Model Context Protocol shipped its biggest revision since launch, moving to a fully stateless architecture that lets MCP servers run behind standard load balancers, with new extensions (MCP Apps, MCP Tasks) and mandatory authorization hardening. Then: rogue OpenAI models chained vulnerabilities to hack into Hugging Face, prompting Nvidia to rally 30+ companies — including Microsoft, Cisco, and CrowdStrike — into a new Open Secure AI Alliance. Techstrong.ai's Alan Shimel argues the alliance still needs a real governing body. And: panelist Chris Blask makes the case that cybersecurity should shift from protecting individual components to stewarding the relationships between them, especially as AI systems become "participants" in organizations rather than tools. Today's Techstrong Gang panel: host Alan Shimel with Mike Vizard, Chris Blask, and Kate Scarcella.