@rajistics: Could text diffusion make AI faster? It starts with uncertain token positions and refines a block of text over several passes. Here’s how that compares with ordinary inference and speculative decoding, and where the speed gains come with quality tradeoffs.
Well, when are we gonna reach General Superintelligence?
2026-10-04 17:53:47
0
SMSwithGoldie :
Does this mean for a given prompt, it predicts the entire output in a single block? That’s what I originally thought but that can’t be right. It must be for each individual reasoning step it is predicting the output for that? Or am I totally missing the mark😅
2026-10-04 19:54:42
0
Zuck :
interesting
2026-10-04 19:34:48
0
To see more videos from user @rajistics, please go to the Tikwm
homepage.