Nvidia has introduced GR00T 2, a new model designed to enhance robot learning by combining human and robot demonstrations. The model aims to predict future actions in both video and robotic contexts, with the underlying principle that physical intelligence is improved by learning both task structure and control. AI
IMPACT This model aims to improve robot learning by integrating human and robot demonstrations, potentially accelerating the development of more capable physical AI systems.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →