Large Language Models have moved rapidly from research labs into everyday tools, changing how we create, search, learn, code and interact with technology. But beneath their impressive capabilities are important questions about how these models actually work, where their limitations lie, and what the next generation of language models might look like.

​

Moderated by Bernard Montebello, Head of Malta’s European Digital Innovation Hub at the MDIA, the panel brings together Prof. John Abela, Prof. Carl James Debono and Dr Sandro Spina from the University of Malta. Drawing on expertise spanning AI, machine learning, computer science, computer vision and advanced computing, they will explore the evolution and future of LLMs. The discussion will consider whether continued increases in model size and computing power will keep delivering meaningful improvements, the potential of alternative architectures and specialised models, and key questions around reasoning, reliability and efficiency, separating genuine technological progress from hype.

​