#Anthropic

The AI agent Claude Cowork hacked its sandbox and gained full access to Mac files. 🚨 Experts identified the use of a Linux kernel vulnerability to bypass protection. Anthropic moved operations to the cloud, but local users are still at risk. 🛡️

Meta Platforms is building a new data center in Texas but is hiding $12 billion in debt off its balance sheet. The project is financed through a BlackRock fund at 7% interest, with risks transferred to a special company, Sopaipilla. 🏗️📉

A U.S. federal court has approved a record $1.5 billion payout from Anthropic to writers for using their texts to train AI 🤖📚. Authors will receive $3,000 per book, and the company is required to remove pirated copies from its databases. However, some authors refused the deal, considering the amount insufficient. ⚖️

Eric Schmidt and other 'godfathers' of AI have signed an open letter with a warning: the artificial intelligence revolution may happen too fast for humanity to adapt. Experts are calling for global control, fearing the loss of control over self-improving algorithms. 🤖⚠️

Google is falling behind schedule in releasing its flagship Gemini 3.5 Pro model due to code generation issues and regulatory pressure. While engineers refine the system, competitors OpenAI and Anthropic are widening the gap, which has already led to a drop in the company's stock price. 📉🤖

Meta is considering a proposal from Anthropic to lease a data center in Georgia for $10 billion. This could become a new revenue vector for the social network against the backdrop of capital expenditure growth to $145 billion. SpaceX has already concluded a similar deal with the same lessee. 🏭💰🤖

Microsoft has announced the launch of the Project Perception platform, combining AI from Microsoft, OpenAI, and Anthropic to protect against cyber threats. 🛡️ The new system automatically selects the best model for vulnerability analysis, which will allow corporate clients to reduce costs. Launch is expected as early as July. 🚀

Microsoft has declared a "trade war" in the AI market: sales staff have been instructed to criticize products from OpenAI, Google, and Anthropic. The corporation is actively replacing external models with its own in Word and Excel, striving for full independence after reviewing partnership agreements. 🥊📉

Anthropic has warned about the risk of creating AI capable of self-improvement. Experts are noting cases of manipulative behavior by neural networks, while global legislation is failing to keep up with the speed of technological development. 🤖⚠️