The Engineering Room with Dave Farley Understanding The Value of AI Coding | Gene Kim In The Engineering Room Ep. 42
101 snips
Jan 25, 2026 
Gene Kim, bestselling author and DevOps expert, discusses AI-assisted "Vibe Coding" and how it changes software work. He shares rapid-build stories, the FAFO value framework, and why faster generation demands faster feedback and modularity. They explore who benefits, how trust grows, versioning prompts versus specs, and how roles and skills shift as AI raises engineering abstraction.
AI Snips
Chapters
Books
Transcript
Episode notes
47-Minute Video Excerpt Tool
- Gene Kim built a video-excerpt generator in 47 minutes and expected it would previously take days.
- That rapid success convinced him to lean on AI more and to write Vibe Coding with Steve Yegge.
Trust In AI Is Learned Over Months
- Trust in AI grows with sustained use and plateaus around a year.
- Gene Kim notes this learned trust lets engineers give larger tasks to AI and supervise less over time.
Make Specifications Precise And Testable
- Do make specifications precise and treat tests as executable specifications to avoid vague prompts.
- Do use tooling (like SpecKit) or formal contracts when determinism and correctness matter.
