AI is smarter than you think with Shane Legg of Google DeepMind

In this episode, Guy Raz speaks with Shane Legg, co-founder of DeepMind, about his lifelong journey in artificial intelligence (AI). Legg describes his early fascination with computers and programming, which began at age ten. He studied mathematics and statistics in New Zealand and became involved in machine learning projects, eventually leading to the development of DeepMind in 2010 with the goal of creating artificial general intelligence (AGI).

DeepMind gained fame by mastering the game of Go, which required a sophisticated approach beyond brute-force computation. Legg explains the innovative combination of deep learning and search techniques that allowed their AI to defeat world champion players. After being acquired by Google, DeepMind applied its technology to various challenges, including protein folding, dramatically reducing the time previously required to understand protein structures and paving the way for advancements in drug discovery.

Legg also reflects on the state of AI research, noting the significant impact of large language models and the accelerated pace of AI development. He discusses the merger of DeepMind with Google's AI division and the implications for AI safety, emphasizing the importance of thoughtful governance as society navigates the complex challenges posed by advanced AI. Throughout the conversation, Legg underscores the transformative potential of AI and the importance of addressing risks while pursuing its benefits.