

Agentic AI Slop vs. AI XP Excellence? Iteration, Batch Size, Testing, and the Future of Dev Work
18 snips Sep 24, 2025
Explore the clash between Agentic AI systems and Agile practices like Extreme Programming. Discover if AI can enhance productivity or just speed up chaos. The hosts provide insights on managing AI complexity, emphasizing small iterations and testing. They discuss the risks of large-batch changes and the importance of maintaining engineering rigor. With real dev experiences and practical tips, this conversation delves into the evolving role of AI in software development, from humor-filled interactions to the challenges of team adaptation.
AI Snips
Chapters
Transcript
Episode notes
Live Agentic Coding With Test Feedback
- Chris describes using agentic systems (Claude + VS Code orchestrator) to read code, run diffs, and iterate while watching tests run in real time.
- He checks diffs, watches Wallaby.js or ncrunch run tests, and refines prompts as the agent updates code.
Small-Batch XP Loop Wins Over Big Batches
- Austin shares his comfort with an XP-style small-batch AI workflow that joins a mob and TDD loop to solve small tasks quickly.
- He contrasts that with large agentic runs which produced a production bug and pushed him back toward small-batch practices.
Big Batches Amplify Risk Even If Mostly Right
- Austin notices big-batch agent runs often deliver many mostly-correct changes that still leave clustered, hard-to-audit errors.
- He finds small TDD-driven batches easier to validate and trusts that loop's higher perceived quality.