AI Summary
→ WHAT IT COVERS Catherine McGarvey, SVP of Developer Experience at Capital One, explains how the company enables 14,000 technologists to move faster while maintaining security and compliance through standardization, AI-powered coding assistants, and continuous deployment practices. → KEY INSIGHTS - **Make compliance easy:** Capital One defaults developers to pre-approved, secure tools and databases rather than requiring manual security reviews for each choice. Teams can request exceptions when needed, but the default path is both fastest and most secure, reducing friction while maintaining controls. - **Measure outcomes, not output:** Track time between deployments and user satisfaction scores rather than lines of code or PR counts. Focus on whether tools actually solve developer problems and enable continuous deployment. Metrics like time-to-first-commit for onboarding reveal if teams have what they need to contribute quickly. - **AI coding assistants provide asymmetric value:** Junior developers gain the most lift from AI tools for understanding codebases, learning new languages, and getting unbiased answers. Senior developers see more value in automated migrations, test generation, and eliminating low-value tasks like dependency updates rather than core development work. - **Standardize selectively for leverage:** Only standardize tools when multiple teams adopt similar solutions or when open standards exist. In rapidly evolving areas like LLMs, use abstraction layers and consistent evaluation criteria across pilots rather than locking into one vendor, enabling teams to switch models as technology improves. → NOTABLE MOMENT McGarvey compares resisting AI tools to taking an open-book test but choosing not to use the book. She argues engineers who fear these changes likely enjoy tasks that automation now handles, and should find roles aligned with their preferences rather than missing productivity gains. 💼 SPONSORS None detected 🏷️ Developer Experience, AI Coding Assistants, Enterprise DevOps, Developer Productivity