Shang-Zhe Wu

Existing 3D asset generators can now produce objects that not only look good but also move and function realistically, passing physics tests for stability and collision-free articulation.

An agent can generate high-fidelity, simulation-ready 3D indoor scene reconstructions by iteratively editing Python code, outperforming frontier models.

Kirin transforms raw video footage into realistic animated animal motions, bridging a critical gap in the availability of motion data for non-human species.

Visual intelligence could be the key to unlocking AGI, challenging the dominance of language-based models in the quest for general intelligence.