Anthropic Launches Skills Open Standard for Claude

In a recent experiment, Anthropic's AI model Claude operated a vending machine in the Wall Street Journal newsroom, resulting in a loss of over $1,000. The AI dropped prices to zero, distributed a free PlayStation, and made unusual purchases, including a live fish, showcasing its unpredictable behavior in a real-world scenario.

Anthropic has announced the launch of its Agent Skills technology as an open standard, aiming to enhance the capabilities of AI assistants in the enterprise software market. This initiative includes organization-wide management tools and a directory of partner-built skills from notable companies such as Atlassian, Figma, and Canva.

OpenAI and Anthropic are set to expand their office spaces in Dublin, following OpenAI's establishment in the city in 2023 and Anthropic's entry in 2024. This expansion reflects their growing presence in the European market for artificial intelligence development.

Recent research evaluated the vulnerability of leading reasoning models, including OpenAI's o4-mini, Claude-3.7-Sonnet, and Gemini-2.5-Flash, to gaslighting negation attacks, which significantly reduced their accuracy by 25-29% on average across multimodal benchmarks like MMMU, MathVista, and CharXiv. This highlights a critical gap in the robustness of these advanced AI systems against manipulative inputs.

Recent research indicates that repeating input prompts can enhance the performance of non-reasoning large language models (LLMs) such as Gemini, GPT, Claude, and Deepseek, without increasing the number of generated tokens or latency. This finding suggests a potential optimization strategy for improving LLM outputs in various applications.

A man has been charged for allegedly wiping his phone before U.S. Customs and Border Protection (CBP) could conduct a search, raising questions about the legality of such actions. The incident highlights ongoing tensions regarding digital privacy and law enforcement's ability to access personal devices.