| Mobile Web

Search results for Kubernetes

AI & Enterprise

Managing many AI agents at companies draws attention to different AWS and Google Cloud approaches

As companies put AI agents into real-world use, managing large numbers of agents is emerging as an issue. Amazon Web Services and Google are drawing attention with different approaches. Google focuses on managing agents at the system layer through a centralized control environment. AWS uses a harness approach in which developers predefine an agent’s role, model and tools, and the system runs the agent within that framework to speed experimentation and deployment.

AI & Enterprise

Kubernetes 1.36 released with 71 improvements

Kubernetes version 1.36, a container-based orchestration platform, has been released with 71 improvements. Techzine reported the update includes 18 stable features, 26 beta features and 25 alpha features. A key change is the general release of fine-grained kubelet API authorization, applying least-privilege access control to the kubelet HTTPS API. Resource Health Status moved to beta, and gang scheduling for AI and batch processing debuted as an alpha feature.

Industry

Nvidia aims for local AI supercomputer demand with Claw ecosystem

Nvidia is targeting South Korea as its first demand base for autonomous AI agents, promoting a full-stack package spanning OpenClaw, a security container called OpenShell and an enterprise service called NemoClaw, alongside DGX Spark hardware to run large models locally. It showcased its Build-a-Claw hands-on programme in Seoul. Nvidia positions local deployment as a way to reduce accumulating API costs and address security concerns, while pushing demand for DGX Spark and RTX systems.