New AI Chronicles episode just dropped! This one’s with Daniel Marques from Google DeepMind — and it’s a goldmine for anyone curious about where research meets real-world AI systems. Daniel and I talked about how even without new research, the number of places we can already build new things with today’s models is insane. We also got into: • LLMs as judges and why speed might matter more than perfection • How research turns into production-ready AI • Why understanding current models still has a decade of work left in it 🎧 Give it a listen if you’re into agent memory, evaluation, or just curious about how DeepMind engineers think. YT: Spotify: https://open.spotify.com/episode/1jJhReiHzxp0lutfAUjAy8?si=wM7dYNN0QT2ZYSBcCuwTTQ