bowers@
LightNews LightNews
banner
srewob.bsky.social
bowers@
@srewob.bsky.social
310 followers 90 following 3.9K posts
Programmer. Nerd. Introvert. Probable dumbass.

This machine blocks Luddites.
Posts Replies Media Videos
srewob.bsky.social
Maybe try something like “An illustration of the stages of life” as the prompt?

E.g. (via MAI-Image-1):
December 13, 2025 at 1:19 AM
srewob.bsky.social
Via MAI-Image-1. This is slightly more dynamic with the shots, as the subjects aren’t exactly back to back. The kid is sitting cross-legged on the floor, while the adult is lounging back on what looks like a sofa.
December 12, 2025 at 7:37 AM
srewob.bsky.social
Via gpt-4o. If one were to take the child as an earlier edition of the adult, these two certainly are the same person, who literally hasn’t learned anything in several decades. (Both the child and the adult are holding their respective devices screen outward, looking at the back of the device.)
December 12, 2025 at 7:35 AM
srewob.bsky.social
#aiart prompt 1097: Split screen image of two people sitting back to back: in warm tones, a young child intently playing on an original Gameboy; in cool tones, a 40 year old playing a game on a modern tablet.
These four generated by Dall-E 3; while some of their details are a bit shoddy, they generally adhere to the prompt text. Though the older individual in this image, having white hair, is unlikely to be in his 40s.
December 12, 2025 at 7:30 AM
srewob.bsky.social
Via Dall-E 3. These are both much less literal and far more abstract. Generally: a spherical type of structure with regular pointed extrusions, somewhat like an abstract star, 3d snowflake, or the NYC Times Square Near Year’s ball. This likely is a closeup, given the other images, of the previously described pointy sphere, but no outer curve or background is present, so it sort of resembles a wireframe topological study. Pointy Sphere! There’s some symmetry going on, mirroring right and left across the center, and somewhat top to bottom. Kaleidoscopic. This pointy sphere is lit from within, down near the bottom of the image; this is weak light, as the top of the image appears lit almost completely from without.
December 11, 2025 at 4:42 AM
srewob.bsky.social
Via gpt-4o. About 11 crystals in this shot (though, the surface they are sitting on looks slightly reflective, so somewhat hard to say); the crystals toward the right seem clear, well faceted, polished, and sparkly. They become progressively more opaque and rough leftward.
December 11, 2025 at 4:36 AM
srewob.bsky.social
#aiart prompt 1096: Overlapping crystalline conical grid; the crystals are of variable opacity, refractivity, and reflectivity. Cooly lit from offscreen.
Generated by MAI-Image-1. We’re above a bluish-gray flat surface, perhaps a table top. Across most of the image, numerous cone-shaped crystals, pointy ends up. A small number of the set are milky white; others appear scuffed or otherwise translucent; most are clear and well cut. A few have rainbow refractions dancing through. All together, it somewhat resembles, abstractly, a well kept Christmas tree farm.
December 11, 2025 at 4:34 AM
srewob.bsky.social
#DragonTuesday
These generated by Dall-E 3. Generally, the show forested and brush-strewn hills, awash in low-lying fog. The dragon in each has a very dark, distinct silhouette, which produces an interesting effect: something akin to a double exposure used in older movies to composite a large movie monster in the background of a live action shot. The dragon is in the upper right, wing raised, facing leftward. A more centrally located dragon, near the top, facing left. Also with raised wing. Top-center-left, facing rightward. Definitely appears to be composited behind the hillside. It has a tiny t-Rex forearm stretched foreward. The dragon is pretty much dead center in this image, facing left; a bit larger than the others, too.
December 10, 2025 at 5:23 AM
srewob.bsky.social
#DragonTuesday
Via MAI-Image-1. A much broader, further away shot than GPT’s. It is still foggy, but the dragon is much more distinct, despite being smaller and distant. It’s climbing uphill toward the right, from the center.
December 10, 2025 at 5:17 AM
srewob.bsky.social
#aiart prompt 1095: Extreme fog cloaking the hillside; out of focus shrubbery silhouette the foreground; merging with the fog, lost in it, occasionally willowing into focus: the dark silhouette of a prowling dragon.

#DragonTuesday
Generated by gpt-4o. A very desaturated, out of focus photograph of a hill, slanting upwards toward the left. A couple of silhouetted bushes sit in the foreground at the bottom of the shot. The hill is very foggy: its upper ridge line fades into gray cloud. Taking up the middle third of the image, climbing the hill from right to left, the indistinct form of a non-winged dragon, walking on all legs. Its head is darkest, likely implying closest, while the rest of its body disappears into the fog.
December 9, 2025 at 8:20 PM
srewob.bsky.social
Via MAI-Image-1. The most photorealistic of the renderings, and the most sloppy. A slightly raised image, looking down toward a hillside. Many of the forks are short and blend into very desaturated grass. Only a few, top center left, are reasonably designed to resemble tuning forks.
December 9, 2025 at 6:48 AM
srewob.bsky.social
These four generated by Dall-E 3. They are all a bit more photorealistic than GPT’s offering, though only just. They also play with both light and shadow: the tuning forks are back-lit, casting shadows along the ground toward the camera. In the first, the “forks” look a bit more like metal straws. A singular fork with a curved base and no handle/post; it is comprised of five tines. The landscape appears a bit more painted with the interplay of light. Three “forks”, each a mostly straight pillar, in a row trailing into the distance. The closest is casting two shadows towards the camera, offset by about 30°. There does not appear to be multiple light sources to cause this. A field of relatively squat, impossible object style forks. Each has either three or four tines, with the inner tines cleaving off one of the outer tines, usually the left most.
December 9, 2025 at 6:45 AM
srewob.bsky.social
#aiart prompt 1094: A surreal field in muted tones, out of which grow curved metal tuning fork tines.
Generated by gpt-4o. A desaturated image, in a painted or drawn style, primarily in beiges and grays. Within a yellow field on an overcast day, numerous two-tined, curvy tuning forks stand. What is either a tree line or very distant mountains edges the horizon.
December 9, 2025 at 6:38 AM
srewob.bsky.social
Via MAI-Image-1. More of a chalk or oil painting vibe: an orange background, smeared or splattered with cyan and purple. The cyan predominantly is used in a sort of crosshatch pattern in the upper left of the image, while the purple makes a sort of bubble pattern lower center right. Some splatters in both colors are spread otherwise around the image.
December 8, 2025 at 5:27 AM
srewob.bsky.social
Via gpt-4o. Very abstract: on what looks like either thick piled carpet or possibly grainy concrete, a pattern of intermeshed curves, densely arranged. The background seems to be orange, layered with purple and cyan.
December 8, 2025 at 5:25 AM
srewob.bsky.social
#aiart prompt 1093: Bright, nuclear orange, overlaid with neon cyan and amethyst: an abstract pattern.
These four generated by Dall-E 3. This image is evocative of a floral pattern: a cyan colored flower in the lower right spirals out through shades of purple and orange in elegant organic swooshes with curlicues. Center-right, a star/flower in orange, with abstract, curved/triangular leaves beneath. Very ornate in appearance. Mostly orange in the upper right, while the left segues toward purple and cyan. Even more abstract. A core, centrally located nexus that might be floral, out of which reverberate numerous waved repetitions. Cyan and purple at the core, mostly orange with some purple left, bottom, and top right. A fractal-like pattern in the prompt colors. A large nexus on the lower left looks like nested concentric circles, some of which are orbited by beads of circles. Part of this set creates the illusion of a partial moon. Swooshes and curls spread out in repetition outward from this nexus.
December 8, 2025 at 5:23 AM
srewob.bsky.social
These much brighter images generated by Dall-E 3. They all have a bit of soft focus, and suggest a room the village is sitting in decorated with lights and garland. The first has a road with a large Christmas tree on an island on it. Shops line the street on the far side, with figurines loitering outside. Similar to the first, but at a slightly different angle. The trees are primarily diminutive wire brush conifers, about twice the size of a figurine. A larger, decorated tree sits at the end of a street in this image on the right. Fewer noticeable trees, although a slightly larger, undecorated tree stands in the center of the image in front of one of the store fronts. This resembles a bit of a European wall-to-wall high street, with a central island in the street on the right of the image, upon ‘em which sits an undecorated tree.
December 7, 2025 at 7:36 AM
srewob.bsky.social
Via gpt-4o. A bit more spartan than the first model. “Snow” is modeled on the ground, likely with felt. The figurines, oddly, don’t seem to have bases to stabilize them. While some out of focus buildings can be seen in the background, the focus is on a building with a sign reading “TOY VILLAGE”: a brick and mortar with a curved front window and a triangular prism roof.
December 7, 2025 at 7:30 AM
srewob.bsky.social
#aiart prompt 1092: A detailed miniature Christmas village, near a multi-shop plaza, full of ceramic figurines and wire brush trees. At street level, looking toward one of the shops. Low lighting, the buildings lit with bulb lights. Low depth of field.
Generated by MAI-Image-1. We’re at street level in a rectangular plaza, the buildings on the sides just visible; mostly, the plaza is defined by two rows or “conifer” wire brush trees. A number of chromed or possibly porcelain figures decorate the plaza. At the far end, what looks like two buildings side by side. As there is no air gap between them, in reality this would likely be a single set piece rather than two individual ceramic buildings. Rather dark; mostly lit from off scene, and from the bulbs in the buildings.
December 7, 2025 at 7:26 AM
srewob.bsky.social
Currently thinking about getting an 8bitdo Sn30 Pro for retro and cloud gaming on my iPad.

It looks like a very smexy controller: Hall effect thumbsticks, motion controls and rumble, reasonable suite of buttons, snes-style aesthetic.
A retro style gamepad, visually similar to a classic SNES controller: white with a black dpad, gray thumbsticks, and purple face buttons. It sort of resembles a dog bone.
December 6, 2025 at 7:25 AM
srewob.bsky.social
These generated by Dall-E 3. They are a bit more fanciful: ice as shards of translucent glass, overlapping, laying in the water. The background of this image has some mountainous terrain; a fjord, perhaps? Similar in structure to the first. The sun hangs low on the horizon, center top of the image. Thicker flows of ice, with more of a top down perspective; no noticeable land masses. The sun reflects off the water at the top. Mountainous land and sun in the background; the foreground is dominated by clear ice shattered into a complex web of shards still connected.
December 5, 2025 at 11:58 PM
srewob.bsky.social
Via MAI-Image-1. A photorealistic image of ice flows, densest in the center of the image, open water around. Some of the ice is submerged, giving cyan hues; some in open air, giving white. The water nearer the camera is dark blue, while further away is more blue-gray.
December 5, 2025 at 11:54 PM
srewob.bsky.social
#aiart prompt 1091: In white, cyan, steel gray, and navy: fractured ice on the water.
Generated by gpt-4o. A loose grid of irregular chunks of ice, with a nice gradient running from white along the left to dark blue on the right, transitioning through the other colors of the prompt.
December 5, 2025 at 11:51 PM
srewob.bsky.social
These aren’t me trying to sneak genitalia past their GAN or violence or drugs or anything.

The following image, rendered by the gpt model, had its prompt (in alt text) rejected by the other two models.
The dirt path wove through the forest, meandering near a nearby river it occasionally intersected; a dusting of frost tinged the remnants of fallen autumn leaves scattered across the path. A wildcat cautiously bent over a puddle on the path, lapping at the scant moisture contained therein.
December 5, 2025 at 8:16 AM
srewob.bsky.social
Via MAI-Image-1. A much subtler image: very little visual distortion, mostly in focus. The subject, a woman, is seemingly reaching toward her hair at her right temple. Her eyes have a peculiar digitization effect, as though they were replaced with two small LED screens displaying eyes, which have just gone on the fritz. Some of this distortion bleeds out onto her cheeks and nose, but mostly as colorful highlights.
December 5, 2025 at 5:01 AM