Florian Michahelles’ Post

Massimo Villari from University of Messina Just delivered the closing keynote at #iotconference on why AI has accelerated so rapidly and where it’s heading next. He explained how AI took off when machine learning met cloud-scale computing around 2015, enabled by distributed processing (MapReduce). Recent breakthroughs came from attention mechanisms and large generative models trained on massive datasets. With models now reaching trillions of parameters, the focus has shifted to fine-tuning and running lighter versions at the edge. As new trends he identified continual learning, embodied LLMs for robotics, agentic workflows, and distributed intelligence across cloud, containers, and IoT boards. Massimo wrapped up his talk with the business implications and the big question: Are we in an AI bubble? Opinions remain mixed—between fears of job displacement and clear limits of current LLMs. Exciting times ahead for edge and agentic AI. Schahram Dustdar Stefan Nastic