davidfowl: AI is proving most developers are pretty awful testers
iamKierraD · x · 2026-09-15
David Fowler, a prominent figure in the .NET ecosystem, argues that AI coding assistants are exposing a long-standing weakness in the industry: most developers are poor testers. As AI generates more code that humans must verify, the gap in testing skills becomes glaring. The quip has sparked discussion about what core engineering skills matter in the AI era.
More from coding & agent
- GPT-6 Astra says you can delete scaffolding — just not the control kind — Slight_Republic_4242 · 2026-09-15
- You Don't Need Ultra: Power User's Rule of Thumb Says Low/Medium Tiers Suffice for Daily Tasks — soumitrashukla9 · 2026-09-15
- Stop Asking AI Agents 'Can You Build This?' — Use This Inspection-First Prompt Instead — taherdhanera · 2026-09-15
- Claude Cowork 101: A Beginner Tutorial Written for Academic Researchers — JeremyNguyenPhD · 2026-09-15
- ASC: a fast Android decompiler front-end built for AI agents, 1K stars in days — MG1937 · 2026-09-15
- AI agent toolkit pi hits 105K GitHub stars with unified LLM API and coding agent CLI — earendil-works · 2026-09-15