Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation Link: https://arxiv.org/abs/2610.14162 Discussion: https://news.ycombinator.com/item?id=50422168