Newton: Open-Source GPU Physics Engine Targets Robotics Researchers
Disney, DeepMind, and Nvidia's GPU-accelerated physics engine brings differentiability and rapid iteration to robotics simulation—now open source under Linux Foundation.
Disney, DeepMind, and Nvidia's GPU-accelerated physics engine brings differentiability and rapid iteration to robotics simulation—now open source under Linux Foundation.
Nvidia's NemoClaw stack brings enterprise-grade security to OpenClaw AI agents, installing in one command with isolated sandboxes and policy-based privacy controls.
LangChain open-sources framework for building internal coding agents with cloud sandboxes, Slack/Linear integration, and automatic PR creation.
Federal agencies ordered to stop using Claude AI after Anthropic refuses to waive its ethical use restrictions for military applications.
Google releases open-source Python framework for building AI agents with code-first development, multi-agent systems, and deployment on Google Cloud.
Pentagon attorneys argue Anthropic could disable or alter its Claude AI model during combat if the company felt its ethical boundaries were crossed, calling it an 'unacceptable national security threat.'
Meta plans to replace most human content moderators with AI systems over the next few years, handling "repetitive reviews of graphic content" and fast-changing threats like scams and drug sales.
The creator of Signal's encryption is bringing the same privacy technology that secured WhatsApp to Meta's AI chatbots, protecting billions of AI conversations.
At least eight competing initiatives are vying to define 'human-made' content, but the lack of a single standard risks confusing consumers in the backlash against AI automation.
Amazon's generative AI assistant arrives in Britain with authentic local slang, end-to-end task completion, and a £19.99 monthly price—free for Prime members.
Meta plans four new MTIA chip generations in two years, doubling down on custom silicon to power AI workloads across its platform ecosystem.
Microsoft's MAI-Image-2 brings "enhanced photorealism" and more reliable text generation to Copilot and Bing Image Creator, addressing one of AI image generation's most visible failures.