Is image generation or text generation more impactful?
Is image generation or text generation more impactful?