"Do not fall into the trap of anthropomorphizing Larry Ellison. You need to think of Larry Ellison the way you think of a lawnmower. You don't anthropomorphize your lawnmower, the lawnmower just mows the lawn - you stick your hand in there and it'll chop it off, the end. You don't think 'oh, the lawnmower hates me' - lawnmower doesn't give a * about you, the lawnmower can't hate you. Don't anthropomorphize the lawnmower. Don't fall into that trap about Oracle." --Bryan Cantrill
What about a guy with a lawnmower intentionally chopping off your foot? And what about a CEO making decisions on behalf of the corporate entity called Oracle?
It doesn't matter if the lawnmower is chopping off feat intentionally or not. If it's presenting a public hazard, it should be stopped. Even though the ocean couldn't care less it doesn't prevent us from building breakwalls. And Larry Ellison isn't unstoppable as the ocean.
Rather than harmless, what the lawnmower analogy achieves is to make him seem blameless. "How can you assign blame to a thing that's just doing what it's built to do without any intent?" It's important to push back against that. Intent isn't a necessary condition for something to be dangerous, and so it's not necessary to ask whether it is intentional or not. A rabid dog will be put down regardless of who is to blame for it.
I think the point of the lawnmower analogy is to subjectively characterize Larry Ellison and Oracle from the viewpoint of someone who has had the unfortunate experience of having to deal personally with either of them in some way. You feel powerless against the machine, as it were, and the best course of action is to keep yourself far from the blade.
I think you both are missing the point. Hes sayin larry is inhuman, hes closer to an apex predator or a serial killer freak of nature than your ordinary human
We're not missing the point, we simply don't agree. You're putting him too much on a pedestal. He's an ordinary human bound by the same laws anyone else. If he's criminally insane and can't be held accountable for his actions, he should be locked up in an mental institution, not running a corporation.
Don not fall into the trap of anthropomorphizing Larry Ellison. You need to think of Larry Ellison the way you think of a 1958 Plymouth Fury named Christine. You don't anthropomorphize your killer car, the killer car is just out for blood - you rebuild it and it kills your friends and family, the end. You don't think 'oh, Christine hates me' - Christine doesn't give a * about you, Christine can't hate you. Don't anthropomorphize a 1958 Plymouth Fury named Christine. Don't fall into that trap about Oracle.
Never thought to hear someone here mention Maximum Overdrive. Indeed a hilarious movie and the only movie for which AC/DC ever recorded an (instrumental) soundtrack.
The flaw in the analogy is that oracle is made of people. It's not a lawnmower, it's a group of people with human motivations and human thoughts. It's true they don't HATE people, but it's also true they absolutely have no conscience in negotiations and have zero concern about relationships.
Maybe so, but perhaps the most successful part of the AI marketing pitch has been exploiting management's hatred of labor (and vice versa).
A significant motivator for a lot of the nonsensical AI layoffs and initiatives over the last three years has been that it's a great opportunity for management to get ill on their slaves.
Tech labor got a little too big for their britches during the hiring spree of 2021, and AI was a great opportunity to take them down a peg, even if it offered little in the way of ROI.
He's talking about the owner of Oracle, not the people who work there. It doesn't matter how many smart and talented real humans work at Oracle, the guy running it acts like a fleshy paperclip optimizer.
How many bureaucracies in history have behaved in deeply misanthropic ways, especially when led by misanthropic leaders?
Don't underestimate the power of the desire to keep one's job to cause a large organization to behave towards its own self-perpetuation despite the moral lines that must be crossed by individuals to do so.
> How many bureaucracies in history have behaved in deeply misanthropic ways, especially when led by misanthropic leaders?
Lots!
> Don't underestimate the power of the desire to keep one's job to cause a large organization to behave towards its own self-perpetuation despite the moral lines that must be crossed by individuals to do so.
For reference, Walter Lippmann is the guy from the NYT who covered up the Holodomor in Ukraine (1930s) in the west as well as Stalin's gulags and purges. I wouldn't call him the voice of the proletariat in any way. He was the voice of the Stalinist in the west and the term tankie was invented for him and people like him.
That’s an interesting story to bring up in this context, because I have always wondered how a fundamentally customer-hostile business like Oracle could be so successful. It really puts the lie to libertarian cliches about the invisible hand of capitalism that ends up being philanthropic in spite of itself.
I think free society in general absolutely requires its people to be generally moral and civil. Freedom inevitably evaporates as morality and civility decline.
And that's why it's important to go after morally corrupt, uncivil monsters.
The problem today is that the press and the DAs/courts are complicit--they're less interested than ever in pursuing public morality and civility as ends in themselves. They're mostly only interested in partisan politics.
The lawnmower exists, you might disagree, but it’s not your lawnmower, you can’t uncreate it. But you can and should take care never to put your hand near it.
Well (to make an analogy), nature produces psychopaths - some people's brains just aren't capable of empathy. Whether you characterize that observation as "legitimizing" or just "being realistic" is merely a philosophical concern.
What I meant was that there is no moralizing of natural force, which this metaphor tries to equivocate. Oh, it's just like the wind, don't get in Larry's way!
It’s been established that LLM-generated code is not copyrighted so I can fully understand the company living from copyrighted data to not accept LLM-generated contributions.
> It’s been established that LLM-generated code is not copyrighted
If that's a reference to Thaler v. Perlmutter, the only thing that's been established is that an LLM can't be considered an author under the Copyright Act, only a human being can. It says nothing about the consequences of a human claiming authorship of LLM-generated code, which would be relevant here.
My understanding (belief) is that it's going to depend on how much human involvement is there.
If you write a prompt and one-shot a problem and share the source code, that source code is probably not covered by copyright.
If you substantially edit or modify the generated code you would own the copyright.
It's like with a camera. If I set a camera and carefully aim it and somehow trigger the shutter then make adjustments in Photoshop, I own the copyright on that image.
If I stick a Flock camera on a pole somewhere and post the live output, there's been no meaningful human creative involvement in producing those images and so nobody can claim copyright on them.
I think if I as a human use an llm to do something technical that would qualify copyright, it should still qualify for copyright. How do you decide how much human is copyrightable. If I use a package that writes code or use a library for some piece of it, I could still copyright.
I don't like this idea that llm code can't be owned by a human, copyrighted. It's just code.
I think your last example with flock camera is relevant here - I can take a picture of a public football as a reporter or something (or a fan I guess) and I can copyright and sell that picture. Newspapers do it every day.
So if I stand on a street corner and take a pic, it's copyrightable. If I take a pic using a flock camera it should also be copyrightable, just like if my nest camera at home takes a pic of something, I can use that.
I guess you are saying "someone else owns the flock camera" so you don't get to own pictures. What if I buy the flock-like camera and put it up, I should own that.
You seem to have misunderstood an is/ought distinction. You may hold the (fairly extreme, as far as copyright goes) position that surveillance footage should be subject to copyright, but it's well established that it's not. Who owns the camera is irrelevant. At least in the US; I'm not aware of any jurisdictions that hold otherwise. This is why Wikipedia articles on world events in the past few decades are full of stills from surveillance cameras: it's one of the few sources of imagery of an event that are unambiguously legal to include, because unlike a photo or intentionally made video of something specific, it's not a creative work. It's also pretty firmly established that human authorship is required for something to be subject to copyright, and having an idea that lead to some particular expression is itself not sufficient; see, e.g.: https://en.wikipedia.org/wiki/Monkey_selfie_copyright_disput...
I appreciate your informed take. I follow the reasoning but I am amazed it works this way. I found some articles that supported what you said, and also said there's a follow-on industry that figured out how to alter and edit videos just enough for a revised video to have creative contribution and make it copyrightable.
This is a great case to study, but no determination of copyright was made. The only actual lawsuit was filed by PETA arguing that the monkey should have copyright, which led to a settlement with the human photographer and nothing else because obviously that's not possible.
For various reasons (mostly $$$) the guy never actually got a judgement. I think the chances are good that he could have prevailed in court; there is significant creative input to setting up cameras and triggers in a way to convince a wild monkey to take a selfie. It's not like he just left his camera sitting somewhere on accident and came back to find a photo in it.
I'm also not a lawyer, but I did do a lot of work in copyright for a company you've heard about.
> If you write a prompt and one-shot a problem and share the source code, that source code is probably not covered by copyright.
We will have to see about that! This is the kind of boundary that's still being figured out in court; it's going to depend on how hard you worked on the prompt. I highly doubt that even most slop was generated with a single half-ass prompt, and the bar is not as high as you might expect.
> If I stick a Flock camera on a pole somewhere and post the live output, there's been no meaningful human creative involvement in producing those images and so nobody can claim copyright on them.
It really depends on what pole, where, and why. In a parking lot in rural Wisconsin? Probably not. A recorded livestream of a political march? You likely have copyright.
I think that by virtue of the sheer amount of time spent using AI tools, it's pretty clear that these outputs have enough creative input to be copyrightable.
No, that’s not what they are saying. They’re saying that the code generated by a human with help from an LLM may potentially be. This is what I hope we are going to arrive at, eventually.
How could you establish what parts of the code was produced by an LLM vs updated by a human afterwards?
The LLM will output different results over time as the models get updated. Are we heading towards needing to retain a full prompt history that can be replayed against a specific LLM model version to prove what the output was for copyright purposes?
For those wondering what the difference is: consider what happens when an LLM regurgitates its training data. It's copyrighted... but not by the person who generated it.
This is a false narrative based on a (IMO often intentional) misunderstanding. It has by no means been established by any court that LLM-generated code is not copyrightable.
Thaler v. Perlmutter stands for a much narrower proposition and at any rate is not binding nationally, SCOTUS having denied certiorari.
This is not necessarily true. While no court has explicitly come out and said that copyright does not apply to AI-generated works of authorship, the US copyright office has[0]:
> Based on an analysis of copyright law and policy, informed by the many thoughtful comments in response to our NOI, the Office makes the following conclusions and recommendations:
> • Questions of copyrightability and AI can be resolved pursuant to existing law, without the need for legislative change.
> • The use of AI tools to assist rather than stand in for human creativity does not affect the availability of copyright protection for the output.
> • Copyright protects the original expression in a work created by a human author, even if the work also includes AI-generated material.
> • Copyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements.
> • Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis.
> • Based on the functioning of current generally available technology, prompts do not alone provide sufficient control.
> • Human authors are entitled to copyright in their works of authorship that are perceptible in AI-generated outputs, as well as the creative selection, coordination, or arrangement of material in the outputs, or creative modifications of the outputs.
> • The case has not been made for additional copyright or sui generis protection for AI-generated content.
> The Office will continue to monitor technological and legal developments to determine whether any of these conclusions should be revisited. It will also provide ongoing assistance to the public, including through additional registration guidance and an update to the Compendium of U.S. Copyright Office Practices.
Congress or the courts could, of course, override the stance of the copyright office, but I think it would be highly unusual for them to do so (particularly for something like this). It would however be a lot better if congress just stepped in and said no outright, but until then this will have to do.
Do you own the copyright to a painting you commissioned or does the artist? You may have described what you wanted, but the creative work is the output of the artist. Same with an LLM.
Sure, but I would be incredibly shocked if the courts overturned these conclusions. These kinds of determinations are within the remit of the USCO, so a court does not need to come out and say it if the USCO has already done so. Obviously, as I said it would be better if congress weighed in and solved this problem, given that the USCO is free to publish a new NOI to change it's practices/policies, but we all know that congress is too gridlocked atm for that to happen
I am trying really hard not to accuse you of not having read what you posted, because your conclusions are in strong tension with what it plainly says.
> Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis
It says a plain prompt is not enough but that is not the reality of real software development. People aren't one-shotting complex business apps. The vast majority of software development will trivially pass that bar and end up in the "requires case by case analysis".
You are a false narrative. I’m just repeating what I read. The thing is, it’s murky waters. Someone’s going to challenge it but do you want to be the guy who takes it on a chin?
The sibling comment lays this out and my original comment above is based on exactly the same link.
Also it's a different story intranationally for those of us who live in countries with much more restrictive/no fair use. Are you geolocking your software to the USA?
I highly doubt that that's the last word on that matter, but even if: Even before LLMs you could combine individual non-copyrighted components into something copyrighted.
"Do not fall into the trap of anthropomorphizing Larry Ellison. You need to think of Larry Ellison the way you think of a lawnmower. You don't anthropomorphize your lawnmower, the lawnmower just mows the lawn - you stick your hand in there and it'll chop it off, the end. You don't think 'oh, the lawnmower hates me' - lawnmower doesn't give a * about you, the lawnmower can't hate you. Don't anthropomorphize the lawnmower. Don't fall into that trap about Oracle." --Bryan Cantrill
This quote is why I have my Claude Code "done" notification sound be the sound of a lawnmower revving.
The lawnmower can't hate you. Don't anthropomorphize the lawnmower, no matter how clever it sounds. (Actually not that hard with Opus5)
What about a guy with a lawnmower intentionally chopping off your foot? And what about a CEO making decisions on behalf of the corporate entity called Oracle?
There is no guy, only the lawnmower, mowing lawns. Sometimes it chops off feet.
It doesn't matter if the lawnmower is chopping off feat intentionally or not. If it's presenting a public hazard, it should be stopped. Even though the ocean couldn't care less it doesn't prevent us from building breakwalls. And Larry Ellison isn't unstoppable as the ocean.
This seems to make Larry more harmless. A lawnmower does not think. Larry thinks, so it seems to white-wash Larry here.
It's more like that movie with Stephen King where the cars and other machines turn actively evil. Hilarious movie too, not great but hilarious.
Rather than harmless, what the lawnmower analogy achieves is to make him seem blameless. "How can you assign blame to a thing that's just doing what it's built to do without any intent?" It's important to push back against that. Intent isn't a necessary condition for something to be dangerous, and so it's not necessary to ask whether it is intentional or not. A rabid dog will be put down regardless of who is to blame for it.
I think the point of the lawnmower analogy is to subjectively characterize Larry Ellison and Oracle from the viewpoint of someone who has had the unfortunate experience of having to deal personally with either of them in some way. You feel powerless against the machine, as it were, and the best course of action is to keep yourself far from the blade.
I think you both are missing the point. Hes sayin larry is inhuman, hes closer to an apex predator or a serial killer freak of nature than your ordinary human
We're not missing the point, we simply don't agree. You're putting him too much on a pedestal. He's an ordinary human bound by the same laws anyone else. If he's criminally insane and can't be held accountable for his actions, he should be locked up in an mental institution, not running a corporation.
Don not fall into the trap of anthropomorphizing Larry Ellison. You need to think of Larry Ellison the way you think of a 1958 Plymouth Fury named Christine. You don't anthropomorphize your killer car, the killer car is just out for blood - you rebuild it and it kills your friends and family, the end. You don't think 'oh, Christine hates me' - Christine doesn't give a * about you, Christine can't hate you. Don't anthropomorphize a 1958 Plymouth Fury named Christine. Don't fall into that trap about Oracle.
Wrong movie
Never thought to hear someone here mention Maximum Overdrive. Indeed a hilarious movie and the only movie for which AC/DC ever recorded an (instrumental) soundtrack.
Maximum Overdrive. Even the electric turkey carver gets revenge.
you and what army are going to stop him?
If there were such an army, let's call it a union, would you join it?
A lawnmower doesn't decide to do something bad by itself. The case with L.E. is much more sinister.
The flaw in the analogy is that oracle is made of people. It's not a lawnmower, it's a group of people with human motivations and human thoughts. It's true they don't HATE people, but it's also true they absolutely have no conscience in negotiations and have zero concern about relationships.
> It's true they don't HATE people
Maybe so, but perhaps the most successful part of the AI marketing pitch has been exploiting management's hatred of labor (and vice versa).
A significant motivator for a lot of the nonsensical AI layoffs and initiatives over the last three years has been that it's a great opportunity for management to get ill on their slaves.
Tech labor got a little too big for their britches during the hiring spree of 2021, and AI was a great opportunity to take them down a peg, even if it offered little in the way of ROI.
Yep, it's been a huge tool in the control games management likes to play.
Oh you fell into the trap... Just don't anthropomorphize Oracle...
I'm not, I'm anthropomorphizing people.
So they're not human then. The quote holds up.
The original was not even about Oracle the corporation, it was about understanding the psychology of Larry Ellison specifically.
But I wasn't replying to the original, I was replying to another comment.
You replied to a quote about Larry Ellison personally with an unrelated remark about Oracle the corporation.
> It's true they don't HATE people, but it's also true they absolutely have no conscience in negotiations and have zero concern about relationships.
Sure. Like a lawnmower.
It won't be made of people for long given how many Ellison is culling to prop up his gamble on AI data centres.
It's a joke that you're missing.
He's talking about the owner of Oracle, not the people who work there. It doesn't matter how many smart and talented real humans work at Oracle, the guy running it acts like a fleshy paperclip optimizer.
I agree, but Larry doesn't hate people, he just enjoys crushing them for his own benefit. If you can't benefit him he doesn't care about you at all.
How many bureaucracies in history have behaved in deeply misanthropic ways, especially when led by misanthropic leaders?
Don't underestimate the power of the desire to keep one's job to cause a large organization to behave towards its own self-perpetuation despite the moral lines that must be crossed by individuals to do so.
> How many bureaucracies in history have behaved in deeply misanthropic ways, especially when led by misanthropic leaders?
Lots!
> Don't underestimate the power of the desire to keep one's job to cause a large organization to behave towards its own self-perpetuation despite the moral lines that must be crossed by individuals to do so.
I don't. I don't put anything past people.
Yes, lots, look at the current US government.
An AI powered lawnmower though...
more and more it seems like he's referring to billionaires more or less.
That's because you've had years of consuming media that blames everything on billionaires.
Ah, yes. The media. Famously the voice of the proletariat and the oppressed
Walter Lippmann is spinning in his grave
For reference, Walter Lippmann is the guy from the NYT who covered up the Holodomor in Ukraine (1930s) in the west as well as Stalin's gulags and purges. I wouldn't call him the voice of the proletariat in any way. He was the voice of the Stalinist in the west and the term tankie was invented for him and people like him.
That's a strange take. But okay
He also wrote the book "Public Opinion" in 1922 [1] which explained how the US mass media apparatus could be used to influence the common man.
[1] https://en.wikipedia.org/wiki/Public_Opinion_(book)
Billionaires own the media. Why would they blame themselves?
well, plus the evidence of billionaires pushing for things that aren't that great after all. Maybe good for their bankroll though.
Who does this metaphor serve? To me it seems to try to legitimize the actions as a force of nature.
It's a twist on the Scorpion and the Frog story.
In other words, Larry Ellison is going to do what Larry is going to do, and there's no use wondering why.
Usually, that's to make money and sue people, at any expense. It's in his nature.
https://en.wikipedia.org/wiki/The_Scorpion_and_the_Frog
Abusers say "that's just how I am" which is my point. Just because it's his nature doesn't mean it is at all natural.
That’s an interesting story to bring up in this context, because I have always wondered how a fundamentally customer-hostile business like Oracle could be so successful. It really puts the lie to libertarian cliches about the invisible hand of capitalism that ends up being philanthropic in spite of itself.
I think free society in general absolutely requires its people to be generally moral and civil. Freedom inevitably evaporates as morality and civility decline.
And that's why it's important to go after morally corrupt, uncivil monsters.
The problem today is that the press and the DAs/courts are complicit--they're less interested than ever in pursuing public morality and civility as ends in themselves. They're mostly only interested in partisan politics.
those libertarians almost certainly don't believe in asinine things like 'copyright'
It serves you.
The lawnmower exists, you might disagree, but it’s not your lawnmower, you can’t uncreate it. But you can and should take care never to put your hand near it.
Weird, we don't absolve murderers of guilt by saying they are an lawnmower.
I read it as a pretty obvious jab at him being inhuman.
That's exactly my point. Labelling him as inhuman gives him a pass, like how we don't criticize hurricanes as malevolent.
Well (to make an analogy), nature produces psychopaths - some people's brains just aren't capable of empathy. Whether you characterize that observation as "legitimizing" or just "being realistic" is merely a philosophical concern.
Since when are human actions categorized as being a consequence of nature? Does that let me steal or murder without consequence?
The lawn mower is operating within the laws that constraints it (laws of physics in this case) and not a greater sense of morality.
Lessons:
- The only way to constrain what the lawnmower does, is make sure what do desire is enforced by law (physics)
- A reminder, the lawnmower will chop your hand off and think nothing of it
- The lawnmower is not "evil" but if the laws that constrain it allow for evil behavior, it may do "evil" things.
Law does not constrain at all. Law only operates after the fact.
It’s nonsensical to define the actions of nature as legitimate or illegitimate.
What I meant was that there is no moralizing of natural force, which this metaphor tries to equivocate. Oh, it's just like the wind, don't get in Larry's way!
It’s been established that LLM-generated code is not copyrighted so I can fully understand the company living from copyrighted data to not accept LLM-generated contributions.
> It’s been established that LLM-generated code is not copyrighted
If that's a reference to Thaler v. Perlmutter, the only thing that's been established is that an LLM can't be considered an author under the Copyright Act, only a human being can. It says nothing about the consequences of a human claiming authorship of LLM-generated code, which would be relevant here.
You lay out my words in different order and claim I’m not correct.
No, your claim is "LLM generated code is not copyrighted." His claim is "LLM generated code is eligible for copyright."
My understanding (belief) is that it's going to depend on how much human involvement is there.
If you write a prompt and one-shot a problem and share the source code, that source code is probably not covered by copyright.
If you substantially edit or modify the generated code you would own the copyright.
It's like with a camera. If I set a camera and carefully aim it and somehow trigger the shutter then make adjustments in Photoshop, I own the copyright on that image.
If I stick a Flock camera on a pole somewhere and post the live output, there's been no meaningful human creative involvement in producing those images and so nobody can claim copyright on them.
I think if I as a human use an llm to do something technical that would qualify copyright, it should still qualify for copyright. How do you decide how much human is copyrightable. If I use a package that writes code or use a library for some piece of it, I could still copyright.
I don't like this idea that llm code can't be owned by a human, copyrighted. It's just code.
I think your last example with flock camera is relevant here - I can take a picture of a public football as a reporter or something (or a fan I guess) and I can copyright and sell that picture. Newspapers do it every day.
So if I stand on a street corner and take a pic, it's copyrightable. If I take a pic using a flock camera it should also be copyrightable, just like if my nest camera at home takes a pic of something, I can use that.
I guess you are saying "someone else owns the flock camera" so you don't get to own pictures. What if I buy the flock-like camera and put it up, I should own that.
You seem to have misunderstood an is/ought distinction. You may hold the (fairly extreme, as far as copyright goes) position that surveillance footage should be subject to copyright, but it's well established that it's not. Who owns the camera is irrelevant. At least in the US; I'm not aware of any jurisdictions that hold otherwise. This is why Wikipedia articles on world events in the past few decades are full of stills from surveillance cameras: it's one of the few sources of imagery of an event that are unambiguously legal to include, because unlike a photo or intentionally made video of something specific, it's not a creative work. It's also pretty firmly established that human authorship is required for something to be subject to copyright, and having an idea that lead to some particular expression is itself not sufficient; see, e.g.: https://en.wikipedia.org/wiki/Monkey_selfie_copyright_disput...
(Not a lawyer, just a Wikipedia editor.)
I appreciate your informed take. I follow the reasoning but I am amazed it works this way. I found some articles that supported what you said, and also said there's a follow-on industry that figured out how to alter and edit videos just enough for a revised video to have creative contribution and make it copyrightable.
https://www.techdirt.com/2020/02/24/can-you-license-video-yo...
> https://en.wikipedia.org/wiki/Monkey_selfie_copyright_disput...
This is a great case to study, but no determination of copyright was made. The only actual lawsuit was filed by PETA arguing that the monkey should have copyright, which led to a settlement with the human photographer and nothing else because obviously that's not possible.
For various reasons (mostly $$$) the guy never actually got a judgement. I think the chances are good that he could have prevailed in court; there is significant creative input to setting up cameras and triggers in a way to convince a wild monkey to take a selfie. It's not like he just left his camera sitting somewhere on accident and came back to find a photo in it.
I'm also not a lawyer, but I did do a lot of work in copyright for a company you've heard about.
> If you write a prompt and one-shot a problem and share the source code, that source code is probably not covered by copyright.
We will have to see about that! This is the kind of boundary that's still being figured out in court; it's going to depend on how hard you worked on the prompt. I highly doubt that even most slop was generated with a single half-ass prompt, and the bar is not as high as you might expect.
> If I stick a Flock camera on a pole somewhere and post the live output, there's been no meaningful human creative involvement in producing those images and so nobody can claim copyright on them.
It really depends on what pole, where, and why. In a parking lot in rural Wisconsin? Probably not. A recorded livestream of a political march? You likely have copyright.
I think that by virtue of the sheer amount of time spent using AI tools, it's pretty clear that these outputs have enough creative input to be copyrightable.
No, that’s not what they are saying. They’re saying that the code generated by a human with help from an LLM may potentially be. This is what I hope we are going to arrive at, eventually.
How could you establish what parts of the code was produced by an LLM vs updated by a human afterwards?
The LLM will output different results over time as the models get updated. Are we heading towards needing to retain a full prompt history that can be replayed against a specific LLM model version to prove what the output was for copyright purposes?
For those wondering what the difference is: consider what happens when an LLM regurgitates its training data. It's copyrighted... but not by the person who generated it.
Congrats on having the worst take in a thread full of them.
This is a false narrative based on a (IMO often intentional) misunderstanding. It has by no means been established by any court that LLM-generated code is not copyrightable.
Thaler v. Perlmutter stands for a much narrower proposition and at any rate is not binding nationally, SCOTUS having denied certiorari.
This is not necessarily true. While no court has explicitly come out and said that copyright does not apply to AI-generated works of authorship, the US copyright office has[0]:
> Based on an analysis of copyright law and policy, informed by the many thoughtful comments in response to our NOI, the Office makes the following conclusions and recommendations: > • Questions of copyrightability and AI can be resolved pursuant to existing law, without the need for legislative change. > • The use of AI tools to assist rather than stand in for human creativity does not affect the availability of copyright protection for the output. > • Copyright protects the original expression in a work created by a human author, even if the work also includes AI-generated material. > • Copyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements. > • Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis. > • Based on the functioning of current generally available technology, prompts do not alone provide sufficient control. > • Human authors are entitled to copyright in their works of authorship that are perceptible in AI-generated outputs, as well as the creative selection, coordination, or arrangement of material in the outputs, or creative modifications of the outputs. > • The case has not been made for additional copyright or sui generis protection for AI-generated content. > The Office will continue to monitor technological and legal developments to determine whether any of these conclusions should be revisited. It will also provide ongoing assistance to the public, including through additional registration guidance and an update to the Compendium of U.S. Copyright Office Practices.
Congress or the courts could, of course, override the stance of the copyright office, but I think it would be highly unusual for them to do so (particularly for something like this). It would however be a lot better if congress just stepped in and said no outright, but until then this will have to do.
[0]: https://www.copyright.gov/ai
The copyright office does not determine the standards for copyright. They are an advisory and notary organization.
Only Congress and the courts do. Copyright exists from the moment a work is created, and does not need to be registered with the copyright office.
The law isn't that complicated; if a work was created with a human being with intent, it's probably eligible for copyright protections.
As long as you can convince a court that you did this, the tools you used are not relevant. The vast majority of LLM art falls in this bucket.
Do you own the copyright to a painting you commissioned or does the artist? You may have described what you wanted, but the creative work is the output of the artist. Same with an LLM.
The courts are 100% going to have to interpret what is "sufficient human control" at some point.
Sure, but I would be incredibly shocked if the courts overturned these conclusions. These kinds of determinations are within the remit of the USCO, so a court does not need to come out and say it if the USCO has already done so. Obviously, as I said it would be better if congress weighed in and solved this problem, given that the USCO is free to publish a new NOI to change it's practices/policies, but we all know that congress is too gridlocked atm for that to happen
I am trying really hard not to accuse you of not having read what you posted, because your conclusions are in strong tension with what it plainly says.
But there are no conclusions. It literally says:
> Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis
It says a plain prompt is not enough but that is not the reality of real software development. People aren't one-shotting complex business apps. The vast majority of software development will trivially pass that bar and end up in the "requires case by case analysis".
You are a false narrative. I’m just repeating what I read. The thing is, it’s murky waters. Someone’s going to challenge it but do you want to be the guy who takes it on a chin?
The sibling comment lays this out and my original comment above is based on exactly the same link.
Also it's a different story intranationally for those of us who live in countries with much more restrictive/no fair use. Are you geolocking your software to the USA?
I highly doubt that that's the last word on that matter, but even if: Even before LLMs you could combine individual non-copyrighted components into something copyrighted.