Vibe Coding is Dead, Long Live Viable Coding - The Blue-Red Framework for Trustworthy AI Development
-
11
-
- Write review
AI Adoption - Drive Business Value and Organizational Impact
Master Finance Tools - 35% Off CFI (Code CFI35)
Overview
Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore why AI-generated "vibe coding" failed in production and discover a structured approach to trustworthy AI development in this 35-minute conference talk. Learn from Itamar Friedman and Nnenna Ndukwe from Qodo AI as they demonstrate the critical flaws in AI-generated code through live demos, showing how dropped logging and breaking changes can slip through passing tests. Discover the blue-red framework for viable coding, where blue agents handle planning and code generation while red agents catch quality issues before production deployment. Watch demonstrations of agent-to-agent communication, CLI review workflows that eliminate context switching, and structured development modes including plan mode, code mode, and review mode. Understand the hidden complexities beneath code generation and master techniques for building software carefully with AI assistance, moving beyond crossing your fingers to implementing systematic quality assurance in AI-powered development workflows.
Syllabus
- Introduction: Who's vibe coding?
- Live demo: Blue agent vibe codes with Qodo AI
- Committing and creating the pull request
- Hold up: What about planning and review?
- Red agent: Breaking down what AI missed
- The logging and validation you dropped
- CLI review workflow: No context switching
- Agent-to-agent communication
- Welcome to viable coding
- The iceberg beneath code generation
- The blue-red framework explained
- Plan mode, code mode, review mode
- Building software carefully with AI
Taught by
All Things Open