Luci for Chai Tea on Nostr: i feel like a usage for generative neural networks that i keep waiting for, but sadly ...

i feel like a usage for generative neural networks that i keep waiting for, but sadly don’t see a whole lot of energy for; is real time or near real time fluffing up of 3d renders. Like, imagine google earth zooming in and doing an image 2 image transform of a heightfield using street view images as a reference for generating novel views.

Or creating a lofi 3d model and “texturing” it by painting over several views, and rather than uv mapping the paintings, “deep faking” them onto new renders