OpenAI's Image Generator Has Great Tech But Terrible Fashion Sense

Impressed with OpenAIs new image mode, not impressed with its fashion choices.

That said, it's definitely a solid step up from Google's Nano Banana model. Competition in AI image generation is heating up and we're all winning from it.

But let's be real - my AI-generated face still looks like a slightly melted version of myself, and apparently the algorithm thinks I should be wearing ... this ;)

The tech progress is impressive though. Each iteration gets us closer to truly realistic outputs, which opens up massive opportunities for product development and marketing.

Makes me wonder: when will we hit that sweet spot where AI-generated visuals are indistinguishable from reality AND actually have decent style sense?

Here is the prompt if you want to try it out as well:

Freeform fashion-editorial collage of me in 8 distinct full-body spring outfits, arranged organically on a clean cream studio background. Keep my face consistent across all 8 looks, and make my proportions read around 6'00" tall without mentioning height in the text. Add neat handwritten arrows and labels for the key clothing pieces. No grid, border, or boxed panels. Keep all 8 full-body figures at the same visual scale and camera distance. Arrange them in a balanced two-row layout. Use a portrait 2:3 composition.

Whats your take on OpenAIs fashion choices?