Reaction
Kevin Zakka accuses Animesh Garg of backpedaling on Supermimic
Garg praises Supermimic’s execution, but Zakka accuses him of changing his tune after dismissing the research.
TLDR
Animesh Garg says he contacted Supermimic’s authors, praises their execution, and welcomes progress in model-based reinforcement learning with first-order methods. He argues that research ideas often need iteration and the right timing. Kevin Zakka says Garg and another poster previously dismissed the work as an old idea with better marketing, and accuses Garg of backpedaling.
Combined views
406
2 Sources, first seen ago
Comments