Full-text search across 17894 articles. Combine with topic and date filters; results sorted by relevance.
Rails' Agents on Rails benchmark added Grok 4.6, GLM 5.3, Gemini 3.7 Flash, and Claude Opus 4.8, with Claude Opus 5 still leading at 58 of 63 tasks. The benchmark now includes 792 runs with full traces published in rails…
Ray Direct Transport (RDT), a feature in Ray Core, enables RDMA-backed weight transfer in reinforcement learning for LLMs, achieving up to 7.5x speedup over naive RDMA implementations on GB200 nodes using Multi-Node NVLi…
A Gallup Korea poll of 2,674 students found that 67 percent had used AI in the previous week, with usage highest among high school students at 82.3 percent, as Generation Alpha, born between 2010 and 2024, comes of age i…
Three.js r186 adds native 3D Gaussian Splatting support with a built-in GaussianSplat mesh type and loaders for .spz, .ksplat, .splat, and glTF formats, enabling photorealistic rendering of captured objects and scenes di…
AI Avatar, a free app that lets VRoid avatars cheer users in VS Code and browser side panels, has been updated to version 17 with Catbot integration, Cat Mode, and AI Office. The developer added whiskers to the avatar us…
The Model Context Protocol (MCP) 2026-07-28 specification removes the stateful session model, eliminating the initialize handshake and Mcp-Session-Id header, requiring all requests to be self-contained. This breaks legac…
Stripe Inc. has agreed to terms to acquire OpenRouter Inc., a developer of a single access point to more than 400 AI models, for over $7 billion, according to Bloomberg. OpenRouter, founded in 2023 by Alex Atallah and Lo…
San Mateo County's Board of Supervisors approved a resolution on August 11th directing county staff to draft a permitting ordinance for commercial humanoid robots, potentially requiring a trained human supervisor on-site…
The American Civil Liberties Union (ACLU) criticized Flock Safety's recent changes to its AI-powered traffic cameras, calling them 'retreads' that fail to address privacy harms. The ACLU urges default retention periods b…
GitHub's repository for Binaryify/open-kimi-ppt-skill has been archived and cleared of all content except a README citing copyright reasons, leaving its core claims unverifiable. The project page still lists features lik…
OpenAI has released Multi-Agent v2, enabling its GPT-5.6 Sol and Terra models to delegate tasks to the speed-optimized Luna variant, resolving a version mismatch that previously blocked interoperability. The update allow…
Model inference, the process of running a trained large language model to generate output tokens, is projected to exceed $50 billion in 2026, with inference now representing 55% of AI infrastructure spending in early 202…
According to new data from the Washington Post, 40% of current U.S. races involve candidates who list AI and data centers as issues on their websites, making AI more prominent than cryptocurrencies, Israel, manufacturing…
A GitHub commit to the modelscope/ms-swift repository removed the Qwen 35B MoE model, frustrating developers who relied on it for local deployment. The removal suggests the model may be scrapped or pivoted, leaving a gap…
SambaNova Systems is hiring an Inference Systems Performance Architect in San Jose, California, with a salary range of $245k–325k per year, to own end-to-end performance for large-scale LLM inference. The role involves w…
A developer has released DeepRead, an open-source tool that transforms articles, books, and PDFs into structured reports centered on claims, evidence, data, and relationships. The tool links important claims to supportin…
Anthropic CEO Dario Amodei said in a rare post on X on Saturday that the AI industry can win over the public by delivering tangible scientific breakthroughs, such as actually curing cancer, rather than making big promise…
Unsloth, the open-source fine-tuning library, released Unsloth Desktop, a free beta app for macOS, Windows, and Linux that runs and trains LLMs, diffusion models, and audio models locally, directly competing with Ollama …
A developer built a web app using Google's Gemini 2.5 Flash vision model to analyze dog photos and return a 'judgment percentage,' with the technical implementation highlighting the use of a responseSchema to enforce str…
Google's Gemini 3.7 Flash, released August 13, built a playable browser game from a single prompt in 2 minutes and 13 seconds, a task Gemini 3.6 Flash failed three weeks earlier. The model runs at 75 cents per million in…