AI + QA Insights

A shell alias saved me from typing the same flag hundreds of times. Here's how Claude built it for me.

Early impressions of Claude Code's /btw side questions and /voice input — from someone who already pays for Superwhisper.

My army of AI agents was supposed to test an app, generate automation, and run it live. Nothing worked. That turned out to be the best thing that could have happened.

After 15 years in QA leadership, AI tools let me start building the solutions I'd always imagined. This is the story of how a tester became a builder.

Learn a practical 5-layer testing strategy for AI-generated code, built from real QA experience and backed by research from Anthropic and OpenAI.

I orchestrated multiple Claude agents into an autonomous dev pipeline. The research explains exactly why it failed — and what actually works instead.

AI is generating code faster than teams can test it. If QA still starts after development, you're not catching bugs — you're documenting them. Shift left went from best practice to survival strategy.

Vibe coding feels like a superpower. You describe what you want, the AI builds it, and suddenly you have a working feature. The problem is you don't stop. You keep going until your simple idea becomes an over-engineered mess nobody asked for.

My team was spending too much time writing test cases manually. A ChatGPT experiment turned into Sarah — a Custom GPT for test management, refined through 13 versions.