> Maybe it was because this was my first attempt, but with 100 more posts to go, I hoped I could get better with practice. It would be super cool to just feed DALL-E a whole blog post and have something great pop out, but even with some GPT-3 magic we probably aren’t there yet.

I see a business opportunity here. Feed text into GPT-3 and have it generate DALL-E prompts to make appropriate images.

Then you have it do the same thing for a children's book.