Jailbreaks, sandboxes, and the limits of AI safeguards
Matt walks us through how the attacks actually work, from jailbreaks found on small open weights models that transferred straight to frontier systems, to training AI attackers with reinforcement learning
Top insurer CNA disconnects systems after cyberattackcyberscoop.com
China-based hackers used front companies to hack Uighurs, Facebook sayscyberscoop.com
Federal advisories detail bitcoin payments to ransomware gangs, urgency of threatcyberscoop.com
US racing to address Microsoft vulnerabilities, especially for small businessescyberscoop.com
Energy giant Shell impacted in Accellion hackcyberscoop.com
Thousands of Exchange servers breached prior to patching, CISA boss sayscyberscoop.com
Sidecars: A low-latency trust boundary for SandboxesModal
How Hunch supercharged AI workflows with Modal SandboxesModal
Policy as Code for Any Cloud ProviderPulumi
Introducing Origin Shield Concurrency Limits and QueueingBunny
Introducing Usage Limits for Pulumi NeoPulumi
Modal Sandboxes are generally availableModal
Scaling to 1 million concurrent sandboxes in secondsModal
Runtime Roundup: VM Sandboxes, Multi-node clusters, and moreModal
Coding Agent Horror Stories: The rm -rf ~/ IncidentDocker
Run OpenClaw Securely in Docker SandboxesDocker
Migrate to AI Gateway Using Your Coding AgentVercel
Running Container Images in AWS LambdaPulumi
Why MicroVMs: The Architecture Behind Docker SandboxesDocker
Why I joined Docker: security at the center of the software supply chainDocker
