God this shit is exhausting; having a functioning legislature would actually be really helpful right now by passing a bill that spells out whether or not ML on copy-written content is a copyright violation and/or what conditions it can become a violation. But instead we have lawsuit after lawsuit leaving a blatantly ambiguous legal situation to be interpreted in as many different ways as there are courts.
I wish our court system had a pathway by which a judge can rule "not enough information to make a ruling" and force the legislature to pass a bill disambiguating the situation.
> passing a bill that spells out whether or not ML on copy-written content is a copyright violation and/or what conditions it can become a violation.
how would legislators know an answer to this. lawsuits provide a platform for discussion based on which a policy can emerge. Guessing wrong answer in form of legislation will be detrimental to the society.
Legislation can say, 'an AI/llm trained on copyrighted works needs to respect copyright/ ask or pay for permission to read/process a work'
Or,
It can say, if I can read it or watch it at a library, and the llm uses publicly accessible access methods, its not different then if a person had read and watched all those works and then made thoughts and determinations out of them.'
Obviously, there's some complexity in both of those arguments, and its hard to see what exactly is right, but making the rules we play by would be really helpful about now.
They don't "know" the answer, they decide the answer. Judges in these cases aren't pulling the answers from the void, they're interpreting the existing body of legislation and case law that governs copyright which was written and passed by the legislature.
It's funny you say that lawsuits provide a platform for discussion that leads to policy because you're describing the legislative process. I will grant you that the structure of lawsuits and amicus briefs does lend itself to quality discussion about issues and I think the legislative process could do with a bit of that.
It's time to let (most of) copyright go. Ideas are copied around the world at light speed and so much content is being produced that no individual has a chance to consume all the media the world produces in a single minute.
The value is in the copying not the material. In having a mechanism to spread your material to as many eyes as possible. Hanging onto Micky Mouse's anachronistic paradigm does not encourage more art, it only performs an oligopolistic role for artists already at the top, already with the copy mechanism and attention to spread their ideas.
But in these scenarios, it is overwhelmingly large, powerful corporations using the work of individuals. Sarah Silverman has been successful as a comedian, but when it comes to her versus Microsoft, OpenAI, AWS, et al. she is the one clearly being taken advantage of by someone far wealthier and more powerful.
I'm no fan of OpenAI's aggressive tactics, but I'd rather not risk harming open-source AI projects just to deal with them. Regardless of how any lawsuit comes out, OpenAI/Microsoft/whatever isn't going anywhere: like you said, they're powerful. They can afford to work around whatever laws or rulings that happen.
OSS projects won't have this luxury. So while I have my issues with OpenAI and the other tech companies, I think advocating for more copyright is like cutting off your nose to spite your face.
What makes you think this would strongly harm open-source AI projects? Do you mean e.g. small "companies trying to make money?" If so, I see that danger, but I also kind of don't care.
What I don't think it would harm is actual "free/free-ish open source, just OUT there" AI projects in terms of pure "software that's available for everyone to mess around with, regardless of legality," -- and (as a lawyer I can't recommend doing this at all of course, but) I think it would end up mostly being a good and safer thing (in the same way that having e.g. Kali Linux out there is a good thing).
The models are out there and it's hard to think that any single player could have a real big advantage here. I say let the suits commence.
It's pretty hard to develop an open-source project in the open if your activity is seen as illegal. We don't need AI projects lumped in with Popcorn Time and movie torrents.
Interesting examples you put there; FWIW you named the things that typically strongly favor the rights and desires of the consumer, despite the law, and give them a significantly superior experience to their legal alternatives (at least in the foreseeable short term.)
> Do you mean e.g. small "companies trying to make money?"
No they mean open source projects.
The overwhelming majority of all generative AI related open source projects depend on, use, or are related to publicly available weights that were trained without permission on other people's data.
The entire open source AI community wouldn't exist if those weights weren't available for everyone to use for basically any reason.
One could argue that the existing weights aren't going to disappear as thats impossible to enforce.
But what about the updated weights? If nobody is allowed to spend 50 million dollars training a new model and then releasing that model publicly for anyone to use, without getting permission from the entire internet to make that model, well then there aren't going to be anymore new models coming out.
The government isn't going to be able to confiscate everyone's gamer PCs in order to ban inference, but they can stop someone from spending that 50 mil training the next cutting edge open source model.
Which means that the existing models will be all that the open source community has. Yeah they could fine tune them, but cutting edge base models are still extremely important for open source AI progress.
Right. I think you hit on the biggest question -- what is the likely MARGINAL improvements in weights that your hypothetical 50 million buys? -- even if you accept that we wouldn't be here without that first "50 million" that already went in?
A reasonable argument, a possibility, is "very little." Perhaps compare to "so and so is going to put 50 million in improving the Linux Kernel?" I'm not sure this would improve anything at ALL, in fact, maybe make it worse -- given that someone will want to get something out of 50 million?
> what is the likely MARGINAL improvements in weights that your hypothetical 50 million buys
Wrong question.
The marginal improvements of a specific model might be marginal, but what matters is new models for different usecases.
For example, we have great image generation models 6 months ago, and one coming out now isn't going to be that much better. But, what is starting to get good right now is generative video models. And the video models are a huge marginal improvement over what was available 6 months ago.
Now do the same thing for any other use case. We have reasonable good videos now, but what about 3d models? Those mostly suck right now. Lots of improvement opportunities that wouldn't exist if training new models was banned.
There might be a dozen such major usecases that don't exist yet, where we can't repurpose old models, and instead need to train an entirely new model with different data. Making sure we still get that innovation is still important.
No one actually knows "what matters" yet, and also no one knows whether or not you actually need "new models."
Everything you're saying right now is extremely hypothetical.
The tech is very shiny, no doubt about that. But it's wildly conclusory to think "Yes, there will definitely be a demand for new use cases AND ALSO you definitely will be unable to create those new models without big money investment."
Again, it sounds to me exactly like someone clamoring for "big investment in a new, revolutionary, operating system." Nothing at all inevitable about that for MANY reasons, most having to do with "incremental building on the past thing that already works is probably the most feasible."
Only if we first let go of patent laws. Copyright is the only law which protects the rights of regular citizens because patents are often prohibitively expensive.
How does this follow? I take a picture (or make any other work of art), I have the copyright. This is outside of what is patented or not. They are not equivalent. Even in absence of the other, patents and copyrights do not offer the same protections.
They have differences but they are both forms of IP protection.
Citizens need some form of automatic IP protection or else anybody could (for example) copy someone's proprietary code and start offering a competing software service without any legal repercussions.
That would be completely wrong since the person who wrote the code invested a substantial amount of effort to write it. Someone who has better business connections shouldn't be able to just copy it and use it to launch a competing service... Such competitor could leverage their business connections to secure funding to drive the original author of the software out of business using their own work against them.
We’ve had enough concentration of resources in the hands of those who merely build technology that connects valuable things together, after having exploited market power to siphon off all the profit from the actually valuable things.
We need a special AI tax. Profits made by foundational models (not the application layer) get an additional tax.
We're entering a new paradigm that copyright is not sufficient to handle. These AI models have sucked up the output of humanity itself, the entire commons. The public vs private distinction isn't that important. They're going to cause job losses across many fields, so we need a new solution. Relying on the court system to enforce copyright claims for the benefit of a single claimant is not a solution that scales to solve the actual problem here. Another thing is, foundational models are prone to monopoly because the barriers to access compute are so steep.
The barrier of entry to building a "foundational model" is less than $400 of compute time (72 hours) on a handful of rental A100 using widely available open source data sets. Open source projects by ICs are already on par with GPT3. It's gonna be really hard to regulate stuff in its current form, which is already commercially beneficial
But hasn't the concentration you're describing been created under the current copyright system? Those winning from this system appear to be exactly those groups that are concentrating power the way you describe.
Currently there is little concentration compared to what it would be, without copyright you would have a single big corporation where people get all their content. You either get paid by them to create content or they distribute your content anyway without paying you. Under such a system no way youtube would pay creators 55% of revenue and so on, big corporations gains way more power when you remove copyright.
That wasn't much of a problem before copying was so cheap and easy, but today that would absolutely happen.
Nah, it's time to remember the spirit of it, not the letter. It was designed primarily to give leverage to the little guys against the big companies who had more power to copy, and it's needed now more than ever.
I think the original law was: Manual registration required for 22 years of protection. Manual re-registration required for a 22-year extension. Public domain after 44 years at most. I’d support that. Make something at 18. Sell it until you are 62.
“Author’s life + 70 years” is much too slow for the pace of culture in the past 70 years.
I think this would be an equitable compromise. AI can use the older stuff to build a foundation and license the newer stuff. Artists alive today get paid for their work, and we as a society can reap the benefits of art in a reasonable time frame. I have a speech I am delivering on this issue in a few months, would you mind if I adapt your comment and cite you?
Original copyright in the US was 14 years with the option to extend for 1 additional 14 year term. In 1831 it was changed to 28 years for the base term and the option to renew it for an additional 14 years or 42 years total. That was the copy right law for nearly 200 years.
Today it is actually 95 years or 70 years after the death of the author, whichever one is sooner.
I think a return to the 1831 copyright regime is pretty fair. Even just 28 years no extensions is enough time to have a full career in something.
If you believe the value isn’t in the material, then it should be a simple matter for you to avoid using it, and naturally doing so won’t be an inconvenience for you because there’s no value in it.
Not really clear why the means of distribution are valuable without anything of value to distribute, though…
> The value is in the copying not the material. In having a mechanism to spread your material to as many eyes as possible.
Sorry friend but this is a really bad take. I sink my time and energy into creating something and then when I get ripped off I should be happy because of the exposure?
Agreed, I feel like this attitude comes from people who consume content but never have tried to make and disseminate anything themselves. I genuinely feel so bad for anyone making music the last 10/15 years, it's been so undervalued socially and from within the recording/streaming industry itself.
> Agreed, I feel like this attitude comes from people who consume content but never have tried to make and disseminate anything themselves.
Yeah, I agree. I think most absolutist rejections of copyright and intellectual property are, at their core, a reaction against attempts to prevent piracy the person wants to engage in. Maybe mixed in with some misplaced Free Software idealism (that ignores stuff like the fact that copyright is the literal foundation of Free Software).
That's because the musicians flooded the market. Good music isn't rare and it is not hard to find it. Spotify has 100M tracks - even if 99% of them were crap - that would still mean that there are 1 million good songs. More than you can listen in a lifetime.
>Good music isn't rare and it is not hard to find it.
I think you and I may have different standards and taste. I don't disagree that there's plenty of good stuff being made, but a lot of it is hard to find and sorting through all the chaff is a gigantic pain.
This just points to the broader issue I was getting at in my previous comment. Record labels have never been great for artists, but they've pretty much abdicated any value they brought as a sort of filter awhile ago and have handed that off to algorithms that are questionable at best.
Quality also isn't always rewarded "fairly" or given a fair chance to catch on with a different market than whatever insular scene it exists within- this is compounded further via the insular nature of online groups too. Music labels and streaming platforms offer basically no support to musicians and groups who need it the most and who could benefit from a more hands on approach, and instead defer all their resources onto a few golden geese, because its safe and pays the bills.
This is short sighted though and creates an atmosphere of stagnation- a similar process is happening in film as well, culture is just being mediated by algorithms and markets, which is why culture has become so stagnant, predictable, boring, and circular.
Yeah it read to me like the original comment is discounting the importance of the actual material for content, which would be the most hilariously tech bro thing I've read in a long time.
Generally speaking I do but when I see a top comment like that I feel compelled.
It’s honestly somewhat schizophrenic on HN sometimes. You’ll read comments on one article and everybody goes full-bore Bay Area tech bro libertarian - no regulation, ever!!! And then in literally the next comment thread everybody is up in arms because why isn’t there government protection around this vitally important issue?!? Come on Uncle Sam!
I’m old enough to know that I’m dealing with people and so there’s no reason to expect anyone to make sense or be consistent. But I’ll be darned if I don’t feel the cognitive dissonance from time to time.
I think you mistake my meaning. I'm saying the HN audience is not a hive mind. Different articles attract different commenters which have different opinions.
Copyright as we know it didn't exist before the Statute of Anne in 1709.
Shakespeare. Leonardo. Michelangelo. Aristotle.
All of them produced their timeless works without copyright.
Copyright was a workable hack as long as making copies required buying a printing press (or equivalent), but those days are over and they aren't ever coming back.
Rather than continuing to bail out the sinking ship of copyright, artists need to start thinking about (and working on) new ways to get paid.
> Do you know anything at all about how all those artists got paid?
Do you?
Besides, that's not the point. Your previous way of making money is untenable, and has been since the invention of the Xerox machine.
What are you doing about it?
The manual drafters either learned CAD or got out of the trade.
The blacksmiths either learned how to be machinists or got out of the trade.
The farmers either learned how to use tractors or got out of the trade.
Your current business model is utterly dependent on the assumption that making copies is difficult and/or expensive. That's not true, and hasn't been true for quite some time. You can either deal with that fact or find yourself in the position of whalers after Rockefeller started cranking out fuel for kerosene lamps by the trainload.
> Why is it on the exploited to figure it out?
No one is being "exploited". No one is standing over you with a whip and forcing you to work.
Other people get paid for doing a job at the end of a week or month, and that's it. They don't continue to get paid for 90 years for one piece of work. I mean, yeah, that's a sweet deal, but sweet deals rarely last forever.
The stated reason for copyright (at least in the US) is to improve society by "promot[ing] the Progress of Science and useful Arts", not to enrich the authors per se.
> “Sorry bud, I know I keep holding you up for money. You really ought to figure out a way not to get robbed.”
Casual copyright violation is not, and never has been, prosecuted as theft or "robbery". The very concept didn't even exist until recent times, while real theft and robbery go all the way back to the beginning of history, likely long before.
With respect, I find this position quite baffling.
The ability to use your years of experience and hard-earned skills in your particular craft to make something and then have ownership of the thing you made allowing you to make money from that thing, or sell it, should be something we absolutely protect.
It’s a foundational principle behind capitalism and the very startup culture that HN/Ycombinator is built around.
Saying that we should abandon copyright is a bit like saying we should abandon “ownership”.
Too often I see this framing referring to the likes of Disney but this is to ignore the vast majority of individual small creators, who are not huge corporations, and absolutely need the protections that copyright provides in order to make a living from their craft.
FYI, one of the lawyers representing Sarah Silverman is an HNer of some note, Matthew Butterick. [1] He is the creator of Practical Typography, [2] which is posted about on HN from time to time. [3]
It's wonderful he's part of the local conversation.
But it's irrelevant to my not being wrong (am I?) =D
----
Thanks for the tangental information. Hope Sarah makes her point not too expensively, while the lawyers are fairly compensated.
I personally don't see how she can win, just because of the sheer openness / unregulatability of the technologies (cat's already out of the bag). I wish them well.
I don't see why we need any new laws to deal with it. Existing 'fair use' legislation should suffice.
The only issue I see is if the AI quotes the text word-for-word and allows the user to craft prompts in such a way that they can extract a substantial amount of copyrighted text in a predictable, organized manner.
If the AI spits out short quotes from the copyrighted text in a way that is relevant to the prompt and it doesn't allow the user to use prompt-engineering to reliably obtain the full text, then I don't see a problem.
If the prompt engineering required to extract the full text is highly complicated and difficult to execute in a reliable way, then I think it should not be an issue.
If some third party app decides to use AI with clever prompt engineering on the back end to extract copyright text in order to bypass copyright laws (and automatically reconstruct the full text to present to their users), then the third party app should be held responsible, not the AI provider.
The AI provider's responsibility should be limited to ensuring that prompts cannot be used in a predictable, repeatable way to extract a substantial, contiguous portion of the copyrighted text.
The copyright violation occurs at the moment that the material is presented to the end user (a human). Merely storing the material and not presenting it to any human should not be a copyright violation... Otherwise you might as well sue Seagate and Western Digital for copyrighted material being stored on hard drives manufactured by them. It would not make sense.
Some libraries loan out books for free. That's how a lot of people learn. It's fine.
It depends on whether or not the book's text was accessed illegally.
71 comments
[ 2.8 ms ] story [ 149 ms ] threadI wish our court system had a pathway by which a judge can rule "not enough information to make a ruling" and force the legislature to pass a bill disambiguating the situation.
how would legislators know an answer to this. lawsuits provide a platform for discussion based on which a policy can emerge. Guessing wrong answer in form of legislation will be detrimental to the society.
Or,
It can say, if I can read it or watch it at a library, and the llm uses publicly accessible access methods, its not different then if a person had read and watched all those works and then made thoughts and determinations out of them.'
Obviously, there's some complexity in both of those arguments, and its hard to see what exactly is right, but making the rules we play by would be really helpful about now.
It's funny you say that lawsuits provide a platform for discussion that leads to policy because you're describing the legislative process. I will grant you that the structure of lawsuits and amicus briefs does lend itself to quality discussion about issues and I think the legislative process could do with a bit of that.
The value is in the copying not the material. In having a mechanism to spread your material to as many eyes as possible. Hanging onto Micky Mouse's anachronistic paradigm does not encourage more art, it only performs an oligopolistic role for artists already at the top, already with the copy mechanism and attention to spread their ideas.
OSS projects won't have this luxury. So while I have my issues with OpenAI and the other tech companies, I think advocating for more copyright is like cutting off your nose to spite your face.
What I don't think it would harm is actual "free/free-ish open source, just OUT there" AI projects in terms of pure "software that's available for everyone to mess around with, regardless of legality," -- and (as a lawyer I can't recommend doing this at all of course, but) I think it would end up mostly being a good and safer thing (in the same way that having e.g. Kali Linux out there is a good thing).
The models are out there and it's hard to think that any single player could have a real big advantage here. I say let the suits commence.
No they mean open source projects.
The overwhelming majority of all generative AI related open source projects depend on, use, or are related to publicly available weights that were trained without permission on other people's data.
The entire open source AI community wouldn't exist if those weights weren't available for everyone to use for basically any reason.
One could argue that the existing weights aren't going to disappear as thats impossible to enforce.
But what about the updated weights? If nobody is allowed to spend 50 million dollars training a new model and then releasing that model publicly for anyone to use, without getting permission from the entire internet to make that model, well then there aren't going to be anymore new models coming out.
The government isn't going to be able to confiscate everyone's gamer PCs in order to ban inference, but they can stop someone from spending that 50 mil training the next cutting edge open source model.
Which means that the existing models will be all that the open source community has. Yeah they could fine tune them, but cutting edge base models are still extremely important for open source AI progress.
Remember all those "can it run crysis?" memes? Those GPUs are now slower than mobile phones.
And the AI progress time gap we are seeing right now, of "amount of time between major AI breakthroughs" is measured in weeks, not years.
Going back to waiting years instead of weeks for entire technologies to be revolutionized is a major slowdown.
A reasonable argument, a possibility, is "very little." Perhaps compare to "so and so is going to put 50 million in improving the Linux Kernel?" I'm not sure this would improve anything at ALL, in fact, maybe make it worse -- given that someone will want to get something out of 50 million?
Wrong question.
The marginal improvements of a specific model might be marginal, but what matters is new models for different usecases.
For example, we have great image generation models 6 months ago, and one coming out now isn't going to be that much better. But, what is starting to get good right now is generative video models. And the video models are a huge marginal improvement over what was available 6 months ago.
Now do the same thing for any other use case. We have reasonable good videos now, but what about 3d models? Those mostly suck right now. Lots of improvement opportunities that wouldn't exist if training new models was banned.
There might be a dozen such major usecases that don't exist yet, where we can't repurpose old models, and instead need to train an entirely new model with different data. Making sure we still get that innovation is still important.
Everything you're saying right now is extremely hypothetical.
The tech is very shiny, no doubt about that. But it's wildly conclusory to think "Yes, there will definitely be a demand for new use cases AND ALSO you definitely will be unable to create those new models without big money investment."
Again, it sounds to me exactly like someone clamoring for "big investment in a new, revolutionary, operating system." Nothing at all inevitable about that for MANY reasons, most having to do with "incremental building on the past thing that already works is probably the most feasible."
All the more reason to not make it illegal to train these new models.
There are massively improved models coming out once a month.
There could be many more to come.
As you said, we don't even know where this stops! and the current innovation is very large right now.
Therefore, it is premature to just ban all of that, when we don't even know how much more innovation there is to go.
So the point stands. The massive amount of innovation that has helped open source has all been driving by these new models being released.
And there could be many more to come.
> Everything you're saying right now is extremely hypothetical.
Looking at the fact that we come out with a powerful new model every month is not hypothetical.
I am referencing existing reality here.
Granted, the US patent system is broken.
Citizens need some form of automatic IP protection or else anybody could (for example) copy someone's proprietary code and start offering a competing software service without any legal repercussions.
That would be completely wrong since the person who wrote the code invested a substantial amount of effort to write it. Someone who has better business connections shouldn't be able to just copy it and use it to launch a competing service... Such competitor could leverage their business connections to secure funding to drive the original author of the software out of business using their own work against them.
We’ve had enough concentration of resources in the hands of those who merely build technology that connects valuable things together, after having exploited market power to siphon off all the profit from the actually valuable things.
We're entering a new paradigm that copyright is not sufficient to handle. These AI models have sucked up the output of humanity itself, the entire commons. The public vs private distinction isn't that important. They're going to cause job losses across many fields, so we need a new solution. Relying on the court system to enforce copyright claims for the benefit of a single claimant is not a solution that scales to solve the actual problem here. Another thing is, foundational models are prone to monopoly because the barriers to access compute are so steep.
That wasn't much of a problem before copying was so cheap and easy, but today that would absolutely happen.
“Author’s life + 70 years” is much too slow for the pace of culture in the past 70 years.
Today it is actually 95 years or 70 years after the death of the author, whichever one is sooner.
I think a return to the 1831 copyright regime is pretty fair. Even just 28 years no extensions is enough time to have a full career in something.
Not really clear why the means of distribution are valuable without anything of value to distribute, though…
Sorry friend but this is a really bad take. I sink my time and energy into creating something and then when I get ripped off I should be happy because of the exposure?
https://theoatmeal.com/comics/exposure
Yeah, I agree. I think most absolutist rejections of copyright and intellectual property are, at their core, a reaction against attempts to prevent piracy the person wants to engage in. Maybe mixed in with some misplaced Free Software idealism (that ignores stuff like the fact that copyright is the literal foundation of Free Software).
This is why the music industry gave up and now we have services like Tidal, Spotify and Youtube.
I think you and I may have different standards and taste. I don't disagree that there's plenty of good stuff being made, but a lot of it is hard to find and sorting through all the chaff is a gigantic pain.
This just points to the broader issue I was getting at in my previous comment. Record labels have never been great for artists, but they've pretty much abdicated any value they brought as a sort of filter awhile ago and have handed that off to algorithms that are questionable at best.
Quality also isn't always rewarded "fairly" or given a fair chance to catch on with a different market than whatever insular scene it exists within- this is compounded further via the insular nature of online groups too. Music labels and streaming platforms offer basically no support to musicians and groups who need it the most and who could benefit from a more hands on approach, and instead defer all their resources onto a few golden geese, because its safe and pays the bills.
This is short sighted though and creates an atmosphere of stagnation- a similar process is happening in film as well, culture is just being mediated by algorithms and markets, which is why culture has become so stagnant, predictable, boring, and circular.
What they want is for you to be very happy about you getting ripped off by them whilst they infringe and grift on your work for free with zero credit.
It’s honestly somewhat schizophrenic on HN sometimes. You’ll read comments on one article and everybody goes full-bore Bay Area tech bro libertarian - no regulation, ever!!! And then in literally the next comment thread everybody is up in arms because why isn’t there government protection around this vitally important issue?!? Come on Uncle Sam!
I’m old enough to know that I’m dealing with people and so there’s no reason to expect anyone to make sense or be consistent. But I’ll be darned if I don’t feel the cognitive dissonance from time to time.
Shakespeare. Leonardo. Michelangelo. Aristotle.
All of them produced their timeless works without copyright.
Copyright was a workable hack as long as making copies required buying a printing press (or equivalent), but those days are over and they aren't ever coming back.
Rather than continuing to bail out the sinking ship of copyright, artists need to start thinking about (and working on) new ways to get paid.
Why is it on the exploited to figure it out?
“Sorry bud, I know I keep holding you up for money. You really ought to figure out a way not to get robbed.”
Do you?
Besides, that's not the point. Your previous way of making money is untenable, and has been since the invention of the Xerox machine.
What are you doing about it?
The manual drafters either learned CAD or got out of the trade.
The blacksmiths either learned how to be machinists or got out of the trade.
The farmers either learned how to use tractors or got out of the trade.
Your current business model is utterly dependent on the assumption that making copies is difficult and/or expensive. That's not true, and hasn't been true for quite some time. You can either deal with that fact or find yourself in the position of whalers after Rockefeller started cranking out fuel for kerosene lamps by the trainload.
> Why is it on the exploited to figure it out?
No one is being "exploited". No one is standing over you with a whip and forcing you to work.
Other people get paid for doing a job at the end of a week or month, and that's it. They don't continue to get paid for 90 years for one piece of work. I mean, yeah, that's a sweet deal, but sweet deals rarely last forever.
The stated reason for copyright (at least in the US) is to improve society by "promot[ing] the Progress of Science and useful Arts", not to enrich the authors per se.
> “Sorry bud, I know I keep holding you up for money. You really ought to figure out a way not to get robbed.”
Casual copyright violation is not, and never has been, prosecuted as theft or "robbery". The very concept didn't even exist until recent times, while real theft and robbery go all the way back to the beginning of history, likely long before.
I'd still recommend that you work on alternative methods of getting paid.
The ability to use your years of experience and hard-earned skills in your particular craft to make something and then have ownership of the thing you made allowing you to make money from that thing, or sell it, should be something we absolutely protect.
It’s a foundational principle behind capitalism and the very startup culture that HN/Ycombinator is built around.
Saying that we should abandon copyright is a bit like saying we should abandon “ownership”.
Too often I see this framing referring to the likes of Disney but this is to ignore the vast majority of individual small creators, who are not huge corporations, and absolutely need the protections that copyright provides in order to make a living from their craft.
1: https://fortune.com/2023/07/12/sarah-silverman-lawsuit-chatg...
2: https://practicaltypography.com/
3: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
But it's irrelevant to my not being wrong (am I?) =D
----
Thanks for the tangental information. Hope Sarah makes her point not too expensively, while the lawyers are fairly compensated.
I personally don't see how she can win, just because of the sheer openness / unregulatability of the technologies (cat's already out of the bag). I wish them well.
The only issue I see is if the AI quotes the text word-for-word and allows the user to craft prompts in such a way that they can extract a substantial amount of copyrighted text in a predictable, organized manner.
If the AI spits out short quotes from the copyrighted text in a way that is relevant to the prompt and it doesn't allow the user to use prompt-engineering to reliably obtain the full text, then I don't see a problem.
If the prompt engineering required to extract the full text is highly complicated and difficult to execute in a reliable way, then I think it should not be an issue.
If some third party app decides to use AI with clever prompt engineering on the back end to extract copyright text in order to bypass copyright laws (and automatically reconstruct the full text to present to their users), then the third party app should be held responsible, not the AI provider.
The AI provider's responsibility should be limited to ensuring that prompts cannot be used in a predictable, repeatable way to extract a substantial, contiguous portion of the copyrighted text.
The copyright violation occurs at the moment that the material is presented to the end user (a human). Merely storing the material and not presenting it to any human should not be a copyright violation... Otherwise you might as well sue Seagate and Western Digital for copyrighted material being stored on hard drives manufactured by them. It would not make sense.
Your honor, I rest my case.
OpenAI isn't just storing the data like Seagate but all its answers are derived from that data.