Content_machine_challenge Day 3 · The AI Identity Lab

Content_machine_challenge Day 3

Because Claude really wasn’t cooperating, I ended up doing the exercise myself first in ChatGPT. I used screenshots of Claude’s questions as a starting point, so I didn’t have the training documents there and had to rely entirely on ChatGPT. It took quite a bit of time, and I actually went through the process three times before I got something that felt right.

Then I generated the photo from the prompt. My first reaction was: OMG, this feels so serious and so dull. The colors were flat, the whole vibe felt lifeless, and it just didn’t feel like me at all. So I gave ChatGPT my own FluxLora, plus one for my cat, and that already made a huge difference.

I also asked ChatGPT to add the quote it had chosen onto the photo. Even that took two tries, because in the first version the text was much too big and nearly sat on top of my hair.

The photos you see first now were made in ChatGPT based on the result from Lindsay’s app. My personal favourite is the second one, because it has a bit more colour and feels more alive. I’m also really happy with the texts. Since I’m working in English here for all of you, I can’t share the original ones just yet - but they’ll come later. Which one do you like most?

0 comments