
High-performance software engineering is undergoing a fundamental shift as AI agents transform from simple autocomplete tools into powerful partners for building complex, production-ready systems. Peter Mattis, CTO of Cockroach Labs and veteran of early Google projects like Gmail and Colossus, illustrates how deep domain expertise in distributed systems—specifically the use of B-trees and resilient storage architectures—remains critical even as coding becomes automated. By treating AI as a sparring partner rather than a replacement, engineers can amplify their output, rapidly iterate on design experiments, and enforce rigorous quality standards through automated testing and adversarial review. This evolution demands that developers move beyond basic coding to become "F1 drivers" of agentic workflows, focusing on architectural vision and disciplined guardrails to ensure that the massive increase in software production maintains the high reliability required for mission-critical applications.
Sign in to continue reading, translating and more.
Open full episode in Podwise