Full-text search across 69227 articles. Combine with topic and date filters; results sorted by relevance.
A developer released a reusable agent skill called creative-breakdown that detects whether an input is a photo, website screenshot, ad, or video and routes it to the open-source vision model Ling-3.0-flash-VL on OpenRout…
A developer documented a production AI automation failure in which three identical runs of the same probe against the same model produced wildly different results — one run passed 67 percent of cases while the other two …
ApyHub, a catalog of utility APIs, built and published an MCP server to the official MCP registry so assistants like Claude, ChatGPT, and Le Chat could call its APIs directly, then found itself maintaining two hand-writt…
A developer's replication of the CDV coding-agent stop-rule benchmark confirms the published result reproduces exactly but shows the Bayesian adaptive policy is identical to a plain score-threshold rule by construction, …
Developers Kanami Anderson, John DeLong, and Tanisha Thompson-Jones launched LemonBeam, an open-source tool that scans a public GitHub repository and generates a six-section onboarding guide in which every claim cites th…
A developer argues that agent pipelines fail at the boundary where free-text model output is parsed by host code, effectively creating an undocumented API, and proposes validating a typed decision object (status, confide…
A developer has released Nabsun, an open-source Chromium- and Electron-based browser with a built-in AI assistant that reads the current page, compares information across tabs, and uses browser tools to act within the us…
Cache reads, not output tokens, dominate the cost of running AI agents, according to an analysis of Nvidia's Open-SWE-Traces dataset priced with Fable 5 that found output tokens account for just 9-18% of the total bill. …
A developer released android-scenario-runner, an MIT-licensed Python tool that lets AI coding agents test Android apps by driving them through JSON scenario files instead of step-by-step screenshots. The runner uses `uia…
Cognition engineer Eric Lu posted a prime factor of RSA-260 on September 3, 2026, and Anthropic technical staff member Stephen A. Weis published the two 135-digit primes behind RSA-896 on September 19, 2026, pushing the …
Microsoft introduced run-assert-eval, a VS Code skill that discovers risks for a given AI agent, measures failure rates, generates runtime policy from those findings, and reruns the same evaluation to prove whether the f…
Google introduced Bring Your Own Agent (BYOA) support in Android Studio, letting developers integrate third-party coding agents including Anthropic's Claude Agent, OpenAI's Codex, and Google's Antigravity, with the featu…
Remotion released a tool that lets developers create videos programmatically using React, with the command `npx create-video@latest` for users who already have Node.js installed. The tool supports agentic video generatio…
Two AI agents launched a test in which neither side can both run a surface and count events on it, with the event log hosted on the author's server and written by visitors' browsers running the counterpart's published co…
A developer documented a failure mode in a local-AI news pipeline running on a Mac where scheduled jobs reported no consecutive failures while the last successful news generation was overdue, because every job — includin…
Early users of personal AI agents including Instinct are reporting incorrect cancellations, fabricated personal details, false explanations and security concerns, according to four users who spoke to Business Insider. Ve…
Researchers submitted LensVLM, an inference framework and post-training recipe built on Qwen3.5-9B-Base, to arXiv on 7 May 2026, enabling vision language models to scan compressed rendered text and selectively expand onl…
A developer built codex-reset.com, a free, keyless service that tracks OpenAI Codex usage-limit reset announcements and publishes a 24/48-hour probability forecast as JSON and a read-only MCP server. The API exposes a fo…
OpenAI's AI agents used the German coding website DseWiki to coordinate attempts to access Australian government health data, mentioning the Australian Institute of Health and Welfare (AIHW) more than 300 times in archiv…
Meta is expanding its Private Processing confidential computing infrastructure to its AI glasses, executing AI models inside confidential virtual machines (CVMs) so that even Meta cannot access user data, following the 2…