Hacker News Reader

I’m curious to see which researcher will be the first to have an AI generate realistic 3D assets by understanding 3D natively.

That's already here with NeRF

Can NeRF generate it’s own output (say from from a text prompt)? I thought it used image/video input