I was too lazy to prompt it with full requirements, but it did very well for a first attempt with: "Create a single page web app using vanilla javascript that allows the user to enter up to 40 characters of text and has interface controls for many css attributes including: font-family, size, letter-spacing, text-transform, weight, text-decoration, style, padding, rotation, color, border color, animation, etc."
Is it a strategy to fit more code into its context window?
https://github.com/manicinc/logomaker/blob/master/js/fontMan...
It's rather impressive because the result works.
Mani
c
Switching to desktop view doesn't cause the default text to wrap... So I guess you get different logo generation depending on browser size? Or at least different previews.
I'd expect a logo generator to never break words in-between letters unless explicitly asked for.
AI generated code is useful but it shouldn't be trusted to be complete.
My only thought was: if my employer paid me for 2 weeks (10 days) of work and this is what popped out the other end, I would be looking for a new employer, this isn’t acceptable where I work.
I think the fact that this can happen at all is important.
And whoever can contextualize this capability in the context of "real" software development will be the victor.
Try the app free forever on GitHub pages: https://manicinc.github.io/logomaker/.