raising at hundreds of millions to billions of dollar in valuation at the peak of the current hype cycle. At these valuations, companies may not have a choice, but to optimize for liquidity via public markets