AN OBJECT IS WORTH 64X64 PIXELS: GENERATING 3D OBJECT VIA IMAGE DIFFUSION

Summary from the TLDR AI newsletter

Treating 3D objects as UV wrapped images makes it simple to use a DiT model to synthesize novel objects based on textual inputs.