What the Dev? 336: The parts of the software development life cycle that can't be automated by AI (with OutSystems' Woodson Martin)
8 snips
Dec 16, 2025 
Woodson Martin, the CEO of OutSystems and a seasoned leader in enterprise software, discusses the complexities of the software development life cycle. He highlights that while AI can expedite coding, many tasks still require a human touch, particularly in system design and collaboration. They explore the evolving role of developers, emphasizing job satisfaction and the need to measure engineering success by impact rather than output. Woodson also shares insights on OutSystems’ platform as a low-code solution that integrates AI and enhances application development.
AI Snips
Chapters
Transcript
Episode notes
AI Helps Code But Not The Whole Job
- AI is a superpower for generating code but actual coding is only about 20% of enterprise developer work.
- The remaining 80%—architecture, integration, testing, coordination—still needs deep human involvement.
Rapid Code Generation Creates New Bottlenecks
- Faster code generation by AI may create more downstream work like cleanup and refactoring.
- Teams worry aggregate delivery speed hasn't improved despite rapid code creation.
Redesign Your Development Lifecycle
- Reimagine the software development lifecycle to fit AI-native workflows and learn iteratively.
- Spend time identifying where AI reduces churn versus where it increases it.