Skip to content
Agientry
← Back to Videos

A-0 Trains Robot Policies on 16,000 Hours of Human Video

Loading video
Loading video

A-0 builds a complete robot-learning system from 16,000 hours of YUBI handheld-gripper demonstrations with no robot data, running one set of weights on four embodiments.

Human demonstrationsRobot LearningCross-EmbodimentVLA
Category: research
Author: @yh_kris
Date: 2026-11-04T00:00:00
Duration: 39.0s
Source: https://x.com/yh_kris/status/2118010399748726977
Reference: https://arxiv.org/abs/2611.03815

Related Videos

RoboTok: Human Demonstration Data Engine
2:06

RoboTok: Human Demonstration Data Engine

@xiynfi1520580 · 870 views · 2026-09-04
RoboTokRICE
RL Post-Training Generalizes Robot Policies Beyond Seen Tasks
0:13

RL Post-Training Generalizes Robot Policies Beyond Seen Tasks

@MichalAKorniak · 3 views · 2026-11-05
Robot LearningRL Post-Training
Astra Edits Four Weights to Lift VLA Success
0:32

Astra Edits Four Weights to Lift VLA Success

@Lingxiao234 · 58 views · 2026-11-04
VLAWeight Editing
Same VLA, three data sources: Human Archive OTS wins on task success
0:43

Same VLA, three data sources: Human Archive OTS wins on task success

@babugi28 · 192 views · 2026-10-15
VLAPre-training Data
Long-WAM: Scaling World-Action Model Context to 19.2s in Real Time
2:38

Long-WAM: Scaling World-Action Model Context to 19.2s in Real Time

@AaronWeiHuang · 531 views · 2026-10-08
World-action modelsVLA
FLUX 3 Action: Open-Weights 7B World Action Model
0:05

FLUX 3 Action: Open-Weights 7B World Action Model

@HuggingPapers · 852 views · 2026-09-23
World-action modelsVLA

As an Amazon Associate, we earn from qualifying purchases.