This made me realize that recent openAI tech is able to cut through misleading or confusing images, even finding the potentially inferred humour or absurdity layered in the images. I wonder how we’ll find a way to trick that kind of technology as well.