Full-text search across 70861 articles. Combine with topic and date filters; results sorted by relevance.
Google released Gemini 3.8 Flash on September 2, its third Flash-tier model in six weeks, undercutting rivals on price while beating GPT-5.6 Terra and Claude Sonnet 5 on the Terminal-Bench 2.1 coding benchmark. The rapid…
A Linux user running Ubuntu 24.04.4 LTS with an Intel Arc B580 GPU, Ryzen 5 7600, and 32GB RAM details struggles and workarounds for running local LLMs on Intel Arc hardware, documenting issues in a GitHub repository (sp…
GitHub announced that enterprise-managed settings now support any default model for GitHub Copilot, allowing administrators to set a preferred default model for new conversations and customize defaults by enterprise team…
CodeYam CEO and cofounder Nadia Eldeib outlined two installation paths for CodeYam: a CodeYam MCPB (Model Context Protocol Bundle) desktop extension for Claude Desktop aimed at newer or non-technical builders, and a Code…
Coder announced Coder Agent Relay, a self-hosted execution environment for cloud coding agents, with SpaceXAI as its launch partner, aiming to bring Cursor agents into regulated enterprises.
FLOSS Weekly Episode 880 features Jonathan Bennett interviewing Benjamin Samuels of Trail of Bits about Patch the Planet, an initiative to help open-source projects address vulnerabilities introduced by AI coding tools. …
GitHub announced that content exclusions are now generally available in the GitHub Copilot app and Copilot CLI, allowing enterprise, organization, and repository administrators to configure policies that prevent Copilot …
A developer added an agentic interview mode to their STRIDE threat modeling tool, P2 Threat Model Generator, which reads Docker Compose, OpenAPI, and Kubernetes manifests. The new mode uses an LLM to interview users turn…
Google released Gemini 3.8 Flash, its third budget model in six weeks, which matches Claude Opus 5 on some agentic coding benchmarks at a lower cost, according to The Decoder. However, the model's reasoning burns about 3…
Francisco Carlos Serra released Agent Voice Kit, an open-source tool that lets users talk to AI agents like Claude Code, pi, Codex, and Gemini CLI locally, with dictation and optional text-to-speech, running entirely on …
Panic Room, an open-source MIT-licensed tool, provides a zero-dependency sandbox for AI agents, giving each agent a cloned repo, fake credentials, a throwaway database, and a container, with an audit trail and a diff for…
Trail v1.25.2, an issue tracker and durable project memory tool for coding agents, was published to the official MCP registry on 2026-09-02 and is distributed as usetrail on npm. The registry listing declares 2 environme…
Windrunner has launched an AI-powered project management platform that uses an LLM agent to monitor project flows, automate status updates, and detect bottlenecks proactively, aiming to reduce administrative overhead for…
Black Duck has launched its Signal vulnerability scanning engine as an MCP server in the Claude Directory, enabling developers using Anthropic's Claude Desktop to scan code for security flaws without switching tools. The…
An engineer's experiment building an LLM-powered support agent concludes that prompt-based restrictions are ineffective against injection attacks, advocating instead for architectural enforcement through scoped tools and…
GitLab 19.3, released August 20, introduces a native Secrets Manager in limited availability and brings the Dedicated AI Gateway to general availability, enabling finance, healthcare, and defense teams to use Duo Agents …
Researchers introduced VeriOCRBench, a 1,800-sample human-verified benchmark with 1,600 trap-injected invalid tasks across 8 trap types and 200 trap-free controls, to test whether multimodal large language models can det…
Researchers propose H2Table, a hierarchical hypergraph-enhanced framework for complex table reasoning, which represents tables as nested hypergraphs and uses a tailored encoder with learnable query vectors to bridge stru…
Researchers propose DISTAL, a dual-prior framework for structure-agnostic materials property prediction that combines self-supervised compositional pretraining with knowledge distillation from a pretrained ALIGNN teacher…
Researchers propose AnySearch, a reinforcement learning framework that enables a single LLM-based search agent policy to adapt to any budget constraint at deployment, outperforming baselines across seven QA benchmarks an…