TallSilhouette 2 days ago
Looking at Nvidia's presentation I'm impressed that both single and multi-frame gen now essentially add no extra latency to the experience. Maybe Reflex 2 has something to do with it?
More exciting to me though are the memory and performance improvements of AI compression along with the improvements coming to hair, skin, and faces:
It sounds like you'll be able to natively deepfake faces past the uncanny valley like they recently started doing in Hollywood vfx.
Last edited by TallSilhouette - 2 days ago
Essay Pro