Every story we have kept, newest first.
Looking for the daily editions? → Past editions
Page 1037 of 1375
ELITE bridges the gap between VLMs' static knowledge and embodied tasks through experiential learning from interactions. It employs self-reflective knowledge construction to build reusable strategies and intent-aware retrieval for task adaptation.
ArXiv AI · 189d ago
DUPLEX is a neuro-symbolic architecture that confines LLMs to information extraction for PDDL-based planning in robotics. A fast system uses lightweight LLMs for quick NL-to-PDDL mapping; a slow system activates on failures for iterative repairs.
ArXiv AI · 189d ago
DeepXube is a free open-source Python package automating pathfinding via machine-learned heuristic functions and tailored search algorithms. It combines deep reinforcement learning, heuristic search, and answer-set programming for goals.
ArXiv AI · 189d ago
AnalogAgent is a training-free framework integrating LLM-based multi-agent systems with self-evolving memory for analog circuit design automation. It coordinates a Code Generator, Design Optimizer, and Knowledge Curator to adapt from feedback without expert input.
ArXiv AI · 189d ago
AI-Supervisor is a multi-agent framework for end-to-end AI research supervision, maintaining a persistent Research World Model as a Knowledge Graph for shared understanding. It features structured gap discovery, self-correcting loops to analyze failures and biases, and self-improving loops via cross-domain searches.
ArXiv AI · 189d ago
As AI handles homework, Onion Academy's AI system shifts evaluation to capturing students' thinking processes via multi-AI agents. It records behaviors across learning stages to quantify abilities like logic and persistence.
虎嗅 · 189d ago
Lin Junyang post-Alibaba reveals Qwen3's failed merge of thinking and instruct modes due to data mismatches. Advocates shift from reasoning chains to 'agentic thinking' for action-oriented, environment-interactive AI.
虎嗅 · 189d ago
Young creators on Zhihu develop AI to make friends and organize social events for users. AI handles 'group building' to facilitate real human connections.
量子位 · 189d ago
Fujian Suijun Intelligent Tech Co. newly established with 100M RMB capital, legal rep Tang Feng.
36氪 · 189d ago
Alibaba Cloud's Bailian platform debuts third-party video generation model PixVerse V5.6. Supports text-to-video, image-text-to-video, first-last frame, and reference generation modes.
量子位 · 189d ago
Direava, a startup from Keio University School of Medicine, announced the development of Surgical VLM, an AI that supports surgical operations. It analyzes intraoperative images to provide advice.
ITmedia AI+ (日本) · 189d ago
Google announced a new feature to simplify switching from ChatGPT, Claude, and other AI services to Gemini. It allows importing user settings via 'memory' and past chat history.
ITmedia AI+ (日本) · 189d ago
China’s top chip foundry SMIC has unveiled an action plan to optimize existing businesses and pursue new growth opportunities by 2026. Released with its 2025 annual report, the plan aims to solidify SMIC's role as the backbone of China's semiconductor self-sufficiency drive.
SCMP Technology · 189d ago
Tencent Cloud unveiled its first Agent product panorama at the Shanghai Summit, upgrading MaaS to TokenHub and launching enterprise Agent Infra governance. It spans Infra, models/ecosystem, scenarios/applications, and security dimensions for personal, office, and R&D efficiency.
36氪 · 189d ago
Local governments compete fiercely in cultural tourism for traffic. The article questions if AI can take over roles of tourism directors.
钛媒体 · 189d ago
Embodied AI's trillion-dollar market enters elimination phase. Data and real-world scenarios are pivotal for breakthroughs.
钛媒体 · 189d ago
Zhihu reported its first full-year Non-GAAP profit of 37.9M CNY in 2025 despite 23.6% revenue drop to 2.75B CNY. Paid reading became top revenue at 51.8% in Q4 via IP derivatives like AI short dramas.
虎嗅 · 189d ago
Tunee from Qfun Tech launches MV Agent with OpenClaw integration (exclusive). Unlike generate-and-done tools, it offers fully editable workflows for creative ideas, characters, scripts, and storyboards.
钛媒体 · 189d ago
Opinion piece depicts employee exit where work skills are archived as AI-callable 'ability packages' in team knowledge base. Highlights shift to knowledge precipitation, SOPs, and digital employees replacing tacit expertise.
虎嗅 · 189d ago
Yarbo, a modular yard robot maker, raised nearly 200M RMB from investors like Songhe Capital. The new lightweight M series launched on Kickstarter, surpassing $1.5M in 6 hours.
36氪 · 189d ago
Xiaomi Group hit record annual revenue but Q4 showed revenue up without profit growth. Automotive segment succeeded while smartphones faltered.
钛媒体 · 189d ago
China is unhappy with its top AI experts seeking opportunities abroad, whether to attend conferences or join foreign companies. Authorities are urging scientists to avoid a major international conference.
The Register - AI/ML · 189d ago
Zhixing Robotics completed B+/B++ rounds totaling nearly 100M RMB for dexterous hands and grippers. Launched LingSi (4-finger, 5kg payload) for industrial tasks and ShuangQiao (5-finger rope-driven) for unstructured scenes.
36氪 · 189d ago
Narwal co-founder Wu Yihao (Hower) confirmed departure to start a secretive new venture while retaining chief strategy advisor role. He led software, supply chain, ops, and global marketing during Narwal's growth to 10B+ RMB valuation.
36氪 · 189d ago
Lin Junyang, former Alibaba Qwen training lead, argues AI is shifting from reasoning (o1-style) to agentic thinking via actions and environment interaction. He critiques Qwen's reasoning efforts and praises Anthropic's task-adaptive approach.
虎嗅 · 189d ago
China's A-share major indices opened lower across the board, with Shanghai Composite down 0.95%, Shenzhen Component 1.34%, and ChiNext 1.1%. Storage memory and optical modules led declines, with Changguang Huaxin and Biwin Storage dropping over 6%.
36氪 · 189d ago
Google has rolled out the 'Search Live' feature in Japan and all regions with AI mode enabled. Powered by Gemini 3.1 Flash Live, it enables real-time voice and camera interactions with improved audio nuance understanding and emotion-adapted responses.
ITmedia AI+ (日本) · 189d ago
Hang Seng Index opened down 0.35%, Hang Seng Tech Index 0.44%. Semiconductors and tech led losses, with GigaDevice over 4% drop, SMIC and Hua Hong near 2%.
36氪 · 189d ago
GitHub is revising its Copilot privacy policy, effective April 24, to default to collecting user interaction data for AI training. This covers code snippets and context from real-time use, including in private repositories when Copilot is enabled.
钛媒体 · 189d ago