Great perspective—especially the point that speed ≠ quality. AI can generate test cases fast, but it doesn’t question assumptions or connect edge cases across systems. That’s where real QA thinking matters. Loved the “QX Notes” idea—feels like the right way to combine human insight with AI scale.

This is a solid direction—you're solving a real pain point. Big takeaway: agent capabilities are moving faster than infra/security. Remote browser relay = huge UX win 135k exposed instances = serious problem Token optimization = practical value Feels like the real winners in AI will be infra-first, not model-first. Curious how you handle security + observability at scale.

I think most companies are still shipping AI-flavored features, not true AI-native products. We’re in a transition phase—some are moving toward workflow-level integration, but very few have fully rethought their product around AI. The main blockers aren’t just tech, but also mindset and legacy systems. The real shift will happen when companies start building for outcomes, not features.