#reinforcement-learning

2 posts

manipulation physical-ai robot-security ros2 open-source-hardware benchmarks delivery-robots privacy reinforcement-learning tactile-sensing

Force models, backdoored planners, open arms

Eka bets on force instead of vision, a trigger word hijacks LLM robot planners, and open hardware gets a dual arm platform and five hands.

October 21, 2026

physical-ai ros2 navigation simulation manipulation quadruped foundation-models reinforcement-learning nvidia aws

Physical AI toolchain, a walking hand

AWS open sources a physical AI toolchain, an ETH Zurich hand walks on its fingers, and EasyNav 0.5.0 gives ROS 2 a new navigation stack.

October 9, 2026