I just published “Basic LLM performance testing of A100, RTX A6000, H100, H200 Spot GPU instances from DataCrunch”
I had a great experience with Crossplane as a cloud-native multicloud provisioing control plane. It seems that this approach become important for Google as well, so they created their own solution - Kube Resource Orchestrator (kro)
Abacus AI CodeLLM coding experience
H200 spot instances become more affordable at datacrunch.io . Do you know of any more cost-effective services?
Grok 3 vs Deepseek r1 impressions
vCluster: Multi-tenancy in Kubernetes
LLM reasoning abilities comparation (lineage relationship quizzes) from lineage-bench
I just published “Basic LLM performance testing of GPU-powered Kubernetes nodes from Rackspace Spot”
AI Strategy Roadmap from Gartner
I just published “LLM context size support in Ollama Helm-chart”
Architects’ favorite topic for discussion is the roles
To read: Building Medallion Architectures
It’ll be interesting to see the comparison of AMD Ryzen AI Max+ 395 and NVIDIA DIGITS for 70b LLMs when it comes out.
To think: Zero knowledge design patterns for Hyperledger Fabric
Database Architectures collection
KubeVPN: Kubernetes port-forwarding replacement
kcp: Kubernetes Virtual Workspaces