This post was deleted or flagged for the following reasons:
- [DELETION] Irrelevant to site (AI assisted/generated) - Lafcadio -
Description
I would like my apology. Oh and there is a video of this piece being made: https://www.youtube.com/watch?v=h6GY5Pbxod4 . Quality is a bit worse, but what you expect from a fast one done under an hour
furaffinity.net/view/61619730/
SolusInvictus
MemberWe can see the edge of the PureRef window at 0:43 in the video on the top of the second canvas, right when you move it away to stop tracing and you start actually moving around the canvas.
It was a nice attempt though.
user 2281675
Memberthe fuck you are on about?
SolusInvictus
MemberAt 0:42 you drag something out of the way of the canvas which also makes the recording glitch out, Pureref has an option to become transparent which allows to interact with the canvas below.
Exactly at that time you actually start moving the canvas around for the first time, rotating, flipping, zooming in and out, but at the same time the lines become sloppy and the anatomy starts falling apart, if I had to guess I'd say this is the point where you stopped tracing and decided to do the rest by yourself.
Your CSP looks like a fresh installation too, with every setting enabled, even the animation timeline below, the ruller, the asset folders, quite odd. I've never seen an artist use CSP at "factory settings" for so long.
user 2281675
MemberSo unless I wave the window all the time people like you won't be satisfied eh? Noted for the next stream
SolusInvictus
MemberNah, I am done with you, I know what you are.
Grab a pencil.
user 2281675
MemberIm using CSP since 2014, have EX PRO version and yes, when I went to version 4 it reset most of the settings. I can give you that.
But for the draft part it is easier for me to go about it with actual (physical) ruler for the base on the screen for some perspectives. The slopiness is because I actually start to make normal lines so they actually have to be in motion and cannot be choppy. Or I guess I can do a more choppy lines to make it more "unique" and have less stress about fixing it... BUT it would make painting harder as there wouldnt be enclosed area. I guess that is pros/cons of that approach
user 2281675
MemberI did - not even that is convincing for grifters like you. So I'm done with you too
hhrff
MemberThis seems like an AI witch hunt. Give the man back his art posts.
Lafcadio
JanitorThis artist is duplicitous and they will lie to you.
This constant fuss about getting an apology, and fighting with people in the comment section, is clearly intended to sway public opinion without privately addressing accusations. I have not once received a single message from them, despite having personally handled all of their deletions.
So, then, how do I know that they're tracing AI images?
Several posts show highly consistent details with other traces from AI-generated outputs, and inconsistent with the artist's prior submissions. There is also a noticeable jump in output that is highly unusual (in theory there could be a bunch of pieces that weren't shared or were intentionally deleted), and a casual review of the composition of pieces reveals yet more AI consistencies and non-AI inconsistencies: a lot of more "samey" poses versus prior works that show more variation, canvas sizes are suddenly very regular, etc., and after 4,913 deletions we have a really good idea of what standard dataset furries look like! delreason:*generate*
After their sudden reappearance in May 28th, a full 32 different Furaffinity submissions have either 1440x1024 or its rotation 1024x1440 as their resolutions, and 7 more are 1748x1181 instead. AI-generated images usually fit into highly regular resolutions since they start with a canvas, fill it in, and then never adjust the canvas size between/during generations.
This particular case is notable, however, as we have samples from this artist that predate AI-generated images! The newest 35 images before May 28th have 13 completely unique resolutions with no repeats, 10 are used only twice, there are five pieces at 1280x865, and only seven instances of 1748x1141. A lot of these completely precede high-quality furry image generation. It would be literally impossible for these to be traced from AI! And yet, throughout this entire year, johnnywolf21's uploads are very consistent in size from an artist who traditionally hasn't been very consistent.
The accused has multiple upload descriptions that playfully hint at using AI references.
https://www.furaffinity.net/view/61433576/
https://www.furaffinity.net/view/61289218/
I am strongly of the opinion that this artist will either modify these descriptions or delete these posts, so for the next month, you can download their posts and descriptions from furarchiver.net. You don't have to trust me: you can download Furarchiver's backup and have tangible proof that johnnywolf21 admitted to using AI references, even if johnny deletes the evidence!
The speedpaint is also recorded in such a way that parts of the operating system are hidden. On its own, this means that other windows and the taskbar aren't being drawn, but in addition, parts of the program are being drawn incorrectly around 0:43, where a window above CSP is being resized/relocated, creating a few rectangles that consist of pure grey or pure white. Also, for the first 14 and a half minutes (since this video is at 19x speed) of your recording, you literally do not erase anything, and the lineart revisions are minimal. You have a reference image you won't show to anybody.
I am extremely disappointed that you turned to these means to pad your portfolio. It seems like your prior 17 years of experience (your first DeviantArt upload is from 2008) have amounted to nothing. You should apologize for manipulating everybody while demanding an apology.
July 24th Addendum:
Johnny's reply (visible at the Furaffinity source) does not actually address any of the points I have made. The sudden change in style is not discussed. The sudden change in resolutions is not discussed (of course, now that I've revealed how the sausage is made, any future trends in resolution or aspect ratio are more or less pointless.) The technically issues related to having an overlay are not discussed (in fact, their proposed solution to flip and zoom the canvas more often would change literally nothing!) It exists to try and curry favor by making a bold show for the crowd: all style, no substance.
"Lots of BS reasoning"? Okay, then produce something substantial instead of just insisting I'm wrong. What's your excuse for admitting to the use of AI for references and then insisting that this current gallery isn't AI-assisted?
Seriously, how was Multi-tasking's description not the end of this farce!? You're literally admitting to the exact thing we deleted your gallery for! You literally gave us the most important evidence for free! We didn't even ask for it! I didn't even know you existed by the time that post was uploaded!
They repeatedly insist that they have a mysterious livestream with no documented proof that will solve everything; up until their recent falsified recording, their YouTube channel was completely devoid of any activity for 5 years, no Twitch channel is known to exist, and any private Discord stream might as well not exist if not intentionally recorded. They want Good Boy Points from being able to offer evidence, but will not actually provide any evidence.
If they were doing regular art streams, the proof they claim to have would already exist. There straight up is not an obvious link anywhere. Not on Furaffinity, nor Pixiv, nor YouTube, nor Twitter, nor Deviantart. All the proof seems to suggest that it doesn't exist. This is not the kind of behavior that an artist does if they want people to tune in to a stream and pay commissions. And yet, they insist that they have an audience: "I can even stop the audience engagement in-between so the CLIP never looses focus."
Fucking HOW?
Is this the world's first case of a public artist privately streaming for the illuminati? Some mysterious undefined crowd that can find that which was not meant to be found?
I know for a fact you can do better than this. You can draw without these shortcuts. You can draw without picking these arguments. Your gallery before the release of Stable Diffusion is completely undeniable proof of this. But it's completely obvious that you are not using the same workflow that produced images like https://twitter.com/thejohnnywolf21/status/1454507527505256448. You have entirely legitimate pre-2022 art pieces that you could reshare here and I would approve them without even a single complaint!
Updated
Chrno
MemberHoly fuck, there were 2 updated/edits to this post.
I weep for Your sanity at this point.... :((
Login to respond ยป