What I find while testing software, working with AI, and breaking things along the way.
Featured
I was talking through a post on AI as a Swiss Army knife. About two minutes later, the voice session forgot it and taught Victorinox history.
Read the article →
Developers were never supposed to test their own code. Now AI writes it, reviews it, and tests it too — and the same blind spot just moved up a level.
They don't write code. They interrogate the problem in plain English until it's precise.
A line from an old manager still holds up: assumption is the mother of every mistake. AI just gave it a new place to hide.
When devs treat AI like it replaces QA, the blame doesn't disappear. It just lands on QA for missing something QA was never asked to check.
QA depends on questioning confident claims, including answers from AI tools.
