|
|
|
|
|
by pembrook
871 days ago
|
|
As evidenced by any example I've seen from these "AI logo generators," image-generation models simply cannot do logos well. They always create insanely detailed illustrations (not logos) that feel like clip art or anime. Is this is a data-set issue (too much art station/deviantart scraping) or just a taste-issue on the part of the creators of these tools (bad prompting)? Or can image-generation models simply not do "simple" well? |
|