A live stream discussion between a DevOps practitioner and the president of Embrace (a mobile/frontend observability company) exploring how AI-generated code is outpacing QA and observability practices. Key themes include: AI productivity gains creating more code without proportional QA investment, the risk of AI agents writing and reviewing their own PRs with no human in the loop, the limitations of current frontend/mobile observability tools (crash reporting, core web vitals) versus measuring actual user flow success and abandonment, and how AI struggles to interpret observability data without business context (e.g., marketing campaigns causing traffic spikes). The discussion also covers Embrace's approach of measuring user journey completion rates rather than just error counts, and the challenge of bringing mobile developers into the cloud-native observability ecosystem.