New research from Springboards shows humans can barely tell leading LLMs apart—and the models are training creativity out of themselves.
I recently found the same when using a few different models to create abstract superhero names. The models gave me identical outputs in most cases, riffing in similar ways from known sets.
I recently found the same when using a few different models to create abstract superhero names. The models gave me identical outputs in most cases, riffing in similar ways from known sets.