1 comment

[ 3.9 ms ] story [ 11.6 ms ] thread
Quick description

I asked OpenAI's GPT3 Playground (Davinci) for a three page children's story with illustration suggestions in square brackets. It invented the topic, the title, the story and suggestions for illustrations.

I then fed the suggested illustrations to Midjourney, an image generation service (?) currently in beta. (You can see the process in #newbies-10 if you're on there, same handle)

I then asked for a cover page and ending page as well as a background image for the text. The only one I really iterated on was the background image trying to find the right prompt.

I then assembled it on Google Drive and posted it here. My creative input was the font selection.

The main thing I learned in the process is that 'AI' doesn't care about happy endings. :)