Revisiting the Toyota Production System in the Age of Coding Agents

As AI coding agents dramatically increase software production capacity, the industry faces a new bottleneck: verification. Remo Jansen explores how principles from the Toyota Production System (TPS)—specifically Jidoka (automation with a human touch) and Just-in-Time manufacturing—can be adapted to software engineering. By treating AI-generated code as a manufacturing process, developers can implement 'Andon cords' to halt production when anomalies are detected, establish 'poka-yoke' (mistake-proofing) environments, and enforce strict Work-in-Progress (WIP) limits to manage verification debt. The article argues that as the cost of generating code approaches zero, the primary challenge shifts from speed to quality control. Success in the age of autonomous agents will not depend on building perfect AI, but on designing robust production systems where defects are contained, visible, and corrected through automated, tiered inspection pipelines and continuous process improvement.