2026

• Why Bounded Residual Actions Are Not a Safety Guarantee in Robot Learning

  Why limiting a learned residual action is useful in robot control, but does not by itself guarantee closed-loop safety.

  4 min read  ·  August 12, 2026

   2026  ·   robotics  ·   reinforcement-learning  ·   safe-control  ·   residual-learning  ·   technology  ·   robotics

  

• Diffusion Policies in Robotics: Why Robot Actions Are Becoming Generative

  A short engineering note about diffusion policies in robotics, why they are useful, and why control still matters.

  8 min read  ·  June 1, 2026

   2026  ·   robotics  ·   robot-learning  ·   diffusion-policy  ·   imitation-learning  ·   technology  ·   robotics

• Classical Control Is Not Dead: Why Robots Still Need Physics

  A short engineering note about why robot control still needs physics, even as AI and reinforcement learning become more capable.

  5 min read  ·  May 19, 2026

   2026  ·   robotics  ·   control  ·   reinforcement-learning  ·   physics  ·   technology  ·   robotics