Who the fuck gives Anthropic the right to use all that training data for their model?! And if other companies distill their models, THEY are to be protected by laws?! Cry me a fucking river.
And don’t even get started on how he says that China is a represive regime (which it is) while leaving the US as it it wasn’t. Like they won’t “CCCP” the fuck out of the population using Anthropic’s models.
No idea. The whole copyright topic is a clusterfuck when it comes to LLMs. The matter of the fact is: most AI companies steal data to train their models and they don’t care about licenses or copyrights.
In most countries, the output of a single prompt will not be copyrightable. However, once someone starts arranging the output into something more, then copyright can be applied.
Overall, it depends on the country, but in some they are asking their legislators for clarification. Many are trying to apply pre-existing copyright laws to LLMs.
Yes, as the other commenter said, in most jurisdictions (i think China is the only exception?), generative AI outputs (this covers more than LLMs, like diffusers and other modern image generators) are not eligible for copyright. However, while a human made arrangement of the outputs gets copyright protection, in the US at least, it is a very narrow scope of protection, literally only a direct copy/paste of large swaths, and doesn’t extend to the individual AI generated components.
There’s nothing else they could say. They have to justify the hundreds of billions of dollars that have been poured into this thing. Anything less than world domination and complete control will make it impossible to ever make up for all that money.
AI is great (sometimes), but the centralized command-and-control cloud model is not. It needs to die. So we should all use the opensource models and ditch the hyperscalers.
I agree. I’m an open source maxi and go out of my way to only use open source models.
And I don’t mean open weight models, I mean fully open source models where you also have access to the training data and checkpoints and stuff like that. In my opinion, go open source or GTFO.
This guy is full of shit.
Let me rephrase his points:
Who the fuck gives Anthropic the right to use all that training data for their model?! And if other companies distill their models, THEY are to be protected by laws?! Cry me a fucking river.
And don’t even get started on how he says that China is a represive regime (which it is) while leaving the US as it it wasn’t. Like they won’t “CCCP” the fuck out of the population using Anthropic’s models.
Hasn’t the output from LLM already been ruled ineligible for copyright protection?
No idea. The whole copyright topic is a clusterfuck when it comes to LLMs. The matter of the fact is: most AI companies steal data to train their models and they don’t care about licenses or copyrights.
Copyright should not exist
Are you for real?
“The whole copyright topic is a clusterfuck”, full stop.
Except for use in llm
In most countries, the output of a single prompt will not be copyrightable. However, once someone starts arranging the output into something more, then copyright can be applied.
Overall, it depends on the country, but in some they are asking their legislators for clarification. Many are trying to apply pre-existing copyright laws to LLMs.
The U.S. Copyright Office wrote up a report about this and included a list of how other countries are handling it.
https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Copyrightability-Report.pdf
Yes, as the other commenter said, in most jurisdictions (i think China is the only exception?), generative AI outputs (this covers more than LLMs, like diffusers and other modern image generators) are not eligible for copyright. However, while a human made arrangement of the outputs gets copyright protection, in the US at least, it is a very narrow scope of protection, literally only a direct copy/paste of large swaths, and doesn’t extend to the individual AI generated components.
There’s nothing else they could say. They have to justify the hundreds of billions of dollars that have been poured into this thing. Anything less than world domination and complete control will make it impossible to ever make up for all that money.
AI is great (sometimes), but the centralized command-and-control cloud model is not. It needs to die. So we should all use the opensource models and ditch the hyperscalers.
I agree. I’m an open source maxi and go out of my way to only use open source models.
And I don’t mean open weight models, I mean fully open source models where you also have access to the training data and checkpoints and stuff like that. In my opinion, go open source or GTFO.