
Bottlenecks Beyond Code Authoring

Data Beats Hype: Measuring Your AI Adoption Impact - Laura Tacho
Topic 11 of 14
Published Sep 29, 2025 · Clip 52:30–55:45 (3:15)
Laura explains AI started with coding tasks because they're complex, fun, and have abundant training data, but tools are a small percentage of developer bottlenecks. Prioritization, shifting priorities, production debugging, security, and compliance dominate. She cites an AWS study that developers spend only about 20% of their time writing code, and code generation is just a fraction of that. She believes the biggest productivity gains will come from outer-loop work, noting tools like GitLab Duo and Atlassian Rovo are extending AI across the SDLC.