I agree Diffusion Models are very powerful too, yet I still think they were less of a revolution than StyleGAN - but that's my personal opinion. Regarding being concerned maily with Deep Learning - I do believe this is (and should be) the main focus of data science today.