Beyond Coding

← Beyond CodingNew · 1 h 15 min

Lead Architect Reveals His Rules for Coding Agents

Lead Architect Reveals His Rules for Coding AgentsNew1 h 15 min

<p>Lead Architect Carlos Kelkboom shares his rules for coding agents and how great engineers architect software with AI. One of the great ones is: he throws away generated code the moment he doesn&#39;t understand it. He breaks down why coding agents fail at iterative changes, why his main agent never touches files, and how he scaffolds projects for one-shot execution.</p><p>In this video, we cover:</p><ul><li>Why AI coding agents struggle with iterative changes (SlopCodeBench) and excel at one-shot execution</li><li>How to manually scaffold code, types, and folder boundaries before letting sub-agents run</li><li>Designing a distributed Markdown and indexed-search graph harness for progressive context disclosure</li><li>Using Git worktrees to run parallel coding agents and merge the best implementations</li><li>Why dynamic memory and the heap need to disappear in favor of deterministic programming languages</li><li>How AI bubbles and solo coding impact junior engineers and team collaboration</li></ul><p>This episode is for software engineers, tech leads, and system architects who want to move beyond &quot;vibe coding&quot; and build maintainable, deterministic production software with AI agents.</p><p>Timestamps:<br>00:00:00 - Why Carlos Throws Away Working AI Code<br>00:00:53 - Why Pull Requests and Team Collaboration Are Dying<br>00:03:50 - The Trap of Starting Projects You Never Finish<br>00:06:06 - Escaping the Solo AI Developer Bubble<br>00:11:54 - How to Scaffold Code Before Using AI Agents<br>00:15:54 - Learning New Programming Languages With AI Agents<br>00:21:14 - Why We Must Get Rid of the Heap<br>00:28:32 - Building a Shared Language Across Engineering Teams<br>00:31:31 - Vibe Coding vs Deep Software Engineering<br>00:40:12 - Code Is Cheap: Git Worktrees for Parallel Agents<br>00:41:45 - How Deep to Scaffold Before the Agent Runs<br>00:45:09 - Democratizing Software, Missing Juniors, and the AI Bubble<br>00:53:33 - Designing a Graph Harness for AI Agent Context<br>00:58:18 - Why His Main Agent Never Touches Files<br>01:03:07 - Throw Away Code You Don&#39;t Understand<br>01:06:25 - Rethinking Code Reviews and Mutation Testing for AI</p><p><a href="#CodingAgents">#CodingAgents</a> <a href="#SoftwareEngineering">#SoftwareEngineering</a> <a href="#AgenticEngineering">#AgenticEngineering</a></p>