Coding AI controlling hardware sparks robotics breakthrough
Huggingface's lerobot 101 and openclaw agent demonstrate that large language models can write code to control physical robot arms, overcoming traditional calibration hurdles.
a recent experiment using huggingface's lerobot 101 hardware and an openclaw agent demonstrated that coding models can successfully configure, calibrate, and control robot arms to perform specific tasks. researchers from uc berkeley, nvidia, carnegie mellon, and stanford introduced the cap-x benchmark and cap-agent0 framework to evaluate this 'code as policy' approach. their findings reveal that gemini currently excels in robot programming, likely due to deepmind's focus on physical world grounding. this shift suggests that ai coding agents can lower the barrier to entry for robotics by automating complex setup processes.