Uber's internal AI coding agent now generates 1,800 code changes per week without human authoring, increasing AI-written contributions from less than 1% to 8% of all coding changes at the company. Chief Technology Officer Praveen Neppalli Naga said, "This is a real reset moment for engineering."

Over the last few months, Uber has increased its adoption of AI coding tools across its engineering organization. Engineers at the company review and approve code that is written entirely by AI agents, while 95% of Uber's engineers use AI tools every month.

"The role of the engineer is shifting—from writing every line to architecting systems and reviewing AI-generated code," Naga said. "The results have been phenomenal."

At Uber, agentic AI refers to software that completes tasks autonomously with little or no supervision.

The adoption of AI coding tools extends beyond Uber across the technology industry. A study by engineering intelligence platform Jellyfish surveyed 700 companies and found that 63% of firms use AI tools for most coding, while some companies integrate AI tools into performance reviews.

Rapid advances in AI coding have created massive disruption for software engineering roles. Amazon implemented new guardrails after outages caused by its AI coding tool resulted in nearly 120,000 lost orders, and some software engineers say that using AI tools has increased their workload and led to burnout.