AI / Software

Ever crashed a pricey robot prototype into a wall during testing? Or waited days for one trial-and-error run to wrap up? If you’re into robots—whether for fun, factories, or the future—those headaches are real. But what if you could simulate thousands of bots at once, tweaking their brains in hours instead of weeks? Enter Brax, Google’s free open-source physics engine that’s revolutionizing how we “teach” robots to walk, grab, and navigate. Picture this: No more dusty warehouses or risky resets. Brax runs on your GPU or TPU (fancy hardware for speedy math), letting you parallel-test robot designs like a digital assembly line. Built on JAX (a supercharged Python toolkit), it […] Read More

In robotics, physics engines and simulators serve different but complementary roles. A physics engine handles the mathematics of simulation such as computing forces, collisions, and the dynamics that make objects move and interact realistically. A simulator builds on these engines to provide a complete virtual testing environment, by adding sensors, 3D environments, visualization tools, and interfaces for communication with frameworks like ROS 2. In short, the physics engine provides the underlying math and realism, while the simulator delivers the interactive and visual layer needed for full robotics development and testing. ROS 2 (Robot Operating System 2) is an open-source middleware framework that enables communication between robot components such as sensors, […] Read More

NVIDIA and OpenAI have just entered into a partnership, to build 10 GW of AI data centers with NVIDIA GPUs. As each GW is deployed, NVDIA will invest gradually a total of $100 Billion, starting second half of 2026. The data centers will train OpenAIs next generation AI infrastructure. The detail of the strategic partnership will be finalized in the coming weeks. See press release at: https://nvidianews.nvidia.com/news/openai-and-nvidia-announce-strategic-partnership-to-deploy-10gw-of-nvidia-systems —————————————————— Terms of Use: Copying or republishing our content is not allowed without written permission. We maintain dated records of our content. The content on this website may be incorrect or incomplete, and the user assumes all liability and risk when using it. […] Read More

A neural network is made of layers, and layers are made of nodes (also called neurons or units). It is these nodes that process input data. First, each neuron that receives input. If this is the first layer, the input is raw data. If it is a higher order layer, the input comes from the layer below. The node applies weights to the inputs. Then adds a bias term. Then applies a nonlinear activation function (such as ReLU or sigmoid) and produces its outputs the resulting value to the next layer, which receives it as input. The keyword here is “nonlinear” because in real life things are rarely linear. Everything […] Read More

When designing robots, there is the need for a system in standardized format, to describe physical and kinematic properties of a robot. This is achieved by a robot description format. There are several commonly used formats including but not limited to URDF, SDF, MJCF. URDF (Unified Robot Description Format) is an XML based format which describes physical and kinematic properties of a robot for use in robotics software, and mainly in ROS (Robot Operating System), which is open source and the most widely used robot design software today. Let’s take a closer look at what URDF specifically does: Therefore it is a great tool in visualization of the robot and […] Read More