So, I was playing around, having ChatGPT place Kama Carnatica (a new Hindu superhero) on magazine covers. First, obviously, as person of the year on Time. That went well. Then the cover of the swimsuit issue of Sports Illustrated. No problem. How about the cover of Vogue? That seemed like a natural. But I wan’t a proper “high fashion” background. So I uploaded this photo and asked ChatGPT to use it as the background:
This is what I got:
That’s fine. (Sundari Nova is one of Kama Carnatica’s various incarnations.)
Then I dropped the magazine cover trope and just asked ChatGPT to place Kama Carnatica in this photo, which is a food cart in Manhattan near Herald Square:
This is what I got:
Yikes! There’s the food cart, though the signage is a bit strange and the menu items are odd. But the New York City background is even stranger. Where’d that come from? I’m pretty sure it came from this set of futuristic transformations of scenes from Hoboken, which is set in the year 2225 – notice that number just to the left of her head, along with “Hoboken BNJ.” What’s that?
I didn’t ask for any of that? Why’d ChatGPT “get its wires crossed” in that way. Who knows.
No comments:
Post a Comment