PinnedDaniel Dahlin·Jul 27Your Requirements Document Is an Attack SurfaceHow I defend an AI-powered QA pipeline against indirect prompt injection from Confluence, Jira, and retrieved knowledge
PinnedDaniel Dahlin·Jul 24Prove the Test Can FailWhy an AI-generated test should pass against correct behavior and fail when the defect it claims to protect against is deliberately…
PinnedDaniel Dahlin·Jul 21How I Built an Oracle to Evaluate a QA Test-Generation AgentA defect-seeded reference application, hidden ground truth, and repeatable resets turned “the output looks right” into something I could…
PinnedDaniel Dahlin·Jul 20Green Isn’t VerifiedWhy self-healing tests need evidence, risk tiers, and a hard boundary around assertions
PinnedDaniel Dahlin·Jul 17Who Gets to Decide What the AI Tests?How I encoded human approval, scope control, and auditable exclusions into an AI-powered QA pipelineA response icon1A response icon1
Daniel Dahlin·Sep 18Approaching AGI: Capability Is Not AuthorityWhy more capable AI systems make verification, controls, and human authority more important — not less
Daniel Dahlin·Sep 2How Do You QA an AI QA Agent?Why smaller checks can provide stronger evidence than smarter judges
Daniel Dahlin·Aug 24A QA Agent Needs a Job Description, Not Just a PromptWe’d never hire a QA engineer without defining what they’re responsible for.
Daniel Dahlin·Aug 21Rerunning a Failed AI Test Until It Passes Isn’t Retry LogicIn probabilistic systems, a green rerun can hide the failure rate you actually need to measure.
Daniel Dahlin·Aug 14The future QA engineer is an evaluation engineer.Everyone is debating whether AI will replace testers.