The robot learned to plan in public
In September 2025, Google DeepMind described Gemini Robotics 1.5 as a step toward physical agents that perceive, plan, use tools and act across multi-step tasks. The language was ambitious; the important editorial question is what “plan” means when the world can push back.
In September 2025, Google DeepMind described Gemini Robotics 1.5 as a step toward physical agents that perceive, plan, use tools and act across multi-step tasks. The language was ambitious; the important editorial question is what “plan” means when the world can push back.
Physical planning is not a list of instructions. It is a loop involving observation, action, error and recovery. A robot that succeeds only when objects, lighting and timing cooperate has demonstrated a narrow condition, not general work.
Embodied systems deserve the same restraint applied to software agents: publish the task boundary, intervention rate, failure modes and operating environment. The future of robotics will be decided in the gap between a capable demo and a maintained worksite.
