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.
Nvidia and Microsoft Launch Open Secure AI Alliance — Without OpenAI, Google or Anthropic
Nvidia, Microsoft, SpaceX and IBM formed the Open Secure AI Alliance to build open AI security tools — without the three biggest labs: OpenAI, Google and Anthropic.
Moonshot's Kimi K3: Why China Is Giving Away Its Best AI Models for Free
Moonshot AI has released Kimi K3, a Chinese model that reportedly beats top US systems at a fraction of the cost, and it is opening the weights for free with US…
Meta AI Comes to Threads DMs: The Chatbot Is Now Available in Private Messages
Meta has built its Meta AI assistant into Threads direct messages, so users can now chat with the bot right in their DMs, just as they would with another person.
DeepSeek Founder Liang Wenfeng: What the Leak of His Closed-Door Meeting Revealed
OpenAI Opens ChatGPT Health to All US Users, With Apple Health and MyFitnessPal Integration
OpenAI and Anthropic Guardrails Are Getting in the Way of Cybersecurity Researchers
Nvidia and Mistral Urge US Not to Impose Broad Restrictions on Open-Weight Models
Silicon Valley Splits Over Chinese AI: Big Startups vs. Small Ones
Multibillion-dollar AI startups in Silicon Valley are sounding the alarm over Chinese AI, while smaller companies see open Chinese models as an opportunity, not a…
Moonshot AI vs. Anthropic: Wired Podcast Examines Accusations of Copying Claude
Wired's Uncanny Valley podcast covers China's Moonshot AI accused of copying Anthropic's tech, OpenAI losing control of two models, and the risks of AI in the US…
Moonshot AI's Kimi Has Spooked Silicon Valley: Why Investors Are Panicking Over Chinese AI
PEARL: A 4B Model Beats DeepSeek-V3.2 685B at Optimization Modeling
Thinky Releases Open Inkling 975B-A41B Model Under Apache 2.0 License
Evidence Chain Evaluation: An AI Fact-Checker That Is Allowed to Say "Not Sure" Instead of Issuing False Verdicts
D2VBench: A 10,000-Dilemma Benchmark Tests LLM Value Alignment
Researchers have introduced D2VBench, a benchmark of 10,000 everyday dilemmas and 158 value concepts that tests how language models resolve conflicts between values.
TriAgent: A Committee of Small Models Instead of an Expensive LLM Saves $9.3M a Year
Researchers unveiled TriAgent, a multi-agent financial sentiment system where cheap models replace a cloud LLM, saving $9.3M a year at 10M users.