I tried generating several text prompt based A.I. experiments today using the DALL-E mini GitHub project. It's an unofficial (and extremely simplified) version of the DALL-E 2 software I was talking about earlier. Check out my 'creations' below.
And keep in mind, I didn't touch any virtual brush, Photoshop- or 3d modeling tool in order to make these. Just text, literally a few words. Around 50 seconds, that's how long it took on average to produce each and every one of those uniquly generated images that didn't exist before.