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.
OpenAI Publishes 'Abundant Intelligence' Strategy: AI to Become More Accessible
OpenAI outlines 'abundant intelligence': through a full-stack approach, it aims to make AI more powerful, more affordable and more widely useful, all at once and…
Google's SynthID Holds Up Against Tampering, but It Doesn't Solve AI Disinformation
Ars Technica's tests confirm Google's SynthID watermarks are technically robust, but labeling AI content won't stop disinformation without a single industry-wide…
Moonshot Is Using 20,000 Nvidia Chips From Alibaba to Train Its Kimi Model
Chinese AI startup Moonshot has leased a cluster of about 20,000 Nvidia chips from Alibaba to develop Kimi, despite US export restrictions on high-performance…
Meta open-sources SAM 3 and DINOv3 for Genesis Mission: scientific image analysis in 15 minutes
Cursor: AI Agent Swarm Builds 80% of SQLite in 4 Hours — at a Fraction of the Old Swarm's Cost
Tencent Open-Sources AngelSpec, a Framework for Speeding Up LLM Inference With Speculative Decoding
Vercel's DeepsecBench: GPT-5.6 Leads at Finding Vulnerabilities in Application Code
DeepSeek Reaches 22.6% of AI Gateway Tokens, Open-Weight Models Hit 29% — Vercel Report
Vercel AI Gateway data shows open-weight models handled 29% of all tokens in June 2026 on under 4% of spend. DeepSeek rose to third place, and Anthropic took 61% of…
Moonshot AI Open-Sources MoonEP, the Library Used to Train the MoE Model Kimi K3
Moonshot AI released MoonEP, an expert-parallelism library, under the MIT license. It gives every GPU rank exactly S×K tokens and sped up Kimi K3 scaling by 2.5x.
Fireworks Nexus: Routing Code Tasks to Open-Weight Models and Cutting AI Costs 3–5x
Xiaohongshu's UltraEP Balances MoE Experts in 300 µs, Reaching 94.3% of Ideal
VAST at SIGGRAPH 2026: Chinese 3D Startup Gives Keynote Alongside NVIDIA and Disney
Nvidia and Microsoft Launch Open Secure AI Alliance — Without OpenAI, Google or Anthropic
OpenAI, Anthropic and Google Employees Urge US Government to Step In on the AI Race
Employees of OpenAI, Anthropic, Google, Meta and Microsoft signed an appeal to the US government: the race to automate AI research should be slowed or brought under…
Zuckerberg: Meta's Enterprise AI Opportunity Extends Beyond Agents
On the Q2 earnings call, Mark Zuckerberg said Meta's enterprise AI business spans agents, APIs, compute and internal software, not just agents. *Meta is designated…