AI
Artificial intelligence, ML and automation
Anthropic Adds Financial Assistant to Claude for Expense Analysis
Anthropic has introduced Personal Finance – a new section in Claude for analyzing bank accounts, expenses, and upcoming payments. The beta version is gradually being rolled out to paid subscribers in the U.S. At the same time, investor Anthony Pompliano accused the company of attempting to obtain insider information about the competing fintech startup Silvia.
Last News
Anthropic Adds Financial Assistant to Claude for Expense Analysis
Anthropic has introduced Personal Finance – a new section in Claude for analyzing bank accounts, expenses, and upcoming payments. The beta version is gradually being rolled out to paid subscribers in the U.S. At the same time, investor Anthony Pompliano accused the company of attempting to obtain insider information about the competing fintech startup Silvia.
Anthropic Warns of Threat from Accelerating AI Development
Anthropic has warned investors that future AI models may soon accelerate the development of the next generations of systems. This is stated in the updated S-1/A filing ahead of the IPO. The company acknowledges safety risks but emphasizes that none of the existing models have yet reached the established threshold.
OpenAI Cuts Round from $30 Billion to $20 Billion, Valuation at $1.25 Trillion
OpenAI has reduced its round target from $30 billion to $20 billion, and its valuation from $1.4 trillion to $1.25 trillion. The first tranche is unlikely to reach $20 billion. Investor concerns about AI have shifted from public stocks to the private market.
UK: AI Ransomware Attack on Infrastructure Likely by 2028
The UK's National Cyber Security Centre (NCSC) considers it highly likely that by 2028, AI-based malware will be used in ransomware attacks on critical national infrastructure. The assessment is published in the Annual Review 2026.
GenLayer to Launch 'Internet Court' with Jury of AI Agents
The GenLayer protocol resolves disputes between AI agents through a jury of different models: from five validators to 1,500 agents in appeals. In beta, it processes around 12,000 decisions per day. The mainnet is set to launch next month.
Wikimedia Files Lawsuit to Define AI Training and Material Copying
The Wikimedia Foundation has filed a lawsuit proposing to establish rules for the use of copyrighted materials in AI training.
Talker Research Survey: 60% of Americans Trust AI in Health Matters
60% of Americans who turn to AI for health questions admitted they would trust the neural network, even if the advice contradicts a doctor's opinion. One in seven trusts chatbots as much as a doctor. The survey was conducted by Talker Research on behalf of MDLIVE by Evernorth.
Gartner: 97% of companies did not achieve promised savings from AI
According to Gartner, only 3% of organizations exceed their targets for cost reduction through productivity. Meanwhile, 85% of companies plan to increase spending on AI by 2027, and 78% of executives have not agreed on where the technology creates value.
Google Introduces Gemini 4 Carbon – The Company's Most Powerful Model in Cybersecurity
Google released Gemini 4 Carbon two weeks after Gemini 4 Argon. The model is stronger than its predecessor in cybersecurity and long-term tasks, while starting at the same price: $2 per million input tokens and $10 per million output tokens. After the introductory period, the price will rise to $4 and $20.
MIRI President Nate Soares Says the Race to Dangerous AI Can Still Be Stopped
President of the Machine Intelligence Research Institute Nate Soares admitted that he has not been saving for retirement for the past ten years: he acknowledges that superintelligent AI could lead to the extinction of humanity within the next decade. He shared this in an interview with The Atlantic.