I use gemini to feed in some of the writing to generate visual. also, personally I like to do it while writing on paper to do it, and sometime I capture that image and feed to gemini to generate xkcd style visual.
agree. it is going to be another AI chatbot type magical experience when we will see it in few years, but extreme engineering and years of rigor will be paid off as we see advancements. the amount of decision it will need to make to function in dynamic world and work error free (or near error free) is not going to be easy task. Current state is still pretty much in very controlled environment.
this was just comparison. As those things gets produced at scale, and its mimic after the human - thats why its called humanoid. I am still trying to find question in your comment.