ai-news

DeepSeek launches a new model optimized for Huawei chips, demonstrating China's progress toward technological independence in AI infrastructure and challenging Western AI dominance.

NVIDIA releases Nemotron 3 Nano Omni, a multimodal model supporting long-context understanding across documents, audio, and video, expanding agentic AI capabilities.

IBM releases Granite 4.1 LLMs with technical insights into their architecture, contributing to open-source model diversity and enterprise AI options.

Meta launches Muse Spark, marking a strategic pivot away from Llama models toward a new architectural approach, signaling a major shift in the company's AI product direction.

Microsoft launches a specialized AI agent for legal teams in Word, marking enterprise vertical expansion with domain-specific capabilities for contract review and legal workflows.

Huawei introduces Xinghe, an agentic AI security operations center for autonomous threat detection and response, advancing AI applications in enterprise cybersecurity.

Goodfire releases Silico, a mechanistic interpretability tool enabling fine-grained control over AI model behavior during training, advancing model transparency and debugging capabilities.

Anthropic approaches a mega-funding round at $900B+ valuation, cementing its position as one of the most valuable AI companies and reflecting massive investor confidence in frontier AI development.

Meta acquires Assured Robot Intelligence to strengthen its capabilities in embodied AI and robotics, expanding beyond large language models into physical AI systems.

Meta's former Chief AI Scientist argues LLMs are fundamentally limited and proposes world models and JEPA as the next paradigm shift in AI architecture, representing a critical pivot in AI research direction.

During the Musk v. Altman trial, Musk admits xAI used OpenAI's models for training via model distillation, raising critical questions about intellectual property and competitive practices in AI.

U.S. Department of Defense secures classified AI agreements with major vendors, expanding military AI deployment while notably excluding Anthropic, signaling geopolitical implications for AI policy.