
This is Fine! A podcast about resilience engineering and software The Messy 9 and Coding with AI - A Panel Discussion
12 snips
Feb 1, 2026 David Woods, resilience engineering founder and Professor Emeritus, brings foundational perspectives on the Messy 9 and socio-technical risks in AI systems. Shiri Cabral, enterprise architecture leader with experience at MongoDB and Salesforce, explains using AI for diagnostics and knowledge retrieval. They discuss AI in coding workflows, de-skilling risks, automation pitfalls, observability with AI, and designing collaborative human–AI systems.
AI Snips
Chapters
Books
Transcript
Episode notes
Always Verify AI Summaries
- Validate AI outputs against trusted sources like official docs and source links before acting.
- Treat AI suggestions as starting points and investigate rather than accept them blindly.
Judgment Beats Raw Detail
- Expertise provides judgment beyond facts; AI often supplies details but not that judgment.
- Dr. David Woods emphasizes know-how and judgment remain essential when tools give details.
Encode Good Habits Into Agents
- Encode your habitual safety and hygiene rules into prompts or agent rules when using LLMs.
- Require tests and validations in generated code to avoid breaking changes during edits.



