Comment by Krastan

LMM is a large multimodal model. So it does more than just language, in this case interacting with UI, in others using voice and video