GPUs, TPUs, & The Economics of AI Explained | Gavin Baker Interview
In this conversation, Patrick O’Shaughnessy and Gavin Baker discuss the state of frontier AI, how to interpret rapid model releases, and what recent developments may mean for investors. The excerpt focuses on the tension between Nvidia’s GPU stack and Google’s TPU strategy, the role of scaling laws in pre-training, and why reasoning models, reinforcement learning with verified rewards, and test-time compute have driven recent progress. It also touches on the economics of AI infrastructure, from chip transitions to low-cost token production, and briefly explores how leading investors and researchers are following the field in real time.