Jòan Challenges With Deploying AI in Web Dev
summary
Deploying an AI-centered feature fails when its predictability drops: internal tests and deterministic code pass, but at scale the model’s nondeterminism produces bad outputs for 5% of users, making the experience unacceptable. The claim is that traditional unit and integration testing no longer suffices, so teams adopt different QA techniques — large-scale monitoring, user segmentation, staged rollouts and rapid UX iteration — to find and fix the surprising edge cases.
AI-written summary · about
related talks
32mQuality engineering in the AI age: From testing AI to leveraging GenAI
41mSurviving and Thriving with AI in QA
14mAI Sprawl: The Hidden Risk Threatening Your Enterprise Now
55mSurviving the AI Testing Apocalypse
24mHow to Test AI Features Before Your Users Do
43mTaming the Unpredictable: Reliability in Chaos
37mThe AI Factor: The Blueprint for Intelligent Deployment
20mKeynote: Testing In The Era of AI - From QA to Trust Engineering