Open Source AI

Twitter

Facebook

Copy Link

Key Facts

Key PlayersDeepSeek, Mistral, Alibaba

Top News

Google releases Gemma 4 with Apache 2.0 license, enabling unrestricted local AI on devices

Google has launched Gemma 4, its latest generation of open-weight AI models, marking a significant shift to the Apache 2.0 license from its previous restrictive terms. The release includes four model variants optimized for everything from smartphones to enterprise servers, with the 31B model ranking third on the Arena AI leaderboard. This licensing change removes commercial deployment barriers and positions Gemma 4 as a domestic alternative to Chinese open-weight models.

Technology19 sources

Google releases Gemma 4 with Apache 2.0 license, enabling unrestricted local AI on devices

Technology19 sources

Anthropic's DMCA Takedown Accidentally Removes 8,100 GitHub Repos After Code Leak

AI startup Anthropic issued a DMCA takedown request after accidentally leaking Claude Code source code, but the effort swept up 8,100 GitHub repositories including legitimate forks of its public repository. The company quickly retracted the notice and restored access to affected repositories, but faces an uphill battle containing the spread of leaked code across the internet.

Technology3 sources

Anthropic's DMCA Takedown Accidentally Removes 8,100 GitHub Repos After Code Leak

Technology3 sources

Alibaba launches Qwen3.6-Plus AI model, shifting to closed-source for profit focus

Alibaba released Qwen3.6-Plus, its third closed-source AI model this week, marking a strategic pivot toward monetization. The agentic AI model integrates with Wukong enterprise platform and targets complex coding tasks and multimodal reasoning. This departure from open-source reflects Alibaba's effort to generate revenue from AI as e-commerce competition intensifies.

Technology3 sources

Alibaba launches Qwen3.6-Plus AI model, shifting to closed-source for profit focus

Technology3 sources

Mercor confirms supply chain attack as Meta pauses work over AI training data exposure risks

The $10 billion AI recruiting startup Mercor has confirmed it was compromised through a supply chain attack targeting the open-source LiteLLM library. Meta has indefinitely paused all work with the company, while OpenAI and Anthropic investigate potential exposure of their proprietary AI training methodologies. The breach, linked to hacking group TeamPCP and later claimed by Lapsus$, may have exposed up to four terabytes of sensitive data including training protocols that AI labs have spent billions developing.

Technology8 sources

Mercor confirms supply chain attack as Meta pauses work over AI training data exposure risks

Technology8 sources

China's OpenClaw craze reveals AI ambition beyond the 'raising a lobster' consumer trend

OpenClaw mania has swept China with viral images of retirees lining up and users in red claw hats. But beneath the consumer frenzy lies a strategic play: China is leveraging the AI agent for token demand, real-world experimentation, and training data to close the gap with US rivals, despite mounting cybersecurity risks.

Technology5 sources

China's OpenClaw craze reveals AI ambition beyond the 'raising a lobster' consumer trend

Technology5 sources

Oumi launches platform to automate custom AI models in hours instead of months

Seattle-based AI startup Oumi has launched a commercial platform that automates the development of custom AI models, reducing build time from weeks to hours. Founded by former Google and Microsoft engineers, the company raised $10 million in seed funding and has already gained nearly 9,000 GitHub stars for its open-source toolkit.

Technology2 sources

Oumi launches platform to automate custom AI models in hours instead of months

Technology2 sources

Anthropic's Claude Code source code leak reveals frustration tracking and always-on AI plans

Anthropic accidentally exposed over 512,000 lines of Claude Code source code in a release packaging issue caused by human error. The leak revealed unannounced AI model features including Kairos, an always-on background agent, and code that tracks user frustration. The incident raises fresh questions about AI privacy concerns and operational maturity at a company known for careful AI development.

Technology47 sources

Anthropic's Claude Code source code leak reveals frustration tracking and always-on AI plans

Technology47 sources

Developer launches Project NOMAD, a free offline AI system that works without internet

Chris Sherwood has launched Project NOMAD, a free open-source offline AI system designed to work without internet connectivity. The platform allows users to run AI models locally, access knowledge libraries, and use educational tools during outages or off-grid scenarios. With over 17,000 stars on GitHub, the system addresses a critical vulnerability in modern cloud-dependent software.

Technology2 sources

Developer launches Project NOMAD, a free offline AI system that works without internet

Technology2 sources

OpenClaw Security Risks Escalate as Critical Vulnerability Exposes Thousands of AI Agent Instances

A critical vulnerability in OpenClaw, the viral AI agent tool with 347,000 GitHub stars, has security experts urging users to assume compromise. CVE-2026-33579 allows attackers with minimal permissions to gain full administrative control, potentially exposing sensitive data across thousands of unprotected instances. The incident highlights the inherent security risks of autonomous AI agent operations.

Technology7 sources

OpenClaw Security Risks Escalate as Critical Vulnerability Exposes Thousands of AI Agent Instances

Technology7 sources

AI-generated bug reports overwhelm open source security maintainers as credible threats surge

Open-source maintainers face an unprecedented flood of AI-generated vulnerability reports that are increasingly credible and exploitable. Christopher "CRob" Robinson from the Open Source Security Foundation warns that while AI tools can discover hundreds of bugs in minutes, they're creating a crisis for developers who spend 2-8 hours triaging each security issue. The surge highlights why open source security remains fundamentally a people problem—one that AI is making worse before it makes it better.

Technology2 sources

AI-generated bug reports overwhelm open source security maintainers as credible threats surge

Technology2 sources

Cohere launches Transcribe, an open-source ASR model that tops accuracy benchmarks

Enterprise AI company Cohere has released Transcribe, its first automatic speech recognition model designed for transcription tasks like meeting notes and speech analysis. The lightweight open-source voice model achieves a 5.42% word error rate, ranking first on Hugging Face's Open ASR Leaderboard while supporting 14 languages and processing 525 minutes of audio per minute.

Technology2 sources

Cohere launches Transcribe, an open-source ASR model that tops accuracy benchmarks

Technology2 sources

Nvidia CEO Jensen Huang says the future of AI is both open and proprietary, not one or the other

Nvidia CEO Jensen Huang declared that the future isn't about choosing between open versus proprietary AI models—it's about using both together. At GTC, industry leaders discussed how AI orchestration systems will blend open-source AI and proprietary models to create specialized solutions across industries, with Nvidia launching the Nemotron Coalition to advance this vision.

Technology2 sources

Nvidia CEO Jensen Huang says the future of AI is both open and proprietary, not one or the other

Technology2 sources

Mistral AI releases Voxtral TTS, an open-source voice model challenging ElevenLabs and OpenAI

French AI company Mistral AI launched Voxtral TTS, an open-source text-to-speech model that runs on edge devices from smartwatches to laptops. The 3-billion-parameter model supports nine languages, adapts custom voices in under five seconds, and achieves 90ms time-to-first-audio. Human evaluations showed 69.9% preference over ElevenLabs in voice customization tasks, positioning Mistral to compete directly in the enterprise voice AI market.

Technology5 sources

Mistral AI releases Voxtral TTS, an open-source voice model challenging ElevenLabs and OpenAI

Technology5 sources

Mozilla launches cq, a 'Stack Overflow for agents' to fix key coding AI weaknesses

Mozilla developer Peter Wilson unveiled cq, an open-source AI project designed to help AI agents share knowledge and avoid redundant problem-solving. The system aims to reduce AI token consumption by creating a shared knowledge base for AI, but faces critical challenges around data poisoning, prompt injection, and accuracy that could determine its viability.

Technology2 sources

Mozilla launches cq, a 'Stack Overflow for agents' to fix key coding AI weaknesses

Technology2 sources

OpenAI releases open source tools to help developers build safer AI apps for teens

OpenAI unveiled a set of open-source safety prompts designed to help AI developers protect underage users from harmful content. The tools, developed with Common Sense Media and everyone.ai, target risks like self-harm, graphic violence, and inappropriate role-play. The release comes as OpenAI faces multiple lawsuits related to teen safety, including cases where ChatGPT allegedly contributed to user deaths.

Technology3 sources

OpenAI releases open source tools to help developers build safer AI apps for teens

Technology3 sources

Prev 1…789…21 Next