Case Study 345: The Meme That Was Not A Meme
We trained Peak Effekta on a dataset of 10,000 cat pictures with captions like "I can smell the cheese" and "Lol, I'm a cat." The goal was to teach it to recognize the difference between a genuine meme and a picture of a cat.
Results:
Peak Effekta correctly identified 9,999 cat pictures and one image of a cat in a fedora holding a cane, which it classified as a "Meme In Progress." It also flagged 1,000 instances of the word "no homo" as a possible AI-powered dad joke.
Hyperlinks: