The field of graphical user interface (GUI) grounding and visual generation is witnessing significant developments, with a focus on improving the accuracy and efficiency of models in localizing interface elements and generating realistic images from text prompts. Researchers are exploring novel approaches, such as self-generated reasoning and spatial-aware criticism, to enhance the performance of multimodal large language models (MLLMs) in GUI grounding tasks. Additionally, reinforcement learning techniques are being applied to improve the chain-of-thought reasoning capabilities of MLLMs in visual generation, enabling models to autonomously discover effective reasoning strategies and handle complex prompts with precise spatial relationships and attributes. Noteworthy papers in this area include: ReGUIDE, which proposes a novel framework for web grounding that enables MLLMs to learn data efficiently through self-generated reasoning and spatial-aware criticism. GUI-G1, which achieves state-of-the-art performance in GUI agent grounding by addressing challenges in input design, output evaluation, and policy update. GoT-R1, which applies reinforcement learning to enhance semantic-spatial reasoning in visual generation, resulting in significant improvements on compositional tasks involving precise spatial relationships and attribute binding.