Abstract page for arXiv paper 2610.14162: Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation