Chasing the Main Thread - Adventures in AI Assisted Coding
summary
Daily hands-on output: a workflow that produces 400+ posts so far by chaining ideation in ChatGPT/Claude with implementation in Gemini, Anthropic, Cursor and IBM Bob, review via GitHub and Quarkus-guided four-pass technical reviewer, and publishing on Substack. Tutorials are authored as executable step-by-step plans, validated with tests and guarded by ~1,000 lines of skill files that encode modes, constraints and formatting rules. Expect model drift, hallucinations (e.g., non-existent Quarkus extensions), deliberate human editing, and no full-agent, unattended autonomy.
AI-written summary · about
chapters (15)
- 0:00Introduction and background
- 3:05Daily blogging challenge
- 8:04AI workflow setup
- 13:24Evolution of the process
- 17:41Tutorial structure and skills
- 23:21Human involvement and authenticity
- 27:45Model drift and version stability
- 31:16AI failure modes and risks
- 34:47Multi-step tutorials and hallucinations
- 37:37Security and trust in automation
- 39:04Automation spectrum and workflows
- 42:15Complexity and problem breakdown
- 44:01AI-enhanced process benefits
- 45:25Physical toll and burnout
- 46:09Q&A and closing
related talks
45mPanel Discussion on AI for Code Generation
58mAI, Determinism, Code Generation and Refactoring
50mIdentifying and fixing Issues in Code using AI based tools
48mMastering the AI Shift: Adaptive Strategies for Software Evolution
51mAI-Powered Development: Hands-On Techniques for Immediate Impact
26mI’m Not Writing Code Anymore—Now What? | The 7-Step Workflow
30mBeyond the Hype: Using AI Coding Agents Today
34mUnsession: "AI coding tools"