Michael Tomlinson · Nov 2, 2026 · 10:33 PM UTC

Michael Tomlinson

31Media

Manda retweeted

Michael Tomlinson

@mikejt4Nov 2

A peg-insertion policy that scores 49/50 in MuJoCo succeeds 11/50 in PhysX. Same task, different physics engine. They also found the same authored foot friction resolves to 0.8/0.6 in PhysX and 1.0 in Newton, and PhysX-trained ANYmal-D policies use about 35% more joint torque. Running a policy in a second engine looks like a cheap check before trying it on hardware. References mandarobotics.com/blog/polic…

Manda

@MandaroboticsNov 2

We trained robot policies in different physics engines (MuJoCo, PhysX, Newton, & Genesis) and tested how well they transfer. Even with identical assets, the choice of engine changes learned behavior and success rates. 1. The same policy might fail out of its original engine 🧵

Chris Paxton · Oct 6, 2026 · 1:14 PM UTC

Manda retweeted

Chris Paxton

@chris_j_paxtonOct 6

Simulating realistic physics is hard with traditional methods; the question is how much work it takes to something good enough to be useful? I like that the one which looks most accurate to me here (mujoco) stops because of an invalid geometry error

Manda

@MandaroboticsOct 5

Comparing major physics engines on robot tasks, we found stark differences in how Newton, native MuJoCo and Genesis handle deformable objects. We compared behavior and forces with matched geometry across increasingly complex scenarios. 🧵

9,119

Emerson S · Oct 5, 2026 · 10:46 PM UTC

Manda retweeted

Emerson S

@Em_NomadicOct 5

Deformables are such a good stress test for simulation. Small differences in the physics engine can completely change the interaction.

Manda

@MandaroboticsOct 5

3,641

Akira Sasaki · Sep 18, 2026 · 1:56 AM UTC

Manda retweeted

Akira Sasaki

@gclue_akiraSep 18

GPT-6 Astra Directで、リアルタイムに動いているな。これはどうやってるんだろう....

Manda

@MandaroboticsSep 17

Astra is dominating existing robot policies. But what can Astra do that existing policies could not? We analyzed 10 tasks, 5 Episodes each on Astra, Pi0.5, Nvidia’s Cosmos3 Nano and Gr00t 1.7 on RoboLab. 1) Astra is by far the most reliable model in terms of object recognition:

9,974