"Hah! I finally found this stupid temple! The five bits I paid that dweeby mare to translate the scroll was worth it. All that remains is to go inside and collect my rightful reward!

The Temple Challenges are marked "Pegasus Only", so a well-trained griffin will surely handle them even better! Easy treasures, dweeb!"

After creating the previous image of Gilda, I decided to try a new style.
Conclusion: it's not worth it. AI can generate human feet nicely, but it still struggles with creating good-looking paws. Correcting such detailed images manually is pure hell.

Perhaps I should move from single images to generating comics?