Pet-Friendly Interstellar Vacation lol
- I went for an ironic interpretation here, thinking it would be fun to recreate the Sigourney Weaver “Alien” chestburster scene, but with a cat.
- Here is my sketch:
- For image generation, I went with NightCafe
- Prompt 1: “make a simple cartoon resembling this photo, except it should be featuring a cat lying on its back making a silly expression of pain, replacing the main character in this photo. An alien is about to break through the cat’s belly.” *I included with my prompt a reference photo of Sigourney Weaver in this scene
- Result:
- My reaction: wow, the woman is clearly more of an alien than whatever is inside her. Cat companion= cute version of Puss in Boots made creepy by strange expression and added appendages. Ok, on to the next one…
- Prompt 2: make a pen sketch resembling this photo, with a cat (instead of a woman) lying on its back and screaming, as an alien is about to break through the cat’s belly
- Result:
- My reaction: this woman is clearly deriving too much pleasure from this experience. this image feels voyeuristic, with the combo of her expression and the vicious cat img on the wall, telling the viewer to back off/go away. ok, next:
- Prompt 3: make a cartoon sketch with a cat lying on its back, making a ridiculous face while screaming, as an alien is about to break through the cat’s belly
- Result:
- My reaction: I was so disappointed with this weird, straightjacketed, chicken-legged, angry cat-alient with a white elephant trunk outgrowth, that I tried one more, using my own sketch as a source image
- Prompt 4: make a cartoon sketch with a cat lying on its back and making a ridiculous face while screaming, as an alien is about to break through the cat’s belly
- Result:
- This was a little closer to the end result I was after, but still pretty weak, overall.
- Takeaways from this process, Using NightCafe:
- 1. I liked that you could add some specificity to the image generation process by using a source image, but in this case, the source image seemed to frequently add confusion to the process of image generation.
- 2. The minor adjustments that could be made in the side panel- prompt weight, refiner weight, noise weight, etc- also made for a more customizable user experience, giving a greater sense of control, though you don’t necessarily get the outcome you would expect by making those adjustments.
- 3. Overall, I can see how this AI could be useful in helping generate some initial ideas and/or visuals for projects, but there are quite a few issues with it, keeping it from achieving a polished, curated, refined result that would be acceptable as a final product or project.
- TL;DR- yet another quasi-helpful AI tool, but without the vision behind it and a lot of attempts and adjustments, won’t be terribly helpful.