AI Now Writes Most Code at Uber, Engineers Shift to Oversight and System Design

Uber's Chief Technology Officer, Praveen Neppalli Naga, has revealed a substantial transformation in the company's software development approach, where AI is increasingly taking over code generation. This pivotal change is redefining the roles of engineers, moving them from hands-on coding to more strategic tasks like architecture design, reviewing AI-produced code, and ensuring the reliability and scalability of systems. According to Naga, about 95% of Uber's engineers now use AI tools monthly, indicating a widespread adoption of these technologies. The company's internal AI coding agent autonomously generates approximately 1,800 code changes each week. The proportion of AI-generated code has rapidly escalated from under 1% to roughly 8% of all code changes in just a few months. This evolution involves engineers delegating entire tasks to autonomous AI agents, shifting beyond mere suggestions or autocomplete functions. This "agent-style workflow" means AI systems operate more independently. Even within traditional Integrated Development Environments (IDEs), AI now generates approximately 70% of committed code. This organic adoption by engineers, rather than top-down mandates, marks a "reset moment" for engineering at Uber.