Just found out my AI art prompt trick was actually making things worse
My buddy Dave who works at a tech startup in Austin told me to stuff my prompts with 15+ adjectives for better results. Been doing that for 2 months and getting weird muddy images. Finally tested his advice against just using 3-4 keywords and the simpler prompts came out way cleaner. Has anyone else found that more detail in prompts just confuses the model?
Blame Dave for making you waste two months of electricity and prompt tokens. I tried the "adjectives galore" method once and got a cat with six legs wearing a top hat that looked like a melted crayon. Models seem to handle short and sweet way better. Maybe your buddy secretly works for an image generator company and gets paid per prompt.
Wait, did anyone else see that one guy who tried the adjective thing and got a horse wearing a wedding dress? My friend Jenna actually fell for this whole thing a while back. She spent like three weeks loading up her prompts with "vibrant majestic ethereal" and got back pictures that looked like a kid's art project after drinking too much soda. She finally ditched the adjective overload and went with "yellow cat" and got something that actually looked like a yellow cat. I think Dave might just be messing with people, honestly.