AI Token Costs Are Spiraling — Rob Hirschfeld of RackN on Hybrid Infrastructure | TFiR
RackN CEO Rob Hirschfeld makes the case for hybrid AI infrastructure — explaining when self-hosting makes business ...
Cloud Native
Deep-dive interviews and analysis on cloud-native platforms: Kubernetes, containers, microservices, service mesh, serverless, and GitOps. We talk to the engineers and leaders building enterprise platforms—covering real-world architectures, security, reliability, and operations
AI Agents Fail in Production Without Workflow State Recovery | Mark Fussell, Dapr | TFiR
Most AI agents fail in production due to state loss and crash recovery issues. Mark Fussell of ...
How Does JDK 26’s HTTP/3 API Transform Microservices Performance Using UDP | TFiR
JDK 26 introduces HTTP/3 API support that switches from TCP to UDP as the transport protocol, ...
Why NVIDIA Donated Its DRA Driver to KubeVirt Community | Ryan Hallisey at KubeCon EU | TFiR
AI workloads expose a fundamental limitation in Kubernetes: static GPU allocation. KubeVirt maintainer Ryan Hallisey discusses ...
GPU Costs Are Killing AI Budgets—Volcano’s Unified Scheduling Cuts Waste | Jesse Stutler, Volcano
Volcano 1.14 evolves from batch scheduling to unified AI-native platform, handling training, inference, and agent workloads on ...
Why AI-Era Infrastructure Needs Control Planes | Bassam Tabbara, Crossplane
Bassam Tabbara of Crossplane discusses declarative infrastructure, platform engineering, CNCF graduation, and why control planes are becoming ...
Multi-Cloud Chaos Is Killing Enterprise Margins | Dirk Alshuth, emma
emma's new brownfield onboarding capability discovers and governs existing infrastructure without migration. Dirk Alshuth, CMO at emma, ...
Why Crossplane Fails at Kubernetes Scale: Julian Fischer, anynines | TFiR
Julian Fischer, CEO of anynines, explains why Crossplane alone breaks down for Kubernetes data service orchestration at ...
Your Container Images Ship With 300 Vulnerabilities Before You Deploy | John Morello, Minimus
Enterprises pull the latest Python image and start with 300 vulnerabilities before writing code. John Morello, ...
Split-Brain Explained: The Application-Level Failure Multi-Region Can’t Prevent | Philip Merry, SIOS Technology | TFiR
What is split-brain and why can't multi-region deployments prevent it? Philip Merry from SIOS Technology explains the ...