Ever wondered how systems like Udio and Stable Audio turn a text-prompt into a full-fledged song? In this talk, we’ll pull back the curtain on the technology behind text-to-music generation, focusing on latent diffusion models. We’ll compare popular model architectures, break down key concepts with intuitive visuals, and explore the “why” behind their design choices. No deep learning background needed - just curiosity! We’ll end with a short interactive quiz to recap and test your understanding.