- Joined
- Dec 3, 2018
Doing a mass archival atm so I kinda need something in the background. Thank you lol
Furry Fandom and Drama General
- Thread starter Holdek
- Start date
CIA Nigger
Not a fed, just a random object on the street.
Retired Staff
True & Honest Fan
kiwifarms.net
- Joined
- May 15, 2017
So putting the thread shitting aside, here's a followup to the rhetorical question of how original and commoditized fursonas are.
Lately the craze with techies working on toy projects has been the "This X does not exist" form of website, including This Person Does Not Exist, This Cat Does Not Exist, This Article Does Not Exist, and even the weeb version This Waifu Does Not Exist. They're pretty much the typical deep machine learning content generators, an AI analyzes thousands of images of something and generates something based on these images.
Anyhow it's well documented that there are plenty of furries in the tech industry, so a furry decided to create the latest one of these. This Fursona Does Not Exist was trained on thousands of safe for work images from e621 (I'm just as shocked as you are that there's that many on it) to generate new furry images.
As usual with this genre of site there are plenty of disturbingly badly moshed images (this is evident on images it tries to make based on famous characters). But it's being seen as interesting by many furries:
Some furries are throwing a tantrum over it and thinking it's wide art theft for kids stealing fursonas, while others are worrying about the implications of this new technology. Then there's furries that think it'll reduce kids stealing icons instead since they can "generate one instead".
archive.fo
This furry is having a huge ass meltdown:
archive.fo
Artist asked about if the creator got permission:
archive.fo
Another furry asks how long before this is used by trolls:
archive.fo
There's going to be an interesting response from some furries as time goes on.
Lately the craze with techies working on toy projects has been the "This X does not exist" form of website, including This Person Does Not Exist, This Cat Does Not Exist, This Article Does Not Exist, and even the weeb version This Waifu Does Not Exist. They're pretty much the typical deep machine learning content generators, an AI analyzes thousands of images of something and generates something based on these images.
Anyhow it's well documented that there are plenty of furries in the tech industry, so a furry decided to create the latest one of these. This Fursona Does Not Exist was trained on thousands of safe for work images from e621 (I'm just as shocked as you are that there's that many on it) to generate new furry images.
As usual with this genre of site there are plenty of disturbingly badly moshed images (this is evident on images it tries to make based on famous characters). But it's being seen as interesting by many furries:
Some furries are throwing a tantrum over it and thinking it's wide art theft for kids stealing fursonas, while others are worrying about the implications of this new technology. Then there's furries that think it'll reduce kids stealing icons instead since they can "generate one instead".
Retro-beep (tm) on Twitter: "So i just found this https://t.co/L2UiQQ…
archived 6 May 2020 20:25:07 UTC
This furry is having a huge ass meltdown:
Social Distance Warrior Arvu on Twitter: "Hey uh... Am I not getting …
archived 6 May 2020 20:26:54 UTC
Artist asked about if the creator got permission:
King of Tails on Twitter: "I don't know how this site works exactly b…
archived 6 May 2020 20:29:51 UTC
Another furry asks how long before this is used by trolls:
Nystre (20XL) on Twitter: "how long until This Fursona Does Not Exist…
archived 6 May 2020 20:30:41 UTC
There's going to be an interesting response from some furries as time goes on.
- Joined
- Jun 20, 2019
If anything, an AI which digests and recycles thousands of fursonas will be more original than the average fur artist
Welcome to the 20s
Welcome to the 20s
- Joined
- May 10, 2019
- Joined
- Dec 28, 2014
Cry some more furfag
GrayWater
kiwifarms.net
- Joined
- Mar 14, 2017
This furry is having a huge ass meltdown:
View attachment 1275089View attachment 1275099View attachment 1275100View attachment 1275101
Social Distance Warrior Arvu on Twitter: "Hey uh... Am I not getting …
archived 6 May 2020 20:26:54 UTC
- Joined
- Nov 1, 2016
These twits are sorely misguided about it being a vehicle for art theft, since most of the avatars are warped in weird ways and a 'thief' could just run a hue slider and some Photoshop filters to get better results.
Arvu is still arguing with the creator as of writing because he thinks it is a literal face crop
Arvu is still arguing with the creator as of writing because he thinks it is a literal face crop
CIA Nigger
Not a fed, just a random object on the street.
Retired Staff
True & Honest Fan
kiwifarms.net
- Joined
- May 15, 2017
Furry still doesn't understand fair use
archive.fo
lol he's interacting with Carcin:
archive.fo
Social Distance Warrior Arvu on Twitter: "I don't give a damn about t…
archived 6 May 2020 21:33:45 UTC
lol he's interacting with Carcin:
Social Distance Warrior Arvu on Twitter: "I wish someone would just p…
archived 6 May 2020 21:38:58 UTC
Leonard Helplessness
kiwifarms.net
- Joined
- Nov 28, 2018
Would an image generating AI like this one work even better if it was trained specifically on the work of one artist who’s created thousands of pictures that all look pretty much the same? I don’t know the technical underpinnings of AIs like this but it‘d be hilarious if somebody brewed up a Marci-Badge-O-Matic.
- Joined
- May 26, 2019
I mean, potentially? You'd have to start quantifying the trends between given lines and the deviations therein. Definitely plausible, but IDK if turning a Machine Learning algorithm loose on the color pallettes would end up any better than leaving them up to the furries using it.
ExsanguinateHorizon
kiwifarms.net
- Joined
- Dec 16, 2019
Imagine being so fucking devoid of identity you throw a spastic fit over your 100% unique avatar merely being analyzed. I get that people are naturally protective over their stuff, but man do they expose their ignorance and insecurity. Furry designs are rarely unique or stand out, shit this might even give people ideas on how to improve their own designs.
Wow never seen this design before, definitely the only one to have existed and ever will exist.
Wow never seen this design before, definitely the only one to have existed and ever will exist.
- Joined
- Apr 25, 2020
- Joined
- Dec 28, 2014
They're just mad there's any possibility at all that they won't be able to make enormous amounts of money from the shittiest and most degenerate fandom in the world by doing paint-by-numbers bullshit with absolutely zero creativity to it any more.
It somehow decreases the value of the shit they bought that shit that is essentially randomly generated is of comparable quality.
Leonard Helplessness
kiwifarms.net
- Joined
- Nov 28, 2018
That's a good point. A friend of mine brought up another important point about this: Training an AI exclusively on Marci babyfur pictures is objectively abusive and would probably lead to AI suicide.
- Joined
- Apr 11, 2018
I was planning on reporting on Kavaeric, myself!Furry still doesn't understand fair use
View attachment 1275389
Social Distance Warrior Arvu on Twitter: "I don't give a damn about t…
archived 6 May 2020 21:33:45 UTC
lol he's interacting with Carcin:
View attachment 1275405
Social Distance Warrior Arvu on Twitter: "I wish someone would just p…
archived 6 May 2020 21:38:58 UTC
He seems to be upset over FurAffinity, whic has a 'No Nazis' clause in its TOS, deleting his ubmission that has omeone killing a Nazi.
- Joined
- Sep 29, 2018
Interesting note on Kavaeric which I don't remember (and found no evidence of in searching of this thread) being brought up: He was fully brought forward into the public eye for a near literal five minutes of fame by the National Post and even jerked off by Neil Gaiman for a long-winded rambling tweet thread about living in Canada. See here:
+ https://archive.li/5puxi
The comments on that article are interesting to say the least. I remember this being sent to me at one point and wondering why it was such a big deal what some furry on the internet says about Canada. Slow news day I guess.
- Joined
- Jun 11, 2016
So putting the thread shitting aside, here's a followup to the rhetorical question of how original and commoditized fursonas are.
Lately the craze with techies working on toy projects has been the "This X does not exist" form of website, including This Person Does Not Exist, This Cat Does Not Exist, This Article Does Not Exist, and even the weeb version This Waifu Does Not Exist. They're pretty much the typical deep machine learning content generators, an AI analyzes thousands of images of something and generates something based on these images.
Anyhow it's well documented that there are plenty of furries in the tech industry, so a furry decided to create the latest one of these. This Fursona Does Not Exist was trained on thousands of safe for work images from e621 (I'm just as shocked as you are that there's that many on it) to generate new furry images.
View attachment 1275067
As usual with this genre of site there are plenty of disturbingly badly moshed images (this is evident on images it tries to make based on famous characters). But it's being seen as interesting by many furries:
Some furries are throwing a tantrum over it and thinking it's wide art theft for kids stealing fursonas, while others are worrying about the implications of this new technology. Then there's furries that think it'll reduce kids stealing icons instead since they can "generate one instead".
View attachment 1275078
Retro-beep (tm) on Twitter: "So i just found this https://t.co/L2UiQQ…
archived 6 May 2020 20:25:07 UTC
This furry is having a huge ass meltdown:
View attachment 1275089View attachment 1275099View attachment 1275100View attachment 1275101
Social Distance Warrior Arvu on Twitter: "Hey uh... Am I not getting …
archived 6 May 2020 20:26:54 UTC
Artist asked about if the creator got permission:
View attachment 1275108
King of Tails on Twitter: "I don't know how this site works exactly b…
archived 6 May 2020 20:29:51 UTC
Another furry asks how long before this is used by trolls:
View attachment 1275111
Nystre (20XL) on Twitter: "how long until This Fursona Does Not Exist…
archived 6 May 2020 20:30:41 UTC
There's going to be an interesting response from some furries as time goes on.
Every “fursona” that I have ever seen has just been a recolored dog/fox/generic dragon of some kind. I still don’t understand how some of these mongoloids get so worked up about the whole “OC DO NOT STEAL” thing when all of their fursonas are so similar.
- Joined
- Dec 28, 2014
It triggers them because it makes it really obvious how cookie cutter all their bullshit is. A bot could do as good a job as you paid some mongoloid "artist" to do.
- Joined
- Apr 25, 2020
And yet some artists make a killing selling entire sheets of "adoptables" that are basically just the same lineart with different colors slapped on them. And people are proud to show off the zero-effort art they got for upwards of 50 bucks.
Who said furries have to make sense?
- Joined
- Oct 5, 2018
I have to admit, I'm curious about the whole neural network fursona generator thing because I'm not super familiar with how those programs work and I want to know how it goes about what it's doing.
Could anyone who understands the process of neural network stuff better than me (so, anyone) help explain how this works? From an art perspective and as a total idiot when it comes to tech I'd legit love to know how that works, do they divide the input images into categories, run the results, and then jam the end results together? I know people were talking about seeing repeats so is it preloaded images rather than on the spot generation, or did someone just troll the entire furry fandom by cropping and filtering a bunch of fap material?
I'm mostly curious because art isn't the same as feeding it photographs because it's working with a bunch of images using a different... language, I guess? With photos of people, there's all sorts of variation in color and feature and so on, but it's all photos; the general style of how forms exist in space is the same, the way light acts on the surfaces is more or less the same, etc etc.
With these fursonas, if the output is any indication then the source images are absolutely all over the place in terms of rendering. There's absolutely flat lineart stuff with minimal cel shading, there's very cartoony disney smooth-rendered 3d, there's very painterly stuff with a much more realistic look. Here's a screenshot for an example, you can see a bunch of different 'styles' and they're all pretty distinct.
I could totally see a program spitting out an ungodly mishmash of these different styles because it's treating things like [smooth, pleasing rounded shapes of ears rendered in x style] and [outlined spiky anime hair with no shading in y style] as individual features, but what's wierd to me is that even when they're comically a little skewed, almost all of the generated images pick one style and are consistent in applying it to the design? And it's good at it too, especially for complex structures like more realistically rendered hair or fur where it's maintaining the light source and application.
With these fursonas, if the output is any indication then the source images are absolutely all over the place in terms of rendering. There's absolutely flat lineart stuff with minimal cel shading, there's very cartoony disney smooth-rendered 3d, there's very painterly stuff with a much more realistic look. Here's a screenshot for an example, you can see a bunch of different 'styles' and they're all pretty distinct.
I could totally see a program spitting out an ungodly mishmash of these different styles because it's treating things like [smooth, pleasing rounded shapes of ears rendered in x style] and [outlined spiky anime hair with no shading in y style] as individual features, but what's wierd to me is that even when they're comically a little skewed, almost all of the generated images pick one style and are consistent in applying it to the design? And it's good at it too, especially for complex structures like more realistically rendered hair or fur where it's maintaining the light source and application.
Could anyone who understands the process of neural network stuff better than me (so, anyone) help explain how this works? From an art perspective and as a total idiot when it comes to tech I'd legit love to know how that works, do they divide the input images into categories, run the results, and then jam the end results together? I know people were talking about seeing repeats so is it preloaded images rather than on the spot generation, or did someone just troll the entire furry fandom by cropping and filtering a bunch of fap material?