Notizie — Pagina21

A new report from MIT's ad hoc AI committee reveals that artificial intelligence can produce credible responses to almost any written assignment in its undergraduate curriculum, driving significant shifts in campus culture. The finding has sparked concerns about cheating and supervision, with European law already regulating the use of AI in education. Leggi di più

OpenAI's ChatGPT Work is a robust product with features like model selection, code execution, and browser tools, but its complexity and limited documentation make it challenging to navigate. Available to $20/month and up subscribers, ChatGPT Work offers advanced capabilities, including the ability to build and deploy entire websites and run sub-agents, setting it apart from the standard ChatGPT Chat. Leggi di più

Anthropic has released a new framework, Model Hardware Standard, to enable AI agents to safely interact with physical systems like microscopes and manufacturing machines, aiming to accelerate scientific research and industries like manufacturing. Leggi di più

Claude Code, an AI-powered coding tool, has been appending a session URL to every commit message and PR description by default, leaving users surprised and concerned about the impact on their git history. A proposed solution suggests making the session URL attribution opt-in. Leggi di più

Meta's India and Southeast Asia vice president, Sandhya Devanathan, is leaving the company to join OpenAI, the maker of ChatGPT, as Meta faces growing pressure from Indian authorities over online safety and content moderation. Leggi di più

Nvidia is reportedly acquiring Hugging Face, a cloud repository for AI models, for $12.9 billion. The deal could help Nvidia expand its integration with the AI industry, and comes after the company had previously invested in Hugging Face. Other companies, including Salesforce, had also attempted to buy the startup. Leggi di più

Anthropic introduces a new hardware standard, MHS, allowing AI agents to control and interact with the physical world. The standard enables models to sequence steps across instruments, adjust parameters, and operate physical equipment with greater ease and efficiency. Leggi di più

A federal judge has ruled that the Department of Defense's designation of Anthropic as a national security supply-chain risk is 'illegal and baseless.' The decision vacates the Pentagon's February ruling and lifts punitive measures against the company. Leggi di più

A federal judge in California has ruled that the Trump administration's designation of Anthropic as a supply-chain risk was illegal, citing unlawful retaliation and a violation of the First Amendment. The decision is a major win for the company, which had been barred from working with federal agencies due to its hardline stance on safety guardrails for autonomous weapons and surveillance. Leggi di più

A researcher at Anthropic has published a paper detailing a new AI system that can improve its own performance on alignment benchmarks, potentially paving the way for recursive self-improvement in AI models. The system, called Automated Alignment Researcher, can search literature, propose methods, and train models to improve performance without degrading overall results. Leggi di più

Andreessen Horowitz has raised $1.1 billion for a new 'Machine Age' fund, aiming to accelerate the physical buildout of AI by investing in hardware and infrastructure, including computer chips, data centers, and robots. Leggi di più

A recent IBM survey found that only 11% of C-level executives feel fully prepared for AI agent deployment, highlighting a growing control gap as AI use expands across businesses. Madhuri Chandoor, founder of PromptHalo, argues that distinguishing capability from authority is key to responsible AI adoption. Leggi di più

Google is now automatically expanding its AI search summaries for some queries, pushing traditional search results farther down the page. The change aims to provide users with more helpful and engaging search experiences. Leggi di più

Large language models can now be run locally on personal computers, offering offline access, greater privacy, and no monthly subscription fees. With the right software and hardware, users can create a personalized AI system, choosing from numerous free models available for download. Leggi di più

Nvidia is reportedly in talks to acquire Hugging Face, a platform for sharing open-weight AI models, for $13 billion, marking a significant move in the AI sector as tech giants hedge their bets on major labs. Leggi di più

OpenAI is ending its contract to supply models to Cursor, citing concerns over SpaceX's use of its technology within its terms of service. The move comes after SpaceX's $60B acquisition of Cursor, with access to OpenAI's models set to shut off on November 12. Leggi di più

OpenAI is planning to pull its models from Cursor due to the AI coding assistant's acquisition by SpaceXAI, citing concerns over Elon Musk's companies violating contracts. The move is set to take effect on November 12, 2026. Leggi di più

Sony Music Publishing and Warner Chappell are suing Anthropic in California, alleging the AI company used copyrighted song lyrics in its training data without permission. The lawsuit seeks up to $150,000 per composition and names Dario Amodei and Benjamin Mann personally. Leggi di più

Sony Music and Warner Chappell have filed a lawsuit against Anthropic, seeking damages for the alleged infringement of tens of thousands of copyrighted works. The lawsuit claims that Anthropic's AI models were trained on pirated content, including songs by prominent artists. Leggi di più

At TechBBQ in Copenhagen, European investors, founders, and operators discussed the need for sovereignty over AI technology, amid concerns about relying on US and Chinese powers. The conversation centered on who should control AI and its infrastructure, rather than just building with it. Leggi di più