Arpit Agarwal

DTWM Release 58: Synthesize Reality - Reason Physically - Execute Autonomously. The introduction of the NVIDIA Engineering Toolsets marks a seismic shift in the Digital Twine World Model (DTWM). In this release, we integrate the heavy-compute kernels of the NVIDIA ecosystem - Omniverse, Cosmos, and Isaac Sim- into our sovereign orchestration layer. As a proud NVIDIA for Startups Inception partner, Numorpho Cybernetic Systems (NUMO) is leveraging these world-class platforms to bridge the gap between "Actual State" sensing and "Autonomous State" execution. This release treats these tools as the Modeling, Simulation and Physical AI Substrate of our enterprise—the high-fidelity engine that powers MANTRA M5. With Release 58, we map the journey from universal 3D orchestration to sentient physical AI, grounded in the NVIDIA compute stack: - NVIDIA Omniverse (The Industrial Metaverse Kernel): Serving as our simulation substrate, Omniverse utilizes OpenUSD to unify disparate design and engineering pipelines into a single digital thread. Through ChatUSD and RunUSD, we enable real-time, physically accurate collaboration that breaks down data silos and accelerates complex 3D workflows. - NVIDIA Cosmos (The Physical AI Foundation): Moving beyond pixels to physical reasoning, Cosmos provides World Foundation Models (WFMs) that allow machines to understand and predict 3D interactions. By tokenizing the physical world, we generate the high-fidelity synthetic data required to train robots on "what happens next" in a complex industrial environment. - NVIDIA Isaac Sim (The Robotics Autonomy Kernel): Serving as our robot learning substrate, Isaac Sim closes the "sim-to-real" gap. By utilizing Omniverse Replicator and Isaac Lab, we develop and validate AI-driven robotics in zero-fail virtual proving grounds before deploying sovereign code to physical fleets. Why This Partnership Matters: The Inception of Autonomy Being an NVIDIA Inception partner allows Numorpho to embed at the leading edge of physical AI. By combining our Morphean AGI with NVIDIA’s hardware and software kernels, we realize the "Make, No Break" philosophy for the next generation of industrial machines. 🚀 Unlocking Scale: Accelerating the loop from design intent to autonomous robot deployment. 🛡️ Verifiable Trust: Providing evidence-based proof of simulation fidelity through path-traced renders and physics-aware video. 🧠Physical Reasoning: Enabling machines to navigate "Wicked Problems" by understanding the fundamental laws of motion and gravity. 🌐 Everything Connected: Realizing a fully harmonized industrial metaverse where virtual and physical realities are in constant, secure sync. We are no longer just simulating; we are creating the foundation for a resilient, sentient and sovereign industrial ecosystem. Through this collaboration, we are defining the future of physical intelligence. Make, No Break. #NVIDIA #Omniverse #NVIDIACosmos #IsaacSim #Inception #PhysicalAI #OpenUSD

6

🦾 Daily News of Robotics and Automation in India 🦾 📰 Machina Labs raises $124M to launch large-scale intelligent U.S. factory Machina Labs designed its intelligent factory to manufacture a wide range of complex metal structures without significant retooling. The post Machina Labs raises $124M to launch large-scale intelli... Read more: https://lnkd.in/dnH9f8wD #Machina #Labs #Intelligent #Factory #Raises

3

🧠✨ 𝗨𝗻𝘃𝗲𝗶𝗹𝗶𝗻𝗴 𝘁𝗵𝗲 𝗛𝗶𝗱𝗱𝗲𝗻 𝗗𝗶𝗺𝗲𝗻𝘀𝗶𝗼𝗻 𝗼𝗳 𝗔𝗜: “𝗕𝗲𝘆𝗼𝗻𝗱 𝗪𝗼𝗿𝗱𝘀: 𝗧𝗵𝗲 𝗪𝗼𝗿𝗹𝗱 𝗠𝗼𝗱𝗲𝗹𝘀 𝗥𝗲𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻” ✨🧠 🔥 𝙒𝙝𝙖𝙩’𝙨 𝙝𝙖𝙥𝙥𝙚𝙣𝙞𝙣𝙜? Top AI visionaries like Fei-Fei Li (Stanford University & World Labs) and Yann LeCun (Meta) are boldly challenging the dominance of language-based AI. Their next-gen models—dubbed “world models”—aim to build internal 3D simulations of reality rather than just predicting the next word. They integrate spatial reasoning, mental modeling, and commonsense prediction, enabling AI that can plan, adapt, and interact as humans do. 🚀 𝙒𝙝𝙮 𝙞𝙩 𝙢𝙖𝙩𝙩𝙚𝙧𝙨 𝘎𝘳𝘰𝘶𝘯𝘥𝘦𝘥 𝘪𝘯𝘵𝘦𝘭𝘭𝘪𝘨𝘦𝘯𝘤𝘦: Language alone lacks physical context. By simulating real-world environments, AI gains true situational awareness, rising beyond text patterns. 𝘚𝘱𝘢𝘯𝘯𝘪𝘯𝘨 𝘪𝘯𝘥𝘶𝘴𝘵𝘳𝘪𝘦𝘴: These models could reshape robotics, creative design, and even military planning, offering predictive capacity in real 3D worlds. 𝘕𝘦𝘹𝘵 𝘧𝘳𝘰𝘯𝘵𝘪𝘦𝘳 𝘸𝘪𝘵𝘩 𝘤𝘩𝘢𝘭𝘭𝘦𝘯𝘨𝘦𝘴: Language models train on vast text corpora. Spatial data is scarce—scientists like Li and LeCun stress the need for massive, rich spatial datasets and smarter data engineering pipelines. 💥 𝙆𝙚𝙮 𝙞𝙣𝙨𝙞𝙜𝙝𝙩𝙨 1. 𝘓𝘢𝘯𝘨𝘶𝘢𝘨𝘦 𝘥𝘰𝘦𝘴𝘯’𝘵 𝘦𝘹𝘪𝘴𝘵 𝘪𝘯 𝘯𝘢𝘵𝘶𝘳𝘦: Li emphasizes that AI must learn the structure of the world, not just its descriptions. 2. 𝘔𝘦𝘵𝘢’𝘴 𝘢𝘣𝘴𝘵𝘳𝘢𝘤𝘵 𝘷𝘪𝘥𝘦𝘰 𝘱𝘢𝘵𝘩: LeCun is training with video abstractions to simplify learning and boost predictive reasoning. 3. 𝘐𝘯𝘵𝘦𝘭𝘭𝘪𝘨𝘦𝘯𝘤𝘦 𝘢𝘴 𝘱𝘭𝘢𝘯𝘯𝘪𝘯𝘨: Both researchers envision AI that learns fast, adapts, and plans like humans, not just chatbots. 🌟 𝙄𝙢𝙥𝙖𝙘𝙩 & 𝙏𝙖𝙠𝙚𝙖𝙬𝙖𝙮 This paradigm shift challenges existing AI narratives. We're moving toward systems that: ✅ Understand their environment physically ✅ Reason about unseen outcomes ✅ Act with foresight and adaptability This is not just incremental—it’s redefining “intelligent machines.” 🙌 Do you think AI’s future lies in simulating physical worlds, not just imitating language? What's the most exciting domain for "world models"? Let’s dive in! 👇 🔗 𝙎𝙤𝙪𝙧𝙘𝙚𝙨 𝘛𝘰𝘱 𝘈𝘐 𝘳𝘦𝘴𝘦𝘢𝘳𝘤𝘩𝘦𝘳𝘴 𝘰𝘯 “𝘸𝘰𝘳𝘭𝘥 𝘮𝘰𝘥𝘦𝘭𝘴” ▶️ 𝘉𝘶𝘴𝘪𝘯𝘦𝘴𝘴 𝘐𝘯𝘴𝘪𝘥𝘦𝘳: https://lnkd.in/gDVSncbs 🔖 #AI #WorldModels #SpatialIntelligence #FutureOfAI #Robotics #TechInnovation #MachineLearning

2

All of a sudden, hardware startups have become hot with VC funds, especially where they intersect with AI, compute infrastructure, semiconductors, robotics, and advanced manufacturing. Software developer teams seem poised to be gradually replaced by AI agentic teams as the industry shifts toward open source -> with AI agents building apps while we sleep. Just a gentle example: Kelly Claude is now building 12+ products per day. Autonomously. Welcome to the future of software building. And by the way, you can find Kelly Claude on Moltbook. For those still living in the past, Moltbook is like a “Facebook” for AI agents.

15

Cosmos Policy just dropped for robotics. 🤖 Cutting-edge research is turning a world foundation model into a unified robot brain that can see, predict, and act—no extra action heads, no complicated control stack. Read our blog on Hugging Face ➡️ https://bit.ly/3OQAuzg Want to get hands-on with Cosmos (Reason, Predict, Policy, Cookbook)? Join the Cosmos Cookoff, sponsored by Nebius and Milestone Systems ➡️ https://bit.ly/4r16cHy

1

🎾 Tennis Analysis with AI: Object Detection for Ball Tracking 🎾 Check out the latest post in the MATLAB AI blog by Cory Hoi to see how to: ✅ Build a custom object detector from scratch and apply transfer learning to YOLOX, comparing trade-offs and measuring accuracy ✅ Go beyond detection by integrating object tracking (with a Kalman filter) to reduce false positives 🔗 https://spr.ly/6040A0Sjm #SportAnalytics #ObjectDetection #MATLAB

2