Speaker
Philippe Limantour
(Microsoft France)
Description
In this talk, I will quickly highlight what has changed with generative AI models, and why we should stay with them even if it's early days. I will provide some examples of customer use cases, and how to run these models with confidence: secure private environments and inference infrastructure, well-trained people, and a responsible AI execution framework (not just principles). I will also discuss some current work directions, at both hardware and software levels.
