For aesthetic UI frontend work, even with AI, I am still very much in the development loop. I wonder, if I were coding with a Google model which understands semantics in images much better, if this would already be history.