Mirror author predicts an 'arm-farm' revival as video models make robot learning sample-efficient
vincesitzmann · x · 2026-10-20
Mirror author vincesitzmann argues video models and LLMs open a new path to learning-from-experience: Mirror isn't RL but a form of behavior cloning that needs no video-action pairs—"something different."
He predicts a revival of "arm-farms" for improving robot policies that will actually work this time, since LLMs and video models make exploration sample-efficient. The work ran on both NVIDIA and AMD MI350X GPUs via AMD's university program, contributing PRs to PyTorch nightly for video-generation throughput.
More from Embodied
- "Robots are going underground, figuring out their own routes" — a look past chatbot debates — VraserX · 2026-10-21
- Deepak Pathak: There Is No Perfect Data Source for Robotics — chris_j_paxton · 2026-10-21
- Discrete action representations match or beat continuous ones for robot policies — GuanyaShi · 2026-10-21
- Developer gave his Grok bot $1,000 to build itself a body — parts keep arriving — Baconbrix · 2026-10-21
- Tuesday Labs opens $200 reservations for 999 numbered Voxel Series-1 home robots — testingcatalog · 2026-10-21
- Tuesday Labs' home robot Voxel charms with Pixar-collaborated personality design — cneuralnetwork · 2026-10-20