Odd that the article doesn't mention parties at all, although perhaps this was in an attempt to avoid accusations of partisanship that might ensue from stating facts.
Anyway, a quick look at https://www.congress.gov/bill/119th-congress/house-bill/6028... indicates that all 4 sponsors of the bill are Republicans. The Actions tab seems to indicated that the bill got only 12 minutes of debate before being passed,; I hope this is an artifact of how the page is updated rather than the actual time spent on considering it.
In 2026, a discussion of a bill proposing to make an existing position into a presidential appointee is very different if that bill was proposed by Democrats or Republicans. To pretend otherwise is to ignore virtually all of the current administrations actions.
> In a voice vote earlier this week, the House of Representatives passed H.R. 6028, the “Legislative Branch Agencies Clarification Act.”
wow, i had always assumed actual laws have to pass a recorded vote, but its not true...
from wiki:
> In Congress, "the vast majority of actions decided by a voice vote" are ones for which "a strong or even overwhelming majority favors one side", or even unanimous consent. Members can request a division of the assembly (a rising vote, where each sides rise in turn to be counted), and one-fifth of members can demand a recorded vote on any question, after the chair announces the result of a voice vote.
> It is estimated that more than 95 percent of the resolutions passed by state legislatures are passed by a unanimous voice vote, many without discussion; this is because resolutions are often on routine, noncontroversial matters, such as commemorating important events or recognizing groups.
From what I understand it’s rather true that a lot of Congress’ actual work is incredibly boring and that these procedures were invented to move it along.
You can see a lot of difference in the way congresspeople talk based on whether it’s televised or not as well, especially in committees.
I’m just a little surprised that voice votes haven’t been replaced by some kind of digital process. A voice vote doesn’t save time compare to a modern method of tallying votes. Why avoid making records when records are so “cheap” these days?
Hey, the economy is great and gas is cheap. All we had to put up with is mean tweets.
Whenever anyone complains about Trump, remind them he’s not the cause but the product. Seventy million voted for him, and Republicans in congress let him do illegally what they cannot accomplish legislatively. And all the while they’re busy selling the country for parts, whether through tax policy, or neutering the CAFE standards, or handing copyright to Disney.
He's certainly part of the cause, it seems the reason his handlers were able to use him so effectively is because he was going to jail if republicans and Musk et al. (and Putin?) didn't use him to get the election.
Expecting a convicted felon to live with the consequences of their further crimes, when all they need do is end the republic and steal billions, with the help and financial support of tech bros, ... well.
There's a very simple first patch to a new republic, no convicts in elected offices.
I usually agree with the EFF on things, but after reading their linked https://www.eff.org/deeplinks/2025/05/us-copyright-offices-d... I couldn't disagree more. An LLM is a predict the next word algorithm. If the model is overfitting, it's basically copy paste. There have been several documented instances where that happened and full GPL code, including headers and attribution were copy/pasted by the "AI."
AI is essentially copy paste with more steps. The part that AI companies use to defend this is ?how are we supposed to decide how much each author deserves? They try to wave this away, but their own model can tell them. Their models work off of weights. They can determine how much each work contributed based on those weights, so it's dishonest for them to argue it isn't possible. The way the models are engineered now don't make this possible, but that's intentional and we can all recognize that. They throw up their hands and claim it's not possible because they simply don't want to pay.
The most infurating thing however is how AI companies sidestep the IP rights of authors, but then claim to own those IP rights when their own generated output leaks. Anthropic filed DMCA takedowns on the leaked claude code repos, claiming ownership over something they explicitly have stated is almost entirely AI generated as part of their marketing. They take code, mix it up just enough to scrub away the GPL or whatever license belongs on it, then try to claim ownership of the result, in spite of the Copyright Office repeatedly stating that AI generated works have no copyright protection at all.
Is it actually possible to determine how much the weights were influenced by each work?
I might recall reading some interpretability paper years ago that trained a special model that could attribute each answer to a part of the corpus (like Wikipedia, ArXiV, or "Blogs") but it had a non-zero effect on performance and wasn't nearly as straightforward as weights go in, attribution comes out.
It’s very possible to determine similar works that existed earlier, and from that, recover attribution.
The “downside” is you may attribute similar works that weren’t inspirations, but coincidental. But I think that’s an upside: when someone discovers something novel and great but their work fails because of bad luck or non-novel details, then the discovery is finally recognized in another work, I think they should still be attributed.
>Is it actually possible to determine how much the weights were influenced by each work?
It will be very possible once they become the owners of the intellectual property being infringed. Think about how it was "impossible" to implement DRM on music and movies in the early days of youtube. Now, Google owns the content and platform, and suddenly their "rolling cypher" which involves no encryption at all is supposedly enforcable DRM.
The Silicon Valley tech bros play the same game every time. They violate the law, say it's just too darn difficult to obey the law without stifling progress, and then they get away with it until they kill all the competition. At which point, the law is once again applicable to anyone that might try to challenge them.
Remember how Amazon destroyed all the other retailers when they had a decade of no sales tax while brick & mortar had to obey it. "Calculating sales tax for 50 different states?! That's impossible!!!" What a load of shit...
Now, knowing that they're going to do this playbook again, how do you think it's going to play out? We've already seen it. Anthropic steals your copyrighted code, puts together their claude code project, the code for that project leaks, but now THEY own it! They sent DMCA takedowns on that AI generated code. AI generated code enjoys no copyright protection, it cannot be DMCAed under the law, there's no copyright on it. But Anthropic claims there is, and Github will obey the takedown, and nobody has the money to step up and stop them.
See where this is going? Once they achieve market dominance, they will claim that all the code generated by claude belongs to Anthropic, your prompts belong to you, but THEIR machine generated THEIR code and you only purchase a license to it with your tokens. A limited license. It might be revokable, it might expire, maybe you need to pay an annual fee to keep using THEIR code Claude generated for you. And if you actually just write code on your own, without Claude? Well, prepare to be sued like a network printer is sued by the RIAA because that's going to happen too. They will have their robot scour your code for "fair use" training and discover that it's just too similar to something their machine generated a year earlier. Sorry open source programmer, here's your legalese nasty gram. It appears you owe Anthropic some money.
This is what's known as a category error; an LLM is a 'model', not an algorithm.
It's not even an accurate claim; LLMs predict the next token, not the next word.
> AI is essentially copy paste with more steps
What about when AI creates a limerick about a kubernetes cluster run by Buddhist Monks? Or any number of other novel creations?
Fortunately the courts recognized the transformative use involved in making a model, which is fair use of copyrighted works, in kadrey v meta platforms.
> The most infurating thing however is how AI companies sidestep the IP rights of authors
transformative use falls under fair use, permission from authors is not needed to use legally acquired copyright works for training. Kadrey v Meta Platforms and Bartz v Anthropic.
> but then claim to own those IP rights when their own generated output leaks.
Corporations gonna do corporate things. Blatant hypocrisy is par for the course. Organize and take them to court.
There’s a difference between training a model and using a model.
Training involves copyrighted works but fair use is not just about use of copyrighted works, it’s about whether the use is transformative and substitutes the original market. I struggle to see how is not transformative under these criteria.
The use of the model (being able to output copies of GPL software) is a different question. This depends on the circumstances: if GPL code is exactly reproduced then it very well could be subject to the license of the original work.
I don’t understand the legal objections to the fair use of protected IP. Licenses are legal documents, not moral imperatives. GPL only exists because of copyright law, and you can’t write a license that supercedes copyright law if you don’t like the law.
The Claude Code example is completely different, hosting a repo with the leaked code is clearly not fair use.
So in your consideration the user of training materials for LLMs don't allow the LLMs to compete with the original authors? That's part of the assessment of transformation - does the resulting work compete with the original.
If I no longer need your book to learn its contents, that's direct competition, making that facet of the work entirely untransformed.
> They can determine how much each work contributed based on those weights, so it's dishonest for them to argue it isn't possible.
I don’t know about impossible but it’s definitely not a straightforward read from the post-training weights as you’re implying, unless you’re aware of some technique I’m not aware of.
The closest you could get would be the weight differential from training with a given work. But that’s massively dependent on training order, so that it’s certainly not at all a good measure of “contribution.”
Agreed. Moreover, the authors of copyright law could never have anticipated this type and scale of abuse. Maybe the companies are legally in the right, maybe not, but that's irrelevant for the question of whether it's ethical. The EFF's post definitely goes against their mission to "ensure that technology supports freedom, justice, and innovation for all people of the world."
The question of why US copyright law is administered and, to some extent regulated within the Legislative rather than Executive branch has been raised in a dead thread. The cogent point is made that under the US Constitution, the phrase "checks and balances" generally applies to both the division of powers amongst the three branches (Judiciary in addition to the two previously mentioned), and the principle of review and oversight amongst those branches (e.g., legislation is passed by Congress, approved or vetoed and administered by the President / Executive branch, and subject to interpretation or invalidation by the Judiciary; executive appointments are subject to Congressional approval; and members of both the Executive and Judiciary may be impeached and removed by Congress).
That said ...
... there are other instances in which separation of powers is not strictly followed. Examples which come to mind are:
- Administrative law judges (ALJs), notably in matters concerning Social Security and Immigration law, being a judiciary function under the executive.
- The Sergeants at Arms of the US Senate and US House, both legislative bodies, but performing executive functions. Recent history suggests that the Executive cannot be entirely relied upon to provide this function.
- Judicial Review is probably the biggest appropriation of powers, in which the US Supreme Court arrogated the right to rule on, interpret, and invalidate legislation. This is a power arguably derived absent any constitutional, legislative, or executive foundation.
And of course the present Administration has increasingly expressed a philosophy not only of Unitary Executive, but increasingly of Unitary Government, enacting law by decree, executing citizens without due process, and openly flouting courts. H.R. 6028 could be seen as part of this expansion of the Executive.
Which still leaves us with the question of how Congress ended up administering copyright.
I don't have a full history, and have only been exploring the question for the past hour or so.
The US Copyright Office itself has a history page noting that:
On July 8, 1870, Congress centralized the administration of copyright law in the Library of Congress at the encouragement of Librarian of Congress Ainsworth Rand Spofford.
Which remedied the previous arrangement in which Copyright was administered by ... the Judiciary.
Why Congress ended up regulating copyright is probably largely a set of historical accidents and conveniences. The Library of Congress does in fact serve Congress (and IIUC the Judiciary, to which it is also proximate) as a legislative research tool. I've read enough of the annual reports in the latter half of the 19th century to know that the Library was growing rapidly at this time, and was constantly pressed (literally) for space, culminating in the commissioning, construction, and opening of the separate Library of Congress Jefferson Building, in which the main collection is now housed. (As I'd recently commented, there were concerns at the time of how merely moving to an adjacent building might affect retrieval time for materials.)
Arguably, the US Library of Congress had, and still has, more expertise in the management of large corpora of physical publications than virtually any other institution on Earth. Copyright registration itself served the interests of Congress by growing the collection. And as of the late 19th century, the overall size of the US government, though growing, was still comparatively small. The Executive would possibly have had neither the interest nor capacity to administer the Library, or even the Copyright office sufficiently, nor the convergence of goals in growing the Library's collection noted here. Given numerous issues with other areas of intellectual property which are administered under the executive (patents and trademark, though my criticisms are largely of the former), its also possible Things Could Have Gone Badly Wrong, though arguably as the EFF piece notes they have already. Though the House legislation seems likely to worsen that.
The present situation though is that the Library of Congress and Copyright Office do strongly blur the separation of powers principle, affording a complex set of legislative, executive, and even judiciary roles, all under the Legislative branch.
That just my own nonexpert nutshell summary. If anyone has further information on the history of the US Copyright Office, legislation, and judicial rulings, please pitch in.
It's in the constitution. Article I, Section 8, clause 8:
> The Congress shall have Power [...] To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries;
Unless your question was more, "why was that written into the constitution". In which case the answer basically boils down to the fact that the framers intended for Congress to be the most powerful branch. The modern de facto running of the country places far more power under the executive than the framers intended.
The Constitutional power of Congress is to enact laws, but not to execute them. Article I generally lists other authorities of Congress, e.g., "to coin money", but the Treasury is an Administrative office, not a Legislative one. Another interesting example would be the establishment of a Post Office, which is now an independent agency, though under the Executive.
You'll also find: "The executive Power shall be vested in a President of the United States of America." Art II, Sec 1. That establishes the Separation concept, though neither "separation of powers" nor "checks and balances" are explicitly stated in the US Constitution. They are part of the political discussion in which the Constitution was framed, however.
It suppose I can see "executive power should be part of the executive branch" as a facile argument because it does seem basic and a bit tautological, but it is still quite a strong point. It needs to be addressed rather than just identified and dismissed.
No surprise that you'd show up to shill for it.Your argument boils down to 'if it looks like an executive branch agency, then the Executive branch should have control over it' rather than accepting that Congress is free to set things up as it sees fit within the Constitutional constraints.
> accepting that Congress is free to set things up as it sees fit within the Constitutional constraints.
The unitary-executive crowd worships a golden calf, namely the first three words of Article 2: "The executive Power shall be vested in a President of the United States of America."
One could argue that the "executive Power" is whatever the Congress says it is. Suppose that a future Congress were to decide that the president is to be, basically, the butler for the federal government — think Mr. Carson on Downton Abbey, in charge of the household staff but still very much a servant himself — with of course the other specifically-enumerated presidential powers (read: duties) in the remainder of Article 2. It's hard to say in advance that such a hypothetical congressional action would be categorically inconsistent with the Framers' intentions.
"Shilling" would require me to care about the policy, which I don't. The genius of the founders is that they realized that structure and power allocation was more important than policy, so that's what I'm commenting about.
On that point, Congress cannot "set things up as it sees fit." The constitution goes to great lengths to create a complex, three-branch system of government with specific powers allocated to each branch. Anytime Congress creates something new, it has to fit it into the three-branch model in a way that is consistent with the principles of that model. It's like a "pure" microkernel in computer science: there is a framework that dictates what goes in kernel space versus user space. Except with the constitution, the structural principles are legally binding. You can't delegate executive functions to mere employees of the legislative branch, just like in a pure microkernel you can't put the GUI into the kernel.
In this case, the DMCA creates civil and criminal liability. Creating exceptions for that is the exercise of a quintessential executive power--enforcement discretion. That power must be allocated to an executive-branch agency.
Hypothetically, if Congress passed legislation saying "it looks like an executive branch agency, the Executive branch should have control over it" you'd consider that a generally reasonable position all else equal?
If you concede that it looks like an executive agency then it actually seems quite proper that the executive control it.
I don't think you've at all addressed why moving anything there towards the executive is desirable, especially given the capriciousness of the current executive.
FYI: though EFF articles have individual named authors, they go through an extensive collective editing process. Every post will have had at least one domain-specific lawyer reviewer who signs off on it.
That is a rather narrow definition of checks and balances. The term can be applied to any group of organizations where each organization has power and interest to limit the power of the others.
The article is talking about a bill that restructures a body in the U.S. federal government. In that context, “checks and balances” has a specific, well-known meaning. It’s like writing an article about Fedora 42 and using the term “kernel.” In that context, readers expect the term to be used in a specific way.
Such an extreme and emotional statement makes me think you've never really thought it through. For instance: without copyright the GPL is nothing. Also without copyright, all of the profit made on creative works (of a perhaps smaller pie) would get be kept by distributors like Amazon or Netflix. Authors wouldn't get a dime anymore, it'll all go to the likes of Bezos.
That’s ok, GPL’s entire purpose and only restriction is to prevent other copyrights.
> without copyright, all of the profit made on creative works (of a perhaps smaller pie) would get be kept by distributors like Amazon or Netflix
This is already true in most cases: companies own everything their employees create for them. And without copyright, studios would still pay artists, because that’s the only way art is created (which even rich people want, although you probably and I think their taste mostly sucks, so does everyone else’s…)
> without copyright, all of the profit made on creative works (of a perhaps smaller pie) would get be kept by distributors like Amazon or Netflix
Assuming copyright gets dismantled is a good-faith way, Netflix/Amazon remaining as gatekeepers sounds unlikely, IMO. Free software clients like Popcorn Time provide a better experience and would be able to exist without threats from copyright trolls.
It's also much more robust regarding cultural preservation (as users and organizations can keep DRM-free local copies) and censorship (being torrent-based makes it much harder to delete a movie from existence).
Besides the unfairly long duration of protection, intellectual property also is unfairly used to squash small firms via frivolous lawsuits.
I won't use an argument in favor of AI training here because AI can probably still be trained by fair-use information extraction from copyrighted works.
Without copyright, we can return to a patronage based system. Both rich and poor consumers gladly offer proportional patronage for authors they truly believe in.
Humanity will progress just fine via its scientific works which don't really require a copyright. Arxiv proves it.
The cost imposed by GPL not working will be negligible compared to the benefit of free use.
I understand it risks adding unpredictable political corruption to the process, but I feel that such unpredictable corruption is exactly what it takes to gradually destroy something in an indirect way.
It is not clear to me what their political agenda is. Overall it might be good for AI if the goal is to scrape freely and use it for AI training.
Last time someone uttered something similar, I didn't get an answer, so I'll ask it to you: what entitles you to free access to any song, movie or book?
This: The basic idea of freedom, that I should be able to generally do things including accessing media without interference from a third party.
Someone using a physical property can possibly deprive others of its use. This applies to the physical mediums of songs, movies, or books, but not the songs, movies, or text of the books themselves.
Intellectual property isn't real, it's a concept that exists to support copyright, which exists for this exact purpose stated in the Constitution:
"[the United States Congress shall have power] To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries."
I'm ok with accepting a temporary limitation on my freedom to support those who make songs, movies, or books, but life of the author + 70 years, plus the ability to assign the right to corporations which don't die, is not reasonably "limited" these days. It should be something like 5 years today.
No one is entitled to be a songwriter, movie director, or author; society needs people doing other things too.
Provided a friend of mine agrees to let me borrow (and copy) their media containing music / movie / book / whatever, what entitles you to interfere with such agreement?
Especially since that agreement didn't involve you.
There's no $deity-given right to control what happens to stuff you wrote / designed etc, once it's been published. Copyright is, sorry was, a legal construct meant to promote people creating artwork.
Once it overshot that intent bigtime, there's no justification for keeping it around. At least not in its current form.
I want copyright to be completely abolished and patronage to re-become normal and common. Most of my favorite artists already distribute most of their work for free and rely on the latter.
>what entitles you to free access to any song, movie or book?
Does this sound profound to you? When you see yourself type it out, does it seem like you've really came up with a zinger?
What entitles them to come in and police my hard drive platters with "you can't write that sequence of bits to storage, that's our sequence of bits"? It's sort of a weird idea, sounds kind of medieval. Like King Cnut has granted them license to "the birds in the forest, and the timber, and the water that runs through the meadows".
Your question is a loaded question founded on a false premise that the author of the content has an innate right to its viewership. There is no such innate right.
Also, the argument that you made elsewhere about "damages" is nonsense because there is no damage from someone viewing what they were never going to pay for anyway, and there also is no deprivation.
This is a one-sided article which does not discuss the opposing view, or the reason why they thought congress should appoint. Ironically, if this became law then it might have prevented Trump from removing the librarian as he attempted in 2025 (still pending in the supreme court). It also includes a term limit of 10 years.
The plain language of the bill’s summary on the bill’s web page (ignoring the EFF article) explains it quite clearly:
1. Gives power to Congress to appoint/remove the librarian rather than the president (cool, great)
2. Strips the copyright power held by the Library of Congress away, library of Congress becomes a supporting resource like a consultant
3. Reassigns that same power to a different position that’s politically appointed by the president.
What you are saying is technically true, but the deck chairs have been shuffled around in a way that seems to at least partially negate the positive change.
I also find it odd that this was passed in a voice vote. It’s hard for me to tell if that means it has strong bipartisan support? I guess I’d have to watch a video recording of the proceedings to know. If I am recalling correctly, congresspeople can call for a tallied vote if they think the voice vote was too ambiguous.
I don't really understand the hypothetical problems here. "The copyright office head would be a presidential appointee, which could make the copyright office more political". I mean, I guess? Are people worried they're going to start selectively enforcing copyright law? But they don't enforce copyright law right now...
Yes, so what this does is centralized that selective enforcement directly under politicized control, so that it can be weaponized against political enemies.
Are you kidding? If there’s something in there they don’t like I don’t put it past this administration to break it internally and then make a case for shutting it down. This whole thing sounds very similar to the postal service situation…
They will break the system and use it for their friends. No way they are shutting it down. There is way too much money to be made in selective enforcement.
> Are you kidding? If there’s something in there they don’t like I don’t put it past this administration to break it internally and then make a case for shutting it down.
Might be a win? The copyright system is one of the major suspects for why US industry ended up crippled and replaced by Asian labour refusing to respect US IP laws to their significant advantage. To say nothing of the corrosive influence on culture of locking down music and stories. The biggest IP success in the last 50 years seems to have been Open Source because they built a framework inside the copyright system to achieve the opposite outcome and build a thriving industry despite the lawyers trying to encourage them in alternative directions.
The people defending the copyright system should have to keep making their case until they come up with something persuasive for how they're helping.
Odd that the article doesn't mention parties at all, although perhaps this was in an attempt to avoid accusations of partisanship that might ensue from stating facts.
Anyway, a quick look at https://www.congress.gov/bill/119th-congress/house-bill/6028... indicates that all 4 sponsors of the bill are Republicans. The Actions tab seems to indicated that the bill got only 12 minutes of debate before being passed,; I hope this is an artifact of how the page is updated rather than the actual time spent on considering it.
The article doesn't mention parties because it's irrelevant. A bad bill is bad on its merits, not because of who has brought it about.
It’s relevant, because you shouldn’t vote for politicians who make bad policies, and most party members tend to vote with their party.
Unfortunately, the Democrats haven’t demonstrated themselves to be much better (at least, I’m not aware of them opposing copyright).
2 replies →
The identity of the people who crafted the bill is the second most relevant thing besides the bill itself.
1 reply →
Ignoring power politics doesn't make them go away
5 replies →
In 2026, a discussion of a bill proposing to make an existing position into a presidential appointee is very different if that bill was proposed by Democrats or Republicans. To pretend otherwise is to ignore virtually all of the current administrations actions.
For those of us at home who need to decide which team to root for its very much relevant when and what bills a party sponsors.
1 reply →
wow, i had always assumed actual laws have to pass a recorded vote, but its not true...
from wiki:
https://en.wikipedia.org/wiki/Voice_vote#United_States
> Members can request a division of the assembly (a rising vote, where each sides rise in turn to be counted)
Isn’t this the important bit? IIRC this can be demanded by anyone. If it passes by a voice vote, assume your representative voted yay or abstain.
Oh and the biggest bullshit about this is it removes one’s ability to hold their local representatives accountable. I just assume the worst!
From what I understand it’s rather true that a lot of Congress’ actual work is incredibly boring and that these procedures were invented to move it along.
You can see a lot of difference in the way congresspeople talk based on whether it’s televised or not as well, especially in committees.
I’m just a little surprised that voice votes haven’t been replaced by some kind of digital process. A voice vote doesn’t save time compare to a modern method of tallying votes. Why avoid making records when records are so “cheap” these days?
2 replies →
Nope. If your local representative disagreed, he or she would have called for an actual vote. Your local representative agreed.
Hey, the economy is great and gas is cheap. All we had to put up with is mean tweets.
Whenever anyone complains about Trump, remind them he’s not the cause but the product. Seventy million voted for him, and Republicans in congress let him do illegally what they cannot accomplish legislatively. And all the while they’re busy selling the country for parts, whether through tax policy, or neutering the CAFE standards, or handing copyright to Disney.
He's certainly part of the cause, it seems the reason his handlers were able to use him so effectively is because he was going to jail if republicans and Musk et al. (and Putin?) didn't use him to get the election.
Expecting a convicted felon to live with the consequences of their further crimes, when all they need do is end the republic and steal billions, with the help and financial support of tech bros, ... well.
There's a very simple first patch to a new republic, no convicts in elected offices.
I usually agree with the EFF on things, but after reading their linked https://www.eff.org/deeplinks/2025/05/us-copyright-offices-d... I couldn't disagree more. An LLM is a predict the next word algorithm. If the model is overfitting, it's basically copy paste. There have been several documented instances where that happened and full GPL code, including headers and attribution were copy/pasted by the "AI."
AI is essentially copy paste with more steps. The part that AI companies use to defend this is ?how are we supposed to decide how much each author deserves? They try to wave this away, but their own model can tell them. Their models work off of weights. They can determine how much each work contributed based on those weights, so it's dishonest for them to argue it isn't possible. The way the models are engineered now don't make this possible, but that's intentional and we can all recognize that. They throw up their hands and claim it's not possible because they simply don't want to pay.
The most infurating thing however is how AI companies sidestep the IP rights of authors, but then claim to own those IP rights when their own generated output leaks. Anthropic filed DMCA takedowns on the leaked claude code repos, claiming ownership over something they explicitly have stated is almost entirely AI generated as part of their marketing. They take code, mix it up just enough to scrub away the GPL or whatever license belongs on it, then try to claim ownership of the result, in spite of the Copyright Office repeatedly stating that AI generated works have no copyright protection at all.
Is it actually possible to determine how much the weights were influenced by each work?
I might recall reading some interpretability paper years ago that trained a special model that could attribute each answer to a part of the corpus (like Wikipedia, ArXiV, or "Blogs") but it had a non-zero effect on performance and wasn't nearly as straightforward as weights go in, attribution comes out.
It’s very possible to determine similar works that existed earlier, and from that, recover attribution.
The “downside” is you may attribute similar works that weren’t inspirations, but coincidental. But I think that’s an upside: when someone discovers something novel and great but their work fails because of bad luck or non-novel details, then the discovery is finally recognized in another work, I think they should still be attributed.
>Is it actually possible to determine how much the weights were influenced by each work?
It will be very possible once they become the owners of the intellectual property being infringed. Think about how it was "impossible" to implement DRM on music and movies in the early days of youtube. Now, Google owns the content and platform, and suddenly their "rolling cypher" which involves no encryption at all is supposedly enforcable DRM.
The Silicon Valley tech bros play the same game every time. They violate the law, say it's just too darn difficult to obey the law without stifling progress, and then they get away with it until they kill all the competition. At which point, the law is once again applicable to anyone that might try to challenge them.
Remember how Amazon destroyed all the other retailers when they had a decade of no sales tax while brick & mortar had to obey it. "Calculating sales tax for 50 different states?! That's impossible!!!" What a load of shit...
Now, knowing that they're going to do this playbook again, how do you think it's going to play out? We've already seen it. Anthropic steals your copyrighted code, puts together their claude code project, the code for that project leaks, but now THEY own it! They sent DMCA takedowns on that AI generated code. AI generated code enjoys no copyright protection, it cannot be DMCAed under the law, there's no copyright on it. But Anthropic claims there is, and Github will obey the takedown, and nobody has the money to step up and stop them.
See where this is going? Once they achieve market dominance, they will claim that all the code generated by claude belongs to Anthropic, your prompts belong to you, but THEIR machine generated THEIR code and you only purchase a license to it with your tokens. A limited license. It might be revokable, it might expire, maybe you need to pay an annual fee to keep using THEIR code Claude generated for you. And if you actually just write code on your own, without Claude? Well, prepare to be sued like a network printer is sued by the RIAA because that's going to happen too. They will have their robot scour your code for "fair use" training and discover that it's just too similar to something their machine generated a year earlier. Sorry open source programmer, here's your legalese nasty gram. It appears you owe Anthropic some money.
1 reply →
> An LLM is a predict the next word algorithm.
This is what's known as a category error; an LLM is a 'model', not an algorithm.
It's not even an accurate claim; LLMs predict the next token, not the next word.
> AI is essentially copy paste with more steps
What about when AI creates a limerick about a kubernetes cluster run by Buddhist Monks? Or any number of other novel creations?
Fortunately the courts recognized the transformative use involved in making a model, which is fair use of copyrighted works, in kadrey v meta platforms.
> The most infurating thing however is how AI companies sidestep the IP rights of authors
transformative use falls under fair use, permission from authors is not needed to use legally acquired copyright works for training. Kadrey v Meta Platforms and Bartz v Anthropic.
> but then claim to own those IP rights when their own generated output leaks.
Corporations gonna do corporate things. Blatant hypocrisy is par for the course. Organize and take them to court.
I could not agree more with EFF.
There’s a difference between training a model and using a model. Training involves copyrighted works but fair use is not just about use of copyrighted works, it’s about whether the use is transformative and substitutes the original market. I struggle to see how is not transformative under these criteria.
The use of the model (being able to output copies of GPL software) is a different question. This depends on the circumstances: if GPL code is exactly reproduced then it very well could be subject to the license of the original work.
I don’t understand the legal objections to the fair use of protected IP. Licenses are legal documents, not moral imperatives. GPL only exists because of copyright law, and you can’t write a license that supercedes copyright law if you don’t like the law.
The Claude Code example is completely different, hosting a repo with the leaked code is clearly not fair use.
So in your consideration the user of training materials for LLMs don't allow the LLMs to compete with the original authors? That's part of the assessment of transformation - does the resulting work compete with the original.
If I no longer need your book to learn its contents, that's direct competition, making that facet of the work entirely untransformed.
> They can determine how much each work contributed based on those weights, so it's dishonest for them to argue it isn't possible.
I don’t know about impossible but it’s definitely not a straightforward read from the post-training weights as you’re implying, unless you’re aware of some technique I’m not aware of.
The closest you could get would be the weight differential from training with a given work. But that’s massively dependent on training order, so that it’s certainly not at all a good measure of “contribution.”
Agreed. Moreover, the authors of copyright law could never have anticipated this type and scale of abuse. Maybe the companies are legally in the right, maybe not, but that's irrelevant for the question of whether it's ethical. The EFF's post definitely goes against their mission to "ensure that technology supports freedom, justice, and innovation for all people of the world."
The question of why US copyright law is administered and, to some extent regulated within the Legislative rather than Executive branch has been raised in a dead thread. The cogent point is made that under the US Constitution, the phrase "checks and balances" generally applies to both the division of powers amongst the three branches (Judiciary in addition to the two previously mentioned), and the principle of review and oversight amongst those branches (e.g., legislation is passed by Congress, approved or vetoed and administered by the President / Executive branch, and subject to interpretation or invalidation by the Judiciary; executive appointments are subject to Congressional approval; and members of both the Executive and Judiciary may be impeached and removed by Congress).
That said ...
... there are other instances in which separation of powers is not strictly followed. Examples which come to mind are:
- Administrative law judges (ALJs), notably in matters concerning Social Security and Immigration law, being a judiciary function under the executive.
- The Sergeants at Arms of the US Senate and US House, both legislative bodies, but performing executive functions. Recent history suggests that the Executive cannot be entirely relied upon to provide this function.
- Judicial Review is probably the biggest appropriation of powers, in which the US Supreme Court arrogated the right to rule on, interpret, and invalidate legislation. This is a power arguably derived absent any constitutional, legislative, or executive foundation.
And of course the present Administration has increasingly expressed a philosophy not only of Unitary Executive, but increasingly of Unitary Government, enacting law by decree, executing citizens without due process, and openly flouting courts. H.R. 6028 could be seen as part of this expansion of the Executive.
Which still leaves us with the question of how Congress ended up administering copyright.
I don't have a full history, and have only been exploring the question for the past hour or so.
The US Copyright Office itself has a history page noting that:
On July 8, 1870, Congress centralized the administration of copyright law in the Library of Congress at the encouragement of Librarian of Congress Ainsworth Rand Spofford.
<https://www.copyright.gov/history/copyright-exhibit/history-...>
Which remedied the previous arrangement in which Copyright was administered by ... the Judiciary.
Why Congress ended up regulating copyright is probably largely a set of historical accidents and conveniences. The Library of Congress does in fact serve Congress (and IIUC the Judiciary, to which it is also proximate) as a legislative research tool. I've read enough of the annual reports in the latter half of the 19th century to know that the Library was growing rapidly at this time, and was constantly pressed (literally) for space, culminating in the commissioning, construction, and opening of the separate Library of Congress Jefferson Building, in which the main collection is now housed. (As I'd recently commented, there were concerns at the time of how merely moving to an adjacent building might affect retrieval time for materials.)
Arguably, the US Library of Congress had, and still has, more expertise in the management of large corpora of physical publications than virtually any other institution on Earth. Copyright registration itself served the interests of Congress by growing the collection. And as of the late 19th century, the overall size of the US government, though growing, was still comparatively small. The Executive would possibly have had neither the interest nor capacity to administer the Library, or even the Copyright office sufficiently, nor the convergence of goals in growing the Library's collection noted here. Given numerous issues with other areas of intellectual property which are administered under the executive (patents and trademark, though my criticisms are largely of the former), its also possible Things Could Have Gone Badly Wrong, though arguably as the EFF piece notes they have already. Though the House legislation seems likely to worsen that.
The present situation though is that the Library of Congress and Copyright Office do strongly blur the separation of powers principle, affording a complex set of legislative, executive, and even judiciary roles, all under the Legislative branch.
That just my own nonexpert nutshell summary. If anyone has further information on the history of the US Copyright Office, legislation, and judicial rulings, please pitch in.
It's in the constitution. Article I, Section 8, clause 8:
> The Congress shall have Power [...] To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries;
https://www.archives.gov/founding-docs/constitution-transcri...
Unless your question was more, "why was that written into the constitution". In which case the answer basically boils down to the fact that the framers intended for Congress to be the most powerful branch. The modern de facto running of the country places far more power under the executive than the framers intended.
The Constitutional power of Congress is to enact laws, but not to execute them. Article I generally lists other authorities of Congress, e.g., "to coin money", but the Treasury is an Administrative office, not a Legislative one. Another interesting example would be the establishment of a Post Office, which is now an independent agency, though under the Executive.
You'll also find: "The executive Power shall be vested in a President of the United States of America." Art II, Sec 1. That establishes the Separation concept, though neither "separation of powers" nor "checks and balances" are explicitly stated in the US Constitution. They are part of the political discussion in which the Constitution was framed, however.
*
[flagged]
[flagged]
[flagged]
Yes, but saying just "congress" implies both chambers passed it
6 replies →
[flagged]
The author, Joe Mullin, is a policy analyst:
<https://www.eff.org/about/staff/joe-mullin>.
He's been working in that capacity with the EFF since at least 2018: <https://www.eff.org/deeplinks/2018/02/ipr-process-saves-80-c...>.
Your further objections are ... facile.
It suppose I can see "executive power should be part of the executive branch" as a facile argument because it does seem basic and a bit tautological, but it is still quite a strong point. It needs to be addressed rather than just identified and dismissed.
11 replies →
No surprise that you'd show up to shill for it.Your argument boils down to 'if it looks like an executive branch agency, then the Executive branch should have control over it' rather than accepting that Congress is free to set things up as it sees fit within the Constitutional constraints.
> accepting that Congress is free to set things up as it sees fit within the Constitutional constraints.
The unitary-executive crowd worships a golden calf, namely the first three words of Article 2: "The executive Power shall be vested in a President of the United States of America."
One could argue that the "executive Power" is whatever the Congress says it is. Suppose that a future Congress were to decide that the president is to be, basically, the butler for the federal government — think Mr. Carson on Downton Abbey, in charge of the household staff but still very much a servant himself — with of course the other specifically-enumerated presidential powers (read: duties) in the remainder of Article 2. It's hard to say in advance that such a hypothetical congressional action would be categorically inconsistent with the Framers' intentions.
"Shilling" would require me to care about the policy, which I don't. The genius of the founders is that they realized that structure and power allocation was more important than policy, so that's what I'm commenting about.
On that point, Congress cannot "set things up as it sees fit." The constitution goes to great lengths to create a complex, three-branch system of government with specific powers allocated to each branch. Anytime Congress creates something new, it has to fit it into the three-branch model in a way that is consistent with the principles of that model. It's like a "pure" microkernel in computer science: there is a framework that dictates what goes in kernel space versus user space. Except with the constitution, the structural principles are legally binding. You can't delegate executive functions to mere employees of the legislative branch, just like in a pure microkernel you can't put the GUI into the kernel.
In this case, the DMCA creates civil and criminal liability. Creating exceptions for that is the exercise of a quintessential executive power--enforcement discretion. That power must be allocated to an executive-branch agency.
3 replies →
Hypothetically, if Congress passed legislation saying "it looks like an executive branch agency, the Executive branch should have control over it" you'd consider that a generally reasonable position all else equal?
If you concede that it looks like an executive agency then it actually seems quite proper that the executive control it.
4 replies →
I don't think you've at all addressed why moving anything there towards the executive is desirable, especially given the capriciousness of the current executive.
FYI: though EFF articles have individual named authors, they go through an extensive collective editing process. Every post will have had at least one domain-specific lawyer reviewer who signs off on it.
That is a rather narrow definition of checks and balances. The term can be applied to any group of organizations where each organization has power and interest to limit the power of the others.
The article is talking about a bill that restructures a body in the U.S. federal government. In that context, “checks and balances” has a specific, well-known meaning. It’s like writing an article about Fedora 42 and using the term “kernel.” In that context, readers expect the term to be used in a specific way.
1 reply →
Anything that destroys copyright is a good thing. It is a societal evil.
> [Copyright] is a societal evil.
Such an extreme and emotional statement makes me think you've never really thought it through. For instance: without copyright the GPL is nothing. Also without copyright, all of the profit made on creative works (of a perhaps smaller pie) would get be kept by distributors like Amazon or Netflix. Authors wouldn't get a dime anymore, it'll all go to the likes of Bezos.
> without copyright the GPL is nothing
That’s ok, GPL’s entire purpose and only restriction is to prevent other copyrights.
> without copyright, all of the profit made on creative works (of a perhaps smaller pie) would get be kept by distributors like Amazon or Netflix
This is already true in most cases: companies own everything their employees create for them. And without copyright, studios would still pay artists, because that’s the only way art is created (which even rich people want, although you probably and I think their taste mostly sucks, so does everyone else’s…)
7 replies →
> without copyright, all of the profit made on creative works (of a perhaps smaller pie) would get be kept by distributors like Amazon or Netflix
Assuming copyright gets dismantled is a good-faith way, Netflix/Amazon remaining as gatekeepers sounds unlikely, IMO. Free software clients like Popcorn Time provide a better experience and would be able to exist without threats from copyright trolls.
It's also much more robust regarding cultural preservation (as users and organizations can keep DRM-free local copies) and censorship (being torrent-based makes it much harder to delete a movie from existence).
RMS will happily tell you that he'd trade enforcability of the GPL for the non-existence of copyright.
1 reply →
Besides the unfairly long duration of protection, intellectual property also is unfairly used to squash small firms via frivolous lawsuits.
I won't use an argument in favor of AI training here because AI can probably still be trained by fair-use information extraction from copyrighted works.
Without copyright, we can return to a patronage based system. Both rich and poor consumers gladly offer proportional patronage for authors they truly believe in.
Humanity will progress just fine via its scientific works which don't really require a copyright. Arxiv proves it.
The cost imposed by GPL not working will be negligible compared to the benefit of free use.
This bill very much does not do that. It does the opposite, in fact. I encourage you to re-read the article.
I understand it risks adding unpredictable political corruption to the process, but I feel that such unpredictable corruption is exactly what it takes to gradually destroy something in an indirect way.
It is not clear to me what their political agenda is. Overall it might be good for AI if the goal is to scrape freely and use it for AI training.
3 replies →
Last time someone uttered something similar, I didn't get an answer, so I'll ask it to you: what entitles you to free access to any song, movie or book?
This: The basic idea of freedom, that I should be able to generally do things including accessing media without interference from a third party.
Someone using a physical property can possibly deprive others of its use. This applies to the physical mediums of songs, movies, or books, but not the songs, movies, or text of the books themselves.
Intellectual property isn't real, it's a concept that exists to support copyright, which exists for this exact purpose stated in the Constitution:
"[the United States Congress shall have power] To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries."
I'm ok with accepting a temporary limitation on my freedom to support those who make songs, movies, or books, but life of the author + 70 years, plus the ability to assign the right to corporations which don't die, is not reasonably "limited" these days. It should be something like 5 years today.
No one is entitled to be a songwriter, movie director, or author; society needs people doing other things too.
2 replies →
Provided a friend of mine agrees to let me borrow (and copy) their media containing music / movie / book / whatever, what entitles you to interfere with such agreement?
Especially since that agreement didn't involve you.
There's no $deity-given right to control what happens to stuff you wrote / designed etc, once it's been published. Copyright is, sorry was, a legal construct meant to promote people creating artwork.
Once it overshot that intent bigtime, there's no justification for keeping it around. At least not in its current form.
Because it’s effectively free to copy.
I want copyright to be completely abolished and patronage to re-become normal and common. Most of my favorite artists already distribute most of their work for free and rely on the latter.
1 reply →
What entitles you to use force to stop me creating a copy of something?
4 replies →
>what entitles you to free access to any song, movie or book?
Does this sound profound to you? When you see yourself type it out, does it seem like you've really came up with a zinger?
What entitles them to come in and police my hard drive platters with "you can't write that sequence of bits to storage, that's our sequence of bits"? It's sort of a weird idea, sounds kind of medieval. Like King Cnut has granted them license to "the birds in the forest, and the timber, and the water that runs through the meadows".
1 reply →
Your question is a loaded question founded on a false premise that the author of the content has an innate right to its viewership. There is no such innate right.
Also, the argument that you made elsewhere about "damages" is nonsense because there is no damage from someone viewing what they were never going to pay for anyway, and there also is no deprivation.
2 replies →
This is a one-sided article which does not discuss the opposing view, or the reason why they thought congress should appoint. Ironically, if this became law then it might have prevented Trump from removing the librarian as he attempted in 2025 (still pending in the supreme court). It also includes a term limit of 10 years.
https://www.stoneslaw.net/legislative-branch-agencies-clarif...
The plain language of the bill’s summary on the bill’s web page (ignoring the EFF article) explains it quite clearly:
1. Gives power to Congress to appoint/remove the librarian rather than the president (cool, great)
2. Strips the copyright power held by the Library of Congress away, library of Congress becomes a supporting resource like a consultant
3. Reassigns that same power to a different position that’s politically appointed by the president.
What you are saying is technically true, but the deck chairs have been shuffled around in a way that seems to at least partially negate the positive change.
I also find it odd that this was passed in a voice vote. It’s hard for me to tell if that means it has strong bipartisan support? I guess I’d have to watch a video recording of the proceedings to know. If I am recalling correctly, congresspeople can call for a tallied vote if they think the voice vote was too ambiguous.
I don't really understand the hypothetical problems here. "The copyright office head would be a presidential appointee, which could make the copyright office more political". I mean, I guess? Are people worried they're going to start selectively enforcing copyright law? But they don't enforce copyright law right now...
It’s not hypothetical at all. The FCC is currently being used for political attacks: https://www.cnn.com/2026/05/28/media/abc-fcc-disney-licenses...
Those who are under attack happen to also be the biggest copyrighter holders, so this would open up a new avenue of attack.
It's not hypothetical nor an unintended consequence. Most likely this is the point
> Those who are under attack happen to also be the biggest copyrighter holders, so this would open up a new avenue of attack.
Don't threaten me with a good time
Conversely you're already not dealing with that, so the letter and spirit of the law are both being ignored and the American voter doesn't care.
1 reply →
It's really hard for me to feel sorry for Disney here. Is it possible for both sides to lose a lawsuit?
1 reply →
>Are people worried they're going to start selectively enforcing copyright law?
Yes. Not only that, but to grant copyright protection only to those that are allied with/loyal to/bribe the current administration.
This would have massive, far reaching effects.
Never gotten any emails from lawyers, I see.
Copyright laws are heavily enforced, only selectively.
Yes, so what this does is centralized that selective enforcement directly under politicized control, so that it can be weaponized against political enemies.
> Are people worried they're going to start selectively enforcing copyright law?
Yes.
Are you kidding? If there’s something in there they don’t like I don’t put it past this administration to break it internally and then make a case for shutting it down. This whole thing sounds very similar to the postal service situation…
They will break the system and use it for their friends. No way they are shutting it down. There is way too much money to be made in selective enforcement.
1 reply →
> Are you kidding? If there’s something in there they don’t like I don’t put it past this administration to break it internally and then make a case for shutting it down.
Might be a win? The copyright system is one of the major suspects for why US industry ended up crippled and replaced by Asian labour refusing to respect US IP laws to their significant advantage. To say nothing of the corrosive influence on culture of locking down music and stories. The biggest IP success in the last 50 years seems to have been Open Source because they built a framework inside the copyright system to achieve the opposite outcome and build a thriving industry despite the lawyers trying to encourage them in alternative directions.
The people defending the copyright system should have to keep making their case until they come up with something persuasive for how they're helping.
4 replies →