US quietly tightens grip on Chinese AI models
Washington is reportedly using sanctions and soft pressure to slow the adoption of Chinese AI models, favoring domestic firms without an outright ban.
Artificial intelligence
US public health agencies to trial AI models via PULSE program
Public health agencies across the U.S. will test generative AI tools through the PULSE initiative, aiming to shape implementation guidance and assess real-world impact.
Local AI at 1B scale: MiniCPM5 fine-tuned on Claude traces ships in 657MB
A community dev has released a 1-billion-parameter model that runs entirely on local hardware, fine-tuned on Claude Fable 5 reasoning traces and packaged for llama.cpp, Ollama, LM Studio and others.
Alibaba’s Qwen 3.8 Challenges Top AI Models With 2.4T Parameters
Alibaba launches Qwen 3.8, a multimodal open-weight model with 2.4 trillion parameters, positioning itself just behind Fable 5 in performance.
Kimi K3 leads frontend coding but stumbles on advanced math
China’s Kimi K3 tops frontend code rankings yet trails OpenAI and Anthropic by 50+ points on FrontierMath Tier 4.
Open-weight AI models close cyber gap with frontier systems
Open-weight AI models like GLM-5.2 and DeepSeek V4-Pro now rival four-month-old frontier models in cyber tasks, cutting costs sharply.
China launches global AI alliance to challenge Western dominance
Beijing unveils the World AI Cooperation Organization and 5,000 training slots for Global South countries in a bid to shape AI governance outside Western influence.
Pentagon shifts gears: AI adoption beats perfect alignment
The U.S. Navy is prioritizing rapid AI deployment over flawless alignment, aiming to build an "AI-first" fleet with models running on warships and an AI war council for mission decisions.
How AI agents are redefining event operations with memory and real-time context
A new demo shows how MongoDB Atlas, Voyage AI embeddings and LangGraph can power an AI event operator that remembers past events, adapts to live changes and writes outcomes for future use.
Torvalds pushes back on Linux AI critics
Linus Torvalds urges Linux developers to embrace AI tools like Sashiko, dismissing detractors in the kernel community.