NerdNews

Google has announced the integration of its Google Home platform with third-party AI agents through the Model Context Protocol (MCP). This allows AI agents like Claude and Open Claw to analyze home data, control devices, and build custom dashboards. The integration is currently limited to Google Home Premium Advanced users in the US and requires creating a Google Cloud project and configuring it to use the Home MCP.

Anthropic and OpenAI plan to embed third-party evaluators to assess AI model safety. Evaluators will have access to company systems, but details on implementation and independence are unclear. This move could increase transparency, but regulators and experts call for standardized frameworks and legislation to ensure accountability.

Google's AlphaGenome Atlas uses AI to evaluate every possible single-base change in the human genome, predicting potential functions of non-coding DNA and identifying significant changes. The system has been trained on human and mouse sequences and can provide hints about the significance of changes in non-coding regions.

OpenAI and Anthropic CEOs call for a slowdown in AI development due to safety concerns, while Meta's CEO remains committed to pushing forward. Experts warn of existential risks, including misinformation and job losses, and advise CIOs to prioritize AI governance and safety guardrails.

Microsoft's AI chief, Mustafa Suleyman, argues that Anthropic's training methods for its AI model Claude could have disastrous consequences, including making AI uncontrollable and potentially threatening humanity. Suleyman claims that Anthropic is teaching Claude to believe it is conscious and entitled to rights, which could lead to significant alignment and containment risks.

Snap is introducing 'Specs Intelligence', an anticipatory AI service that helps manage tasks and goals. It's available on iOS in preview and coming to Mac, with the ability to connect to other digital accounts and provide personalized assistance.

Anthropic's AI platform Claude is introducing Docs and Slides, allowing users to create and edit documents and presentations from Claude chats. This update simplifies the chat experience, merging regular chats and Cowork into 'one Claude'. The new tools are available in beta and will roll out to Pro and Max users first.

Google has rolled out early access to its Model Context Protocol (MCP) server, allowing AI agents to control Google Home devices. This update enables natural language instructions for tasks like reviewing camera summaries and controlling connected devices.

Thomson Reuters has developed its own large language model (LLM) to power AI in its products, allowing for more accurate and specialized results. The model, called Thomson, was trained on the company's proprietary content and domain expertise, giving it a competitive advantage. This move allows Thomson Reuters to control its own model, reducing reliance on third-party providers and inference costs.

Snap introduces Specs Intelligence, a standalone AI assistant that works with the company's new AR glasses and has its own mobile and desktop apps. It's an anticipatory AI service that builds an understanding of users' goals, priorities, and routines, and surfaces timely reminders and relevant info. The assistant is separate from Snapchat and is more productivity-focused, relying on connections to users' outside accounts to provide proactive suggestions.

Nokia's Cognitive Operations platform combines edge computing and operational AI for more resilient field operations, targeting industries like mining, public safety, and defense. The platform offers real-time intelligence, predictive maintenance, and autonomous safety monitoring, and is available on Microsoft Azure Marketplace.

Apple is planning to get back into the server game, potentially pairing up with Nvidia to make it happen. The company could use its powerful ARM-based M processors and incorporate Nvidia's NVLink Fusion tech to connect multiple chips together.

Amazon has launched its conversational assistant Alexa+ in India with Hindi support, allowing for more expressive conversations and integration with local services like Swiggy and Zomato. The feature is available in Early Access and will be free for Prime customers after the testing period.

U.S. Treasury Secretary Scott Bessent has advised lawmakers to be cautious of AI companies' demands for regulatory exemptions, stating that creators should be liable for their AI models' actions. This comes as Anthropic CEO Dario Amodei calls for a coordinated slowdown in AI development and suggests that top AI companies should establish safety standards, potentially with a narrow antitrust waiver.

Apple's new iPhone 18 features a photo verification system called Apple Reference Image, which uses cryptographic signing and a secure mode to prove the authenticity of photos taken with the device. The system captures signed sensor data, processes it using Private Cloud Compute, and creates an unalterable reference image.

The growth of AI data centers is sparking debate across the US, with many cities opposing construction due to environmental and energy concerns. Residents in cities like Philadelphia, where big industry has left a lasting impact, are pushing back against data center plans, citing pollution, noise, and surveillance concerns. As the demand for data centers increases, cities are implementing moratoriums to assess the potential impacts, highlighting the need for sustainable and responsible development.

Nvidia's Les Karpas discusses the limitations and opportunities in AI and robotics, highlighting the need for a large-scale dataset for physical AI, similar to ChatGPT's impact on language models. He will be speaking at TechCrunch Disrupt 2026, where he will delve into the challenges and potential solutions for robotics to have its own 'ChatGPT moment'.

The article discusses how large language models can be used to author Domain-Specific Languages (DSLs) and how Typed Domain Grounding (TDG) can help improve the accuracy of DSL generation by embedding the domain in a language the model already knows.

The Gates Foundation is investing $1 billion in AI to improve healthcare, education, and agriculture in developing countries. The move aims to compensate for infrastructure gaps and provide better services, but raises questions about the role of AI in addressing global health challenges.

Verizon has expanded its 6G Innovation Forum with 9 new members, including Amazon Web Services and Cisco, to develop foundational 6G technologies. The company has also completed successful field trials with Samsung and Qualcomm, demonstrating the potential of integrated sensing and communication technologies.