The field of 3D reconstruction and generation is rapidly advancing, with a focus on developing innovative methods for detailed 3D shape reconstruction, infinite 3D world generation, and accurate multi-view 3D object reconstruction. Recent research has explored the use of thermal polarization cues, hierarchical frameworks, and generative priors to improve the accuracy and reliability of 3D reconstruction and generation. Notable developments include the integration of reconstruction priors into generative frameworks, the use of diffusion-based methods for image denoising and reconstruction, and the application of human cognitive laws to improve image fusion results. Overall, the field is moving towards more accurate, efficient, and realistic 3D reconstruction and generation methods. Noteworthy papers include: ReconViaGen, which integrates reconstruction priors into a generative framework for accurate multi-view 3D object reconstruction. WorldGrow, which proposes a hierarchical framework for unbounded 3D scene synthesis and achieves state-of-the-art performance in geometry reconstruction.