Large Language Model– tag –
-
Thinking Machines Lab Unveils Inkling, a New Open-Weight 975B MoE LLM, Achieving Superior Benchmark Performance Over GLM-5.2
Sebastian Raschka Germany Overview Thinking Machines Lab has announced Inkling, a new open-weight Large Language Model (LLM) with approximately one trillion parameters. This 975B Mixture-of-Experts (MoE) model demonstrates solid performa... -
Lyzr Secures $100 Million Series B Funding Round, Valuing Company at $500 Million, by Leveraging Proprietary AI Agent ‘Agent Sam’
TNW India Overview Lyzr, a startup developing AI agents for large enterprises, successfully closed a $100 million Series B funding round by leveraging its proprietary AI agent, 'Agent Sam.' This round reportedly valued Lyzr at approximat... -
Effective AI Governance Framework Establishes Permissible Use Policies and Human Accountability Through Legal Design Patterns
Adaptive Security Unknown Overview An effective AI governance framework, starting with an AI governance policy, establishes permissible use of generative software, data security protocols, and human accountability requirements. This fram... -
Large Language Model Agents Show Up to 14.4% Performance Degradation in Dynamically Evolving MCP Server Environments: MCPEvol-Bench Reveals
arXiv Unknown Overview The new paper MCPEvol-Bench addresses limitations of existing benchmarks by evaluating LLM agents' tool-use capabilities in dynamically evolving MCP (Minecraft-like Collaborative Platform) server environments. Benc... -
Robotics AI Development Faces High GPU Costs for Physical Simulation and Synthetic Trajectory Generation, Exceeding LLM Computational Loads
axecompute.com Unknown Overview Physical AI computing for robotics development faces higher computational costs than LLM development, with a single node requiring hours to days just for a robot to navigate uneven terrain. World foundatio... -
Enterprise Generative AI Spending Skyrockets to $37 Billion in 2025, Capturing 50% of Global AI Startup Funding
Fundraise Insider USA Overview In 2025, enterprise generative AI spending surged from $11.5 billion to $37 billion, accounting for approximately 50% of global AI startup funding. This dramatic increase reflects a strong investor consensu... -
AWS Startups Advocates World Models and Multimodal AI Stack to Redefine Robotics Rules, Enabling Risk-Free Simulation-Based Learning
AWS Startups USA Overview AWS Startups proposes that leveraging world models in robotics can fundamentally redefine the field by enabling fault mode exploration, edge case iteration, and capability building in simulation without physical... -
LingBot-VLA 2.0 Robot Foundation Model Achieves Cross-Embodiment Transfer with 60,000 Hours of Trajectory Data and Diverse Robot Configurations
RoboCloud Hub Unknown Overview LingBot-VLA 2.0 marks a new frontier in robot foundation models, trained on approximately 60,000 hours of robot trajectory and first-person human video data. This model extends its action space across 20 di... -
Frontiers Highlights Foundation Models’ Value in Designing Soft, Wearable, and Bio-Inspired Robots by Efficiently Adapting Reusable Representations
Frontiers Switzerland Overview Research in Frontiers emphasizes the critical value of reusable representations generated by foundation models for designing novel embodied systems like soft, wearable, and bio-inspired robots with signific... -
Engineering the AI Era: Why Data Centers Need 5.5-9 Meter Floor Heights for GPU Power and Thermal Management
Kim Kang-han South Korea Overview AI data centers demand significantly greater floor heights—ranging from 5.5 to 9 meters, 1.5 times traditional facilities—and specialized structural designs. This elevated construction is essential to ac...
