Why we need to think about AI and Cloud Native NOW
summary
54% of organizations run AI on Kubernetes and the talk argues that cloud native must change now with concrete work rather than hype For GPUs it showcases dynamic resource allocation to allow partial and shared GPU slices instead of static device plugins It highlights Kubeflow model registry work because 44% of users ask for versioned model artifacts and metadata It demos WasmEdge LLMs for edge inference, pushes OpenTelemetry semantic work for LLM observability, and calls for AI supply chain attestation and contestability tools
AI-written summary · about
chapters (8)
- 0:00Introduction to AI and Cloud Native
- 3:37AI adoption and Kubernetes growth
- 10:22Dynamic Resource Allocation in Kubernetes
- 14:40AI at the Edge with WASE
- 17:55Observability for LLM workloads
- 21:11Responsible AI and ecosystem impact
- 25:53CNCF AI working group resources
- 27:18Q&A session
related talks
23mThe State of Cloud Native: The Shift Towards AI
27mCloud Native Artificial Intelligence - Top to Bottom
06mSponsored Keynote: The Weight of Data: Rethinking Cloud-Native Systems for the Age of AI
05mSponsored Keynote: Towards Clouds of AI Clusters
16mOpportunities for AI/ML with Cloud Native Applications
20mCloud-Native AI: Architectural Patterns for Enterprise Success
20mOperating Frontier Intelligence at Scale
06mAI Inference at the Edge: Beyond the Hype