Podcast Episodes
Back to SearchLight-Loco-Parkour: Learning the Skill Is Only Half the Problem
Light-Loco-Parkour presents a multi-skill distillation approach for learning perceptive whole-body parkour and locomotion skills deployable on real r…
2 weeks, 5 days ago
OM-1: Human Skills, Many Bodies
OM-1 is a robot foundation model trained exclusively on human manipulation data (no teleoperation or robot demonstrations) that achieves near human-l…
2 weeks, 5 days ago
Show-Harness Gives VLMs a Robot Keyboard
Uses discrete semantic actions paired with embodiment-specific interpreters to turn any pretrained VLM into a robot controller without a dedicated po…
2 weeks, 6 days ago
BridgeVLA++: Keep the Geometry, Give the Robot a Memory
Renders point clouds as three orthographic 2D views fed into a pretrained PaliGemma VLM, predicting heatmaps to recover 6-DoF actions with temporal/s…
2 weeks, 6 days ago
The Sim-to-Real Gap Inside the Gearbox
Reinforcement learning (RL) has become a powerful tool for quadrupedal locomotion, and a sim-to-real approach is widely adopted to avoid hardware dam…
3 weeks ago
A Stable Grasp Isn't Enough: Giving 6-DoF Grasping a Purpose
Robotic grasping in cluttered scenes is dominated by methods that optimise either stability or a predefined task category, leaving a gap between 'can…
3 weeks ago
HiDream-O1-Embodied: A World Model Must Earn Its Actions
HiDream-O1-Embodied is a unified native architecture that ingests text, images, and video and directly outputs physical actions, positioning itself a…
3 weeks ago
Axis's Real2Sim2Real Bet: Make the Data Prove Itself
Axis introduced a Real2Sim2Real pipeline that converts real objects and scenes into high-fidelity physics-accurate simulations and validates learned …
3 weeks ago
A Rubik's Cube, Two Robot Hands, and One Category Error
GPT-6 Astra is demonstrated solving a Rubik's Cube using dexterous robot hands in a real physical setup, showcasing a significant leap in vision-lang…
3 weeks, 1 day ago
UnifoLM: Predict the Interaction, Not the Whole World
Unitree's UnifoLM is a 6B-parameter world model trained on 2,500 hours of real robot data that predicts only dynamic pixels and optical flow rather t…
3 weeks, 1 day ago