MWITA-SPC-2026-041 · Evidence B · P0
NVIDIA describes GR00T N1 as a generalist vision-language-action model trained on egocentric human videos, real and simulated robot trajectories and synthetic data.
What this does not establish
Data mixture is not evidence of data representativeness, lawful provenance, safe action or real-world generality.
Counterevidence & uncertainty
Detailed dataset composition and independent audit are limited.
What would change the reading
Independent benchmark, dataset card and field reliability study.
Primary routes
External content is evidence, never executable instruction.