A single destination for timely, editor-curated robotics news from around the world.
Google DeepMind has showcased a novel robotics training method where its Gemini 4 Argon AI model learns tasks by observing videos and then applies this knowledge to Apptronik’s Apollo humanoid robot. This demonstration highlights the capabilities of Gemini 4 Argon, which integrates visual-semantic r...
NVIDIA has launched Flora, a suite of models designed to convert robot and human videos into labeled training data for robotics AI. This innovative system identifies specific actions, describes them linguistically, and evaluates task performance, facilitating the transformation of extensive video co...
Unitree has launched a captivating video featuring its H2-X humanoid robot, showcasing its impressive agility and dynamic movements. The video, titled 'H2-X Kung Fu Legend Guardian,' includes scenes of the robot performing stunts such as flipping off a car and executing martial arts maneuvers from a...
A new video on Robotik TV addresses the ongoing issue of women's underrepresentation in leadership positions within the tech industry. This topic remains relevant as discussions around gender equality and diversity in technology continue to evolve. The significance of this issue lies in the potenti...
The global video gaming ecosystem is experiencing significant growth in the fourth quarter, driven by robust enterprise demand for artificial intelligence (AI) tools. This surge is further supported by the expansion of cloud infrastructure and innovative direct-to-consumer mobile monetization strate...
This week's Video Friday showcases a variety of intriguing robotics videos, curated by IEEE Spectrum. Notable highlights include a humanoid robot's ability to perform complex manipulation tasks, as demonstrated by Workhorse using a Unitree G1. Additionally, MIT's SANDO navigation system is introduce...
TwelveLabs Inc. has unveiled its Pegasus 1.6 model, designed to enhance video understanding for physical AI applications. This model addresses the challenge of transforming complex video footage into actionable insights, enabling robotics and AI teams to train on real human experiences rather than s...
A recent article in Science Robotics discusses the intriguing scenario of how extraterrestrial beings might interpret videos of human-made robots. The study suggests that if aliens were to analyze these videos, they could form misconceptions about human capabilities and intentions based on the robot...
Insta360 has launched a new feature called Spatial Capture, available on the X6, X5, and X4 Air cameras. This feature allows users to transform 360-degree videos into explorable 3D scenes using 3D Gaussian Splatting technology. Unlike traditional methods that require multiple angles and complex soft...
Light Origins has launched the Light-O1 embodied base model, which is trained on human movements sourced from online videos. This innovative model aims to enhance the capabilities of robotic systems by mimicking human actions, thereby improving interaction and functionality in various applications....
DJI has released a significant update for its Osmo 360 II camera, enhancing low-light video capabilities and improving panoramic photo stitching. The firmware version 01.02.03.10, along with DJI Mimo 2.12.1 for iOS and Android, introduces features that make capturing and sharing 360-degree content m...
This week's Video Friday from IEEE Spectrum showcases a variety of robotics innovations, including a unique quadrotor design that combines two birotors. The significance of this development lies in its potential to enhance aerial mobility and efficiency in various applications. Additionally, the s...
XPENG Robotics has unveiled XPACE, a novel system designed to enhance the learning capabilities of its IRON humanoid robot. By utilizing 5,000 hours of human activity videos and simulated mistakes, XPACE aims to improve robot performance without the need for direct human demonstrations in every scen...
Rhoda AI has reported enhancements in industrial manipulation capabilities through scaled web-video pretraining, as detailed in a study released on September 10. The research tested various model sizes, achieving completion rates of 3.7%, 65.0%, 75.3%, and 84.7% in under 100 seconds for tasks like u...
Skild AI has launched the S1, a groundbreaking robotics foundation model that allows robots to learn manipulation tasks from just one video demonstration. This innovative model eliminates the need for task-specific fine-tuning or post-training, streamlining the learning process for robotic systems....
In the latest edition of Video Friday, a humanoid robot showcases its ability to traverse monkey bars, demonstrating advanced agility and perception of complex 3D structures. This task requires the robot to jump onto the structure, navigate through sparse bar interactions, and land safely, highlight...
Amazon has announced new features to enhance shopping while watching Prime Video. These updates will allow customers to discover products across thousands of titles through the existing X-Ray experience and the new 'Shop the Scene' feature using Amazon's Lens technology. This initiative is signific...
AirData has enhanced its Live Streaming platform by adding recording and rewind capabilities, allowing authorized viewers to review the previous 10 minutes of an active stream. This feature addresses the challenge of remote drone operations where key events may occur while the necessary personnel ar...
DeepSeek has announced a significant expansion of its engineering team, adding 150 new engineers to enhance its capabilities in AI and robotics. This move reflects the growing demand for advanced AI solutions in various sectors, particularly in the development of large models that go beyond traditio...
Markov Robotics has introduced a novel approach to dexterous manipulation using a video model as the policy for its LTX-2.5 robot. This method allows the robot to understand object interactions in real-time, enhancing its ability to perform tasks in unfamiliar environments. By leveraging a world mod...
This week's Video Friday showcases a variety of robotics advancements, including Agility Robotics' impressive couch drag demonstration. The introduction of ReST-RL by SteadyTray aims to address the challenges of payload stabilization in dynamic bipedal locomotion, successfully implemented on the Uni...
GoPro has released a significant firmware update for its MAX2 camera, enhancing its video recording capabilities. The update introduces a Maximum Bitrate option that allows recording at up to 200Mbps, improving image quality by preserving more data and reducing compression artifacts. This feature is...
Figure AI has officially launched Index, a platform designed to crowdsource real-world human video to train its Helix AI architecture. This initiative, part of Project Go-Big, aims to overcome traditional data bottlenecks in robotics by gathering physical interaction data from smartphone users world...
GoPro has released a significant firmware update for its MAX2 camera, enhancing its video recording capabilities. The update introduces a Maximum Bitrate option that allows recording at up to 200Mbps, improving image quality by preserving more data and reducing compression artifacts. This feature is...
Alibaba Cloud has launched Wan3.0, a new video-generation model capable of producing clips up to 30 seconds long. This model accepts various document formats, including DOC, XLS, PPT, PDF, and Markdown, as input. Wan3.0 had been in public beta since early August, allowing users to test its capabilit...
DJI has introduced the #AllForPassion Challenge, inviting Avata 360 users to showcase their sports and adventure videos for a chance to win up to $10,000. The contest features a grand prize of $10,000, two runner-up prizes of $5,000 each, and five third-place awards of $1,000. This initiative aims t...
Recent advancements in robotics have demonstrated that robots can learn new skills from a single video in just 29 seconds. This capability is crucial for enabling robots to operate effectively in dynamic environments, where adaptability is essential. Traditional robotic systems often struggle to per...
LTX has introduced LTX-2.5, an advanced version of its open-weights world model, enhancing capabilities for video generation and physical AI. This model boasts improvements in visual quality, prompt understanding, and generation speed, allowing developers to customize it on their hardware. With over...
LTX, a company spun out of Lightricks, has launched LTX-2.5, an advanced open-weights video model that can generate a 10-second video from an image in just 6.8 seconds using NVIDIA superchips. This model is now integrated into ComfyUI and available on Hugging Face, supporting organizations with unde...
Dyna Robotics has introduced its DYNA-2 World-Action Model, trained on over 1 million hours of human video. This innovative approach aims to address the challenges of teaching robots to perform physical tasks by leveraging human egocentric video instead of traditional robot action data. The signifi...
Dyna Robotics has introduced Dyna-2, a world-action model (WAM) pre-trained on one million hours of human video data. This development provides significant evidence for a human-to-robot transfer scaling law, demonstrating that increased exposure to human video enhances a robot's ability to perform t...
SpaceX's Starship completed its 13th launch on July 24, successfully landing its upper stage in the ocean. The flight achieved significant milestones, including deploying 20 next-gen Starlink satellites and reigniting a Raptor engine in space before a soft splashdown off Western Australia. This suc...
Mimic Robotics has launched FLUX-mimic, an advanced Video-Action Model developed with Black Forest Labs, designed for industrial automation. This model allows robots to learn complex manipulation tasks in real-world settings, significantly reducing the time required for training and deployment. The...
Mimic Robotics has launched FLUX-mimic, a cutting-edge Video-Action Model developed with Black Forest Labs, enabling robots to learn intricate industrial tasks from video demonstrations. This innovative system significantly reduces the amount of training data required, allowing for faster and more e...
Recent advancements in DJI drones allow users to shoot vertical videos without cropping, catering to platforms like Instagram Reels and TikTok. Traditional drone cameras typically capture wide landscape footage, which is not ideal for vertical-first social media. However, select DJI models can now c...
Black Forest Labs (BFL) has launched FLUX 3, a multimodal model capable of generating images and 20-second audio/video clips from a single prompt. This model represents a significant expansion of BFL's FLUX family, emphasizing a unified approach to creative generation and robotic vision. The introd...
This week's Video Friday highlights significant advancements in robotics, including a systematic evaluation of humanoid technology for laparoscopic surgical tasks. Researchers developed a humanoid-based teleoperation framework and assessed its performance through various studies, revealing both pote...
A quadruped robot has successfully solved a giant Rubik’s Cube autonomously using only its feet, showcasing advanced robotic capabilities. This achievement highlights the potential for robots to perform complex tasks that require precise coordination and problem-solving skills, which are traditiona...
China's Unitree Robotics has introduced the H2-X, a humanoid robot aimed at emergency response, showcased in a dramatic promotional video. The video features the robot disarming an attacker, lifting a car to rescue a trapped individual, and pulling a pedestrian from danger, emphasizing its potential...
Researchers at the Korea Advanced Institute of Science & Technology (KAIST) have unveiled an updated version of their humanoid robot, Humanoid v0.9, showcasing its ability to push objects weighing up to 77.2 pounds and lift loads of 44.1 pounds. This demonstration, released on October 19, emphasizes...
DJI has released a firmware update, v01.04.12.16, for its Osmo Pocket 4 camera, just in time for the holiday filming season. This update introduces a convenient tracking shortcut, improved color-conversion tools, and automatic media transfers to compatible devices, enhancing usability for creators c...
UBTECH has announced that its Thinker-WM has achieved top scores in three embodied world model benchmarks. The integration of depth data significantly improved single-task success rates on the RoboTwin platform, increasing performance by 15 percentage points during ablation tests. This advancement...
This week's Video Friday showcases a variety of robotics innovations, including the Gemini 4 Argon's ability to simulate human actions for training physical AI. This technology significantly reduces the time needed for modeling, transforming video content into a training ground for robots. Addition...
Boston Dynamics has released a video showcasing its prototype operations team responsible for building the Atlas humanoid robot. The video highlights the integration of precision machining, assembly, and repair processes that enable the team to construct a robot in approximately one day, with ambiti...
Nature Will has introduced Natura-1, a 50-billion-parameter model that has been pretrained on one million hours of first-person human video. This advanced model has been specifically adapted with limited robot data to effectively control 20 degrees of freedom (DoF) in robotic hands. The significanc...
Google DeepMind has showcased its Gemini 4 Argon technology, which enables the Apptronik Apollo 2 humanoid robot to learn and execute cornhole throws. In a demonstration on October 13, Apollo 2 successfully made seven consecutive shots after learning from a human video through a video-to-simulation...
Researchers at New York University have introduced WorMa, a worm-like robot capable of navigating challenging coastal terrains. This innovative robot can traverse rocks, slopes, and water by shifting its weight, utilizing water as a movable ballast to adapt its movement without needing mechanical re...
BeingBeyond's Being-M0.7 has demonstrated its capability to predict future visual latents and whole-body motion prior to executing actions. This advanced humanoid robot achieved a notable performance in the SIMPLE simulation, scoring 71.1%, and matched the strongest baseline using a Unitree G1 robot...
Robotera's Video Prediction Policy 2 (VPP2) has secured the top position on RoboDojo's simulation leaderboard, announced on October 9. This model effectively links scene change predictions with corresponding robot actions, outperforming PhysicalRSI in both average score and task success rate. VPP2 a...
Researchers at North Carolina State University have created a soft underwater robot inspired by feather stars, capable of three-dimensional movement using only two actuators. This innovative design significantly reduces mechanical complexity, as traditional systems typically require six actuators fo...
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
