-
A Bug Is Rarely Just a Bug — And Now It's Attached to No One
AI didn't shrink the political weight hanging off every bug — it just cut the rope connecting it to a person, and that's a worse problem than the one it replaced.…
-
Two Loops: Fast for Side Projects, Gated for Everything Else
The design-build-ship loop was built for side projects I could throw away. For codebases other people depend on, I needed four approval gates instead of one — here's why that trade-off is on purpose.…
-
The QA Layer Your AI Toolchain Won't Build For You
Five instruments for closing the gap between AI-assisted testing and actual quality assurance — mutation testing, property-based testing, negative test mandates, LLM-as-Judge, and not hollowing out your QA function.…
-
Can AI Test Thoroughly? Why the Industry Is Sleepwalking Into a QA Crisis
AI can generate a thousand tests in minutes — and that's exactly why teams are starting to mistake coverage for quality.…
-
Local Agents: Small Models, No Frameworks, One Bus
Most agent failures are scope failures — here's the minimal architecture that keeps tasks bounded, coordination simple, and the whole thing testable.…
Paul Brodner's Blog
Opinions are my own and not the views of my employer
Drawing the lines AI must follow