The End of GPUs? This New Chip is 100,000x More Efficient.

In this episode of Machine Learning Street Talk, Dr. Maxwell Ramstead interviews Guillaume Verdon, also known as Beff Jezos, the founder of Extropic, about thermodynamic computing and its potential to revolutionize AI. Verdon discusses his background in theoretical physics and quantum computing, leading to his current work on physics-based stochastic computers that harness the natural stochastic physics of electrons. He explains how this approach can overcome the limitations of traditional deterministic computers, particularly in probabilistic workloads, and envisions a future where thermodynamic computers power embodied intelligence and personalized AI. The conversation also explores Verdon's philosophical movement, Effective Accelerationism, and its focus on maximizing the growth of civilization through technological advancement and a complex systems approach.