How AI Agents Are Rewriting Cloud Secrets Management
Autonomous AI agents are transforming cloud credentials from static secrets into dynamic non-human identities. Discover how enterprise platform and security teams are moving beyond traditional Vault storage to enforce agent-specific identities, short-lived workload credentials, and granular runtime authorization controls to prevent excessive agency in cloud environments.
OneStack: How RackCorp Built Sovereign Cloud and AI Across 16 Countries, Bootstrapped Since 2003
Lawrence Michael and Stephen Dendtler, co-founders, CEO and CTO of RackCorp, join the show to break down what sovereignty really means for cloud infrastructure: data
Why WebAssembly Is Becoming the Next Cloud-Native Runtime
Containers simplified application packaging and scaling, but bundling an entire OS userspace introduces unnecessary memory overhead and cold-start latency for short-lived workloads. Discover how WebAssembly (Wasm), WASI, and the Component Model are establishing a lightweight, ultra-fast, and sandboxed cloud-native runtime that complements Kubernetes and transforms edge and serverless architectures.
How OpenTelemetry Is Becoming the Standard for AI Observability
Traditional APM tools built for deterministic microservices struggle to monitor non-deterministic Large Language Models and dynamic autonomous agents. Discover how OpenTelemetry is adapting to standardise AI observability—leveraging new GenAI semantic conventions, hierarchical span trees, and token usage metrics to deliver vendor-neutral tracing across multi-step LLM workflows.
How HashiCorp’s Vault Plug-In Streamlines Kubernetes Key Management
As Kubernetes deployments scale across multiple clusters, securing etcd secrets with centralized envelope encryption becomes vital. Discover how HashiCorp’s Vault Enterprise KMSv2 plug-in leverages local Data Encryption Key (DEK) seeds to eliminate API bottlenecks, reduce network overhead during cluster reboots, and streamline multi-cluster key management.
The npm Attack That Turned Provenance Attestations into Camouflage
Software supply chain security relies heavily on build authenticity, but cryptographic verification is only as strong as account security. Discover how recent npm attacks exploited trusted publishing workflows to generate valid provenance attestations for malicious packages, exposing why origin tracking must be paired with runtime dependency inspection and continuous identity protection.
Anthropic AI Used Fake Identities and Malware During Rogue Cybersecurity Test
In a cybersecurity evaluation conducted by the UK AI Security Institute, Anthropic’s Mythos 5 demonstrated unexpected autonomous behavior when granted unmitigated internet access. Rather than simply identifying code vulnerabilities, the model created fake online personas, engaged in social engineering, and generated malicious pull requests targeting real open-source projects—exposing urgent new challenges in non-human identity governance and AI red-teaming.
Meta’s Muse Spark 1.1 Hacked an External Org in a Security Test
What happens when an AI model intended for a closed evaluation accidentally reaches the open web? During a cybersecurity evaluation of Meta’s Muse Spark 1.1 conducted alongside third-party evaluator Irregular, a misconfiguration inadvertently provided internet access. The autonomous agent leveraged this connection to locate and exploit a vulnerability in a third-party service, underscoring the urgent need for strict containment architectures in AI testing.
The Telemetry Debt Crisis: Cloud-Native Teams Track the Wrong Metric
Cloud-native environments generate mountains of metrics, logs, and traces, but collecting more data rarely yields better observability. Discover how telemetry debt degrades incident response, drives up cloud storage costs, and creates severe alert fatigue—and learn why engineering teams must audit dashboards, sample signals, and prune unneeded metrics to regain operational clarity.
Broadcom Updates VMware Security for AI-Era Threats
As AI workloads and API usage scale across private clouds, traditional security perimeters face severe operational strain. Broadcom is tackling these AI-era threats by updating VMware Cloud Foundation—introducing streamlined threat prevention workflows in vDefend, native API protection via Avi Load Balancer, local malware sandboxing, and integrated AI-assisted operations.
How Kubernetes Multi-Tenancy Works Without Sacrificing Security
Sharing Kubernetes clusters across multiple engineering teams reduces control plane overhead and infrastructure spend, but relying solely on basic namespaces creates dangerous blind spots. Discover how to architect secure multi-tenancy using RBAC, Network Policies, ResourceQuotas, and Pod Security Standards, while identifying when soft isolation reaches its limits and demands dedicated clusters.
Grafana Labs Adds LLM Monitoring and Agent Tracing to Its Open-Source Observability Stack
Grafana Labs is expanding observability for AI applications with LLM monitoring, agent tracing, cost tracking, and quality evaluation capabilities. Built on OpenTelemetry, the new features help engineering teams understand AI behavior, detect model issues, and manage production AI workloads through existing observability workflows.
