The ROBOCON competition in mainland China this year concluded successfully from July 9 to July 17. This marks my first year participating in the event as both a referee and a staff member.
This is an interesting paper I saw at ICRA 2026. It talks about how to make a cost-effective 22 DOF robotic hand and goes into a lot of details on how to assemble it and configure it into different configurations.
Boston Dynamics mostró recientemente que Atlas realizó un potente backflip con una precisión notable.
El robot aterrizó limpiamente sobre sus dedos de los pies e inmediatamente se lanzó a un segundo backflip.
Esta capacidad avanzada es impulsada por un sistema de aprendizaje de todo el cuerpo desarrollado por el Instituto RAI y desplegado por Boston Dynamics.
asked claude fable 5 to design a qdd actuator
it also animated the gearbox and inspected collisions as part of the validation loop
~ 30 minutes / 400k tokens
Una foto entra. Un archivo CAD sale.
MIT acaba de liberar GenCAD: un modelo que convierte la imagen de una pieza mecánica en un programa CAD paramétrico editable.
No es text-to-image para hacer renders bonitos.
Es image-to-CAD para tocar una parte mucho más cara del mundo físico: diseño, prototipado y fabricación.
La primera víctima no es SolidWorks.
Es el junior que pasaba sus días traduciendo bocetos en geometría técnica.
Open source.
Se llama GenCAD.
#CVPR2026 Can frontier LLMs write PhD-level 3D vision code?
We introduce GeoCodeBench, a benchmark that asks models to read real 3D geometric vision papers and implement core functions.
Best result so far: GPT-5 reaches only 36.6%.
This suggests that scientific coding in 3D vision remains far from solved.
Paper: arxiv.org/pdf/2603.30038
Project: geocodebench.github.io
When building a custom robot, it is helpful if the same joint position values represent the same physical pose across the entire pipeline: CAD, URDF/MJCF, RL/simulator, robot runtime software, and the motor-controller firmware. Debugging is easier when there are no sign tables or
Video “A” vs. video “B”
Same robot body,
But two different brains.
Do you see the difference?
Model-based control
vs.
Learning-based control
@RoMeLaUCLA
Our External Force-Guided Curriculum Learning (EFGCL) enables robots to learn various dynamic motions with simple reward functions.
The method is inspired by "spotting" in gymnastics, where a coach physically supports an athlete during training.
Here's how it works👇
2 notable changes in new Wuji hand (right)
The 4-bar linkage at the MCP is gone (yellow)
The IP joint axes now intersect the phalanx centerline, hence worm-gears out, bevel gears in
This affords mechanical back drivability and robustness
19K Followers 224 FollowingBuilding open source robots made by open minded humans to explore real-world applications. @Huggingface family
Join our community: https://t.co/KFG0LopcZt
116K Followers 1K FollowingHumanoid Robots: Tech, Business, and Social Dynamics.
Click the “𝕊𝕦𝕓𝕤𝕔𝕣𝕚𝕓𝕖” button on the profile to support.
Run by @dev_and_
2K Followers 981 FollowingAssistant Professor @HKUST | AI for Creativity | ex @Stanford @Meta @RealityLabs @UofT | Works #ControlNet #AnimateDiff @cveu_workshop
1.9M Followers 1K FollowingCo-Founder of Coursera; Stanford CS adjunct faculty. Former head of Baidu AI Group/Google Brain. #ai #machinelearning, #deeplearning #MOOCs
6K Followers 990 FollowingHi, I am the founder of @PokeBot_Global. I like RL, robots, and video games:) I am an Assistant Prof at Tsinghua; Postdoc at Stanford; Ph.D. at Berkeley
2K Followers 67 FollowingOfficial account for OpenDriveLab @hkuniversity and Beyond. We do cutting-edge research in Robotics, Autonomous Driving.
Email: [email protected]
109K Followers 605 FollowingThe official account of the @Stanford Institute for Human-Centered AI, advancing AI research, education, policy, and practice to improve the human condition.
19K Followers 666 FollowingCEO of ZEALS, Omakase Robotics
Building D1, a compact humanoid built in Japan.
準国産コンパクトヒューマノイド「D1」開発🦾
採用中: https://t.co/9gdDtpm9z5
@d1_humanoid
5K Followers 438 FollowingResearch Scientist at BIGAI, Director of Center for Embodied AI and Robotics. Ph.D. in Statistics from @UCLA. Former intern at @DeepMind and @MetaAI.