DALL-E 2 creates amazing images from text. midjourney does it in more or less the same way. the two systems differ in terms of the learned data. that‘s maybe the reason of a better handling of human faces.