Rendered at 23:12:57 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
ForgotMyUUID 2 days ago [-]
> When customers use Claude to generate Outputs that then train competing models, they're essentially using our infrastructure and investment to build direct competitors to our service
We did so, please do not repeat it at home.
pshirshov 1 days ago [-]
Didn't they, ehm, use all the historical infrastructure for free to train the thing?
spwa4 7 hours ago [-]
Without permission, and even after they got the laws changed in their favor they still violated the law so badly they got hit with a $1.3b verdict.
treetalker 1 days ago [-]
Careful: pointing out this distillation hypocrisy (“rules for thee but not for me”) is liable to draw moderation deeming it “a thought-terminating cliché” that is against the HN Guidelines. See, e.g.,https://news.ycombinator.com/item?id=49007792.
xyzsparetimexyz 14 hours ago [-]
Nobody cares about the hn guidelines
1 days ago [-]
cladopa 2 days ago [-]
I have released some of my projects as Open Source, I also have a company with privative software.
Claude and AI partners have taken all what they could from the Open Source projects without giving credit or respecting the licenses. They have increased the traffic on websites in an absolute disrespectful way increasing the hosting cost in inefficient and ridiculous ways.
They have taken all the important books and not asked permission from the authors.
Fair enough. Fair use.
Of course I would create a competitor software to Claude or any others if I could. Using Claude(and others) of course.
I am not paying you 200 dollars/month for you to tell me that I could not create code that competes with you. If you try to go to court in Europe with this you will lose.
It is just the same fair use you proclaim for taking the data from others.
glimshe 2 days ago [-]
They would likely lose in the US too. But until that happens, they can keep living in their fantasy world...
shikck200 2 days ago [-]
They stole all the data, and then dont want you to steal it back. Its basically Robin Hood all over again.
philipallstar 2 days ago [-]
Piracy is not theft.
eoverride 2 days ago [-]
Piracy is theft.
Copyright infringement isn't piracy. Nor theft.
preg_match 1 days ago [-]
Piracy is theft, theft is often justified. Or, at least, that's how I view it.
Laurel1234 1 days ago [-]
What clankers do isn't piracy.
nicbou 2 days ago [-]
I was not asked for permission when they trained their model on my output.
stingraycharles 2 days ago [-]
[flagged]
nicbou 18 hours ago [-]
Most of my output is on a website that I have made and control. The website was made for humans. I don’t recall allowing tech companies to scrape my work and train their LLMs on it.
theshackleford 2 days ago [-]
A few, not all, and in fact I gave many of them away under very specific licenses, of which they have paid exactly zero mind too.
If I allow my friends into my home for a visit, does that mean I should shrug when I return home from holiday and find out they’ve helped themselves to its usage without my knowledge or permission to do so?
lelanthran 2 days ago [-]
Corollary: If you can't use the output to train, then you don't own them.
rubslopes 2 days ago [-]
Apparently, they use "own" in the same misleading way that digital media providers use "buy", but later remove the purchased content from my library.
It's named "yours" at the time of payment, but actually only a license to use under certain conditions.
swat535 1 days ago [-]
If it's under someone else's license, then you can't copyright / patent the code generated from LLM in the commercial codebase.
This basically makes their AI useless.
j45 2 days ago [-]
Yup, either you own it, or you don’t own it and ownership is being redefined as a quasi license.
grumbelbart2 1 days ago [-]
Nobody "owns" it for the most part. Most LLM output can not be copyrighted, because there is not enough human essence in it. Meaning that it is public domain the moment it is produced.
There might be a few exceptions, if you really guided your model in a certain particular way, but that is rare.
What /could/ be is that you violate their ToS. If that ToS is valid and legally and practically enforceable is a different question and will depend a lot on your jurisdiction.
j45 23 hours ago [-]
If something is generated in someone’s private paid account that might not be public domain.
It’s kind of like causing what Google Docs autocomplete sentences helps someone write is still.. theirs.
lifthrasiir 2 days ago [-]
Apart from the usual hypocrisy,
> Safety controls may be lost – models trained on Claude's Outputs won't have our safety measures, potentially leading to harmful or dangerous AI systems. We also have no visibility into deployment, meaning we cannot monitor how these distilled models are used or prevent misuse.
So was that why your models caused three real-world security incidents?
mysterydip 1 days ago [-]
> So was that why your models caused three real-world security incidents?
No, that was just marketing :)
Chris2048 14 hours ago [-]
Isn't that someone else's problem anyway? Someone else's model trained on your outputs is still someone else's model.
> we cannot monitor how these distilled models are used or prevent misuse
which is fine since they are none of your (their) business..
I think the only possible concern is using their models name - making it clear it's a new model merely trained in another one should fix that.
moomin 2 days ago [-]
So can you write GPL content using Claude? MIT? Because if this condition applies to the output, I don't see how it's compatible with FOSS.
grumbelbart2 1 days ago [-]
It might depend on your jurisdiction, but LLM output is usually not copyrightable because it is not made by a human. There might be a few exceptions, if you guided the LLM in very specific and particular ways. But in general, the output of an LLM is public domain the moment it was produced.
Can you never tell anybody and just slap a license on it? Sure.
The issue here is not the license, but that you violate their ToS (if that is valid and enforceable is a different question of course). But if you publish the LLM output on github, and someone else takes it to train their LLM, and you did not actively encourage or help them, it's fine.
ksk23 16 hours ago [-]
Public domain is not an international concept. We do not have it in Europe (afaik. We do not have it Germany)
Steltek 1 days ago [-]
> It might depend on your jurisdiction, but LLM output is usually not copyrightable because it is not made by a human.
Compiler output isn't copyrightable? Is a compiler human?
grumbelbart2 19 hours ago [-]
Compiler output is a derivated work of the source code and thus keeps the copyright of the code.
em-bee 1 days ago [-]
MIT doesn't bar additional restrictions, so you could say this is MIT except you can't train on it. GPL on the other hand does not allow additional restrictions and therefore the code would not be compatible.
but you can't copyright the output anyways. which either makes the restriction on training void or, it means the owners of the model own the copyright, and they only transfer some of the ownership to you. is there such an ownership transfer statement? i haven't seen one yet.
yencabulator 3 hours ago [-]
"MIT doesn't bar additional restrictions" is a bit weird of a statement when the MIT license itself already gave permission to do almost anything with the code. My emphasis.
> Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software *without restriction*, [...]
stingraycharles 2 days ago [-]
Yes, obviously, you own the output. Training on those outputs is a violation of the service agreement, and that’s separate from who owns the outputs.
yencabulator 3 hours ago [-]
Okay so if I own my output I can sell it to my neighbor who then trains on the output?
sourdecor 2 days ago [-]
But LLMs are being trained on GitHub libraries that leveraged Claude?
j45 2 days ago [-]
Transferring actual and complete ownership of the output doesn’t allow any further interpretations of the use of it.
That sentence seems to be in a blurry line between outright “ownership” and licensing.
This type of explanation leans towards the reality being you don’t own the outputs from Claude.
This kind of an explanation is like trying to be half pregnant.
smallerize 2 days ago [-]
The only penalty here is being banned for a ToS violation. Maybe they could sue for fraud or something? That doesn't change who owns the data.
j45 2 hours ago [-]
The other comments here about this policy being done at scale, or by competitors being different seems to be plausible
Hamuko 2 days ago [-]
I hope I see the day I start seeing AI companies suing other AI companies for training their models on GitHub repositories that were written using the AI companies' models. If Anthropic says that I can't train using Claude's output, then surely OpenAI can't train on a repository that's 100% vibecoded with Claude.
riffraff 2 days ago [-]
I think they care about the reasoning traces and such being used for training, not the effectively final output.
Hamuko 2 days ago [-]
Are they even giving out the reasoning these days?
Wouldn't such language and reasoning from Anthropic be an argument that they needed written permission to train their model on data from websites?
Has any individual somewhere around the world tested this in court by now? Sued Anthropic for copyright infringement because Claude can reproduce information that is only available on their website?
It shouldn't be that expensive, right? If you sue them for - say - $10000 then what would the costs of such a court case be?
Personally, I think "learning" is not a copyright violation. But if they themselves make it one, then they should also face the consequences, no?
krona 2 days ago [-]
From my interpretation, in order to get the data you 'own' (it's not theirs to give away since they can't claim the copyright on it), you need to use their services. The agreement the user has with Antropic is for the service, not a restriction on how the data is used.
em-bee 1 days ago [-]
not being able to train in it is a restriction on using the data though.
if i publish that data, and someone else trains on it, am i liable? am i responsible to ensure that noone trains on that? how am i supposed to enforce that?
em-bee 1 days ago [-]
so if i publish that data, and someone else trains on it, am i liable? am i responsible to ensure that noone trains on that? how am i supposed to enforce that?
juancn 1 days ago [-]
Nobody owns Claude's output (nor that of any AI), they're in the public domain, they cannot be copyrighted.
The author must be human, at least for now, in the US this precedent holds as far as I can tell:
> When customers use Claude to generate Outputs that then train competing models, they're essentially using our infrastructure and investment to build direct competitors to our service. Like other software and service providers, we expect that our services won't be used to undermine our product offerings.
I hope Claude wrote this because if a human being typed this we need to kill not just them but everyone that so much as shook their hands.
isusmelj 2 days ago [-]
Any AI safety experts here? I'm wondering if this claim here really holds:
> Anthropic invests significantly in making Claude safe, helpful, and harmless. We conduct rigorous pre-release testing, implement multiple safety layers, and continuously monitor our models' behavior. When Outputs are used to train new models without our oversight, additional risks emerge. Safety controls may be lost – models trained on Claude's Outputs won't have our safety measures, potentially leading to harmful or dangerous AI systems.
From my understanding distillation pretty much copies behaviour. If someone intends to distill from a Model, it can't extract unsafe behaviour, but would learn the same safety mechanisms, no?
____mr____ 2 days ago [-]
> Why we restrict model training
> When Outputs are used to train new models without our oversight, additional risks emerge. Safety controls may be lost - ...
---
> What you can do with Outputs
> You can use Claude's Outputs to train models that don't compete with Anthropic's own models.
Why even include that bullshit at top? It is brazenly obvious to anyone with a brain that Anthropic and other frontier AI Labs have no real way to monetize or recoup their investment unless they strongly guard the usage of their models.
riffraff 2 days ago [-]
I had the same thought.
"We need to control the model to protect you from evil robots, except it's fine if the evil robots are not competing with our business" is hilariously hypocritical.
j45 2 days ago [-]
This means you can train smaller models than Claude’s but not competing.
I’m not sure how an AI company feels that’s safer for their business. Specialization will always best generalization trying to do the same.
Quarondeau 1 days ago [-]
"you own the outputs..."
That's not even the case. No one "owns" the output. Raw generative AI outputs don't come with any new (copy)rights, which are currently only granted to human creative outputs.
Traster 2 days ago [-]
The shear hypocrisy of this is quite staggering. Ok, so we own the outputs, but you get to decide how we use them, and you don't trust us to use them. But the people whose books you stole to build your psychopath machine, they didn't trust you, did they. They didn't really want you to build a machine that can generate thousands of books to compete with their work, but you did do it anyway. Are you going to commit to auditing every input that was used to train your model and attain positive consent for their use for training?
It's also pretty wild to call this standard practice. It's not. I can grab any of the open weights models and train to my hearts content. So it's not standard is it. You'd like it to be standard because you don't want to compete.
And you don't trust us, but it is your company that's been going around telling us how excited you are that your model goes out onto the internet and hacking people.
This continual authoritarian bent from the least trustworthy people in the world is deeply problematic and the only saving grace is their absolute total and complete failure to enforce the restrictions they wish to place on us.
Your ability to build the God machine doesn't magically endow you with the moral authority or judgement to decide how it's used, and the fact that these people believe it does is a great indicator that they aren't to be trusted.
hackmack10 2 days ago [-]
Honestly, the entire dev community should save question and answers, uploaded them to a shared repo anonymously, then just use that data to distill further models and provide them to the public for free.
Theoretically, this should be legal and ethical, when comparing to Anthropic's own behavior. That said, the reason you can't is Anthropic states in their terms that they don't want you to do this.
Anthropic's entire business model is skating on thin ice.
maaaaattttt 2 days ago [-]
If more and more of the web's content is AI generated, AI companies are bound to train on each other's data.
Or, what if I generate content with Claude/ChatGPT/Gemini, warp it in HTML using an open model, put this on my website conveniently dedicated to "Best practices in prompt and AI answers" for example, then train my own model that only scraps my website?
mlvljr 2 days ago [-]
[dead]
woadwarrior01 2 days ago [-]
All this posturing is all for naught, because people can do proper logit distillation into smaller models with open-weight models.
wolvoleo 2 days ago [-]
I don't think they care about the little guy doing a bit of fine-tuning.
What they don't want is DeepSeek training their models with Claude output at scale. That's why they forbid it. Gives them a legal basis to cut off accounts doing that.
Not that it's effective because it's being done anyway.
And yes it's super hypocritical but that's another issue IMO.
neuroticnews25 2 days ago [-]
Can I just sell the outputs to someone with no contractual relation with Anthropic, therefore not bound by their TOS?
Ekaros 2 days ago [-]
"Our Terms do not allow the use of Outputs to train models that are competitive with Anthropic's own. It is also a violation of our Terms to support a third party's attempt to do the same."
So it would be supporting third party's attempts...
Then again. I suppose lot of scrapping could happen by accident and scrappers might ignore such files as allowable-use-cases-for-site-content-must-followed.txt instructing against use in training.
If such is ignored wouldn't be our fault right?
josh-wrale 2 days ago [-]
Isn’t this based in contract law? The terms are between the user and Anthropic, no? If a person chooses to open source their convos, they are in violation? on what grounds?
grumbelbart2 1 days ago [-]
No, they would not be in violation. And an independent third party can pick them up and train with them. But you are not allowed to "support" that third party, otherwise you violate the ToS.
Arnt 2 days ago [-]
But you can give them all to me and I can train my model on them. The restriction on you is not related to your ownership of that days, it's related to the contract you "signed".
Or you can publish them on the web and countless others will do it.
linksku 1 days ago [-]
If competitors like Deepseek need to distill to build similarly powerful models, that's mostly for evals right? I don't see how it can be cost-effective to distill for a significant amount of training data
2 days ago [-]
ferrouswheel 2 days ago [-]
You can, they are just confused about what effective altruism means.
oliwarner 16 hours ago [-]
> you own the Outputs generated from your Inputs
It's amazing how quickly you can spin from that to ownership with limitations.
Obviously it's bullshit. They can't say it's theirs, licensed to you because that would stop people using it commercially... So we have this Gordian knot of logic to explain that we should pay to used their models and infrastructure, and not use the output how we like because their models and infrastructure are theirs and exploiting that would be really mean. The argument makes them look like a petulant child.
But it's just a weak EULA, a flimsy non-compete. They say output is yours? It's yours. Do whatever you like with it. But don't be surprised if they limit access if Anthropic decide it's unsavoury.
ratmilk 2 days ago [-]
This seems like the standard AI company hypocrisy. Hopefully Claude users disregard this nonsense.
sorry LLM output is considered public domain in my country.
jareklupinski 1 days ago [-]
> there are important restrictions on using these Outputs to train AI models which are standard practice across the AI industry
ah yes, this long-running, time-honored traditional industry, where we totally didnt write the rules ourselves
cynicalsecurity 2 days ago [-]
Those who distill are like Prometheus who stole fire from the gods and gave it to humanity.
ButlerianJihad 1 days ago [-]
Look, I think we're past the Prometheus part here. See Ben Franklin and Victor Frankenstein.
The LLM companies are Pygmalion; the LLMs themselves are golems or Frankenstein's monster; we ourselves are merely Igor.
gostsamo 2 days ago [-]
There must be a clear difference between terms of the Anthropic service and the legal standing of the ai output. The output is mine and I'll do with it whatever I want. The service is Anthropic's and they can do business with whoever they want. Everything else is hallucination.
NicuCalcea 2 days ago [-]
Imagine that you bought an axe, but the manufacturer banned you from making more axes with it.
hackmack10 2 days ago [-]
Hmmm... the mods already took this off the front page. What a bunch of bullshit.
ozgung 2 days ago [-]
1st place to 100th place. This might be due to downvotes but currently this post has 65 points and 55 comments in an hour.
I noticed a similar thing for the Antrophic’s previous announcement on open weight models.
You would need a massive amount of claude output beyond anything humanly reasonably to train a model. Namely, consuming a lot of inference resources. If
you don't do that at those scale, the training will probably be quite inefficient
anyways.
xyzal 2 days ago [-]
Nah, just pass the outputs to a person not related to Anthropic in any way.
Or -- maybe better -- just fsck the TOS.
fbrncci 2 days ago [-]
Cute :)
broccoluvr 2 days ago [-]
yikes
theyliesoeasily 2 days ago [-]
You downloaded millions of books from shadow libraries and trained on them.
I don't care about your policies, go fuck yourself.
abuanwar072 2 days ago [-]
Now that is standard AI company hypocrisy
bnj 2 days ago [-]
I guess this is another angle on why they are interested in watermarking the generated content
stavros 2 days ago [-]
Yeah, no. I was OK with them scraping everything if it means we get AI, but, conversely, they don't get to control what happens to their outputs.
Hell, arguably they should release their weights (or at least the weights of their older models), since they trained them on the concentrated knowledge of humankind.
f6v 2 days ago [-]
> I was OK with them scraping everything if it means we get AI
Except there’s no “we” since their models aren’t open.
qsera 2 days ago [-]
And they are probably also training them using the current human interaction as well...
DarenWatson 2 days ago [-]
[flagged]
epicprogrammer 2 days ago [-]
[flagged]
2 days ago [-]
2 days ago [-]
esjeon 2 days ago [-]
One obvious loophole: upload your full outputs publicly. Anyone can feed those to their models.
We did so, please do not repeat it at home.
Claude and AI partners have taken all what they could from the Open Source projects without giving credit or respecting the licenses. They have increased the traffic on websites in an absolute disrespectful way increasing the hosting cost in inefficient and ridiculous ways.
They have taken all the important books and not asked permission from the authors.
Fair enough. Fair use.
Of course I would create a competitor software to Claude or any others if I could. Using Claude(and others) of course.
I am not paying you 200 dollars/month for you to tell me that I could not create code that competes with you. If you try to go to court in Europe with this you will lose.
It is just the same fair use you proclaim for taking the data from others.
Copyright infringement isn't piracy. Nor theft.
If I allow my friends into my home for a visit, does that mean I should shrug when I return home from holiday and find out they’ve helped themselves to its usage without my knowledge or permission to do so?
It's named "yours" at the time of payment, but actually only a license to use under certain conditions.
This basically makes their AI useless.
There might be a few exceptions, if you really guided your model in a certain particular way, but that is rare.
What /could/ be is that you violate their ToS. If that ToS is valid and legally and practically enforceable is a different question and will depend a lot on your jurisdiction.
It’s kind of like causing what Google Docs autocomplete sentences helps someone write is still.. theirs.
> Safety controls may be lost – models trained on Claude's Outputs won't have our safety measures, potentially leading to harmful or dangerous AI systems. We also have no visibility into deployment, meaning we cannot monitor how these distilled models are used or prevent misuse.
So was that why your models caused three real-world security incidents?
No, that was just marketing :)
> we cannot monitor how these distilled models are used or prevent misuse
which is fine since they are none of your (their) business..
I think the only possible concern is using their models name - making it clear it's a new model merely trained in another one should fix that.
Can you never tell anybody and just slap a license on it? Sure.
The issue here is not the license, but that you violate their ToS (if that is valid and enforceable is a different question of course). But if you publish the LLM output on github, and someone else takes it to train their LLM, and you did not actively encourage or help them, it's fine.
Compiler output isn't copyrightable? Is a compiler human?
but you can't copyright the output anyways. which either makes the restriction on training void or, it means the owners of the model own the copyright, and they only transfer some of the ownership to you. is there such an ownership transfer statement? i haven't seen one yet.
> Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software *without restriction*, [...]
That sentence seems to be in a blurry line between outright “ownership” and licensing.
This type of explanation leans towards the reality being you don’t own the outputs from Claude.
This kind of an explanation is like trying to be half pregnant.
Has any individual somewhere around the world tested this in court by now? Sued Anthropic for copyright infringement because Claude can reproduce information that is only available on their website?
It shouldn't be that expensive, right? If you sue them for - say - $10000 then what would the costs of such a court case be?
Personally, I think "learning" is not a copyright violation. But if they themselves make it one, then they should also face the consequences, no?
if i publish that data, and someone else trains on it, am i liable? am i responsible to ensure that noone trains on that? how am i supposed to enforce that?
The author must be human, at least for now, in the US this precedent holds as far as I can tell:
https://law.justia.com/cases/federal/appellate-courts/cadc/2...
I hope Claude wrote this because if a human being typed this we need to kill not just them but everyone that so much as shook their hands.
From my understanding distillation pretty much copies behaviour. If someone intends to distill from a Model, it can't extract unsafe behaviour, but would learn the same safety mechanisms, no?
> When Outputs are used to train new models without our oversight, additional risks emerge. Safety controls may be lost - ...
---
> What you can do with Outputs
> You can use Claude's Outputs to train models that don't compete with Anthropic's own models.
Why even include that bullshit at top? It is brazenly obvious to anyone with a brain that Anthropic and other frontier AI Labs have no real way to monetize or recoup their investment unless they strongly guard the usage of their models.
"We need to control the model to protect you from evil robots, except it's fine if the evil robots are not competing with our business" is hilariously hypocritical.
I’m not sure how an AI company feels that’s safer for their business. Specialization will always best generalization trying to do the same.
That's not even the case. No one "owns" the output. Raw generative AI outputs don't come with any new (copy)rights, which are currently only granted to human creative outputs.
It's also pretty wild to call this standard practice. It's not. I can grab any of the open weights models and train to my hearts content. So it's not standard is it. You'd like it to be standard because you don't want to compete.
And you don't trust us, but it is your company that's been going around telling us how excited you are that your model goes out onto the internet and hacking people.
This continual authoritarian bent from the least trustworthy people in the world is deeply problematic and the only saving grace is their absolute total and complete failure to enforce the restrictions they wish to place on us.
Your ability to build the God machine doesn't magically endow you with the moral authority or judgement to decide how it's used, and the fact that these people believe it does is a great indicator that they aren't to be trusted.
Theoretically, this should be legal and ethical, when comparing to Anthropic's own behavior. That said, the reason you can't is Anthropic states in their terms that they don't want you to do this.
Anthropic's entire business model is skating on thin ice.
Or, what if I generate content with Claude/ChatGPT/Gemini, warp it in HTML using an open model, put this on my website conveniently dedicated to "Best practices in prompt and AI answers" for example, then train my own model that only scraps my website?
What they don't want is DeepSeek training their models with Claude output at scale. That's why they forbid it. Gives them a legal basis to cut off accounts doing that.
Not that it's effective because it's being done anyway.
And yes it's super hypocritical but that's another issue IMO.
So it would be supporting third party's attempts...
Then again. I suppose lot of scrapping could happen by accident and scrappers might ignore such files as allowable-use-cases-for-site-content-must-followed.txt instructing against use in training.
If such is ignored wouldn't be our fault right?
Or you can publish them on the web and countless others will do it.
It's amazing how quickly you can spin from that to ownership with limitations.
Obviously it's bullshit. They can't say it's theirs, licensed to you because that would stop people using it commercially... So we have this Gordian knot of logic to explain that we should pay to used their models and infrastructure, and not use the output how we like because their models and infrastructure are theirs and exploiting that would be really mean. The argument makes them look like a petulant child.
But it's just a weak EULA, a flimsy non-compete. They say output is yours? It's yours. Do whatever you like with it. But don't be surprised if they limit access if Anthropic decide it's unsavoury.
Now what.
ah yes, this long-running, time-honored traditional industry, where we totally didnt write the rules ourselves
The LLM companies are Pygmalion; the LLMs themselves are golems or Frankenstein's monster; we ourselves are merely Igor.
I noticed a similar thing for the Antrophic’s previous announcement on open weight models.
https://news.ycombinator.com/item?id=49076057
Or -- maybe better -- just fsck the TOS.
I don't care about your policies, go fuck yourself.
Hell, arguably they should release their weights (or at least the weights of their older models), since they trained them on the concentrated knowledge of humankind.
Except there’s no “we” since their models aren’t open.
Wait, we're already doing it. /s