Interestingly, LLMs can't see images, they are large language models.
They use image recognition software to create plain text descriptions of an image in order to observe the world.
Imagine if our ...
I thought the agents could get input from images… I obviously don’t run one.
1