Jim Keller, a renowned CPU architect, traces his career trajectory from early engineering influences to leading complex hardware projects at industry giants including DEC, AMD, Apple, and Tesla. His professional philosophy emphasizes the importance of layered design abstractions, the necessity of team-based problem-solving, and the critical role of performance modeling in chip development. Keller highlights the evolution of semiconductor design, noting how shifts from custom, transistor-level engineering to standardized, synthesis-driven methodologies enabled greater scalability. He also reflects on his transition from high-performance computing to AI-focused hardware at Tenstorrent, where he now prioritizes end-to-end system design. Throughout his journey, Keller advocates for first-principles thinking and the continuous integration of hardware, software, and compiler development to overcome the computational walls inherent in modern processor architecture.
Part 1: Early Career, DEC, and Alpha
Part 2: AMD, Networking, and Apple Silicon
Part 3: Zen Architecture and Tesla Autopilot
Part 4: AI Computing and Future Outlook
Sign in to continue reading, translating and more.
Open full episode in Podwise
