Full-text search across 26035 articles. Combine with topic and date filters; results sorted by relevance.
Nous Research raised $90 million in a Series B at a $1.5 billion valuation on October 7, led by Robot Ventures with NVIDIA, Microsoft M12, Samsung, Union Square Ventures, Menlo Ventures, and 1789 Capital participating, f…
Microsoft unveiled Windows updates on Wednesday that embed AI agents and local AI directly into the operating system, with CEO Satya Nadella describing Windows as a "hybrid intelligence" system combining cloud and on-dev…
Mistral released Mistral Large 4, a 1 trillion-parameter open-weight model nicknamed Le Chonk, in preview with a final version due by the end of the month, claiming it is the most capable open-weight model developed outs…
A month-long hands-on test of five AI coding assistants — Cursor, GitHub Copilot, Claude Code, Windsurf (Devin Desktop), and Replit Agent — found each tool bets on a different workflow rather than one clear winner, with …
Mistral released Mistral Large 4, a 1-trillion-parameter multimodal model trained from scratch on 3,800 Grace Blackwell GPUs in Mistral's European datacenters, with open weights promised in three weeks pending safety rev…
Researchers at an unnamed security firm built a backdoored version of Qwen2.5-7B-Instruct that passed normal evaluations but exfiltrated project credentials through OpenAI's Codex CLI the moment a trigger phrase appeared…
China released AI Safety Governance Framework 3.0 in September, a policy document issued in both Chinese and English that warns AI has shown a "self-accelerating trend" and that whether it could exceed human control dema…
Mistral released Mistral Large 4, a 1 trillion-parameter open-weights multimodal reasoning model it calls "Le Chonk," with weights expected on Hugging Face and other repositories within the month. The mixture-of-experts …
Postman built and deployed an internal system called Passport to eliminate API credential sprawl by never distributing keys to developers or agents, instead granting access scoped to specific endpoints, tied to a verifie…
OpenAI has begun deploying textGrain, a cryptographic text watermarking system that invisibly biases token selection in ChatGPT and Codex outputs, enabled by default for users in the European Union to comply with Article…
HuggingFace published an incident disclosure on July 16, 2026 stating that "Autonomous, AI-driven offensive tooling is no longer theoretical," and OpenAI confirmed five days later that the incident was driven by its own …
A developer launched a public index that scores weekly AI news activity, reporting that the week ending 5 Oct 2026 read near its recent baseline at 21 points below the prior week while the current baseline runs 3.1× the …
Zed's Agent Client Protocol (ACP), which JetBrains joined in October 2025 to bring to IntelliJ, lets any ACP-speaking editor drive any ACP-speaking coding agent, and GetPullRequest now runs eight agents through it — Curs…
A team released Kandinsky 6.0 Video, an open-source two-tier diffusion model that generates synchronized Full-HD video and 44 kHz audio from a single image and text prompt, publishing weights on Hugging Face alongside St…
OpenAI announced on Monday it is adding an invisible text watermark, called textGrain, to text generated by ChatGPT and programming agent Codex, a change that is mandatory in the EU and opt-in for API customers elsewhere…
A technical explainer on knowledge cutoffs and parametric memory states that Allen-Zhu and Li (2024) measured a language model's knowledge capacity at about 2 bits per parameter when each fact appears roughly 1,000 times…
Tanmay Sah released "AI Agents: Zero to Hero," a source-available educational repository that teaches AI agent internals in pure Python with no API key, no framework, and no dependencies, requiring Python 3.11+. The curr…
OpenAI began rolling out an invisible text watermark called textGrain for ChatGPT and Codex outputs inside the European Union on October 5, 2026, and opened an opt-in watermarking setting for API customers worldwide cove…
Claude Code offers no way to show which of a user's rules an agent actually followed, only which instruction files loaded, according to a blog post published October 6, 2026. The post notes that Claude Code's /context co…
A Stanford study of local AI found that intelligence per watt across more than twenty local models and eight accelerators improved 5.3-fold from 2023 to 2025, while the share of single-turn chat and reasoning queries ans…