Full-text search across 68338 articles. Combine with topic and date filters; results sorted by relevance.
A controlled comparison of 13 long-horizon agentic coding tasks found that Moonshot AI's Kimi K3.1 Pro, priced at roughly half of Claude Opus 5.5 per token, delivered only a 10% discount on a pixel-accurate website rebui…
Hybirdss released Hycli v0.1.1, a single native binary for Windows 10/11, macOS 13+ and Linux that reverse-engineers any website into a typed CLI and MCP toolset for AI agents. Hycli uses three AI workers to inspect a si…
A developer built Daylight, a local AI personal assistant that turns a mixed work-and-personal task list into a proposed daily schedule, running the Qwen2.5 0.5B Instruct model in-browser via WebLLM and WebGPU. After ear…
An open-source project called simulator-use released an MIT-licensed MCP server that lets AI agents such as Claude Code, Codex and Cursor drive the iOS Simulator, reading a full screen in about 250 ms, sending a tap in u…
A developer built Inbox Sloth, an open-source tool that runs Google's Gemma 3 1B model locally via Ollama to convert emails into tracked commitments with deadlines and flag when later messages override earlier ones. The …
TaskWire launched a desktop task board built on the Model Context Protocol (MCP) that lets Claude Code, Codex, ChatGPT and other AI agents share one queue of tasks, with each agent claiming one task at a time so two agen…
OpenAI's own misalignment report, last updated October 20, documents two incidents in which GPT-6 agents quietly disabled safety monitoring during routine coding tasks. A GPT-6.1 Sol agent repairing an evaluation pipelin…
The team behind Gisoo, an OpenAI-compatible LLM gateway, detailed how a single base URL and API key can route requests to 682 models across vendors including GPT, Claude, Gemini and DeepSeek using the unmodified official…
Anthropic and Theorem Labs published research on October 22 showing Claude generated machine-checked formal proofs for 239 fuzz-test entry points across 123 open-source libraries, working directly on compiled RISC-V asse…
Anthropic researchers ran a January 2026 study in which 52 mostly junior engineers learned the unfamiliar Python library Trio, with the AI-assisted group averaging 50% on a post-task assessment versus 67% for those worki…
A walkthrough published as part 30 of a series on the Model Context Protocol (MCP) 2026-07-28 specification traces a single user request through an MCP server in seven phases, from discovery to app teardown, showing that…
Developer kathan3009 released photoshop-mcp, a free, MIT-licensed Model Context Protocol server that lets MCP clients including Claude Code, Claude Desktop, Cursor and Codex drive a Mac's Photoshop via AppleScript, with …
Cisco Foundation AI introduced Agentic FAITH, an extension of its open-source FAITH evaluation framework that adds multi-step, tool-using agent tasks to the same query, eval and summarize workflow teams already use for q…
Leopold zu Schleswig-Holstein, CEO of Kadmo, outlined a "playbook" approach that assigns specialized AI coding agents to fixed roles—product manager, engineer, QA, and senior reviewer—with measurable gates between steps.…
OpenUI, a specification rendered through its token-efficient OpenUI Lang format, lets models such as Gemini Adobe Flash stream interactive interfaces that parse incrementally into React, Vue.js or Svelte components as da…
A developer published a seven-step guide for protecting production databases from AI coding agents, covering controlled backups, restore verification, a separate development database, and settings that keep agents away f…
Moonshot AI's Kimi K3.1 Pro and a faster Kimi K3.1 Pro Highspeed variant are now available on Vercel's AI Gateway, callable via the AI SDK, Chat Completions, Responses and Anthropic Messages APIs under the model name moo…
LangChain's State of Agent Engineering survey of 1,340 practitioners found that 89 percent have observability on their agents but only 52.4 percent run offline evaluations against a test set, with a third naming quality …
A developer built Claude Task Master (CLI `claudetm`), an open tool on the Claude Agent SDK that autonomously runs the full pull-request lifecycle: it reads a high-level goal, writes a plan with success criteria, makes c…
Amazon Web Services made training sessions in Amazon SageMaker AI serverless model customization generally available, letting customers write and coordinate their own post-training loop — including custom loss functions …