The field of music generation and source separation is witnessing significant advancements, driven by the development of novel architectures and techniques. Researchers are exploring new approaches to improve the efficiency and quality of music generation, such as leveraging pre-trained diffusion models and integrating rectified diffusion methods. Another area of focus is the development of more sophisticated source separation techniques, including joint latent diffusion models that can simultaneously generate music and extract individual sources. These innovations have the potential to expand the creativity of musicians and improve the overall quality of music production. Noteworthy papers in this area include MGE-LDM, which presents a unified latent diffusion framework for simultaneous music generation and source extraction, and ZeroSep, which achieves zero-shot source separation using pre-trained text-guided audio diffusion models. AudioTurbo is also notable for its fast text-to-audio generation capabilities using rectified diffusion.