Detail on Demand
--
I’m impressed by Midjourney’s new release — V5. With generative AI, version improvements seem to mimic the hardware releases of early gaming systems. NES to SNES was a huge improvement, PS3 to PS4 to PS5, etc… large, meaningful leaps.
What impresses me is that Midjourney’s improvement is able to consistently deliver detail. There are fewer strange artifacts and discontinuities. While this pushes the limit for realism, it can also be used to quickly create vivid illustrations that humans just couldn’t easily produce.
I’ve purchased books in the style of Wimmelbilder for my children and I might enjoy looking at them more than they do. I can get lost in the detail, imaging following works in progress along an assembly line.
What this capability might unlock is taking complex points and converting them into illustrations. I don’t recall which book, but IDEO had once commented that every company should have an illustrator to convert ideas into diagrams that could easily be understood.
Midjourney v.5 and similar image generation AIs might unlock this capability for all companies.
