the process (pt. 2)

We’re trying something different this time…

Consensus feedback on part one of this series was “yawn” & “over my head”. In an attempt to make this content more engaging and accessible, I’m changing the format: pictures (!!) and my voice.

Please take a listen (at least admire the GIF below), and let me know your thoughts on the new approach.

the visual evolution of POPOUT

aspect ratio is fucked but you get the idea !

starting out

This is my very first ask from Claude. I think a crucial bit that folks with less software experience might overlook is to ask for the build + run instructions. Follow these to a T, if they do not work, ask Claude to correct them, and hopefully that can get you “over the hump” of actually running what you’ve built, which I know can seem a dark art to newbies.

After several failed launches (and running into the free usage limit), I got here:

Note the all caps. Addressing known failure points is an important pattern I found throughout this process. Here’s what I got:

about 150 commits later….

the end…????

assorted outtakes

Claude is not so good at drawing:

oh no, claude…. that’s… that’s not a duck…..

If Claude says, “Ah, _____ Understand!” you are probably about to receive garbage:

successive failures of trying to get rich event previews when sharing via iMessage

Details and phrasing matter. I was exploring ChatGPTs ability to generate images and it took my illustrative request very literally:

DISCLAIMER: even if this was better, I think there’s so much value in paying actual human artists for their work that I would not use AI generated imagery on any lasting or monetizable project. Unfortunately, POPOUT isn’t there yet