← Back to Videos

ThinkWorld learns robot futures from human video

Loading video

NVIDIA and KAIST share motion tokens between human video and robot policies, rewarding imagined futures that match reality.

ThinkWorld MotionTokens WorldModel RobotPolicy KAIST

Category: world-model

Author: @robotsdigest

Date: 2026-11-01T00:00:00

Duration: 115.1s

Source: https://x.com/robotsdigest/status/2116970655640174946

RoboJEPA scales latent world models for robots

RoboJEPA scales latent world models for robots

@LeoKharon · 29 views · 2026-11-01

RoboJEPALatentWorldModel

Genie 4: A World You Can Steer in Real Time

Genie 4: A World You Can Steer in Real Time

@GoogleAI · 124 views · 2026-10-27

AW-0: Turning Plans into Physical Action

AW-0: Turning Plans into Physical Action

@robotsdigest · 80 views · 2026-10-27

AW-0World-action models

Psi-0.6: Reasoning Across the Physical World

Psi-0.6: Reasoning Across the Physical World

@khai_loong_aw · 84 views · 2026-10-27

Psi-0.6Multimodal

UniAction Unifies Four Action Types in World Models

UniAction Unifies Four Action Types in World Models

@zizhang_li · 156 views · 2026-10-26

World ModelsUnified Action Conditioning

MUNDO: A Few Photos Become an Interactive, Simulatable World

MUNDO: A Few Photos Become an Interactive, Simulatable World

@zunnanxu · 141 views · 2026-10-23

MUNDOWorld Models

As an Amazon Associate, we earn from qualifying purchases.