about to go to bed after an all-nighter w/ work but i'm slightly more optimistic about my art niche vs the state of AI bullshit, having realized that the clientele for train art will check that the locomotive their fursona is posed next to has the correct air brake plumbing, while AI is firmly on the fucking struggle bus of trying to not make a wheel look like melted swiss cheese
- they/them/moo
- chimoora.zone/
avatar by @citriccenobite
you can say "chimoora" instead of "cow of tailed snake" if you want. its a good pun.
i ramble about aerospace sometimes
I take rocket photos and you can see them @aWildLupi
I have a terminal case of bovine pungiform encephalopathy, the bovine puns are cowmpulsory
they/them/moo where "moo" stands in for "you" or where it's funny, like "how are moo today, Lupi?" or "dancing with mooself"
Bovigender (click flag for more info!)
twitter
twitter.com/awildlupidragontwitch
twitch.tv/lupidragonYou must log in to comment.
in reply to @TomColetti's post:
train furries (me) will really be like "there is an incorrect number of rivets securing the boiler plate
Literally had this thought today. An AI could, perhaps, draw something as complex in shape as a generic steam loco and make it convincing to a layperson (given enough generic data), but if you ask it for a rendition of say a K-5-A from the Chicago, Indianapolis & Louisville in late life condition, it's going to be an absolute no-sale if it can't get the mechanical lubricators in the right place or the shape of the sanding dome shape correct. Of course, maybe that's all a matter of having a good enough data set trained, but one of the funny things about historical objects (esp ones that may not exist any more) is that they really only exist through historical photography. There is only so much of it, and there is no guarantees you'll ever have enough for a replication that would satisfy an enthusiast.
Long way to say, yeah I agree.
I'm pretty confident you'll have a lifetime's worth of gorgeous art to make due to this very specific fact.
The chances of AI getting the fursona right if it's anything defined any further than "An anthropomorphic tiger" is near zero as well anyway.