Technology areas
Data and simulation
Demonstration, teleoperation for data collection, motion retargeting, robot datasets, data quality, physics simulation, synthetic data, digital twins, domain randomization, and sim-to-real transfer. The focus is data, environments, and tools for training and validation. Simulation used only to validate a new learning algorithm does not automatically qualify.
136 selected · 91 in the last 30 days
线性资本 ·
Interview with Tashi Intelligence's Chen Yilun: No VLA, no simulation—an embodied AI company's contrarian bets
Linear Capital reposted LatePost's interview with Tashi Intelligence founder Chen Yilun, who laid out a technical path that differs from the mainstream: he rejects the VLA architecture that grows an action head on top of a VLM, and instead develops its own AWE (AI World Engine), aiming to represent physical quantities such as time, space, and force; it has also abandoned teleoperation data collection in favor of a self-developed collection device combining a wearable glove and a first-person-view camera.
Physical Intelligence ·
Physical Intelligence: VLA models show human-to-robot transfer as robot data scale grows
Physical Intelligence research found that in generalization scenarios where only human demonstrations appear, fine-tuning π0.5 with human first-person videos mixed into robot data alone improved policy performance by about 2x, without using any explicit transfer learning mechanism.
SIGRobotics · X ·
Earth Rover Mini+ low-cost open-source autonomous mobile robot released, with 7TB of real teleoperation data
Earth Rover Mini+ has been released. It is an open-source autonomous mobile robot priced under $300, has actually driven in multiple cities, comes with over 7TB of real teleoperation data, and supports LeRobot integration.
Pieter Abbeel ·
Amazon FAR open-sources Holosoma, a full-stack sim-to-real codebase for humanoid robots
Amazon FAR has released Holosoma, a full-stack solution, and open-sourced a complete codebase covering multiple simulation backends, training, motion retargeting, and real-world inference. Pieter Abbeel shared and added that the infrastructure was built for humanoid robots and can also be readily adapted into a quadruped version (the quadruped part is already included), and that it contains a great deal of infrastructure experience and convenience tools the team relies on daily.
LimX Dynamics ·
LimX Dynamics showcases LimX Oli's human-like walking and TRON 1's extreme terrain capabilities at WRC
Robot Lecture Hall reported on-site from the LimX Dynamics booth at the 2025 World Robot Conference. The full-size humanoid robot LimX Oli, 165 cm tall with 31 active degrees of freedom, made its first WRC appearance and went on sale, demonstrating outdoor live dancing and switching among multiple gaits such as a light quick step, model walk, and goose step, powered by a whole-body motion control algorithm combining traditional control, reinforcement learning, and imitation learning, with prices starting at 158,000 yuan.
Figure AI ·
Figure AI uses reinforcement learning to give Figure 02 humanoid robot a natural, human-like walk
Figure AI released an end-to-end reinforcement learning walking controller that lets the Figure 02 humanoid robot learn a human-like gait in simulation by imitating human reference trajectories through rewards, including heel strike, toe push-off, and arm swing. Training runs in parallel across thousands of robots in GPU-accelerated physics simulation, collecting years of simulated data in hours, and combines domain randomization with kHz-level torque feedback control to achieve zero-shot transfer to the real robot; 10 Figure 02 units run the same neural network without individual tuning.