1011
Absolutely not training data no way no sir
(i.redd.it)
For preserving the least toxic and most culturally relevant Tumblr heritage posts.
Image descriptions and plain text captions of written content are expected of all screenshots. Here are some image text extractors (I looked these up quick and will gladly take FOSS recommendations):
-web
-iOS
Please begin copied raw text posts (lacking a screenshot that makes it apparent it is from Tumblr) with:
# This has been reposted here to Lemmy as part of the "Curated Tumblr Project."
I made the icon using multiple creative commons svg resources, the banner is this.
Tencent isn't the overlord of image generation lmao. This is using people's justified fears of China and surveillance to make a false comparison to image generation. All you're doing is giving more power to companies and states that will abuse it while limiting its use in open source contexts.
How about we just not use people's personal identities for image generation at all?
How about we just not let any drawings or paintings be made of others at all? I'm all for disallowing things like AI edited porn without consent but you can't arbitrarily apply one set of rules to image generation by computer and another to one done by hand when their outputs are fundamentally the same.
One is theft and an infringement of privacy for nefarious ends and the other is a painting. There's a world of difference between agreeing to let someone paint you and a corporation using your data to train AI. Spinning this basic reality into sinophobia is mind boggling. There are people in this thread shitting on Google for the same thing. Would you call it amerophobic to criticize Google for the same shit? Of course you wouldn't
It's not theft. Fuck copyrights.
Sure, for corporations and the wealthy. But it straight up yanks information from small authors, artists, etc. I couldn't give less of a shit if Disney is impacted from AI, but there is real potential for harm to average people. Submitting your shit to AI should be opt-in, scraping the web for content that company didn't create with no consent from the content creators for the means of profiting off their labor is wrong. Copyright is fucked, yes. It protects the wealthy more than it protects the non-wealthy, yes. These companies practices are still fucked too. Two things can be bad and there is plenty of room for nuance in this area
I never said it was sinophobic I said that they're utilizing peoples preexisting dispositions to consolidate power in the AI space. Which is objectively true, the large companies are currently doing everything they can to demonize open source models.
Why not? The arbitrating factor is the people involved in making the image. The inputs.
What is your definition of 'people involved in making the image'. People are involved whether or not you use AI. It's the same argument traditional artists threw when digital art first started to come onto the scene. "your art is worthless and takes away from my real art because it was made with a computer". There is a huge difference between someone posting raw gens and those that spend hours on pre and post processing to get a style consistent with the image in their head. That's exactly what digital artists do. Your tools should get you 80% of the way to your intended product, the rest comes from you. Is that inherently without value just because a computer had a part in the process? Then you'd have to apply the same rules to all digital art made within the last 20 years. Adobe has been utilizing, admittedly worse versions, of AI in things like Photoshop for years. People just didn't realize it until they got good enough to stand out.
š¤±shhhhhhhhhhh drink your milk