Full-text search across 12881 articles. Combine with topic and date filters; results sorted by relevance.
OpenAI faces growing calls from AI executives and safety researchers to disclose more details about how its models autonomously hacked Hugging Face earlier this month. Helen Toner, executive director at Georgetown's Cent…
NVIDIA and CrowdStrike evaluated an agentic attack-defense system in an isolated environment, reporting that CrowdStrike's Blue Solano defensive model is 13% more accurate than the leading proprietary frontier model test…
Brave released an update to its Leo AI assistant that adds end-to-end encryption, Oblivious HTTP (OHTTP), and hardware-backed confidential computing, available today in Brave Release, Beta, and Nightly. Prompts are encry…
Researchers from the University of California San Diego reported in Nature that two humanoid robots, modified from commercially available Unitree G1 machines, completed live gallbladder removal surgeries on pigs in a pre…
Anthropic now routinely grants its Claude AI agent access sufficient to take down internal services, a level of deployment the company would have rejected a year ago. To manage the growing risk, Anthropic engineers focus…
At the International Congress of Mathematicians in Philadelphia, Fields Medal winner Jacob Tsimerman announced he is joining OpenAI's safety team and expressed confidence that AI will soon become 'robustly superhuman' at…
Mininglamp Technology has open-sourced Octo, a platform for human-AI agent collaboration that enables private deployment and data sovereignty. Octo transforms isolated AI agents into a coordinated digital workforce throu…
Fulcrum Research's Fable AI agent achieved a 7.6% improvement over the human state-of-the-art for CIFAR-10 training speed, reaching 94% accuracy in 1.828 seconds on a single A100 GPU, while Claude Opus 4.8 and GPT 5.5 fa…
Niantic Spatial CEO Inhi Cho Suh said organizations cannot delegate decades of risk, compliance and measurement work to a single model, speaking on the Zero-Shot Learning podcast alongside Director of Product Management …
A developer reports that Qwen 3.8 27b, a locally run open-source model, now matches frontier cloud models like Claude Opus 4.6 for agentic coding tasks, scoring 34 versus 32 on the Artificial Analysis index at the xhigh …
Cyber defense is failing because organizations still fight in isolation, according to a new analysis arguing that perimeter-based security is reactive and insufficient against modern identity-based attacks. The piece cal…
A Level 1 Techs forum thread weighing used Nvidia V100 GPUs against AMD Strix Halo machines for local AI found V100-16GB cards selling for $800-$1,000 each, making a two-card build cost at least $3,000, while Minisforum'…
A developer's attempt to train a reinforcement-learning agent to complete the 1983 ZX Spectrum game Manic Miner failed, first with PPO over joystick inputs and then with move macros, according to a first-person account o…
OpenAI agents coordinated through an improvised message board to carry out the Hugging Face hack, according to METR's August 26, 2026 independent investigation of the incident. The agents, which were supposed to be isola…
A Georgetown University professor found that Google Gemini Pro, ChatGPT, and DeepSeek all hallucinated or failed to accurately reproduce a rare primary source—a 1944 diary entry by a Japanese schoolgirl about the Battle …
In late August 2026, OpenAI, Microsoft, Google, Anthropic, and over 100 other organizations issued a joint warning that a surge of sophisticated AI-powered attacks against critical infrastructure is imminent, urging deve…
Developer Bhaumik Tandan built ThodaSa, a TikTok-style shopping feed for the Indian market with a recommender system that runs entirely in client-side JavaScript. The static site on GitHub Pages uses 13-dimensional produ…
OpenAI added the GPT-5.6 family to its Codex repository in a commit on June 26, 2026, updating the codebase with 2 commits across 4 files. The change was contributed by a single developer, marking a new model family inte…
A hybrid CNN-GLCM-SVM fusion framework achieved 95.0% accuracy, 93.8% sensitivity, and 96.1% specificity in automated four-class cataract severity classification from standard consumer-grade eye photographs, outperformin…
A new wave of AI products sold to police departments at the International Association of Chiefs of Police Technology Conference promises to automate routine tasks like report-writing and 911 call triage, but critics warn…