griffin360’s Post

Everything you need to know about Figure 02. F.02 was a ground-up redesign to achieve: - 6x Cameras - 50%+ Battery - Onboard VLM - 3x CPU / GPU - 4th Gen Hands - Integrated wiring - Exoskeleton structure - Speech-to-speech reasoning Technical write-up Speech-to-speech reasoning Figure 02 is capable of speech-to-speech conversations w/ humans Onboard mics + speakers connected to custom AI models trained in partnership with OpenAI The default UI to our robot will be speech Onboard Battery Custom in-house designed 2.25 KWh battery pack in the robot torso Upgraded battery delivers 50%+ more energy than Figure 01 to maximize robot runtime We hope F.02 will be able to achieve upwards of ~20 hours of useful work per day Integrated Wiring This might sound easy but wiring is tough! Integrated cabling for power & compute allows for: > higher reliability > concealed wires > tighter packaging In addition to better wiring, we designed custom wire terminals & connectors to increase robot reliability Cameras Figure 02 perceives and understands the physical world through its AI-driven vision system There are 6 onboard RGB cameras located in the head, front torso, and back torso (robot has super human vision) [08/07, 13:37] Kindle🔭🔬🧑🏽💻: Exoskeleton Structure In order to provide structural stiffness and protect against crash loads Figure 02 was designed as an exoskeleton structure (similar to aircraft where the outer skin bears the load) This was quite a radical design shift from Figure 1.0 4th Generation Hands The latest human-scale hands are equipped with 16 degrees of freedom and human-equivalent strength which enables a wide range of human-like tasks The entire hand was designed in-house: mechanical, electrical, controls, and sensors Onboard CPU & GPU Figure 02 has 3x the computation & AI inference available on-board compared to Figure 01 This enables real-world AI tasks to be performed fully autonomously

Last week, I spoke at NVIDIA GTC and exhibited our Industrial AI applications at the Trifork booth. Danish ITWatch asked me about my observations, and besides the vast interest we received for our AI solutions, I was surprised at how large the GPU-accelerated edge computing space had become in just a few years. By far, it was the most condensed segment at the conference. Many vendors with innovative products. This tells me that intelligent technology is needed in the field, on the manufacturing floor, for trains, cars, and especially robots. I'm happy this space is taking off, as the technology significantly impacts all industrial settings and improves the work lives of all the people we rely so much on. The article is in the comments section, in Danish, and behind a paywall.

Together with NVIDIA, Foxconn is building the largest supercomputer in Taiwan. This will help AI research, healthcare, smart factories, robotics and smart city innovations. Read the Press Release:

🚀 Powerful Edge AI Meets Unmatched Durability—Introducing Premio’s JCO Series! Looking for AI-driven edge computing solutions that are both rugged and scalable? Meet Premio’s JCO Family of NVIDIA Jetson-Powered AI Edge Computers—designed to bring high-performance AI to the harshest industrial environments. Featuring three scalable models, each JCO system is equipped with the power and adaptability to support evolving applications like AGV & AMR, security & surveillance, and machine vision—all crucial in today’s intelligent edge deployments. ✅JCO-6000: Highest performance for intensive AI tasks in demanding conditions. ✅JCO-3000: Mid-range option that balances performance with versatility. ✅JCO-1000: Compact and entry-level, optimized for efficient processing on the go. The JCO Series is a key part of Premio's Rugged Edge AI Matrix—offering customizability, I/O-rich configurations, and out-of-band management for streamlined remote monitoring. 👉Curious to learn more about the possibilities of NVIDIA Jetson at the edge? Dive deeper here: https://lnkd.in/gTQBxW5P #EdgeAI #RuggedEdgeComputing #NVIDIAJetson #IndustrialAutomation #Premio #AI

Gideon 's collaboration with NVIDIA will help address unsolved complex #robotics and #automation challenges in the #supplychain, by combining Gideon’s AI-powered AMR technology and NVIDIA’s accelerated computing. Excited to hear more on this from Josip Cesic at Manifest: The Future of Supply Chain & Logistics https://lnkd.in/ehbWZcf6

Small but powerful: freshly transforming AI at the edge with the AIEdge-X80 The AIEdge-X®80 is an edge AI fanless box PC that couples low power and high performance in a compact form. Powered by the newest NVIDIA® Jetson Orin™ NX, it delivers exceptional processing speeds of 70 to 100 TOPS with a power consumption of only 10W to 25W. It further enhances AI inference capabilities through robust multi-connectivity to other sensors and devices in the field. Additionally, you can build end-to-end, accelerated AI applications with the included NVIDIA® JetPack™ SDK, making the predictive responding of edge AI in security and retail applications, such as sorting machines, parcel tracking, access control, and public safety, a tangible reality. For product details: https://lnkd.in/gFs-BSTk #retails #edgeai

Nvidia has released a new $249 computer kit called the Jetson Orin Nano Super. It's designed for people who want to build AI-powered devices, like robots or smart assistants. The new kit is an upgraded version of Nvidia's previous Jetson Orin Nano kit. It has 70% more power for processing AI tasks, and 50% faster memory. https://lnkd.in/dhtGx-_c

Managing your edge AI devices just got a whole lot easier! Edge Impulse empowers developers to create, optimize, and deploy machine learning models for edge devices, enabling efficient and scalable AI applications. By integrating Edge Impulse's state-of-the-art AI model capabilities with Allxon's dependable Over-the-Air (#OTA) deployment, you can significantly enhance both the performance and manageability of your Nvidia Jetson-powered edge devices. With Edge Impulse OTA model updates on Nvidia Jetson devices, Allxon OTA deployment is your go-to solution. Key features include 🗝️: 💙 Seamless, efficient, and secure portal 🤍 Supports all file types (#BSPs, applications, videos, scripts, security patches, etc.) 💙 Keeps devices up-to-date with the latest features and improvements 🤍 Ensures peak performance at all times Check out this guide to learn how to deploy and manage Edge Impulse models on NVIDIA Jetson devices using Allxon's OTA capabilities here: https://hubs.li/Q02Fc-Rv0 Learn more Allxon OTA Updates solutions here: https://hubs.li/Q02Fc_9X0

🚗💡 AI Innovation: IBM’s NorthPole Chip and the Future of Autonomous Driving? 🚀 IBM’s new NorthPole chip is making headlines for its impressive advances in speed and energy efficiency, and it might just be a great fit for autonomous driving systems in the future. Here is what makes it exciting: ⚡ 46.9x faster latency than leading GPUs, delivering sub-1 millisecond per token performance. 🔋 72.7x more energy-efficient, which could help address the growing need for greener AI solutions. 📊 Capable of processing 28,356 tokens per second on a 3-billion-parameter model in a 16-chip setup, showing its ability to handle large AI workloads efficiently. With its combination of low latency and high efficiency, this chip could potentially bring new possibilities to autonomous driving by enabling quicker, smarter decisions while reducing energy consumption. While it is still early days, this innovation opens up exciting potential for the future of AI in the automotive world! 🌍 #AI #AutonomousDriving #TechInnovation #AIHardware #EnergyEfficiency #FutureOfMobility #MachineLearning #AIInnovation

As I finish up preparations with Intel Corporation's Paul Schneider and Ricky Watts, I'd love to hear from you! What questions would you like to have answered related to Digital Twins, Simulation, Scheduling, Material Handling, AI or Factory Efficiency? I'll do my best to incorporate your questions into what will be shared on Wednesday. See you then! Register here: https://lnkd.in/g4sx83FY #camLine #Intel #DigitalTwin #Simulation #Scheduling #MaterialHandling #AI #OEE #Optimization #Romaric #AMHS #Robotics #Software #SmartSoftware