fandom/personal blog: fandom stuff; personal stuff; etc.
TL;DR:
ME: ROZ; 25↑; 90's kid Sad Adult™
BLOG: MARVEL/AVENGERS (PRIMARILY STEVETONY). MDZS/WangXian. MXTX + 2ha. BATFAM/NIGHTWING. SPN 😔🤡 spn is now in the sideblog... mostly. OTHER STUFF.
My opinions are a constant work-in-progress.
if you'd like me to start tagging something, feel free to hit me up in my ask. Tagging is somewhat arbitrary these days but I'll at least try to remember.
also if you send an ask, lmk if you want it answered privately
New anti-AI tool from the team developing Glaze that looks to poison prompts and ruin data sets entirely.
The tool can go so far as to trick AI into generating things it didn’t ask for. 50 “poisoned” images of dogs into a data set yield typical errors companies have been working to get rid of: weird proportions, extra limbs, etc. When they added 300 samples, it was drastic enough to make a data set that was trained to generate “dogs” into “cats” making the tool basically useless.
It’s not out yet, but they’re adding it to Glaze and they intend on making it open source.
Welcome to the age of the AI arms race.
New Potion just dropped to poison your enemies
Love it. Thanks for posting OP!!!
More details:
A new tool lets artists add invisible changes to the pixels in their art before they upload it online so that if it’s scraped into an AI training set, it can cause the resulting model to break in chaotic and unpredictable ways.
The tool, called Nightshade, is intended as a way to fight back against AI companies that use artists’ work to train their models without the creator’s permission.Using it to “poison” this training data could damage future iterations of image-generating AI models, such as DALL-E, Midjourney, and Stable Diffusion, by rendering some of their outputs useless—dogs become cats, cars become cows, and so forth…
The team intends to integrate Nightshade into Glaze, and artists can choose whether they want to use the data-poisoning tool or not. The team is also making Nightshade open source, which would allow others to tinker with it and make their own versions. The more people use it and make their own versions of it, the more powerful the tool becomes, Zhao says. The data sets for large AI models can consist of billions of images, so the more poisoned images can be scraped into the model, the more damage the technique will cause…
Generative AI models are excellent at making connections between words, which helps the poison spread. Nightshade infects not only the word “dog” but all similar concepts, such as “puppy,” “husky,” and “wolf.” The poison attack also works on tangentially related images. For example, if the model scraped a poisoned image for the prompt “fantasy art,” the prompts “dragon” and “a castle in The Lord of the Rings” would similarly be manipulated into something else…
The research shows that vulnerabilities “don’t magically go away for these new [AI] models, and in fact only become more serious,” Kamath says. “This is especially true as these models become more powerful and people place more trust in them, since the stakes only rise over time.”
“Liminal” means “boundary” as in “the veil is thin here”
Liminal spaces are places that feel dreamlike, fuzzy on reality. Their boundaries are unstable. A bus station at night. A long desert highway. A bridge over a river.
They do, however, have an opposite; the Proximal space. These are places that feel far too real, extremely defined, boundaries that seem to override natural forces like time or space. Large warehouses without windows, identical chain stores at disorienting hours, your home after a nap of indeterminate time.
The traditional liminal space is the crossroads. Liminal spaces are places you travel through, but never stay. Motels in the middle of nowhere. Airports. Places like that.
The traditional proximal space is the monument. Anywhere that exists but isn’t really “used” for anything. Their presence resists use, closer to landscape than location.
“Duolingo is ‘sunsetting the development of the Welsh course’ (and many others)”.
I’ve used Duolingo since 2013. It used to be about genuinely learning languages and preserving endangered ones. It used to have a vibrant community and forum where users were listened to. It used to have volunteers that dedicated countless hours and even years to making the best courses they could while also trying to explain extremely nuanced and complex grammar in simple terms.
In the past two years it feels like Von Ahn let the money talk instead of focusing on the original goal.
No one truly had a humongous problem with the subscription tier for SuperDuolingo. We understood it: if you can afford to pay, help keep Duolingo free for those who couldn’t.
It started when the company went public. Volunteers were leaving courses they created because they warned of differing longterm goals compared to Duolingo’s as a company; not long after it was announced that the incubator (how volunteers were able to make courses in the first place) would be shut down. A year goes by and the forums—the voice of the users and the way people were able to share tips and explanations—is discontinued. A year or two later, Duolingo gets a completely new makeover—the Tree is gone and you don’t control what lesson you start with. With the disappearance of the Tree, all grammar notes and explanations for courses not in the Big 8 (consisting of the courses made before the incubator like Spanish/French/German/etc. and of the most popular courses like Japanese/Korean/Chinese/etc.) are removed with it. Were you learning Vietnamese and have no idea how honorifics work without the grammar notes? Shit outta luck bud. Were you learning Polish and have absolutely no clue how one of the declensions newly thrown at you functions? Suck it up. In a Reddit AMA, Von Ahn claims that the new design resulted in more users utilizing the app/site. How he claims that statistic? By counting how many people log into their Duolingo account, as if an entire app renovation wouldn’t cause an uptick in numbers to even see what the fuck just happened to the courses.
Von Ahn announces next in a Reddit AMA that no more language courses will be added from what there already is available. His reasoning? No one uses the unpopular language courses — along with how Duolingo will now be doing upkeep with the courses already in place. And here I am, currently looking on the Duolingo website how there are 1.8 million active learners for Irish, 284 thousand active learners for Navajo, and even 934 thousand active learners for fucking High Valyrian. But yea, no one uses them. Not like the entire Navajo Nation population is 399k members or anything, or like 1.8 million people isn’t 36% of the entire population of Ireland or anything.
And now this. What happened to the upkeep of current courses? Oh, Von Ahn only meant the popular ones that already have infinite resources. Got it. Duolingo used to be a serious foundational resource for languages with little resources while also adding the relief of gamification.
It pisses me off. It really does. This was not what Duolingo started out as. And yea, maybe I shouldn’t get invested in a dingy little app. But as someone who spent most of her adolescence immersed in language learning to the point where it was literally keeping me alive at one point, to the point where languages felt like my only friend as a tween, and to the point where friendships on the Duolingo forums with likeminded individuals my age and other enthusiasts who even sent me books in other languages for free because they wanted people to learn it, the evolution of Duolingo hits a bitter nerve within me.
THAT FABRIC IS NOT CALLING TO YOU! LEAVE IT ALONE!
boy it’s me the textiles speaking to you inside your head. you need the yarn. you need thread. your soul hungers to participate in the act of creation. you must feed it. you must buy so many beads.
