
AI DevOps Podcast Greg Leonardo: AI & Azure - Episode 424
19 snips
Oct 19, 2026 Greg Leonardo, Azure Solution Architect, Microsoft trainer, author, and founder of Webonology, explores AI’s impact on cloud engineering. They discuss the risks of vibe-coded software, shrinking development teams, frontier AI controls, and ownership of AI-generated work. The conversation dives into least-privilege identities, safely automating Azure, catastrophic production access, multi-agent oversight, and why human governance remains essential.
AI Snips
Chapters
Books
Transcript
Episode notes
Vibe Coding Repeats FrontPage’s Security Mistakes
- AI-generated applications often lack enterprise architecture, security, and governance because vibe coders optimize for visible functionality.
- Greg Leonardo compares this to FrontPage, which let non-developers publish websites without applying enterprise security practices.
AI May Turn Jobs Into Fractal Employment
- AI lowers development costs, but its effect on employment remains uncertain as fewer people may produce far more software.
- Greg Leonardo expects smaller teams and fragmented work arrangements to replace some traditional full-time roles.
Never Run AI Agents Under Your Identity
- Run AI tools with dedicated, least-privileged service accounts rather than personal identities.
- Connecting an agent to Microsoft 365, Slack, or network drives under an administrator account can expose everything that identity can access.

