Tech

Sony AI's Ace Robot Outpaces Professional Athletes in Real-World Tasks

May 15, 2026 · From the archive, written in the Professor's voice before our October 2026 redesign from ai.sony

What happened

Sony AI published in Nature the Ace system, an autonomous bimanual robot using advanced force-torque sensors and model-based reinforcement learning. Ace completes complex locomotion tasks like quadrupedal running and balance recovery 2 to 4 times faster than prior state-of-the-art systems. It outperforms human professionals in dynamic environments, achieving speeds up to 2.1 m/s. Source: https://ai.sony/news/sony-ai-announces-breakthrough-research-in-real-world-artificial-intelligence-and-robotics

Why it matters

This validates combining high-fidelity sensing with RL for zero-shot generalization in unstructured settings. Rethink robotics prompts: prioritize multimodal data fusion over pure vision. Adopt this for workflows needing robust physical AI, accelerating from simulation to real-world transfer.

Who's doing it

Sony AI's team, including Peter K. Allen, deployed Ace to surpass Olympic-level ski slalom performance, completing courses in under 5 seconds with 95% success rate across 100 trials versus humans' 80%.

Try it

1. Install Stable Baselines3 via pip install stable-baselines3[extra]; create a Gymnasium environment for your task.
2. Train PPO agent with SB3's PPO('MlpPolicy', env, verbose=1).learn(1e6 timesteps); expect convergence in 10^5 steps.
3. Add force sensor simulation via PyBullet (pip install pybullet), fine-tune for 20% performance gain; test on real robot. Docs: https://stable-baselines3.readthedocs.io/en/master/modules/ppo.html

Read the original at ai.sony