Episode summary
Hardware is having a moment again — not because it went away, but because the limits of silicon, power, cooling, and data movement are now shaping the future of AI infrastructure. In this episode, Shawn Rosemarin sits down with Hari Kannan, a trained electrical and computer engineer who has spent his career designing processors, systems, and hardware architectures for the data center era. Hari traces the evolution from CPUs and multi-core computing to the rise of GPUs, XPUs, and specialized accelerators built for training and inference. He explains why matrix math is the easy part, why…
Chapters
- 00:00 — — Introduction & The Hardware Renaissance
- 00:53 — — Guest Background: From Apple to Everpure
- 03:15 — — The Cracking of Moore’s Law
- 05:00 — — The AI Revolution and the Rise of XPUs
- 06:37 — — Breaking Out of the CPU “Swiss Army Knife”
- 10:20 — — The Global Power Crisis & System-Level Computing
- 13:51 — — Capital Expenditures & The Reality of Silicon Limits
- 18:26 — — Extreme Innovation: Data Centers in Space
- 20:50 — — Looking Into the Future: Alternative Technologies
- 23:37 — — Career Advice for the Next Generation
Leave a Reply