Truncated: latest AI news

Friday 6 November

Kubernetes, mark II? Why Google's giving away gVisor, a code sandbox used by Anthropic and OpenAI

AI-generated summary Google is donating three core AI agent infrastructure projects, including gVisor, Agent Substrate, and OpenRL, to the Cloud Native Computing Foundation. The technologies are used by companies such as Anthropic and OpenAI to help safely isolate code execution and manage autonomous agent workloads.

Also covered by aipractice.osuslova.pro

Everyone should assume they are under attack: Circle CEO on AI's risk to US infrastructure

AI-generated summary Circle CEO Jeremy Allaire says the rising power of frontier artificial intelligence has dramatically accelerated threat vectors for critical infrastructure, hospital systems, energy systems, utilities, and financial industry participants. Circle reported third-quarter revenue and reserve income of $865 million and announced a $400 million all-stock agreement to acquire Tazapay.

CPython flaw exposes AI inference servers to sandbox escape attacks

AI-generated summary Security researchers at RunSybil say a memory-safety flaw in CPython allows malicious AI model templates to escape code sandboxes and execute arbitrary code inside inference servers. The Python Security Response Team disputed the security implications of the demonstration, stating that executing untrusted Python code within a sandbox is outside CPython’s security threat model.

Also covered by infoworld.com

Russia and US seek to block plan for binding rules on autonomous weapons

AI-generated summary Plans to fill a gap in the global governance of autonomous weapons face opposition as a group of nations led by Russia and the US move to block proposals for a legally binding set of rules at a UN conference. The UN conference in Geneva brings together 128 countries to discuss whether to back a mandate to negotiate a legally binding instrument within a defined timetable.