zaq hack

I've been posting a lot about AI, lately. Here's my perspective ... I use AI every day. I have several of my own personal agents that do work on a variety of systems. My "day job" deals with AI in enterprise systems. I have been building workflows with machine learning for 8+ years. I am an expert in "applied AI," by which I mean using various algorithms/models/et al to do actual work. Not theoretical stuff, but things that companies are actually willing to pay for and that work reliably day-after-day. "But you post anti-AI stuff all the time." YES! Because AI is good at some things and terrible at others. Because certain AI models were made with stolen (and I do mean stolen, as in "trained with copyrighted material that was neither purchased nor duplicated as 'fair use' consideration"). There are copyrighted books in the major AI models, today. There is copyrighted art in image generation models. I want the creators of those things to be compensated for the reproduction of their work. Anyone else would get sued, lose the lawsuit terribly in court, and have to pay. That AI companies are skating by on this issue is completely ridiculous. We need new regulations for the age of AI. In the 1990's, we came up with rules for "sampling music" and using it in a new song. If it is greater than 3 seconds, you have to pay the royalties. The music industry has dealt with parody, reproducing fragments, and rampant piracy not that long ago (although I'm probably double the average X user's age). It seems to me we should be able to set up rules for the use of copyrighted works in AI that look similar in appropriate ways. It's not exactly the same problem, but if we want to value intellectual effort in our society, then we should value intellectual property. If you don't value that, then maybe the algorithms are already doing most of your thinking for you.

@RevyDown @miroyato @VakivaR18 @AyakaMods @ProtonDrive @brave Same people think Android is Linux.

@miroyato @VakivaR18 @AyakaMods @ProtonDrive @brave Although I would go further and say to save the copies of your work on your own drive if they are important that you don't want to lose it. Downside is that drives can still go bad. So you occasionally have to keep on it.

Google just permanently banned a manga artist’s entire Google account, just for uploading his own old manga files to Drive. AI moderation triggered and flagged it, he tried to submit appeal then he got rejected it by Google and now he has lost everything like Gmail, Drive, all linked services is gone. He never even sharing the files publicly, it’s only backing up his own a private work like any creator and artists. This is Google Drive “AI moderation” in action. No human support and no serious to take action. Physical storage or real private alternatives only. Support the artists getting screwed by this. This level of corporate overreach is insane.

My logic is settled law. You are drawing a false equivalence between AI and human learning. The flaw is simple: You can make copies of and distribute AI models easily in digital space, or sell access to the outputs of said models. How many copies of Van Gogh's or Mozart's brains can you make and distribute for money without their consent? How many modern author's or coder's brains are you copying without their consent? You can't. And you can't copy their output without stealing ir consent. Those are the options for trained humans. AI presents a different option for avoiding copyrights or patents

@zaq_hack @thefeldarrian So what's the difference from an artist learning to draw from copyrighted material? They didn't outright steal it, but they rained on it? If your logic cannot be applied equally it is flawed. Your logic, applied equally, invalidates almost all art.

If AI is theft then every artist who has ever used another person's work as inspiration or as a reference is also a thief.

1. That's not at all what I've said. 2. If we have the model, we can see which weights are active compared to training data. That's the only way to know. 3. The violation happens in training. You didn't steal the merchandise, but it was given to you to use with the stolen data embedded.

@zaq_hack @thefeldarrian So what did I steal for this?

@juiceetin Moral: Don't piss mom off ...

same character btw

@relykscarving @thefeldarrian Oh, and additionally, if I really require (some copyrighted material) for use in my next project, I can just pay the royalty. Like "All Summer Long" does for "Werwolves of London." Since I mostly work in code, I can use licensed code -- if I pay for the license.

@zaq_hack @thefeldarrian If your concern is AI using copyrighted material, then you should also be against all artist using copyrighted material as both training and inspiration - even if said material or idea was altered afterword. But that would invalidate the majority of art.

@relykscarving @thefeldarrian 4. Creators that can prove copyrighted material was used in the model training data amd can be reproduced by the model ("in part" at the beginning, but with mamba and other tech, increasingly "whole"). Or maybe we just shouldn't have books or art or code, anymore ...

@thefeldarrian How I see it - there are three types of people who hate AI. I used to be of the first. Ignorant - Does not know anything about AI, hears that it is bad. Incompetent - sees AI as a threat, unwilling to overcome Grifter - sees the common thing, hoping for clout.

@thefeldarrian Ever copyright anything? Should the AI companies get to copy it all for free? Is shoplifting a book theft? There are multiple lawsuits that have proven the theft of copyrighted material for training data. Should nothing happen? No rules for AI that apply to everyone else?

Finally. Of course, I just leveled both of my teams up to 60, so I have zero materials for her ... 🙃

@KuroSekaiAnime It's not a recession ...

Yall think crunchyroll deserves this?

@Daily_hotori I found this one the creepiest. Even a little moreso than the hospital. Great X-Files vibe ...

NTE Anomaly hunts are getting uncomfortable for me. #NTE

Use of copyrighted material as training data without consent is theft. What's the point of copyright if AI companies are blatantly allowed to go around it? Current models can reproduce published articles, code, and art (or close enough that if YOU did it, you would get sued and/or takedown notices). That work is IN the final product as weights and vectors, and is used to produce anything similar that has been requested. If the nodes storing the new representation are activated (i.e., AI training is treated as something like a compression algorithm), then the copyright holder should receive a royalty. Wouldn't you agree that's only fair for using their work as part of the training set?

"Gen ai is theft" What is being stolen? "My work is being used without my consent" If you didn't want people to see and use your art for inspiration then why did you post it? "It takes no work to use Gen AI" This isn't even true. These images require the creative input of humans to exist.

@KyuboxIncubox @Grynoote Plot twist: She's why this dude is headless.

@Grynoote When she rides a bike down the street, her sword clotheslines people

Zankou Carrying that sword around must be a hassle imagine trying to get through a door or in a car #NTE #NTEネバエバ #Zankou

@dolphincafe000 A woman that knows what she wants will always be amazing.

九原さんのスタイルマジで凄いと思う #NTE

@NTE_Ani_Info That was amazingly fun to watch.

I just want to know where this was that it got that kind of dirty. A coal mine? Jeeeeeez ...

@artbymorbid My takes: Haru would be a bard that kept the group spirits up. Creek would be the cleric. Tachyon would be an artificer. Cafe or Rice would have to be rogues. Spe would be a derpy paladin. Suzuka is a horizon walker ranger.

#ウマ娘 #Umamusume Umamusume x Dungeons & Dragons

Cool! This is one I reported on. So, apparently not every issue goes into backlog hell, yet. #NTE

You may want to look up AlphaGo, Move 37. Your entire premise misunderstands what AI is, how it is trained, and how it creates answers. Might want to look up "chain of thought" while you are at it. The future will be made by people applying AI to their field. Not because it is more creative, but because it allows them to do more of what they are good at. Hint: If they weren't posting 7-finger images for the past 20 years, then AI is probably not helping them by making 7-finger images, today.

The CEO of Take-Two, the company behind GTA, just said something the entire AI industry doesn't want to hear. And he said it without being anti-AI. Strauss Zelnick's argument is precise. AI is built on datasets. Datasets are backward-looking. Creativity is forward-looking. A model trained on everything that already exists cannot, by definition, produce something genuinely unexpected. And all hits, by their very nature, are unexpected. Asset creation and hit creation are not the same thing. AI is getting very good at the first one. The second one is what actually makes money, builds franchises, and changes culture. Nobody has shown AI can do that yet. The derivative property problem is real. You can clone GTA with existing technology. You could do it before AI. It would take 3 years and look identical. It still wouldn't sell. Because it isn't GTA. It's a clone of GTA. And consumers, despite what the industry occasionally pretends, can feel the difference between something genuinely new and something assembled from the residue of things that already worked. Thousands of mobile games ship every year. 0 to 5 hits get made. The same studios make them every time. The technology to make more games has been commoditized for years. It didn't democratize hit creation. It just flooded the market with more forgettable product. The Silicon Valley thesis that AI unlocks game creation for everyone is true in the same way that cheap cameras unlocked filmmaking for everyone. They did. And the same 5 studios still make the movies everyone watches. What Zelnick is saying, without quite saying it, is that the thing AI cannot replicate is taste. The instinct for what hasn't been done yet. The cultural antenna that detects the gap in the market before the data can see it. Data tells you what people wanted. Hits tell people what they want next. Those are different jobs.