I’m confused. Do the different modalities compliment each other? Can it learn more from text and images than text alone? Can you ask it to to draw a picture o

I’m confused. Do the different modalities compliment each other? Can it learn more from text and images than text alone? Can you ask it to to draw a picture of a cat with the robot arm?