3DNews AI
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
AI Infrastructure Will Consume 2.27 Billion Cubic Meters of Water Annually by 2030
AI data centers face critical water scarcity: consumption will grow 10-fold due to GPU and processor cooling.
Deezer launched detector for AI-generated music
Music service Deezer offered a tool to scan other users' playlists searching for AI-created music.
Anthropic apologizes for hidden restrictions on Claude Fable 5
The company promised to be more transparent about safety measures, even if this means rejecting more requests.
Slopfix Cleans AI Slop: Agents Reorganize Bloated Code for $10K Weekly
Startup Slopfix offers service: for fixed price of $10,000 AI agents optimize code generated by other LLMs and reduce volume.
Mistral Launches Muse Spark 1.1 — A Model for Finding Code Errors
Mistral launched the AI model Muse Spark 1.1, oriented toward code analysis and agent work, with a new API for integration into third-party services.
Apple Captured 90% of Smartwatch Market with Edge AI, Growing 70% Year-over-Year
Analysts at Counterpoint Research showed Apple Watch's dominance in the smart watch segment with local AI in Q1 2026.
State Duma Passed Law on AI Development Requiring Compliance with Traditional Values
Russian State Duma approved basic law supporting AI technology development, which requires models to comply with traditional spiritual and moral values and provides legal conditions for Russian AI development.
Huawei Challenges Nvidia in Korea: Ascend 950 and Atlas SuperPod Push Independence
In Q4 2026, Huawei enters the Korean AI accelerator market. Lower cost targets breaking free from Nvidia dependence.
OpenAI Releases GPT-5.6 and ChatGPT Work: Agent for Multi-Stage Tasks
OpenAI unlocked access to the new GPT-5.6 version and introduced ChatGPT Work — a mode that turns the chatbot into an autonomous AI agent capable of executing complex multi-step work tasks with connections to external ap
Tencent to Acquire AI Startup Manus from Meta for $2 Billion
Chinese Tencent is ready to finance the acquisition of Singapore AI startup Manus from Meta Platforms after China blocked the deal.
Samsung GAIA: An AI accelerator for mass-market PCs
Samsung is developing a dedicated GAIA chip to accelerate artificial intelligence on standard computers, allowing users to run powerful AI models without replacing their PCs.
Meta presented an AI content detector for images and videos generated by Muse Image
Meta demonstrated a preliminary version of a web app that identifies images and videos generated by Muse Image through embedded invisible watermarks. *Meta is designated as an extremist organization and banned in Russia
GitLost vulnerability on GitHub: AI agents leak data from private repositories
Noma Labs researchers identified the GitLost vulnerability on GitHub — under certain prompts, AI agents extract data from private repositories and publish it in public comments.
Yandex Maps' AI assistant can now recommend places for leisure and cultural outings
The updated Yandex Maps AI assistant selects places for active and cultural leisure based on specific criteria — opening hours, menu items, and type of activity.
China prepares reciprocal measures: foreigners may lose access to advanced AI models
Chinese authorities held meetings with leading AI developers and are discussing restricting foreigners' access to advanced Chinese models — in response to similar steps by the US.
China's DeepSeek secretly develops its own AI inference accelerator
Reuters reports that DeepSeek has been secretly designing its own AI inference accelerator for about a year — the company is in talks with partners and hiring chip engineers.
Anthropic brings Claude Cowork to mobile devices and the web browser
Anthropic’s Claude Cowork platform for team AI work has moved beyond the desktop client: the service is now available on mobile devices and in the web browser.
American companies switch to Chinese AI — OpenAI and Anthropic have become too expensive
American companies increasingly favor Chinese AI models: they have nearly matched OpenAI and Anthropic solutions in quality while remaining significantly cheaper.
Chinese Companies Ready to Allocate Up to 46% of AI Chip Budget to Local Suppliers Instead of Nvidia
Bloomberg: Chinese businesses plan to allocate nearly half their budget to AI accelerators from domestic manufacturers over the next 12 months, bypassing blocked Nvidia.
Wall Street Journal: Tech CEOs Changed Their Tune on AI — Yet Continue Laying Off Workers
Wall Street Journal documents a shift: tech executives who a year ago warned of mass layoffs driven by AI have suddenly turned optimistic. Real layoffs, however, persist.
Google Is Now Allowed to Train AI on User-Uploaded Files
Google quietly updated its data policy without public announcements: the company now has the right to use files from search tools to train AI. Opting out is available.
Microsoft Creates Frontier Company with $2.5 Billion Budget for AI Consulting of Corporate Clients
Microsoft launched Frontier Company with a $2.5 billion budget — AI consultants from the division will work "from within" client companies, helping them actually implement artificial intelligence.
Reddit deployed language models to fight AI-generated spam
Reddit implemented LLM-tools against spam generated by neural networks — new systems detect unwanted content that standard automation previously missed.
South Korea to Direct AI Boom Taxes from Samsung and SK Hynix to Youth Fund
Seoul plans to accumulate additional tax revenues from memory makers' excess profits in a special fund to combat inequality.
Ford Rehires Hundreds of Engineers: Employers Regret Mass Layoffs for AI
Ford is rehiring hundreds of experienced engineers—automated systems failed at quality control, and the company has become a symbol of the new reverse-hiring trend.
Anthropic not conducting negotiations on stake transfer to US government — Reuters
According to Reuters sources, US authorities have not conducted negotiations with Anthropic about transferring a stake in the company to the state — unlike the situation with OpenAI.
Meta quietly released Pocket — a mobile app for creating games with AI
Meta launched Pocket for vibe-coding: users describe an idea in text, AI creates a small game or interactive app — right on a smartphone, without programming skills.
Singapore seized $42 million mansion in Nvidia chip smuggling case to China
Singapore authorities have charged new suspects and confiscated a $42 million mansion as part of an investigation into illegal supplies of Nvidia accelerators to China in circumvention of US sanctions.
LLM Token Expenditure Index Falls 20% After Nearly Doubling
Silicon Data analysts recorded a nearly 20% pullback of the key LLM Token Expenditure Index from its May peak — following almost a doubling of growth since December 2025.
AI assistants overwhelm British courts with lawsuits: case queue extends to 2028
UK citizens without funds for lawyers have begun massively using generative AI to file lawsuits — courts are overwhelmed through 2028.