The Ghibli stuff is really amusing but for me the best are the infographics - you can make amazing stuff with basically 0 2D CG know-how for e.g. a presentation, web page, documentation, blog, etc.
The biggest difference to what was previously available is the accuracy - especially the text. This opens up a plethora of possibilities.
This is what I'm most excited about as well. Prior to 4o image generation, most GPTs I came across really struggled with this type of image generation. The only one that seemed to come close was Ideogram.
The biggest difference to what was previously available is the accuracy - especially the text. This opens up a plethora of possibilities.