In a new study designed to test the co-creative capabilities of large language models (LLMs), internet memes created by OpenAI's GPT-4o model were, on average, rated funnier, more creative and more ...

But what actually stands out to me is how poorly the research relates to even the deliberately broad cultural form it purports to study. The top memes for all three scores are included.

After more than five weeks of monitoring, the study found that those who attended classes with memes incorporated in the ...