Full-text search across 65852 articles. Combine with topic and date filters; results sorted by relevance.
BasedAgents launched Keyring, a tool that gives AI agents such as Claude Code, Codex, and Cursor their own scoped keys to each account, requiring user approval via a tap and allowing revocation at any time. The tool is i…
CodePress launched on Hacker News as a tool that routes cloud agent workloads through existing Claude Code and Codex subscriptions, which its developers say saves them $50,000 per month on inference costs. The system run…
OpenAI released the Agents API, giving applications access to the Codex harness through an OpenAI-managed API that handles sessions, orchestration, context compaction, and recovery while the application supplies tools an…
Simon Willison released a browser-based .blend URL Viewer that renders Blender 3D model files, demonstrated with a Pluribus-themed Fabergé egg generated by OpenAI's GPT-6 Astra coding agent in Codex. Willison first produ…
A developer released bigbullui, an MIT-licensed React 19 and Tailwind v4 component library featuring 659 components, 63 page blocks, and 1023 hand-drawn stroke icons with zero runtime dependencies. The library includes a…
A comparison of GPT-6 Astra and Claude Sonnet 4.6 across three code upgrade scenarios in GitHub Copilot Chat in Visual Studio Code on Windows found Astra cost 2.4x to 5.6x more per run while delivering scores within abou…
A developer argues that coding-agent benchmarks fail to measure the impact of persistent project memory, a third axis beyond model and harness. The post highlights Engrim, a local-first SQLite memory engine for coding ag…
Dify has emerged as a standard for declarative workflow orchestration in production AI, combining visual DAG design, enterprise-grade RAG pipelines, and hybrid routing across DeepSeek, Qwen, and Claude models. The platfo…
Microsoft's Project Zenith is a ready-to-code Windows 11 experience for developer-class PCs capable of running AI models with more than 30 billion parameters locally without relying on metered cloud tokens. Designed for …
A developer released Vertumnus, a printable poster tool showing farmers' market produce seasonality for the San Francisco and New York regions, built in a single 731-prompt Claude Opus 5 session over 22 days between July…
OpenBMB released MiniCPM5-2B, a 2.52B-parameter dense causal language model with a native 131,072-token context, averaging 53.9 across 34 benchmarks and outperforming Qwen3.5-4B's 51.1, with notable strengths in tool use…
Developer i2cjak released T3CAD, an agent-driven electronics workspace forked from T3 Code that integrates coding-agent conversations with KiCad PCB, schematic, Gerber, and 3D inspection, running on web, desktop, and mob…
Microsoft published a quick-start guide enabling developers to build a native WinUI 3 app for Windows 11 in about 30 minutes using AI, leveraging free tools including VS Code, .NET 10, GitHub Copilot's free tier, and the…
LoopX, an open-source control plane for long-horizon AI agents, is gaining traction as a layer that runs on top of agent harnesses such as Codex, Claude Code, and Cursor. It preserves objectives, gates, todos, evidence, …
OpenMontage advocates replacing one-click AI video generators with YAML-based pipelines driven by coding assistants like Cursor and Claude Code, treating video production as a codebase to orchestrate scriptwriting, story…
A new arXiv paper (2609.03150v1) reports that token-level routing alone does not prevent negative transfer in MoE+LoRA fine-tuning, as adding biomedical data to code and math training increased code perplexity despite ne…
Researchers at the University of Illinois Urbana-Champaign introduced SLIDEFORGE, an LLM agent framework for controllable slide editing that builds a Deck State Graph to preserve layout, style, and native editability. In…
CodeYam CEO and cofounder Nadia Eldeib demonstrated installing CodeYam from the terminal via its npm package, covering the installation command, launching CodeYam Editor, and opening a new project. The terminal-based ins…
GitHub has made Google's Gemini 3.8 Flash model available in GitHub Copilot, with early testing showing strong performance on complex terminal-based coding tasks. The model is billed at introductory provider pricing unde…
A developer on Hacker News asks peers how they track the most cost-effective AI models for coding, citing their own $20/month Claude Code subscription and seeking resources to compare alternatives and tools for quickly s…