The Data Stack Show The PRQL: From Theoretical Physics to AI: Misha Laskin on AGI, Superhuman Intelligence, and Autonomous Coding
Feb 17, 2025
Dive into the fascinating journey from theoretical physics to the world of artificial intelligence. Discover how scientific challenges shape AI's evolution and the need for robust evaluation methods. The conversation reveals insights into AGI, superhuman intelligence, and the concept of autonomous coding. It's a captivating blend of science and technology that sparks curiosity about the future of AI.
AI Snips
Chapters
Transcript
Episode notes
From Physics to AI
- Misha Laskin's journey into AI began after witnessing AlphaGo's victory.
- Inspired, he transitioned from theoretical physics, pursuing a postdoc in reinforcement learning before joining DeepMind.
Predicting AI Success
- Evaluating AI's success is crucial before customer deployment.
- Data teams should prioritize setting up robust evaluations to predict real-world performance.
Prioritize Evaluation
- Focus on setting up clear evaluation metrics early in AI projects.
- This helps predict AI performance and is often overlooked.