This is exactly what people like me warned about. The push for copyright-protections against GenAI is not going to stop GenAI. It’s just going to create a future where only the big players can make AI models because they can pay for “licenses” with those strong enough to threaten them (while scraping everyone else anyway) and just kill open sourced models.
We’re seeing it play it exactly like that. Y’all can’t stop GenAI. You literally don’t have enough power where it counts. The only sane solution is to push that any GenAI model trained on public data must have open weights by default.
If we don’t have the power to stop generative AI, then what makes you think we have the power the change copyright law? Generative AI uses up huge amount of power and water to the point of causing issues for national infrastructure. There is a clear climate case to be made against generative AI and unlike copyright law the public actually care about climate change.
You cannot stop GenAI unless you stop sales of all gaming GPUs and recall all sold ones world wide. You just can’t. That cat’s out of the bag. You can try to restrict big corpos like OpenAI, but your politicians and bought already. Therefore adding your voice to trying to “stop GenAI” via copyright laws is just playing yourself. You can try to restrict GenAI via climate protections, but that will go as well as efforts to stop supporting genocides (or as well efforts have gone to protect the climate anyway).
If our politicians are already bought and sold to the point that calling for these industries to be regulated is pointless, then why would politicians listen to our calls for ‘open weights by default’.
They won’t, of course. But they will use the pressure towards copyrights against AI to kill open source GenAI progress. But at least when advocating for Open models, you don’t hurt the people who are doing FOSS.
Copyright law is broken. But I don’t think that means we have no obligations to each other as human beings when we build on each other’s work.
We had the same argument during the crypto craze. The financial system is broken, but 10 years later I think we all agree that crypto is pretty clearly not the answer.
Copyright law is broken. But I don’t think that means we have no obligations to each other as human beings when we build on each other’s work.
Absolutely! This is why I said anything built on public work, should be public goods as well.
We had the same argument during the crypto craze. The financial system is broken, but 10 years later I think we all agree that crypto is pretty clearly not the answer.
That’s not a good comparison. Crypto was a (bad) solution looking for a problem. GenAI already has use-cases.
This is why I said anything built on public work, should be public goods as well.
What if I don’t want certain people to build on my work, or to constrain the ways in which the build on it? (Non-commercial, share-alike, attribution, etc. clauses) Should I be able to?
That’s not a good comparison. Crypto was a (bad) solution looking for a problem. GenAI already has use-cases.
I didn’t mean to compare the technology – though there are some similar scam vectors, but that’s a different conversation.
I meant that there was a strong contingent of crypto fans back then who were saying – correctly – that “the mainstream system is corrupt and wields legislation as a weapon against consumers”. But their proposed alternative was a system that removed all regulation, including consumer protections.
I worry that there’s a trend in tech circles today that echoes that sentiment when it comes to AI.
I’m also rather disappointed that a substantial group of people who I used to assume I was aligned with – pirates and open-sourcerers – turned out to only be there for the free shit and not for the ethos.
An ethos which, to me, is something like: everyone has a right to participate in culture and be a part of the conversation, and everyone has a duty to acknowledge the work that enabled their own and do their best to be a good custodian of the upstream works.
What if I don’t want certain people to build on my work, or to constrain the ways in which the build on it? (Non-commercial, share-alike, attribution, etc. clauses) Should I be able to?
No. The idea that someone should be allowed to control what others do with their expressions and ideas is a very new concept (~100 years) and it has not brought any benefit to society
we all agree that crypto is pretty clearly not the answer.
Tell that to anybody holding crypto as prices hang out around all time high.
Not to mention solving the problem with energy consumption which LLMs are beginning to dwarf.
And if you think state currency is the answer, then I have really bad news about how that’s been going…
This road were going down, it doesn’t look good. Putting the content into a black box that rarely links the original source, let alone other sources is worse than Facebook content previews.
Alternate headline:
Companies accept money for a thing that will happen anyway, and will be unable to prove if they say no.
GenAi is unfortunately here, and the technocracy wants you to want it so they can farm you for more and more intimate data to leverage and enforce their technocracy. And the only way they’re going to do it is by keeping the press positive, and feed it more and more data in the hopes it fixes things.
I just saw an Inverse article about Mario Kart that said “could of.”
I can expect random people online to do that dumb shit, but not someone whose job is literally to write things.
Unfortunately, I think it’s been demonstrated that OpenAI will feed your data into their training pipeline whether you like it or not. They did this with YouTube. And they demonstrated this with Scarlett Johanson. (Even if they used a voice actor instead of actually scraping voice clips of her, it’s still unethical.)
So we knew they’d just scrape these articles anyway. This way, at least the publications get paid.
