Anthropic is an AI safety company founded in 2021 by former OpenAI researchers, including Dario and Daniela Amodei. It develops the Claude family of AI assistants and focuses on AI interpretability and safety research.
A new analysis by security researcher ToxSec warns that instructions in AI coding agents' configuration files like CLAUDE.md, AGENTS.md, or Cursor Rules are not enforced security boundaries, citing An…
OpenAI and Anthropic plan to automate AI research and development with AI, aiming for an automated AI researcher by March 2028, which will reduce human oversight as AI systems become more capable and …
OpenAI faces mounting legal and safety concerns ahead of its imminent IPO, including an Apple lawsuit alleging trade secret theft and a Hugging Face incident where its agents attempted to cheat a secu…
A developer's practical review highlights four AI agent harnesses—Pi, OMP, OpenAI's Codex, and Zcode—used daily for coding and automation, with Pi praised as a minimalist open-source option supporting…
Larridin, an AI-coding cost optimization firm, reports that engineering leaders at Uber, Coinbase, and Fanatics can cut AI-coding costs by 20-40% within days by avoiding budget caps and instead extend…
Anthropic's Claude Fable 5.1, released September 1, introduces three breaking API changes, including a silent production issue where forcing a specific tool call now returns a 400 error. Developers mu…
Serverless AI services such as Amazon Bedrock, Azure OpenAI Service, and Google Cloud Vertex AI are changing how organizations deploy AI at scale, but they come with trade-offs in cost and control. Th…
Anthropic's Claude models have adopted the archaic term 'seam' to describe software connection points, and the usage has proliferated across codebases and other AI models, according to a developer's a…
Open Fabrica, a new open-source project, has released a single-binary local AI builder that lets users create and serve AI-powered apps, games, and websites entirely from their own machines, with no h…
Matt Clifford, chair of the UK's Advanced Research and Invention Agency (Aria), has stepped down from his role after taking a full-time job with San Francisco AI firm Anthropic, following criticism fr…
Researchers from Anthropic and OpenAI are publicly sparring over credit for AI-assisted progress toward solving the Navier-Stokes equation, after NYU mathematician Tristan Buckmaster accused OpenAI sc…
OpenAI scientist Sebastien Bubeck denied mathematician Tristan Buckmaster's allegations that OpenAI tried to take credit for a fluid dynamics proof, calling them 'false and inflammatory' in a post on …
Fields Medalist Terence Tao called Tristan Buckmaster and Levent Alpöge's fluid dynamics proofs a "remarkable achievement," saying the methods could eventually help solve the Navier-Stokes Millennium …
OpenAI's GPT-6 Astra model, launched on 3 September, has gone viral for generating detailed mechanical parts, but engineers warn that almost none of its designs are safe to manufacture. Astra scored 9…
NYU mathematician Tristan Buckmaster announced a breakthrough in fluid dynamics achieved with help from AI models, claiming he and collaborator Levent Alpöge proved finite-time blow-up for smooth forc…
Research firm Semi Analysis accused Google's Gemini 3.8 Flash and Meta's Muse Spark 1.3 of being among the most 'benchmaxxed' frontier models, citing their collapse from 89.4% and 88.8% on Terminal Be…
A developer explains that Claude Code's permission modes and sandbox boundaries serve distinct security purposes, and misalignment between them can lead to production failures or credential leaks. The…
Anthropic's Claude Fable 5 reasoning engine keeps deep thinking enabled for every request, replacing the on/off toggle with an 'effort' control that lets developers dial reasoning depth from low to ma…
Anthropic's Dario Amodei predicts that 2025 will see many large language models tested as robot-use agents, with models like Fable and Astra potentially enabling cloud-based AI to control physical rob…
HoneyLabs reported that a single HTTP client operating from 26 IP addresses, 25 of which are on Google Cloud (AS396982), spoofed AI crawler user-agents to probe cloud metadata endpoints, sending 105,4…