ENLatest AI NewsThe Mastermind Behind Teaching Robots to Cook Tomato and Egg: Genesis AI Open-Source Full-Stack Training PlatformGenesis AI company opens its World 1.0 platform, providing high-performance full-stack simulation infrastructure for robot and physical AI developers, helping agents efficiently train in a virtual training ground, reducing the learning barrier for skills such as robots cooking, and accelerating the implementation of embodied intelligence.1 days ago143.7KUnlock unlimited image generation for 9.9 yuan! Banana 2, Image2 free to create anytimeSkild Art offers a 9.9 yuan 7-day trial, unlocking all premium features for the price of a drink. Includes 150 credits, generating ~150 images, allowing AI visual creators to test the platform stress-free without committing to long-term membership…..1 days ago12.0KGoldman Sachs Analysis of the Second Half of the Market: Funds Are Leaving the Seven Tech Giants, Semiconductor and Other AI Upstream Sectors Are More PreferredGoldman Sachs derivatives expert Brian Garret notes investors are broadly underweight the Magnificent Seven, with funds rotating into AI-benefiting semiconductors. The Seven have underperformed, raising concerns over profitability of massive AI spending and institutional selling. Hedging costs in QQQ options are rising, outpacing those of small-caps…..1 days ago15.6KByteDance Douba AI Phone Project Changes: Hardware Leader Resigns, Project Enters Adjustment PeriodLin Xi, head of ByteDance’s Doubao AI phone hardware, has left, becoming the first core hardware lead to depart since the 2024 AI phone project launch, sparking speculation. Insiders deny the project is cancelled, stating it has entered a new adjustment phase…..1 days ago14.0KAI Research Enters the Autonomous Driving Era: Yang Zhilin Discusses the Third Stage of Large Model TrainingAI research paradigm is undergoing profound transformation. At the 2026 Zhongguancun Forum, Yang Zhilin, founder of Moonshot AI, noted that AI R&D has entered the third stage of ‘AI-led research.’ Starting 2026, past reliance on human-crafted rules and fine-tuning will be overturned, as AI increasingly leads its own development…..1 days ago15.7KNew Era of Scientific Discovery: AI Agent Elements Claw Successfully Completes Superconducting Material DevelopmentAlibaba DAMO Academy, Renmin University and UCAS released the first AI agent for superconducting material discovery, Elements Claw. It achieves independent research, offering an efficient automated paradigm that could replace the trial-and-error long-cycle method…..1 days ago13.6KAI Video Market Landscape Reimagined: Google Gemini Omni Flash Tops Blind Test RankingsGoogle DeepMind’s text-to-video model Gemini Omni Flash climbed to the top of the authoritative blind test ranking Video Arena with 1404 Elo points, demonstrating Google’s multimodal technology strength and confirming that the video generation field is rapidly evolving.1 days ago13.1KNVIDIA Releases Open-Source Dual-Tower AI Model, Text Generation Speed Increased by 2.42 Times, Image Quality Retained at 98.7%NVIDIA released the Nemotron-Labs-TwoTower discrete diffusion language model, solving the problem of slow token-by-token generation speed in large models. The weights have been open-sourced on Huggingface. The model reuses pre-trained weights of existing backbone networks without the need for retraining from scratch, significantly reducing costs. It adopts a 60B dual-tower architecture, with two 30B networks working in parallel. Each tower activates 3B parameters and is equipped with 128 routable expert modules to improve generation efficiency.1 days ago11.7KMicrosoft AI PC Dedicated System Project Aion Exposed, Completely Removes Traditional Start MenuMicrosoft’s internal AI operating system, Project Aion, designed for AI PCs, has been exposed. Built on Edge and lightweight web technologies, it abandons traditional Start menus and desktop icons, with the Copilot-invoked taskbar as the sole interaction hub. Focused on feeds, creation, and real-time info, it introduces ‘Spaces’ to auto-categorize webpages, fundamentally rethinking interaction logic…..1 days ago13.9KAdvertising Governance Embarks on a Visual Evolution: ByteDance Engine Launches Mamoda 2.5 Version to Achieve Comprehensive Video CoverageByteDance Engine launched its self-developed advertising governance large model, Mamoda 2.5, achieving an upgrade in content safety risk control technology. Starting from version 1.0, which could only identify basic prohibited text, the model has continuously evolved, expanding its capabilities. It now provides stronger support for efficiently and accurately identifying and managing prohibited content in the digital advertising ecosystem.1 days ago11.8KEnd of Subscription Benefits? Anthropic Tightens Its Strongest AI Model, Claude Fable5, Will Be Charged by UsageAnthropic will remove access to its top model Claude Fable5 from subscriptions on July 7, 2026, ex