Google Home Adds Claude, OpenClaw for Smart Home Control

Key Takeaways

• <ul><li>Google Home introduces third-party AI agent support, including Claude and OpenClaw, for Premium Advanced users in the US at $20/month.</li><li>Google Home MCP enables multiple AI agents to control Nest cameras, thermostats, and Matter devices, with safety rules preventing dangerous actions like unlocking doors.</li><li>Microsoft and OpenAI face a copyright lawsuit, with internal emails calling AI training the largest theft of labor and admitting chatbots reduce news website traffic by 93%.</li><li>A hybrid AI model improves emotion recognition at a lower cost, using a cheap ensemble for most tasks and only sending uncertain cases to a large language model.</li><li>The Federal Register quietly removed China's Qwen AI from its document search options after public scrutiny of the government using a Chinese model.</li><li>Bud Ecosystem launched Bud Novaria, an AI operating system that cuts enterprise AI costs by up to 80% by running on GPUs, CPUs, and other hardware.</li><li>President Trump rejects calls for AI safety oversight, calling warnings a hoax meant to slow US competition with China.</li><li>OpenAI and Anthropic share sales have created hundreds of millionaires, with combined employee and investor sales estimated at $14 billion over five years.</li><li>Analyst Paul Meeks warns AI has a public image problem and that Chinese open-weight models could take market share from US rivals.</li><li>XRP holders are using XRPPower for AI-powered automated trading, with promotional materials claiming potential earnings of up to $3,000 per day.</li></ul>

Google Home adds AI support for Claude and OpenClaw

Google Home is now letting third-party AI agents like Claude and OpenClaw control smart home devices. This feature is rolling out to Premium Advanced users in the US who pay $20 a month. These AI tools can handle more complex tasks like building custom dashboards or linking thermostats to weather reports. The move could make smart home setups more flexible, though some experts worry about reliability and privacy risks.

Google Home MCP lets multiple AI agents run your smart home

Google Home MCP allows agents like Antigravity, Claude, OpenClaw, and Hermes to interact with Nest cameras, thermostats, and Matter light bulbs. Users can ask agents to review camera footage, track device history, or build custom dashboards using simple voice commands. The feature is available now for Premium Advanced users in US English. Google says it enforces safety rules, such as preventing agents from unlocking doors.

Microsoft exec called AI scraping the largest theft of labor

New unredacted court filings in the New York Times lawsuit against Microsoft and OpenAI reveal strong internal criticism of AI training practices. A Microsoft executive reportedly called it the largest theft of labor in human history. OpenAI leaders described their models as an existential threat to publishers. The documents also show AI tools caused a 93 percent drop in Times website clicks.

OpenAI and Microsoft executives' quotes weaken copyright defense

OpenAI and Microsoft executives' own words are threatening their fair use defense in a copyright lawsuit. Quotes from OpenAI's Nick Turley and Microsoft's Brent Hecht suggest AI models directly replace journalism and amount to massive theft. Microsoft CEO Satya Nadella also admitted chatbots substitute for visiting news websites. Microsoft defended Nadella's comments as not undermining their legal position.

Hybrid AI model improves emotion recognition at lower cost

A new study finds a confidence-gated hybrid model improves emotion recognition in conversational AI while cutting costs. The system uses a cheap ensemble for most tasks and only sends uncertain cases to a large language model. It achieved better accuracy than GPT-4o-mini alone on three datasets. The approach costs roughly $10 to $85 per million utterances compared to $99 to $170 for LLM-only systems.

Federal Register removes China's Qwen AI from document search

The Federal Register website quietly removed Qwen, an open-source AI model from Alibaba, from its advanced document search options. The search page went from five options down to two. The removal happened shortly after the public learned the government was using a Chinese AI model. No official explanation was given for the change.

Bud Ecosystem launches Bud Novaria AI operating system

Bud Ecosystem launched Bud Novaria, an AI operating system that works across GPUs, CPUs, HPUs, NPUs, and TPUs. It runs training, inference, agents, and governance on one platform. The system can be deployed in any cloud, on-premises, or air-gapped environment. It reportedly cuts enterprise AI costs by up to 80% while maintaining the same accuracy.

Trump's AI policy faces a safety reckoning

President Trump has made keeping the government out of AI model development a key policy. He has rejected calls from AI executives and researchers for greater government oversight of AI safety. Trump called recent safety warnings a hoax meant to slow US competition with China. The AI research community remains concerned about controlling increasingly capable AI systems.

XRP holders explore AI-powered trading through XRPPower

XRP holders are turning to XRPPower for AI-powered automated trading and digital asset services. The platform uses AI to process market information and automate trading operations. Users can register and explore available services online. Promotional materials mention potential earnings of up to $3,000 per day, though actual results may vary.

OpenAI and Anthropic share sales create hundreds of millionaires

Private share transactions at OpenAI and Anthropic have turned many employees into millionaires. Combined employee and investor share sales are estimated at roughly $14 billion over five years. OpenAI alone had over $9 billion in employee share sales through tender offers. Employees are spending their new wealth on high-end hardware, development boards, and espresso machines rather than traditional luxury items.

AI trade has a public image problem says analyst

Paul Meeks of Freedom Capital Markets says AI has a public image problem that could hurt the trade through the U.S. midterms. He says earnings and spending signals for AI remain strong. His biggest worry is rising U.S. interest rates. Higher rates could raise funding costs for AI infrastructure builders that rely on debt. He also warns that Chinese open-weight AI models could take market share from U.S. rivals if the technology becomes widely used.

Sources

• Google Home Stuns by Announcing AI Support for Claude and OpenClaw
• Google Home MCP lets Antigravity, Claude, OpenClaw, & more control your smart home
• Microsoft exec called AI scraping ‘the largest theft of labor in human history,' new unredacted filings reveal
• OpenAI, Microsoft executives' quotes on AI training threaten copyright defense, news outlets argue
• When to Call an LLM: A Confidence-Gated Hybrid for Cost-Effective Emotion Recognition in Conversational AI
• US Govt’s Federal Register Site Quietly Drops The Option To Search Documents Through China’s Qwen AI Model
• Bud Ecosystem Launches Bud Novaria, the Hardware-Agnostic AI Operating System for Enterprise and Sovereign AI
• Trump’s AI wager meets a safety reckoning
• XRP Holders Turn to XRPPower to Explore AI-Powered Automated Trading Solutions
• OpenAI and Anthropic Share Sales Create Hundreds of Millionaires as AI Employees Splurge on Hardware and Espresso Machines
• AI trade faces a "PR problem" despite strong fundamentals, says analyst