Audio codec models make surprisingly good action tokenizers for dexterous robot arms

chris_j_paxton · x · 2026-09-27

Original post →

More from Embodied

Embodied channel →