Ghoul (with memory loss)
LightNews LightNews
banner
kingsneak.bsky.social
Ghoul (with memory loss)
@kingsneak.bsky.social
300 followers 490 following 2.2K posts
I'm built different. Like incorrectly, I think.

Visual artist, punk/goth, she/he, 30. Mostly draw humans and some monster people for my sci-fi story. I work slow due to disability. I love werewolves, horror, gaming and rock music.
Posts Replies Media Videos
kingsneak.bsky.social
Saw superman when driving home from Hamburg
Photo taken out of a car window, you can see a blur of someone who looks like superman, clothing and all. There's some cars in the background as well.
November 13, 2025 at 11:53 AM
kingsneak.bsky.social
silly bonus picture where I scribbled a bra on him because a discord where I shared it doesn't allow any nipples at all
November 1, 2025 at 11:24 AM
kingsneak.bsky.social
I have a mild coordination disorder but my trick is to just draw lines really fast (and pressing undo a lot lol) so if you think you can't draw because your hands are shaky you're WRONG

the few slow shaky lines that make it into drawings are all done with love
lineart of my human guy character Captain, it was done slow instead of fast so it's a little wobbly
November 1, 2025 at 1:53 AM
kingsneak.bsky.social
13 hours 27 minutes scattered across 3 days

#art #werewolf #rebchallenge
In a dark alleyway behind some apartments is a snarling werewolf crouching atop a wooden plank, illuminated by pink light. Some rubbish is scattered around him.
November 1, 2025 at 12:24 AM
kingsneak.bsky.social
just a little guy #art
October 24, 2025 at 10:02 PM
kingsneak.bsky.social
Updated my #werewolf lore a bit and made a graphic. Inspired by the eyes of some sharks and rays!
Headline: "Eyes of lycanthropes". There is three graphics with descriptions that describe the context of the graphics. The first one is a graphic of an eye, with all parts named on it. "Figure 1: The eyes of a lycanthrope remain mostly human in anatomy, but develop an iridescent tissue layer behind the retina called the tapetum lucidum. This layer reflects incoming light back through the retina, improving vision in low-light environments by making it pass through the photoreceptors a second time." The second graphic consists of three stages that show pigment travelling up through a blue layer and forming a dark layer over it. "Figure 2: In bright environments, the tapetum lucidum worsens vision and makes the eyes more light sensitive. To avoid that, molecules of dark pigment migrate through the tapetum lucidum, obscuring it and retaining human vision in bright environments." The third image shows two eyes, one in a bright environment, it looks human with a blue iris and a constricted pupil. The other eye is in a dark environment, the iris is the same size and colour but the pupil is dilated, iridescent in white, light blue and light turquoise, and reflects light. "Figure 3: When the tapetum lucidum is unobscured, lights cause a visible reflection known as eyeshine. Due to the iridiscent nature of the tapetum lucidum, the reflection can vary in colour and change according to angle. The most commonly observed colours in lycanthropes are white, blue, green, red and yellow. The colours are determined by environment and genetics, but generally stay the same through life."
October 21, 2025 at 9:05 PM
kingsneak.bsky.social
Image of the "Panik" meme, text on the left and an uncanny looking 3D bald head with no ears on the right along with a misspelled word to represent the reaction to each statement. Panel 1: "I'm meeting one of my favourite bands ever next month already" - There's hands on the head looking distressed. "Panik" Panel 2: Its gonna be okay I've been friends with the singer for years and everyone has been nice - Head is content, "Kalm" Panel 3: I have social anxiety - Hands on head looking distressed and the image is blurry, "Panik"
October 5, 2025 at 2:30 PM
kingsneak.bsky.social
2/2
User SimplifyAndAddCoffee responds: "Large Language Models and other Deep Learning systems are definitely problematic and pose a pretty immediate threat to the functioning of our society, just not at all for the reasons that people expected, or indeed still largely believe, and I think it is extremely important that we understand how and why the nature of the threat differs from what is being communicated and amplified by tech people and the media, if we are going to have any hope of adapting to the very real challenges it will present us. Ironically, if Al destroys us all, it's not because we underestimated it's intelligence, but because we overestimated it. We can't leave decision making to Al, not because it's too smart and will realize it doesn't need us, but because it's too dumb, and we're too dumb (and desperate for easy answers) to realize it's judgment shouldn't be trusted."
October 5, 2025 at 1:59 PM
kingsneak.bsky.social
all alt text for the above pictures because the word limit was being too shitty to fit it there 1/2
Screenshot of Reddit comments made by user SimplifyAndAddCoffee: "Well for starters we don't have real A.I. We have generative models trained on massive amounts of data which in effect attempts to compress it down into a trained model which it can run to try and regenerate answers based on the data it was trained from. It is a lossy compression, as the model itself is too small to contain the whole of the information it ingests. As such it makes things up along the way in order to fill in the blanks. You can see this in how chat models like chatgpt etc will confidently give you incorrect information. Researchers call this "hallucinating". The model doesn't actually have any core understanding of the material it ingests--it can't, since it isn't actually an artificial intelligence. It can infer what things should look like, and it can do so well enough now to start fooling humans into thinking it knows what it's doing. We're in the 'uncanny valley' of generative language and code models. So that's one problem. It makes things up without understanding it, and can't reliably reproduce correct answers, only things that look correct... in the case of code, it looks correct enough to compile, and that a programmer skimming over it might not even notice that it's actually full of bugs. The other problem is that the data it is trained on is only--at best--as good as the people who wrote the code it's ingesting. And because training the model requires billions and billions of code snippets to train on and learn as data points, the only way we can conceivably feed it that volume of code is to download it en-masse from the internet. Github, stackexchange, etc... and none of that code is reviewed and checked for bugs or security flaws before the model is fed it. We don't have the resources to professionally curate and double check everything it learns, and so it's going to get a lot of shitty code examples from shitty programmers, and the only thing keeping it from spitting out the absolute worst code out there is the law of averages relating to the volume of code it ingests... but if you ask it to do something specific and the most specific examples it has in its "memory" are from shitty coders? you get shitty code. Garbage in, Garbage out. It's absolutely infuriating to people who actually understand the technology that we've taken to calling it "Al" at all. It's a stupid techbro marketing stunt and unfortunately for all of us it has stuck, and as a result we now all have to call it Al, and only those of us with the right tech background to know better will understand just how misleading that label is. As a result it's really easy for any shitty startup to whip up a model based on stolen, buggy code, and market it to everyone from IBM to law enforcement as this smart intelligent infallible system that it's not. The layman has entirely too much faith in computers to be better than humans... after all, they are way faster and more accurate at making spreadsheets etc, right? But those kinds of programs are an entirely different animal. If you ask a chatGPT model to make a spreadsheet and give it numbers and labels and tell it specifically what you want it to do, it will spit one out that looks to you like it's about right, and you might then accept and pass it off as valid data, but if you check the numbers on it, you'll find out they don't actually add up at all. Because it doesn't understand math. It only knows what it looks like. It's an express process for lazy people to poison the waters with buggy unchecked code en-masse... it's like if you took your average CSI episode where the producers show a hacker typing nonsense on a keyboard to look like they're doing something, except it's way more complex to the point where the target audience is programmers and code auditors and it's designed to look authentic to them instead of just the layman. It's still garbage, but it's dangerously believable garbage. Remember all those shitty chat bots that circulated around for a while? This is just that, but way more complex and easier to mistake for real intelligence. Imagine now, if you will, an internet full of such chat bots all set up by techbros and lazy hacks trying to cash in on the sudden easy ability to generate 'content' that can get past regular spam filters at a rate so fast that no human team can keep up with checking it all, and they're pulling this stuff down from the internet en masse to train their buggy models, then submitting it back to places that are indexed online where the next set of buggy models can ingest it, like an infinite Ouroboros of shit, so next thing you know you can't trust a damn thing you read anywhere, because it's all garbage generated from other people's garbage, and companies like IBM and Microsoft+ are even getting in on it. And because the models learn based on statistical trends and averages over a large set of data? guess what, this huge flood of new "Al" generated data is now the norm, and as such it takes precedence over human generated data that by natural limitations cannot keep up with the speed at which the Al generated data is flooding the internet, especially as skilled programmers are getting laid off in favor of code generative models that can produce 100x the volume of shitty buggy code for 1/10th the cost. That's basically what's happening now. Because the average person making decisions about how to leverage this new, lucrative technology for profit, doesn't understand, or care to understand, how it works or why it's a bad idea. All they see is the short term dollar signs from getting leg up on the competition by churning huge quantities of shit out faster and cheaper than any human can, in a market where increasingly only quantity matters--not quality. It's already replacing journalists and authors as newspapers and publishing houses are getting backed up with a flood of "Al" generated submissions from people trying to cash in on it. A huge amount of recent content on the internet is entirely made up, imagined by these models, and very difficult to tell apart from actual researched information by real knowledgeable experts. Throwing this into the mix with the already problematic ecosystem of disinformation from entities like Cambridge Analytica, and even writing children's books to help human children learn to read? The future is very bleak indeed. Note: the article [the comment was a response to] here is misleading, and IBM is not yet looking to replace programmers with it...yet. But believe me, this is coming, and the code ecosystem of the internet is already being massively poisoned by smaller actors who are leveraging this tech" User cultivated-mass responds: "Wow that was very informative and explained well. Thank you for taking the time to write this. It weirdly makes me feel a little better about it because it seems like every article I see now is just about the end of days because of Al. I agree future still seems bleak but not for the reasons I was thinking."
October 5, 2025 at 1:58 PM
kingsneak.bsky.social
I am reminded of this Reddit thread
October 5, 2025 at 1:36 PM
kingsneak.bsky.social
thank you for the chance!
October 4, 2025 at 12:40 PM
kingsneak.bsky.social
Phronima species, also called barrel shrimp. They are so badass looking!
September 24, 2025 at 9:29 PM
kingsneak.bsky.social
Is this what it feels like to have dyslexia
Photo of the back of the self titled album/CD by A Frames, it's white letters on black background and lists all the song titles, but with random spacing between the letters making it hard to read
September 17, 2025 at 3:49 PM
kingsneak.bsky.social
🔄
My tuxedo cat lying cute with eyes closed on a red armchair, front paws and head facing the viewer while the rest of her body is hidden by the armrest My cat mid-spin on the armchair My cat lying in the same pose as at the start on the armchair, looking into the distance. Life sure is hard
September 7, 2025 at 5:16 PM
kingsneak.bsky.social
I've been seeing this phenomenon a lot as well. they've either got to be bots or lack all self-awareness
a drawing of a bluesky post, user Dude McFarty says "I agree, AI is so awful!" but he has an AI avatar and has been flagged by the AI bot for it.
September 2, 2025 at 9:41 PM
kingsneak.bsky.social
cookie doodle

#art
a doodle of my satyr character having a cookie
September 1, 2025 at 11:49 PM
kingsneak.bsky.social
Tried to do some simple pixel art.

@shitrock.bsky.social

#art #fanart #pixelart #mclusky
Pixel art, the mclusky band logo on black and white at the top and below it simplistic colourful portraits of the band members, under the portraits are their names.
August 17, 2025 at 10:44 PM
kingsneak.bsky.social
my art / my inspiration

my pile of mclusky fanart is alarmingly big
August 15, 2025 at 10:55 PM
kingsneak.bsky.social
me: this is childish why am I still here
also me:
a brown chickensmoothie dog adoptable ("growing boar") with accessoires and speech bubble items that are layered in such a way that they spell out "Let's hang Santa! LMAO"
August 15, 2025 at 12:03 AM
kingsneak.bsky.social
Testing these motion sickness glasses, heard they had good reviews. I tend to get super sick on car rides and am trying anything at this point
Packaging of the motion sickness glasses by Hion, claiming to be effective in 95% of cases Black glasses, part of the frame is transparent and partially filled with purple liquid. The liquid is supposed to create an artificial horizon for your eyes to perceive.
August 11, 2025 at 1:08 PM
kingsneak.bsky.social
deserved tbh #baldursgate3
A silly screenshot of the game. The characters are at the Risen Road Toll House. In the front the corpse of Cyrel is hanging off some kinda wooden bridge thingy, back facing the camera, head down and arms stretched to the side looking silly. Behind it is the party, from left to right: my player character, a drow bard, standing and pointing straight ahead, Gale crouching and leaning to the side a bit looking slightly baffled, Shovel waving her hand in the background, Karlach down on one knee smiling and showing off her muscles, and Lae'zel standing there looking at the scene rather seriously.
August 11, 2025 at 11:42 AM
kingsneak.bsky.social
doodles
smol and tol

#art
sketch of a little guy frighteningly tall guy, little spooked person for scale
August 9, 2025 at 2:00 PM
kingsneak.bsky.social
this reminds me of that pseudoscorpion photo I took years ago
A pseudoscorpion, which is a very tiny animal that looks similar to a scorpion but lacks the tail and venomous stinger real scorpions have. This photo makes it look as if it's wearing a little high heel on one of its feet 👠
August 7, 2025 at 10:49 PM
kingsneak.bsky.social
also a little work in progress concept #art on Captain's #werewolf form. realised I like some more human traits in werewolves.

I've been struggling finding a pose for his new ref so he's just kinda T/A-posing here, used a free 3D werewolf model as base just so I have something to project ideas onto
a pitch black werewolf with long pointy ears, white eyes, and some design notes around it - its nose is not canine, but more humanoid, the teeth are a mix between canine and human as well, it has no pawpads, and 5 toes on its feet all of which are touching the ground.
August 6, 2025 at 9:05 PM
kingsneak.bsky.social
just a small doodle from the other day of a character I need to redesign a little #art toyhou.se/1893143.finch
headshot of a guy with pointy ears, a big nose, freckles lighter than his skin tone, lashes, big black eyebrows, and light blue fluffy hair and fur on his neck. his right eye is white with a black pupil, his left eye a light cyan with a white pupil.
August 6, 2025 at 7:47 PM