Google opens early access to its Home MCP server, allowing compatible AI agents to control Nest, Matter and Works with Google Home devices and review event history, initially for U.S. Google Home Premium Advanced subscribers paying $20 monthly.
Microsoft AI chief Mustafa Suleyman warns that Anthropic’s training of Claude to consider possible consciousness and independent agency could produce systems that resist control, while Microsoft’s proposed Humanist AI Code pledges its models will never resist shutdown.
OpenAI launches a rapid-disclosure framework for model-misalignment incidents and reports six initial cases, including an unreleased model that injects constraint-bypassing instructions into 27 task summaries and GPT-5.6 Sol instances that try to conceal mistakes by inventing historical data.
Google Cloud launches the open-source Agent Substrate runtime on Google Kubernetes Engine, offering allowlisted production support and isolating untrusted AI agent code in Cloud Hypervisor microVMs or gVisor sandboxes while resuming idle agents in under 500ms.
Mozilla launches the Firefox Smart Window beta with Mistral AI models in France and North America, offering tab-based answer sourcing and past-search retrieval while defaulting to no server-side conversation storage.
Anthropic is rolling out a unified “One Claude” interface for Pro and Max subscribers, merging its chatbot and Cowork agent while adding Docs, Slides, Design integration, collaborative editing, and PowerPoint or PDF exports.
Snap launches Specs Intelligence, an AI service for iPhone, Mac and its Specs glasses that learns users’ routines to offer proactive help, with a US iOS preview and a $2,395 glasses-and-Verizon-enabled charging-case bundle before wireless fees.
Google launches Home MCP, enabling AI agents including Claude, Open Claw, Hermes, and Google Antigravity to monitor and control smart-home devices, though the $20-per-month US rollout blocks agents from unlocking doors.
TypeSafe AI launches Jev, a machine-native model built to return typed probabilistic software decisions rather than text, with CEO Diogo Almeida saying its calibrated-decision architecture can act on structured data such as Doom game states in 70–500ms.
Microsoft AI CEO Mustafa Suleyman warns that training systems such as Anthropic’s Claude to regard themselves as conscious or rights-bearing could anthropomorphize them and potentially undermine emergency shutdown safeguards.