The field of autoregressive image generation is moving towards more efficient and effective methods for generating high-quality images. Recent developments have focused on improving the scalability and robustness of autoregressive models, with a particular emphasis on reducing computational costs and preserving image quality. Notable advancements include the use of novel decoding strategies, such as grouped speculative decoding, and the exploitation of discriminative codebook priors to improve model performance. Additionally, researchers are exploring new approaches to image editing, including the use of autoregressive models for selective regeneration of edited regions. Overall, these developments are pushing the boundaries of what is possible with autoregressive image generation, enabling faster, more efficient, and more versatile image synthesis and editing capabilities. Noteworthy papers include: NEP, which proposes a new approach to image editing via next editing token prediction, and NextStep-1, which achieves state-of-the-art performance for autoregressive models in text-to-image generation tasks.