JUST IN: Google DeepMind just released MuJoCo 4.0.
@lukas_m_ziegler66K followers
29,207 views
12 Comments
Comments from the original post on X (Twitter)

@clankrmedia
Pretty big deal. Still best option for lightweight physics simulation for robots.
2 likes

@antifragiledon
@6carty
1 like

@iitiimii
Thankss!!!
1 like

@yay3d
do I have to learn this too !?
1 like

@poatomik
Tunneling is the old bug where a bullet-fast object skips through a wall because the sim only checks each frame. Sweeping forward to find the hit time fixes it without tiny timesteps. Cloth that keeps creases matters too: robots folding laundry need that.

@mathurutsav
Speculative contacts is the kind of unglamorous fix that quietly saves a thousand sim-to-real projects. MuJoCo 4.0 is a big one.

@thesimaximalist
A robot policy can learn to exploit missed collisions. A “successful” grasp is much less useful when it depends on the fingers passing through a thin object between simulation steps

@pvargasx
Nice! I believe speculative contact is the right call for thin fast stuff, but I believe game engines that use it tend to get ghost collisions though, like objects braking near geometry they’d never actually touch. curious how the solver avoids that here

@yaelmendez
@asimovinc 🤳

@leononrails
plasticity for wires and clay is the part i want to try. cables that actually stay bent

@amanvaibhav
A knife slipping through a timestep is the gap that only shows up once the policy meets hardware. CCD should make contact-rich sim a lot less leaky.

@askdotcharity
@grok is this for cpu users or those rich gpu kids?
