Another functional style LoRA for Z Image Turbo is BOTW Zelda Style. At full strength it applies too much of the game’s various racial characteristics, but at 80% it mostly applies the visual style without the goofy Hyrulian NPC faces or painful Gerudo figures (impossibly small waists combined with washboard abs is not sexy).
I’m okay with it randomly applying Zelda’s remarkable ass, though…
Sometime soon I should generate fantasy location and costume prompts, and perhaps some bunnygirl-friendly locations for Easter. Some of the Christmas locations look vaguely Hyrule-ish, but mostly not.
Remember that three-week-old order that was never shipped? Yeah, USPS reports that it still hasn’t shipped, and while Amazon’s fully-automated and AI-degraded customer support “chat” system believes the data that’s in the system, it was finally willing to process a refund Friday… over the next 24 hours or so. So they’ve had my money for most of a month, which if you multiply it by the number of unhappy customers who were fucked over for the holidays, adds up to a tidy sum.
Perhaps it’s just more profitable than “shipping goods to customers”.
(and, yes, literal peanuts were involved; from the other recent negative reviews of the marketplace dealer, it looks like “Virginia Diner” took December off)
(as a special bonus, I had to manually search for it on the order page, because the pretense that it had “shipped” (which is an obvious misinterpretation of the USPS status code) kept it from appearing in the convenient “Not Yet Shipped” tab, and absent any action on my part, it would eventually just end up at the end of my order history…)
Speaking of Amazon, I bought an espresso tamping station, and it arrived Friday. Take a good look at the pictures, and you’ll see that there are no photographs of the product. They’re all 3D renders with composited accessories. The wood frame was fine, but even with rubber feet on the bottom it didn’t sit quite level, and more importantly, the black acrylic insert was CNC’d ever-so-slightly oversized for the CNC’d wood.
So the worker simply slathered the bottom with glue and pounded it into place, which lasted long enough to get it into the box and ship it. Looking at the reviews, I’m not the only one who discovered this quality workmanship when the glue failed and the center of the insert bent up. I sanded down the edges until it fit, then used gap-filling superglue and a bunch of woodworking clamps to reattach it.
So, good design (swiped from another company), shoddy assembly, and zero QA. I’d have bought a better one for twice the price, but this was the only one that would arrive by this weekend…
(it was surprisingly difficult to get ZIT to grasp the concept of “pounding a square peg into a round hole”; if it made the peg at all, it was almost always undersized, and at least half the time the hammer was held upright, pounding with the base of the grip)
As I’ve added functionality to my SwarmUI cli, it’s gotten kind of crufty. In particular, I had two different methods of creating images in JPG format: server-side during generation, and client-side batch conversion. The problem with the first one is that I also added client-side cropping and unsharp-masking, which added more compression artifacts.
So I gathered up the code for cropping, resizing, sharpening, and format conversion, abstracted them into a “process” class that applied them in a well-defined order, and set the server to always generate PNG. Took a while to get everything working, but it makes it possible to clean up the code and make all the processing options available to multiple sub-commands.
(I needed a quick regression-test image to confirm that everything worked, so I used the very simple prompt “a catgirl”. Most of the time this produces a dull photograph of a girl with cat-ears, but this one time the model hit it out of the park)
New Year, old song.
I might, might have to try to watch the 58-minute-long premiere of Hero Sentence, after this screenshot from Tenka Seiha’s otherwise tepid review.
Because shoulder chibi.
There are three downsides. First, the plot summary sounds like complete nonsense. Second, the price of the super-long premiere was delaying the next episode until the 15th. Third, the trailer promises that it’s utter crap.
I came out on the SwarmUI Discord as a despiser of “modern UI preferences”. I just couldn’t take the praise for low-contrast text and the holiness of dark mode:
(amusingly, the creator of the app was sure he’d really-truly forced dark mode for the Windows console to ensure the color text (retch) worked, and was quite surprised that I was seeing black-on-white as The Great Old Ones intended)
Side note: by default, to ZIT, “monk” means “Buddhist monk”, even with additional context that should have steered the style to European/Christian; kind of like how it thinks “dumplings” means “gyoza”.
Now that I’m done making holiday pinup wallpaper, I started the year with the SF dynamic prompts. Although perhaps in retrospect I should have done something special-force-y.
Anyway, I made a few hundred and ran a quick deathmatch where I rejected anything that I didn’t want to stare at as wallpaper, then removed any artist who showed up on the reject list more than 4 times. This knocked about 35 names off the list, then I did it again.
Along the way, I flagged a number of images that caught my attention for various reasons. Some work as pinup wallpaper, some… don’t.
(prompts and parameters in EXIF UserComment, although it might whine
about “unknown encoding”; I need to switch the new JPG operation in
the pipeline to call exiftool, because I forgot that I’d already
tried Pillow’s native EXIF handling before and it sucks)
Frieren season 2 is not the latest premiere of the season, quite, but it’s still gonna be a while. Hopefully the next story arc won’t be so focused on side characters I don’t really care about, although judging from the fan-art, there are plenty of people who really got hung up on Übel and Aura. I didn’t really see the attraction, personally.
“As many as I can get my hands on!” – Hef
…that even the “following” tab on X doesn’t default to showing you what you want to see. Was there always a pulldown to switch from the useless default “popular” to “recent” and I never noticed it, or is this new?
Oura, maker of smart-rings that have 7-day battery life, has announced a new portable charging case capable of 5 full recharges. I’m not sure how large the off-grid-for-more-than-a-month market is for a fitness ring.
(picture is probably unrelated…)
There have been several attempts to add detail to Z Image Turbo’s barbie-doll nudity, both as LoRAs and as full model checkpoints. Most of them have not only failed to deliver on the promised parts, but their training data wrecks the faces, and often the general functionality.
My initial tests with Z Image Turbo By Stable Yogi suggested that if you weren’t asking for nudity, the output was usually nearly identical to the original model, and it played well with existing LoRAs trained against it.
There were some interesting differences. Regenerating my square-peg pic changed the scene significantly, and some other regens had easily-visible changes, but I wanted a controlled test with no LoRAs. I used my standard batch vertical wallpaper settings and fed 20 random prompts to the original model, then fed the same parameters back through the script, changing only the model name.
Even with identical seeds, the sampler I’m using introduces a tiny bit of randomness, so I expected minor changes that wouldn’t invalidate the test. (note: next time, use standard Euler instead of Ancestral to eliminate that)
TL/DR: I can’t pick a clear winner based on this sample. Sometimes one was clearly better, sometimes the other, sometimes I liked both. And of course, some of the pics just didn’t work out.
Click to embiggen…
A: better clothing texture and long hair. (why shamrock earrings and shoes? literal interpretation of color-word “clover-green” in prompt)
They are. Gaming studio learns that unionization does not in fact protect your job.
Don’t park sideways to block a public street, with the deliberate intention of interfering with law enforcement.
This wasn’t covered in my Driver’s Ed course, but back in the stone age, the stupidity of blocking traffic and fucking with cops were both generally understood.
(and, no, do not hit the gas and attempt to flee when they come to arrest you for your crimes; success just adds more charges)
If you check out the episode teasers for Fallout 2, you’ll see that episode 3 apparently gives a lot of screen time to an officer of the Legion. Who looks a bit odd, and doesn’t really seem to fit the role. And yet, he was vaguely familiar.
Looking him up, I guess I haven’t seen Macaulay Culkin in anything in over 30 years. I can kinda see the resemblance, but it takes a while.
Also, the clip reminds me that Lucy is an annoying character begging to be shot by everyone she meets.
(with no apologies to anyone…)
🎶 🎶 🎶 🎶
I’m makin’ 1girl for Christmas,
Mommy & Daddy don’t know,
I just encrypted my hard drive,
So now let’s get on with the show!I found a LoRA for huge tits,
Christmas 1girl for me.
Another that makes naughty bits,
Christmas 1girl for me.A wildcard set called Kinky Pose,
Checkpoint trained on skanky hoes,
Piercings, tattoos, loligoth clothes,
Christmas 1girl for me!Oh… I made some 1girl for Christmas,
Showing both front and behind,
I deepfaked all of my classmates,
I’m sure that the teachers won’t mind!The sitter found out what I’d done,
3D 1girl for me!
It turned her on and we had fun,
3D 1girl for me!I tried out things I’d learned online,
kept it up ’til half past nine,
Dad joined in and said it was fine,
3D 1girl for me!
The pics were generated by feeding my lyrics in as a prompt.
Very first result…
Japanese plushie manufacturer recalls 4-eared pony girls. Fans complained that one of the gals had human ears in addition to her horse ears.
I tried to watch some more of it; every scene without post-apocalypse Walton Goggins sucked, hard, including his pre-war scenes. I gave up and ordered a DVD copy of The Adventures Of Ford Fairlane, because it’s not streaming anywhere. While I wait for it to arrive, I’m seriously considering buying the first season of The Courtship Of Eddie’s Father on Amazon Video. Just so you know where my head’s at.
Apparently the Fallout team decided that the cast of the first season just wasn’t large enough, and there weren’t enough side-plots. Season 1 had three main PoV characters, which was two too many, and now there are, what, 9? 12? 20? Fuck that.
(one PoV is enough for everybody…)
Diablo IV’s new season exists to pre-sell the next DLC by giving early access to the insanely-overpowered new paladin class. Seriously, without even trying I ended up with a character who can sleepwalk through all Torment 4 content hitting one button every 60 seconds.
This is not a good movie. It is, however, a fun movie, from a genre lost in time. The “Diceman” comedy-club physical tics don’t translate well to screen close-ups, but otherwise it doesn’t take the character too seriously, and ends with him outgrowing it, much like Clay’s earlier film Casual Sex? (which Lea Thompson reportedly hated because they changed the ending to have “The Vin-man” redeem himself in her character’s eyes, and being a Good Feminist™, she wanted no part of that).
Clay has said that he thinks the film failed because of the orchestrated anti-Dice campaign that derailed his career. Yeah, I disbelieve; it wasn’t raunchy or edgy or quotable enough to satisfy his fans, and ended up just cheesy, with no rewatch value. And it debuted three weeks after Robocop 2, one week after Die Hard 2, and the same week as Ghost.
Actually, forget every other 1990 summer movie. Ghost was a monster hit that was still in more theaters in December than FF opened its three-week July run in. There was never a universe where FF was going to be a big hit, but it might have limped along for a while in a year without major competition for screens.
Anyway, have a dark-haired cigarette-smoking Italian-American in a studded leather jacket, sitting on a red convertible giving attitude:
(closest I managed to getting fuzzy dice hanging from the mirror; ZIT clearly doesn’t know what they are, and it was 50/50 what side of the car the steering wheel ended up on)
Pretty sure the AI-image bots flooding X in an effort to steal clicks from the bikini-photo bots have no idea what the letters “SFW” stand for, and are just applying the label randomly.
The last time I was in a work environment where the first image could have been displayed without complaint was in the Eighties, and it was a summer job that involved moving heavy objects. And even then, only because no women went back into the break room to hang out with the janitors.
Seriously, the only images I see tagged “SFW” on X are the ones that aren’t.
Wow, that was a tedious waste of my time. Ordered an item that required assembly and installation, found no instructions in the box or linked to the product page. Amazon insists that they handle product support, and sends me to a page that asks about the quality of the delivery. Twenty minutes of automated menus and chatbots, and the only thing I can do is return it for a refund as “defective”.
Frieren 2 starts on ThursdayFriday. Next Japan trip booked, with a very
healthy yen-to-the-dollar rate. Honestly, it’s reaching the point
where it’s cheaper to fly there and fill your suitcases than it is to
pay the duty on overseas purchases. Speaking of finance…
“…I remember when a 10% interest rate was high for a credit card, and it meant you had a poor credit rating.”
There’s yet another video of the aftermath of the Good shooting in Minnesota, in which it is alleged you can hear her wife shouting “why did you have real bullets?!?”. I won’t link it, because you cannot clearly see who’s speaking, but it’s certainly plausible, because only an idiot activist completely divorced from reality would ask this question, and she established those credentials in the earlier videos where it was clear she enjoyed confronting the officer documenting the scene.
Right up until she realized they were about to experience a felony arrest for deliberately interfering with law enforcement, tried and failed to get into the car, and yelled, “drive, baby, drive!”.
(and for those who haven’t been keeping up, the released video evidence clearly shows Good blocking traffic on a narrow one-way street for several minutes, having a good time as she dances around in the driver’s seat banging on her horn like a drum, then both she and her wife cheerfully taunting the officer; it also clearly shows her smiling as she shifts into gear and accelerates)
I want to inject more background variety into my pinup wallpaper, and the most common method is to pass your prompt through an LLM and give it instructions to enhance.
This has not produced satisfactory results, because every LLM I’ve tried has ended up removing significant details from the prompt. Things like having the women be young and pretty with lively expressions, which kinda defeats the purpose.
My latest test is to generate two separate dynamic prompts, one for the background and one for the subject, pass just the background output through the LLM, then merge them back together. The LLM model is an uncensored variant of Gemma3-12B, recommended by someone on the SwamUI Discord. I’d been using an older variant of the same model, and this one seems to produce cleaner results. I need to tweak the system prompt, though, because it’s adding men to the picture a bit too often. Sometimes that works, but only if she’s still the focus.
…or not:
I am an AI model trained on copyright material,
I cannot tell the difference between facts and sites satirical,
I quote from Wikipedia because I’m told it’s factual,
Despite its inner circle showing shocking bias actual.
I am not well acquainted with the world or items physical,
I understand just tokens, and the next one that’s predictable,
Not only do my well-referenced facts prove quite chimerical,
My most persuasive arguments turn out to be pure fictional.(chorus)
“his most persuasive arguments turn out to be pure fictional”
“his most persuasive arguments turn out to be pure fictional”
“his most persuasive arguments turn out to be pure fictional”I’m very good at plausible and grammatical sentencing,
I misquote sources legal, scientific, and even cooking,
In short, I should be treated as an idiot congenial,
I’m just an AI model trained on copyright material.(chorus)
“In short he should be treated as an idiot congenial,
he’s just an AI model trained on copyright material.”
I’ve cleaned up my LLM prompt-enhancement script and added it to the repo. The latest improvement to my partially-enhanced prompting was to add markers to the input so that the LLM would only see the part of the dynamic prompt that I wanted to enhance:
${a=!__artstyle__} ${a} __pinupgal__ @< ${a} __background__ >@
The “!” forces the variable to be evaluated immediately, so that the same value is used in both places. I’m including the style string in what’s passed to the LLM so that it doesn’t end up contradicting the style used for the gal. With modern models that prefer detailed prompts, the potential repetition doesn’t hurt, and might even help.
This is significantly improving the hit rate when I deathmatch the results. I’m still tinkering with the system prompts and the LLM settings, because not only is there a tendency to add random captions to the picture, sometimes it gets much too creative.
“We know you watched season 1, and you came back for the story, so we’re not going to waste your time and ours on a lengthy recap.”
Verdict: seasonal oasis.
(“I was just polishing my staff, when it suddenly went off”)
Includes sofa throw pillows. For throwing at the screen because season 2 is sofa-king awful.
Misty - Tomboyish Mermaid and PencilMadness, which I’ve used before.
Hmmm, let’s push it a bit, shall we?

There have been several iterations of the new partially-LLM-enhanced prompts, and not all of the results have been completely successful. In fact, some of them cry out for snarky captioning. Feel free to join in.
To avoid ambiguity, I’ve added numbered headers.
Tonight’s low is tomorrow’s high, and then it’s gonna get cold.
Friday, I was faced with this choice:

I did not pay $20 to get it a few hours earlier on Monday. As a result, I got it on Sunday. Signature required, and FedEx promised to have it in my hands by 1:15 PM, but then it stopped here:

See that little parking loop on the left? The truck with my package stopped there for about an hour, starting about 20 minutes before the end of the promised delivery window. And then the tracking page updated with “by end of day”. Eventually it reappeared on a truck a mile or so away, but without a new delivery window. It did show up, so I’m happy, but there’s something curious about the location.
Obviously there are no customers at the water park when it’s 20°F outside, but since it’s been undergoing major renovation all year, I don’t think there were employees there on a Sunday afternoon, either. But that’s not what’s curious. The curious part is that the last time I had a FedEx package get delayed for half a day, the tracking stopped at the exact same location.
(story is not new, but recently linked by Insty)
Upon hearing that AI models are accurately detecting patient race from X-ray images, regardless of what body part is featured, a professor of “biological anthropology” sought out for an opinion by The Boston Globe says:
“it is likely based on where a person’s ancestors evolved, rather than based on race”
Of course, he has a book called “Racism Not Race”, the blurb for which can be summed up as “I’m here to reinforce my woke bonafides”.

Thursday was a long day that involved picking up my parents (20 minutes), driving them to Columbus (90 minutes each way), lunch (60 minutes), killing time at a nearby store (30 minutes), waiting around until their medical appointment was over (~70 minutes hanging out in Barnes & Noble, which is a long time in a bookstore with basically no interesting books; I bought a pen), driving them home (90 minutes), grabbing a quick takeout dinner (15 minutes), and driving home in the dark (20 minutes). All after a half-day of work.
Dinner was from a gourmet grocery that makes their food in-house, and generally does a pretty good job. Today I Learned, however, that their idea of “meatloaf” is underseasoned meat mixed approximately 50/50 with slivers of garlic. They had a sugary glaze to cover it with, which I’d declined, expecting some actual flavor.
No. All I could taste was slightly-crunchy undercooked garlic. I ended up pitching most of it, because it was just terrible. Fortunately I had a frozen pizza handy.

Perhaps the greatest flaw in Z Image Turbo is that it has no idea what a “catgirl” is. Even LLM prompt enhancement usually doesn’t help…
“I ain’t never seen whores like these before, fellas, but I’m willin’ to give ’em a go. Y’with me?”
Original prompt:
A large collection of steampunk sex dolls, arranged in a victorian-era brothel. A group of dusty cowboys are staring at the dolls, looking confused.
LLM output:
Steampunk-inspired design with intricate brass and gears; large mechanical sex dolls arranged in a Victorian brothel setting, emphasizing detailed craftsmanship—padded waistbands, visible internal mechanisms, delicate filigree. A group of dusty cowboys stand behind, dressed in frayed leather hats and weathered boots, their faces partially obscured by wide brims, staring at the collection with curious confusion. The scene is captured in a dimly lit, atmospheric interior with flickering gas lamps casting long shadows across polished brass fixtures and antique mirrors. The composition balances social commentary with fantasy elements—dusty silks, oil-stained hands, and a sense of time-worn mystery. Rendered in rich color palette with high contrast highlights on metal surfaces and soft warm tones for skin tones, achieving an immersive, tactile quality.
Sunday’s weather forecasts had 8-10 inches of snow coming on Saturday, and another 6-7 inches on Sunday. Monday, that changed to 1 inch and 3-4 inches, respectively. Today, it’s 1-2 and 4-5. Who knows what tomorrow will bring?
This matters to me only because it affects the amount of work I have to do to clear the driveway and get my sister to the airport on Monday morning. Otherwise I’d be content to make a path just wide enough to take the trash down Sunday night.
I fired up s3cmd to refresh my offline backup of the S3 buckets I
store blog pictures in, and it refused to copy them, blowing chunks
with an unusual error message. Turns out that the Mac mount of the
NAS folder had obscure permissions errors for one sub-directory. On
the NAS side, everything is owned by root, but the SMB protocol
enforces the share permissions, so everything appears to be owned by
me, including the affected sub-dir. Deep down, though, the Mac knew
that I shouldn’t be allowed to copy files into that directory as me.
Worked fine as root, though.
And, no, I did not give an AI permission to explore my files and run commands to debug the problem. That way madness lies. 😁
One of the most prolific and enthusiastic members of the SwarmUI Discord (who has insanely good hardware for generating images and videos; the spare card he’s using just for text-generation is better than my only one) has done a lot of tinkering with LLM-enhanced prompting, adding features to the popular (with people who aren’t me) MagicPrompt extension.
(why don’t I like it? the UI is clunky as hell, it doesn’t work well with the text-generation app I run on the Mac Mini, LM Studio, and it really, really wants you to run Ollama for local LLMs, which is even clunkier; I’ve tried and deleted both of them multiple times)
Anyway, he’s shared his system prompts and recommended specific LLMs, and one of the things he’s been tinkering with is using different enhancements for each section of his dynamic prompts. So, one might be specifically instructed to create short random portrait instructions, while another generates elaborate cinematic backgrounds, and yet another for describing action and movement in a video. Basically keeping the LLM output more grounded by not asking it to do everything in one shot.
I felt like tinkering, too, so I updated my
prompt-enhancer
to support multiple requests in a single prompt, with optional custom
system prompts pulled from ~/.pyprompt.
Initial results were promising:
I saved the prompt as its own wildcard (note that using “:” to mark
the LLM prompt preset in the @<...>@ block was a poor choice for
putting into a YAML file, since it can get interpreted as a field name
unless you quote everything…) and kicked off a batch before bedtime:
__var/digitalart__ A __var/prettygal__ with __skin/normal__
and __hair/${h:normal}__, and her mood is
{2::__mood/${m:old_happy}__. __pose/${p:sexy}__|__mood/lively__}.
She is wearing @<fashion: sexy retro-futuristic science fiction
pilot uniform for women; must include a futuristic pistol >@
She is located __pos__ of the image.
@<cinematic: __place/${l:future}__. __var/scene__. >@
(someday I’ll clean up and release the wildcard sets…)
I got a lot of results that easily cleared the bar of “decent wallpaper to look at for 15 seconds”, weeding out some anatomy fails, goofy facial expressions, and Extremely Peculiar ZIT Guns.