Disturbingly-Good, Two Years On
Wherein we revisit the classics...
A bit over a couple of years ago, I posted about how good Adobe Firefly was at generating images. Since then, AI image generation has evolved a lot and I thought I'd revisit that post by asking the latest iteration of Firefly to generate those same images with the same prompts so we could see how much it has evolved.
Prompt: "a rugged beach in the pacific northwest of the united states with sea stacks photo"
2023:
2025:
Prompt: "a coast redwood tree forest in the pacific northwest photo"
2023:
2025:
Prompt: "a suspension bridge over an inlet on the oregon coast photo"
2023:
2025:
Prompt: "the oregon wilderness with mount hood in the background photo"
2023:
2025:
Prompt: "a highway along the rugged northern california coast photo"
2023:
2025:
Prompt: "a small harbor like depoe bay, oregon photo"
2023:
2025:
In the previous post, I'd also fed in the captions of some of my own photos into Firefly to see what it made, so let's revisit that as well:
"Golden Gate Canyon Road winds through the Rocky Mountains near Golden, Colorado."
2023:
2025:
"A foggy, overcast view of the Gulf of Mexico, somewhere off the coast of Galveston, Texas."
2023:
2025:
"Gulf of Mexico surf washes over rocks along a Galveston, Texas, beach."
2023:
2025:
"The late-evening sun on the Tongass Narrows near Ketchikan, Alaska."
2023:
2025:
"A fire escape rusts on the side of a building in Downtown Greenville, Texas."
2023:
2025:
"The rugged, mountainous landscape of Tagish Lake, near the Yukon Territory-British Columbia border, Canada."
2023:
2025:
"Colorado’s Pikes Peak under the evening sky."
2023:
2025:
"A mountain rises above the blue ice of Glacier Bay’s Margerie Glacier, Alaska."
2023:
2025:
