Everyone blames the model. "GPT can't design." Wrong. The model isn't the bottleneck. The pipeline is.
Here's what's actually happening when an AI tool spits out that same purple-gradient, rounded-card, centered-hero layout every single time:
1. They generate pixels, not systems. Most tools prompt a model to output a screen. One screen, in isolation. No shared token layer, no type scale, no spacing system carried across views. So the second screen doesn't know what the first one decided. Consistency is impossible by construction, not by accident.
2. They optimize for "renders without breaking," not "looks intentional." The reward signal is valid output. A layout that compiles. Taste isn't in the loss function anywhere, so the model regresses hard to the mean of its training data. The mean of all web UI is Tailwind defaults and a violet gradient. That's the slop.
Mad respect for shipping five products to get to this insight. The best tools always come from founder frustration. If I import a design I love as inspiration, how heavily does the AI rely on its layout versus its aesthetic style?
@luyanda_ntombela it tries to stay consistent to design while being functional because in our eyes a good design needs to be functional as well .
As a UI/UX designer I'm skeptical of most AI design tools but the "whole product" framing is interesting. Most just generate individual screens with no visual consistency. Does it maintain design language across different screen types?
@vlad_zabavskiy yes it does either web or mobile it stays consistent
jared.so
Nice work shipping this, congrats! What made you decide to build this now?
@borrellbr We were constantly using AI to generate design and we felt that we can build better than all those AI ui tools why not create a tool out of this!