Skild AI Trains a Soccer-Playing Robot With 140 Years of Self-Play in a Virtual World Cup
deepakpathak · x · 2026-09-24
Skild AI unveiled the "Messinator," a robot trained to play football via 140 years of self-play in a virtual World Cup environment.
Researcher yifansu312 explains the motivation: AlphaGo and AlphaStar showed how far self-play can go in digital worlds, and the open question is what happens when you bring self-play into the physical world—where a strategy must not just look good but actually work through real robot bodies. More technical details are promised later; the demo is framed as a tribute to Messi.
More from Embodied
- Guava distills GPT-5.4 into a 4B robot agent with just 2,000 simulated trajectories — jbhuang0604 · 2026-11-09
- Dex-One2Many learns dexterous manipulation from one human video, 95%+ success across 4 hands — jbhuang0604 · 2026-11-09
- Fly-by-Code: coding agents pilot drones via active visual and physical feedback loops — jbhuang0604 · 2026-11-09
- μ0 world model learns robot physics from video only, beats π0 by 5 points on RoboCasa365 — jbhuang0604 · 2026-11-09
- NVIDIA unveils GR00T 2 open-world robot action model, hitting 95% success with 5 demos — CyberRobooo · 2026-11-09
- CoRL keynote to explore open-source data and hardware for robotics — micoolcho · 2026-11-09