Chema Alonso talk, the whole history of AI in one arc: algorithms (you explain the how) → machine learning (it finds the vaca/burro boundary itself) → deep learning once Big Data and compute arrive → GANs in 2014 (a generator and a discriminator locked in a duel, the answer isn’t the correct one, it’s the one that beats the discriminator, and that’s where hallucinations come from) → diffusion models, which work like our memory: fuzzy copies, not photographs, which is also where creativity lives. Key concept: human parity, not perfection, just fewer errors than us.