Gemini Robotics Model Test Fails, Highlighting Embodied AI Generalization Gap
chris_j_paxton · x · 2026-08-03
In response to a user's test where DeepMind's newly released Gemini Robotics-ER 2 caused robot arms to fling off the table, robotics researcher Chris Paxton shared his insights.
He noted that even with very, very good underlying policies like Gemini-ER 2, reproducing and transferring robotics results remains incredibly hard. This underscores a systemic challenge in the current embodied AI field regarding sim2real consistency and cross-hardware generalization.
More from Embodied
- Peking University's SimForcing distills simulation motion priors into robot world models — PekingUniversity · 2026-10-16
- NVIDIA and Mercedes-Benz Expand Partnership on Next-Gen Automated Driving — nvidia · 2026-10-16
- ~25,000 Humanoid Robots Shipped in H1 2026, Over 95% from Chinese Vendors — ingliguori · 2026-10-16
- openpilot brings stoplight-stopping autonomy to a 2019 Corolla — k7agar · 2026-10-16
- Hironobu Fujiyoshi publishes deck: from AI agents to Agentic Robotics — HirokatuKataoka · 2026-10-16
- CoDA co-evolves simulation and robot policies via shared failure memory, lifting OOD success from 5.3% to 27.6% — PLAN-Lab · 2026-10-16