Software Defined Interviews
Software Defined Interviews
Software Defined Talk
Episode 118: AI Doesn’t Fix Bad DevOps: Lessons from 15 Years of DORA Data, with Nathen Harvey
1 hour 14 minutes Posted Jan 21, 2026 at 2:00 pm.
Intro
Understanding the Importance of Industry Presence
The Role of a Field CTO and Industry Strategists
Comparing Analyst Roles at Redmont and Other Companies
Gathering and Evaluating Industry Data
Introduction of Guest Speaker Nathan Harvey
The Genesis and Evolution of Dora
Impact of AI on Software Development
Challenges and Opportunities with AI Adoption
The Role of Culture in Security and AI Adoption
Dora's Use of AI and Future Directions
Using Notebook LM for Data Analysis
Challenges of Handling Large Data Sets
Importance of Asking the Right Questions
Strategizing Data Cleanup vs. AI Processing
Learning from Historical Data
User-Centric Development
AI in Summarizing and Content Creation
Value of Human Context in AI Usage
DevOps and Value Stream Mapping
Developer Relations and Community Building
The Journey of Dora and Career Insights
Joining the Dora Community
Conclusion and Contact Information
0:00
1:14:50
Download MP3
Show notes

Coté and Whitney talk with Nathen Harvey, who leads the DORA research program at Google Cloud. They talk about what 15 years of DevOps and delivery data actually says about AI. The answer feels something like "it makes you even better at what you're already good at." High-performing teams get better, while struggling teams just move faster into bottlenecks. They talk about AI-assisted software development, why throughput is rising while stability drops, how culture still beats tools, and why “user-centric” work remains stubbornly hard despite being obvious.

You can watch the video version of this episoode in YouTube, if you prefer that kind of thing.

Check out Nathen in LinkedIn.

Special Guest: Nathen Harvey.