Uber CTO Says 95% of Engineers Now Rely on AI, Internal Agents Push 1,800 Weekly Code Changes

Why It Matters

The adoption of AI agents at Uber illustrates how large engineering organizations can dramatically alter their development processes within months. By moving 95% of engineers onto AI‑augmented workflows, Uber is testing the limits of productivity gains versus code quality risks, a balance that will influence hiring, tooling budgets, and the competitive dynamics of the tech talent market. If Uber's model proves scalable, other firms may accelerate similar deployments, prompting a wave of investment in AI‑code review platforms, security hardening tools, and new engineering roles focused on AI oversight. Conversely, the documented increase in defects from AI‑generated code could drive regulatory scrutiny and push companies to adopt stricter validation pipelines, shaping the future standards for AI‑assisted software development.

Key Takeaways

• 95% of Uber engineers now use AI tools monthly, per CTO Praveen Naga.
• Internal coding agent generates 1,800 code changes per week with zero human authoring.
• Agent's share of total code changes grew from <1% to 8% in a few months.
• 84% of AI users at Uber employ agent‑style workflows, handing off entire tasks.
• Claude Code adoption rose from 32% to 63% in two months; ~70% of committed code is AI‑generated.

Pulse Analysis

Uber's swift pivot to AI‑driven development reflects a broader industry experiment with generative models as co‑developers rather than mere assistants. Historically, IDE autocomplete and static analysis tools offered incremental productivity gains; today, autonomous agents promise to rewrite large swaths of code, compressing development timelines but also introducing new failure modes. Uber's internal data—1,800 weekly changes and an 8% share of total code—suggests that the technology is moving beyond pilot phases into production‑grade usage.

The competitive advantage hinges on how quickly firms can embed reliable review mechanisms. Anthropic's $25‑per‑pull‑request review feature is an early indicator that a market for AI‑code validation is emerging. Companies that can balance speed with defect mitigation will likely capture talent and market share, as engineering teams become increasingly judged on their ability to harness AI responsibly. Uber's organic adoption model, where engineers self‑select tools, may also signal a cultural shift: developers are becoming early adopters of AI, reducing the need for executive mandates.

Looking forward, the next inflection point will be the transition from routine code generation to AI‑crafted architectural decisions. As Uber's CTO notes, engineers will focus on system design and oversight, a role that demands deeper domain expertise and a new set of soft skills. The industry will need to redefine performance metrics, training curricula, and perhaps even regulatory frameworks to accommodate this evolving landscape. Companies that anticipate these changes and invest in AI governance will likely set the standards for the next generation of software engineering.