Open-Source AI
Open models — Llama, Mistral, DeepSeek, Qwen and others — are catching up to closed labs and let you run AI on your own hardware, vendor-free. That shifts the balance of power, the safety debate and the economics of the field. Here is everything on open-source AI: open-weight releases, licenses, local deployment and the race against closed models.
The US closed access to Claude Mythos, China restricts GLM and Qwen: neural networks have become a strategic weapon
Nomagic deployed a VLA model in warehouse robots and halved operator callouts
Five AI rounds in 24 hours: $300 million for quantum computing and Prime Intellect hits unicorn status
Hugging Face: transformers backend in vLLM now runs at native speed
Hugging Face announced performance parity: the `--model-impl transformers` flag delivers native vLLM speed across 450+ architectures without rewriting code.
Veo 3.1, Kling, Sora 2 and Grok: a test of five AI video generators with identical prompts
Habr authors ran Veo 3.1, Kling v3 Omni, Sora 2 Pro, LTX-2.3 Pro and Grok through seven identical scenarios — from a knight to a panicky cat.
Mistral AI introduced a navigation model for robots and is expanding into physical AI
The French startup Mistral AI introduced a navigation model for robotics, expanding its presence in physical AI after signing deals with major European industrial…
China called Anthropic's Claude Code a backdoor and banned it, while the US restricts Chinese models
DeepSeek begins designing own AI chip to bypass U.S. export restrictions
American companies switch to Chinese AI — OpenAI and Anthropic have become too expensive
DeepSeek develops its own AI chip for its neural networks — Reuters
Beijing discusses restricting foreign access to Alibaba and ByteDance AI models
China's Ministry of Commerce held a series of closed meetings with Alibaba, ByteDance, and startup Z.ai, discussing possible restrictions for foreign users of…
Microsoft launched Foundry Managed Compute: thousands of Hugging Face models with one click
Microsoft combined Azure Foundry with the Hugging Face catalog: thousands of open models now deploy in the cloud with one click, with enterprise security and…
NVIDIA Dynamo vs Triton Inference Server: what changed for agent inference
AI for PDF Document Processing: 2026 Tools Guide
Fable creates first GPU megakernel — experts discuss AI self-improvement loop
Hallucinations in AI coding tools: why agents are slower than search-and-replace
Meta Decided to Sell AI Computing Power to Third Parties — Asian Chipmakers' Stocks Crashed
Meta's announcement of plans to monetize AI infrastructure as a commercial service triggered a sell-off of semiconductor stocks in Asia amid fears of market…
Aikido Security acquires Israeli Root: AI fixes open source without breaking dependencies
Belgian cybersecurity unicorn Aikido Security acquired Israeli startup Root — its AI agents eliminate vulnerabilities in open source without disrupting dependent…