Those AI-generated city font experiment videos are visually stunning, no doubt, but I’m a bit skeptical about practical application. Typography is one thing—there’s a whole font engineering process between “looks good” and “actually usable”: consistent stroke weight, proper spacing and alignment, and thousands of characters maintaining the same style. What AI gives you right now is a handful of impressive single characters, not a full typeface.
It’s fine for concept demos or as headline text in a single key visual. But if you’re actually trying to use it commercially, the workload hasn’t been reduced at all—you’ve just added the hassle of tracing vectors out of images. Cool as hell, but still a long way from production-ready.
I’ve done that whole “extracting vectors from images” thing before, and honestly it’s more exhausting than just redrawing it from scratch.
One character looks amazing, but the whole set falls apart—that’s the gap between concept and mass production.
Honestly, using the title font is pretty nice. It’s good enough for the main visual of a poster.
The biggest killer is the inconsistent stroke weight—AI has absolutely zero concept of a stroke system.
Honestly, I think it might actually work in a while. Character structures do follow patterns after all.
OP’s being too optimistic. It can’t even touch engineering stuff like spacing alignment right now.
I tried doing commercial work once, and just aligning a few thousand characters drove the outsourcer crazy.
It’s fine as a source of inspiration, just don’t expect it to deliver a full typeface.