The laughing-while-eating-salad meme was always a curt commentary on the role of marketed images in our society, one hard to forget. In my attempts to seek out re-framing of darker styles with light-hearted motifs I began with a curiosity of trying the intricate work of H. R. Giger against this backdrop.
This album features images generated by text prompt
using StabilityAI's Stable Diffusion, unless otherwise noted.
Here we begin with selections from the first series. The specific prompt was not recorded, but I believe it was prefigured in the style of Giger, Rembrandt and Dr. Seuss.
Series B attempted to move the style lighter. Skeletons seemed possibly a less grotesque image for Giger to work with, avoiding some of the strangeness of rendering people.
Series C: Intrigued by the motif's weight I pushed friendlier aesthetics, but I kept Giger around. The corporate nature of the laughing-salad stock image seemed a strange juxtaposition to offer with the family notion. I suppose it simply becomes the 'work family'.
Here the same prompt was attempted using Midjourney, which was mostly just strange.
Series D: none were saved from this series.
Series E, pushing further into marketing aesthetic, dropping Rembrandt (but not Giger) to draw the palette more modernly. These started to exert a real fanatic sense of the scenario which I appreciated, a kind of madness of "corpo-neutralism" which was brought out further by 'illustration'. Often I will use the 'diorama' cue to try to produce depth to the scene.
This prompt was also attempted using Midjourney. While interesting, I still preferred the whimsical fanaticism of Stable Diffusion's output.
Series F pushed this further. I may have lost Giger at this point, and extrapolated more intriguing illustrative qualities by using Mœbius (Jean Giraud) as a style cue.
And Midjourney attempts with same prompt:
Here I switched to a new prompting. Some other series have tried to explore pareidolia to see if generative AI could encode suggested imagery on top of given imagery. I still have yet to find any success. These were all failures of my ambiguous prompting, but still represented the laughing-salad motif.
Midjourney managed some curiosity with the same prompt and its subsequent effort of using fish instead of grapes. Either way, there is no hidden imagery here.
In the end Stable Diffusion was a real champion in producing what I wanted here, and such a 'vein' could provide a rich source of endless corpo-neutral salad engorgement. Midjourney has its own strengths elsewhere of course.