I haven't told anyone about this yet, so let's go: Last fall, attorney's working for a certain social media company we will call "Schmeta" reached out to me asking for help with a copyright suit specifically because I was an academic who is also an 'expert' with Blender. 🧵
They wanted an expert who would state that 3D models aren't worth anything because they are so easy to make. Evidently Shmeta and an ivy league school we will call "Schmarvard" had scraped data illegally from a certain company's online library and used it to to train their AI...
These attorney's wanted me to show that I could make textured 3D models quickly and and easily in Blender to exhibit how easy they are to make and that the lawsuit against them for stealing hundreds of thousands of 3D models was worthless. They were willing to pay me significantly to do this.
I said I was happy to exhibit how long it takes to make these models, but that it might not be what they want as it is still a time-intensive craft that takes years to learn how to do creatively and competently. They said they didn't want me to lie, but that was not what they were looking for.
After several phone calls they eventually told me that they would not need my services. But I was left with the distinct impression that mega rich institutions like Schmeta and and Schmarvard are gleefully stealing every aspect of our culture to position themselves to resell it back to us.
I also got the impression that our legal systems are in no way fast enough to stem the damage they will do to our creative classes and they know this and are working as fast they can to grab as much as they can before laws catch up to their crimes. https://www.wired.com/story/new-documents-unredacted-meta-copyright-ai-lawsuit/
Meta Secretly Trained Its AI on a Notorious Piracy Database, Newly Unredacted Court Docs Reveal

One of the most important AI copyright legal battles just took a major turn.

WIRED

@brentpatterson "They also allege that internal discussions about using LibGen data were escalated to Meta CEO Mark Zuckerberg (referred to as "MZ" in the memo handed over during discovery) and that Meta's AI team was "approved to use" the pirated material."

If there ever was any doubt about corporations' lack of respect for copyright of others, surely this is the end of it.

@ticho @brentpatterson
Makes it hard to swallow that I spent 20 minutes today using my knowledge of copyright to explain to some poor comms staffer, who just wants to send out a link to an article and have it not be paywalled, "for THESE people, link A will work, but for THOSE people, you'll need to get link B with special tracking cruft from the corresponding author," all so the publisher can have a sustainable business model of charging $5k APC or $40 per download.
@brentpatterson Schmarvard sounds like a big place, do you recall what part in particular? Just curious.
@overholt I was never told which unit of "Schmarvard" was involved. However, there were evidently competitions promoted by Schmeta to use the stolen dataset to train AI to either recognize or create 3D models of furniture.
@brentpatterson Hmm, that kind of sounds like a hypothetical Schmaduate Schmool of Schmesign.
@brentpatterson
This exactly what’s happening before our very eyes.
We have to PAY for necessities like healthcare & then they profit off it. MSM. Agriculture. Water. They’re taking everything.
I feel like their vision is all of us living 100% in servitude to them in horrible conditions.
#Furiosa
#MadMax

@RiaResists @brentpatterson

That's what unrestrained unregulated US capitalism is and does.

@brentpatterson curious, did you get the sense that they did not need your services because you were not giving them what they wanted? Do you think they went to find someone that would?
@brentpatterson It's not the amount of time or effort that determines if something is a work of art or otherwise copyrightable. It can take far longer to paint the halls of an office building beige than it takes an someone to paint an impressionistic painting, but the first isn't art while the second is.
@brentpatterson sounds like the phone calls were the services. They were seeking an opinion and got one. Bill ‘em.
@brentpatterson At last years SIGGRAPH, I attended an NVIDIA talk which basically began with a slide which had WORLD => NVIDIA MACHINE LEARNING => NEW MODELS => SELL BACK TO BUSINESSES. It is hard not to interpret this as companies looting the commons and then packaging it and selling it back to businesses to be sold back to consumers. I'm beginning to consider that perhaps being retired will present me with fewer moral dilemmas.
@brentpatterson as a complete noob I ask: what would be the consequences of accepting the proposition and later just tell the truth during the exposition? Or this was not for a deposition in front of a jury or similar?
@brentpatterson
These attorneys seem a little unclear on the concept of theft.

@brentpatterson funny, so they wanted you to demonstrate that their product provide little to no value, since anyone can quickly pick these skills up and do whatever model they want about as fast as the model could generate them…

I'm glad you did the right thing, though i'm sure it wasn't even a question for you, there are obviously worse people (on the other side of the phone line) who think it's a totally fine thing to do.

@brentpatterson "you're an expert so you know it takes no time at all" what

What fucking logic.

Garbage people.

@KayOhtie @brentpatterson of course they don't actually believe any of that, they just want someone to say it in court to weasel out of trouble