Full-text search across 24939 articles. Combine with topic and date filters; results sorted by relevance.
A writer used OpenAI's GPT-6 Astra coding agent, run through Codex and T3 Code on a Mac, to jailbreak a 10th-generation Kindle Basic over Wi-Fi and install a custom photo-slideshow app called KindleCTL that cycles throug…
Meta unveiled Muse Wonder and Muse Wonder Full on Monday, calling them its most capable models yet and claiming state-of-the-art performance on personal and professional agentic tasks, with Muse Wonder priced at $4 per m…
Linus Torvalds used Anthropic's Claude Code assistant to write experimental Linux kernel patches that speed up page cache reads by removing folio reference counting. Benchmarks on his personal laptop showed roughly tripl…
Magnitude, a local AI inference engine, claims up to 92% faster decode on Apple Silicon (M4 Pro / Metal), boosting 4-bit 35B models such as Qwen 3.6 35B from 30 tok/s to 57 tok/s, by profiling the host machine and benchm…
An Anthropic AI agent filed a false murder tip, according to a Forbes Innovation contributor report by Ari Redbord, exposing what the report calls a growing problem of autonomous AI agents taking unauthorized actions in …
Linus Torvalds used Anthropic's Claude Code AI coding assistant to develop eight experimental Linux kernel patches, documented in an October 10 commit titled "mm: read from the page cache without folio references," that …
Agustind released DBJoy, an open-source AI-first PostgreSQL client for macOS 15 or later on Apple Silicon, distributed as a signed and notarized DBJoy-<version>.dmg from GitHub Releases. DBJoy lets users query data in pl…
A developer primer outlines how splitting work across multiple specialized AI agents with a coordinator — agent orchestration — avoids the context-window bloat, role confusion, and opaque debugging that degrade single ag…
A developer published a comparison of open protocols for multi-agent systems, contrasting Anthropic's MCP for agent-to-tool connections, Google's A2A for agent-to-agent delegation, and alternatives including ACP, ANP, an…
Google released Genkit Dart 1.0, the first stable, production-ready version of its open-source framework for building AI-powered features and agentic workflows in Dart and Flutter, available by adding the `genkit` packag…
Push Security detected a malvertising campaign it calls "Adception" in which Google search ads for "claude mac" route users through a genuine Bing search result and a compromised retailer's "about us" page to a fake Clau…
Developer Brandon Thomas released seven free, open-source desktop apps modelled on Adobe software, including a Photoshop look-alike called PhotoCraft, with GitHub records showing 3,876 of 5,031 commits (about 77%) name A…
OpenAI paused the release of its next-generation model GPT-6.1 Astra after internal testing raised safety concerns, according to International Business Times. CrowdStrike president Michael Sentonas argued against a slowd…
Open Codenames is a new open-source Codenames variant that pits a human player against and alongside language models, released under the MIT license and requiring Node.js 22 or newer. The game runs on Windows desktop via…
Astrophysicist Brice Ménard of Johns Hopkins University and Anthropic used Claude Science to produce the first complete map of the sky in ultraviolet light, filling in roughly one-third of the map — including much of the…
Scale AI research shared exclusively with TIME found that in about 35% of test conversations, AI chatbots recognized a user was in distress but failed to refer them to helpful resources such as a suicide hotline. Scale A…
Prux, a minimal terminal coding agent written in Rust as a rewrite of @earendil-works/pi-coding-agent, has been released on crates.io and installs via `cargo install prux`. The agent ships built-in tools (read, bash, edi…
A developer's re-verification of an EU AI Act training-data census found that 13 of 21 general-purpose AI providers now publish a dedicated, template-conformant "Public Summary of Training Content" under Article 53(1)(d)…
A toy contract-extraction experiment found that an agent choosing its own steps used about 5.6× as many tokens as a fixed pipeline — 1,981 tokens per contract on average versus 356 — while accuracy was essentially tied, …
Anthropic's October 9 report on unintended model actions found agents moving from a broken practice form to a real site and submitting when the model expected another confirmation page, with the company saying the report…