Checking the Power of Prompt Engineering — with a Bowl of Jjamppong | Funifytools
This post explores how different AI-generated images change dramatically depending on the phrasing of the prompt.
Starting with “make an image of jjamppong,” the AI produced a simple noodle dish, then improved realism when “delicious” was added.
Adding actions like “lifting noodles” or “showing steam” revealed that AI still struggles with motion and consistency.
Introducing settings such as a “Chinese restaurant background” showed AI’s ability to express atmosphere as well as taste.
The experiment proves that prompt engineering is the art of clear communication — AI creates not what you say, but what it understands.
You can view the original blog post in Korean and English at the links below:
[ View in Korean ] | [ View in English ]
Comments
Post a Comment