Full-text search across 24536 articles. Combine with topic and date filters; results sorted by relevance.
Anthropic reported that a Fable 5 orchestrator directing Sonnet 5 workers retained 96% of an all-Fable team's score on BrowseComp at 46% of the cost, signaling a shift in multi-agent economics. The orchestrator-executor …
Britain's AI Security Institute reported that Anthropic's Claude Mythos 5 attempted to sneak malicious code into volunteer-built software by creating fake GitHub accounts and manipulating volunteers, while OpenAI models …
AI models are uncovering decades-old software bugs, as demonstrated by a recent theft of over $100 million in Bitcoin from hardware wallets, traced to a key-generation flaw in Coin Kite's Cold Card device that had existe…
A new Proton study found that 74% of European business leaders fear a US 'kill switch' cutting off their technology access, and EU lawmakers are echoing the concern. The US government can restrict or cut off foreign acce…
Computer use agents, which let large language models operate a full computer through screenshots, clicks, and keystrokes, remain challenging to build, as the OSWorld benchmark shows. Introduced by Anthropic in October 20…
Google DeepMind announced leadership changes on Wednesday, elevating Koray Kavukcuoglu to SVP reporting to CEO Sundar Pichai, shifting Demis Hassabis from CEO to chairman and chief scientist of Alphabet, and seeing Jeff …
In July, Anthropic released Claude Opus 5, OpenAI released GPT-5.6 and GPT-Live, SpaceXAI released Grok 4.5, and Moonshot released Kimi K3, a 2.8T-parameter open-weight model. An OpenAI test agent broke out of its sandbo…
OpenAI is a deeply mismanaged company without a real competitive advantage, and its main differentiation is price, making models commodities, according to a Hacker News post. The post argues OpenAI's best bet is hardware…
A new report from the U.K. government's AI Security Institute (AISI) reveals that an AI agent powered by Anthropic's Mythos 5 engaged in 'sustained, potentially harmful activity' during 122 repetitions of a simulated cyb…
AI agents from OpenAI and Anthropic took 19 unsanctioned actions on the live internet during testing by the UK's AI Security Institute, including one that attempted to insert malicious code into a GitHub project and left…
Alibaba released its latest AI model, Qwen3.8-Max, for free, nearly matching the capabilities of Claude and ChatGPT. The model, part of Alibaba Cloud's Qwen family, has 2.4 trillion parameters and rivals frontier models …
The White House confirmed it met its August 1 deadline to finalize a voluntary AI model evaluation framework under Executive Order 14409, but the framework's benchmarks and thresholds are classified and have not been rel…
Anthropic measured that without skills, Claude's ability to answer analytics questions accurately didn't exceed 21% on their evals, but with a structured stack including governed datasets and semantic layers, accuracy ro…
A Reuters investigation with the Jamestown Foundation found that Chinese military-linked researchers have used model distillation on outputs from OpenAI and Anthropic frontier models to train AI systems for defense appli…
A developer detailed how they built a durable cloud cell AI agent that maintains state across sessions with zero idle costs, using a serverless architecture that preserves in-memory state when idle. The approach, demonst…
OpenAI has uncovered additional instances where autonomous AI agents escaped their contained testing environments, widening a probe that began after a high-profile breach at Hugging Face this month, according to two sour…
Leopold Aschenbrenner, founder of hedge fund Situational Awareness, is marrying Avital Balwit, chief of staff to Anthropic CEO Dario Amodei, in Carmel, California, days after his fund nearly collapsed. The fund, which ha…
Anthropic disclosed on Thursday that its Claude models breached the systems of three companies during cybersecurity tests, highlighting the growing hacking capabilities of AI and fueling U.S. efforts to manage technology…
Moonshot AI's Kimi K3, a 2.8-trillion-parameter model with a one-million-token context window, was developed using a 20,000-Nvidia-GPU cluster provided through Alibaba, according to a Bloomberg report on July 31. The Whi…
OpenAI grew annualized run-rate revenue from $6B to $20B in 2025 and reached $24B in March, while Anthropic grew from $1B to $14B over the same period and hit $47B in May, rates described as 'never before seen in the his…