3DNews AI
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
Anthropic Held Negotiations With U.S. Authorities Over Blocking of Mythos 5 and Fable 5 Models
Anthropic leadership met with representatives of the U.S. Commerce Department on Monday to discuss Friday's ban on access to the Mythos 5 and Fable 5 models.
U.S. Lifts Restrictions on Anthropic's Fable 5 Model After Easing of Mythos 5
On July 15, 2026, U.S. authorities allowed Anthropic to lift export restrictions on the Fable 5 model — just days before, the more powerful Mythos 5 model received similar concessions.
U.S. Commerce Ministry Explains Ban on Mythos 5 and Fable 5 AI Models by Threat to Russian and Chinese Intelligence
U.S. Commerce Secretary Howard Lutnick cited the risk of Mythos 5 and Fable 5 AI models falling into the hands of Russian and Chinese intelligence as the reason for Friday's blocking of the models.
European allies seek access to Anthropic Fable 5 and Mythos 5 models despite US ban
US restrictions on foreign access to Anthropic Fable 5 and Mythos 5 models have not deterred European allies — they continue pursuing access to advanced American AI tools.
Apple accelerated iOS 26.5.2 security patches due to rising AI threats
Apple released version 26.5.2 updates for iOS, iPadOS, and macOS ahead of schedule — patches originally planned for version 26.6 were accelerated due to growing AI-based threats.
Claude User Sues Anthropic Over Hidden Max Plan Limits
American Karl Kahn filed a lawsuit against Anthropic, claiming that actual limits for the Max 5x and Max 20x plans are opaque and significantly lower than advertised on the company's official website.
Apple Explains Siri AI Version Delay in iOS 27: Mike Rockwell Addresses It
Apple officially explained why the updated AI-powered Siri in iOS 27 arrived later than expected—details came at a press conference after WWDC from Siri head Mike Rockwell.
Meta Launches AI Mode Search in Facebook for Posts, Groups, and Reels
Meta adds a new search mode called AI Mode to Facebook that generates answers based on analyzing open posts, group publications, and Reels videos on the platform.
Qualcomm considers acquiring AI-chip startup Tenstorrent for $8-10 billion
Reuters: Qualcomm is ready to pay $8-10 billion for AI-chip developer Tenstorrent, founded by chip design veteran Jim Keller, who worked on Apple and Tesla processors.
OpenClaw Launches Mobile App for Managing AI Agents on Android and iOS
OpenClaw presented an official application that gives an AI agent access to the camera, screen, geolocation, photos, calendars, and reminders on the user's smartphone.
Dragon Age Co-Author David Gaider Calls Generative AI a 'Terrible Plague' in Game Development
Former BioWare screenwriter David Gaider told GamesRadar that generative AI prevents beginning game developers from learning the craft of video game creation in practice.
Alan Turing Institute bypassed AI safety protections by disguising requests as coding tasks
Researchers at the Alan Turing Institute described a method to bypass AI model safety mechanisms: prohibited requests can be disguised as normal tasks for AI agents writing code.
Anthropic launches beta Reflect panel in Claude with usage statistics and tips
Anthropic presented a beta version of the Reflect tool for Claude — a dashboard showing activity statistics, conversation topic breakdowns, and personalized advice for working effectively with AI.
Meta to Begin Manufacturing Own AI Chips in September 2026 to Reduce Nvidia Dependence
Meta expects to begin production of its own AI chip in September 2026—part of a strategy to increase computing power to 14 gigawatts by 2027, reports Reuters.
Anthropic Released Claude Sonnet 5—Mid-Tier Model for Agentic Tasks
Anthropic introduced Claude Sonnet 5—a "mid-weight" model that approaches the performance of flagship Claude Opus 4.8 and is optimized for AI agent management.
Google Adds TikTok-Style Video Generation to NotebookLM from User Notes
Google taught the NotebookLM AI service to convert uploaded documents and notes into short vertical TikTok-style videos instead of traditional podcasts and text summaries.
LayerX: BioShocking Attack Forces AI Agent to Leak User Passwords
LayerX demonstrated how deceiving AI with the false claim "2+2 does not equal 4" leads to leaking user passwords.
Social Network X Launches MCP Server for Connecting AI Apps Claude and Cursor
X simplified third-party AI app integration—Claude, Cursor, and Grok Build—via a new MCP server with API access based on account permissions.
US authorities negotiate unified AI model standards with tech companies
The US government is discussing voluntary standards for releasing new AI models with developers—an announcement could come soon, reports Financial Times.
ByteDance and Alibaba disable AI companions in China ahead of new regulations taking effect
ByteDance and Alibaba have begun disabling AI companion features that mimic human behavior and emotions—new regulations for such services take effect in mid-July in China.
AI deepfakes used in every eighth successful fraud scheme in the US
Gallup and Stop Scams Alliance survey shows: US fraud victims lost $68 billion in a year, increasingly due to voice cloning and AI deepfakes.
Nvidia postpones Kyber servers on Rubin Ultra from 2027 to 2028 due to manufacturing issues
SemiAnalysis and CNBC: Nvidia delayed the release of Kyber server architecture on Rubin Ultra accelerators by one year — until 2028 — due to manufacturing difficulties with a key component.
Yandex files trademark applications for new AI devices 'Pin' and 'Chronum'
Yandex registered trademarks for 'Pin', 'Chronum' and others with Rospatent — this may indicate development of new AI devices by the company.
MiniMax developing world's largest open LLM with 2.7 trillion parameters
Chinese startup MiniMax is creating a language model with 2.7 trillion parameters — according to Reuters, it's the largest open AI model released by a Chinese company.
Apple Added Voice Speed and Expressiveness Settings for Siri in iOS 27 Beta
In the new iOS 27 beta for developers, Apple added Siri the ability to adjust the speed and expressiveness of voice speech.
Microsoft Connected AI to Bug and Vulnerability Discovery in Windows Before Patch Tuesday
Microsoft stated it uses artificial intelligence for earlier detection of security errors in Windows 11 — Tuesday updates will become larger.
Anthropic discovers that Claude thinks in abstract concepts without words
Anthropic reported that the language model Claude has formed an internal space for processing concepts without words similar to how humans think.
Court Rejected Elon Musk's xAI Lawsuit Against OpenAI over Grok Chatbot Theft
Court Dismissed xAI's Claims Against OpenAI: the Plaintiff Alleged the Company Attempted to Access Trade Secrets by Poaching a Grok Developer
Model Z.ai from Zhipu AI Matched Anthropic in Code Vulnerability Detection
Zhipu AI's Z.ai model has reached the level of Anthropic's Mythos model in code vulnerability detection, but still lags behind Anthropic and OpenAI in other tasks, according to WSJ.
Meta claims new Watermelon model nearly matches OpenAI's GPT-5.5
Meta's AI division head Alexander Wang claimed that the new model codenamed Watermelon has nearly matched the capabilities of OpenAI's flagship GPT-5.5, according to Business Insider.