
AI DevOps Podcast Jimmy Bogard: AI-Driven Development - Episode 414
22 snips
Aug 10, 2026 Jimmy Bogard, independent .NET consultant and creator of AutoMapper and MediatR, shares views on AI-driven development and tooling. He talks about how AI changes feedback loops and what still requires human requirements. He explains where AI handles routine code vs the complex work developers keep. He discusses architecture, maintaining quality with AI, and practical developer tooling choices.
AI Snips
Chapters
Transcript
Episode notes
Moving From Open Source To A Commercial Product
- Jimmy launched a commercial product around July 1st after years of open-source maintenance and has spent 12+ months learning software licensing and company operations.
- He finds it rewarding to see who actually uses his work rather than only download counts, giving him direct customer connections he lacked with pure open source.
Prompting Mirrors Writing Good User Stories
- Writing prompts and agent instructions is essentially the same craft as writing clear user stories and technical notes for human developers.
- Good requirements still determine output quality: garbage in equals garbage out, but agents now give faster conversational feedback.
Most Routine Code Is Agent Generated
- Jimmy estimates roughly 95% of routine enterprise code is now agent-generated while he focuses on giving agents more context and higher-level design.
- He spends hands-on time on the remaining 5% that's exploratory or architecturally complex.

