29 July 2024
A podcast by the Directorate General for Translation of the European Commission

What is a Large Language Model (LLM)? A neural network? A supercomputer? How and why is DG Translation building a new multilingual LLM?
The new episode of DGT Radio, the podcast of the Directorate General for Translation of the European Commission (DGT), explains it all in simple terms. The AI team in DGT walk us through their LLM project, describing its unique assets: high quality linguistic data, checked by language experts, running on state-of-the-art supercomputers.
They also explain how these supercomputers age and rejuvenate, and give their (rather optimistic) vision of the environmental impact of AI (also available in all podcast apps and Spotify).