Today's major stories, as reported by AWS Machine Learning Blog, TechCrunch AI, The Verge, OpenAI News, and research on ArXiv.org, cover new AI system launches, government actions on AI risk, internal industry warnings about data practices, and research findings on model robustness and alignment.
AWS Adds Moonshot AI's Kimi K3 Model to Platform
Amazon Web Services announced Kimi K3, developed by Moonshot AI, is now available on Amazon Bedrock. Kimi K3 is described as 'the first open model to reach 2.8 trillion parameters' and supports a '1-million-token context window.' It features prompt caching 'to reduce latency and input costs,' with 'zero data retention...while zero operator access prevents even AWS operators from accessing your prompts and completions during inference.' (AWS Machine Learning Blog)
Takeaway: Enables direct use of a high-capacity, open-weight AI model with enhanced efficiency and security controls on AWS.
Anthropic CEO Suggests Plan to Slow Advanced AI Progress
TechCrunch AI reports that Dario Amodei, CEO of Anthropic, has 'outlined his plan to “pace the frontier” of AI development.' The plan involves 'independent safety evaluators and coordination between AI labs in democratic countries.' Some leaders support it, while Nvidia’s Jensen Huang has opposed it.
Takeaway: The proposal spotlights ongoing industry debate over the pace and safety of advanced AI systems.
Google's CC AI Agent Now Targeted at Household Tasks
TechCrunch AI reports Google is 'refocusing its CC AI agent on household coordination, letting families share emails, schedules, and tasks so the AI can manage calendars, fill out forms, make shopping lists, plan meals, and more.'
Takeaway: Google positions its AI agent as a tool for managing daily household routines.
Virginia Issues New Controls on Data Centers, Launches AI Task Force
The Verge reports Virginia Governor Abigail Spanberger issued an order that 'bans executive branch officials from signing nondisclosure agreements for data center projects, requires expedited noise regulations, and requires a review of backup-generation operations.' The order also 'establishes an AI task force responsible for evaluating how the state government can address risks to Virginians.'
Takeaway: The state moves to strengthen data center oversight and study AI risks at the government level.
OpenAI and Microsoft Warned of AI Data 'Doom Loop,' Court Documents Show
The Verge says unsealed court documents from the New York Times lawsuit reveal OpenAI and Microsoft warned internally their data scraping was starting a 'doom loop' that would 'damage the web,' calling it the 'largest theft of labor in human history' and undermining 'the idea of fair use.'
Takeaway: Internal industry concerns recognized the scale and risks of their own model training practices.