A generative AI adventure by Robbie Jones.

1. So I learned a new programming term: SCREAMING_SNAKE_CASE (as opposed to snake_case or camelCase).

2. I thought this was funny, and with a few sentence prompt, Chat GPT gave me an image to visualize it.

3. I liked that image so much that I kind of wanted to know more about the show it came from.

4. I had Chat GPT propose 8 ideas for 2000's era Cartoon Network shows that could use this in it's plot.

5. I picked my favorite of those and had it write a plot summary for the episode.

6. I then fed the snake image and the first plot point into Chat GPT image model via Adobe firefly and generated an image of the characters for the show.

7. I fed this show characters image and the next plot point and a few other instructions into firefly to make the next image.

8. I repeated for each plot point, sometimes generating 2 or 3 until one was good.

9. I opened cursor in a folder that contained those 9 images.

10. I gave it the full plot summary text and asked the agent to make a md file.

11. I asked it to make an HTML page using the md file as content, the images, bootstrap, and a simple 2-column responsive layout.

12. I went back and forth with it a few times to add color and make other tweaks.

13. I had it make a fonts page and show me 5 possible google fonts to fit a cartoon network vibe. I made it edit this page a few times until I found a font I liked okay, then add it to the main page.

14. I zipped up the files and deployed them to AWS amplify on our nitty gritty staging site domain. Within a few minutes, Amplify had the custom subdomain working.

15. Here's the finished web page that overviews the pilot episode (The Screaming Snake Case) of a show called Casey and the Artifact Club.

16. All this took about 3.5 hours.