What if one AI could seamlessly understand and generate both text and visuals, adapting to any length? Zebin You and team from Renmin University of China and Ant Group present LLaDA-o! Their new LLaDA-o model uses a clever "Mixture of Diffusion" approach, blending discrete

→ View original post on X — @jiqizhixin