> The strike by Alibaba is described as a "distillation" effort, which Anthropic has said involves training a less capable model on the outputs of a stronger one.
I don't see what's wrong about this.
> Anthropic said the campaign was conducted between April 22 and June 5, 2026, and generated more than 28.8 million exchanges with Claude through almost 25,000 fraudulent accounts.
What makes the accounts fraudulent? If they have paid the agreed price, surely it's fine? If they haven't paid, why did Anthropic provide them service?
Because Anthropic has terms of service with more stipulations than just "you must pay and can use the service for any purpose"?
Oh, Anthropic, the company that hoover'd up everyone else's data, and is now unhappy when others are doing to it what it did to others? The same Anthropic?
Yes, this joke/point has been made 10,000 times in this thread in almost every comment, and on every other previous thread. Thank you!
If Anthropic doesn't like people repeating this point, Anthropic should stop repeating that they are somehow entitled to keep what they have rightfully stolen.
Hey, are you unaware of the settlement they paid or the change in how they get this information that they made?
I don't see how the settlement changes the fundamental dynamics.
Anthropic argues that since they paid for the book they use in training, the can now distill as many copies of the book from their AI.
Nevermind what the author wants - anthropic doesn't need a permission to do so, and the author cannot opt-out their books from training. But somehow that is fair use now.
Now some chinese have bought access to anthropics AI, and anthropic is arguing that the Chinese shouldn't be allowed to distill from them, the same way they trained from others without consent..
"The settlement"? Not gonna cut it.
$1 billion and a huge change to how they gather information? What do you know about what they changed?
Well you seem to clearly be in the know. Care to share?
The problem with this line is that it becomes an argument against me, rather than self education. There are many comments about this on this very post, and it's recent headline news!
The reason I asked in that way was to see if GP actually knew anything about what they had strong opinions about.
On the other I think asking someone for a citation to their rather strong claim isn't incorrect. Digging through unstructured HN comments isn't my idea of a good time and being sent off on a treasure hunt by fellow commentors isn't why I come to this website. Collaboration is how we move forward!
It's not a joke. It pointing out hypocrisy and associated loss of all moral rights.
And yet it seems to continue to need to be repeated. If the shoe fits and all.
I'm sure all the artists and creators they stole from had stipulations too.
The artists had actual laws to protect them, not just vaguely enforceable terms of service. And look where that got them. I have zero empathy for the huge company getting a taste of their own medicine.
Anthropic paid one billion in a copyright settlement. That's a lot of money considering they never distributed the pirated books they trained on.
Nowadays they buy copies of books, train on them, and then destroy them.
And it looks like the companies distilling Claude are paying for tokens using the subscription Anthropic provides. Seems like fair play to me.
They are almost certainly paying orders of magnitude less than a billion dollars. According to another comment, they instead buy tokens resold from subsidized subscription accounts, which is against Anthropic's TOS.
Well, anthropic can just catch these and cancel their subscription - what's the problem?
It's almost like websites also have their robots.txt files that anthropic blatantly ignored. What's the problem, that now a US company is getting out-venture capitalismed by a Chinese company?
>> I'm sure all the artists and creators they stole from had stipulations too.
> Anthropic paid one billion in a copyright settlement.
Because a judge determined Anthropic was engaged in piracy.
> That's a lot of money considering they never distributed the pirated books they trained on.
This is "fruit of the poisonous tree" as it were. Distributing content derived from pirated content ("pirated books they trained on") is why Anthropic had to pay what they paid.
> Nowadays they buy copies of books, train on them, and then destroy them.
There is a case one could make that this practice could be seen as unauthorized redistribution of a derivative work intended to deprive copyright holders of legitimate revenue.
There no hypocrisy in saying that it is unethical and immoral to rob a former robber, especially if he was already harshly punished for his crime. Two wrongs don't make a right. I can't believe have to even point this out.
They were not "harshly punished". As always, they get slap on the wrist, proceed to benefit from the crime and that is it. You and I would be harshly punished. Anthropic was not.
1.5 billion dollars are not a "slap on the wrist". That's a huge amount of money. It exceeded the amount they would have legally paid for those books by orders of magnitude.
It's fucking nothing. The copyright industry used to threaten individual citizens with $250,000 fines per violation for willful commercial infringement. Where are they now?
Why aren't these big tech CEOs in cuffs with rifles pointed at their faces while SWAT seizes all of their computers?
Anthropic paid a billion dollars? Ridiculous.
> Because Anthropic has terms of service
Not following terms of service doesn't necessarily constitute a fraud. It just means Anthropic can close an account that breaks the terms of service.
Robots.txt are also ToS of sorts.
violating their terms of service doesn't make it fraudulent?
Are you not violating terms of service? Have you ever read any? I wouldn't call you fraudulent though!
Laws define fraud, terms of service define what a company would like you to do, usually in the narrowest and most abusive and extractive way possible.
The idea that anyone would side with a company doing more to support the ToS con than (at most) terminating an account they find it violation is sickening.
Really if we had competent, uncompromised government, most of these terms should illegal and result in Anthropic (and basically every other tech company) being hauled up in front of a regulator and fined heavily until they rewrite them to be less sociopathic.
So does a lot of the owners of data that Anthropic used for training. Anthropic preceeded to ignore said terms under the guise of fair use. Yet now they cry faul? Cry me a river.
To be clear: In principle I'm on Anthropic's side here. But Anthropic et al. have been very clear that they want to take a huge dump on those principles, so here we are.
[dead]
I mean they could read the traces and learn it themselves right? /s
> What makes the accounts fraudulent?
Fake identity? And general deception about the use
Nobody cares about the feelings of the trillion dollar corporation.
Terms of use is local US fiction of wishful thinking. Nobody cares. You make something available, it's up to the consumer to decide how are they gonna use it. You don't want people to use your stuff how they please? Get off the market.
Obviously wrong on all counts: the company cares. Just like isn't not up to the consumer since the provider can restrict said consumer's access, report those actions to the authorities etc. Lastly, don't like it? Get off the provider?
They ignore the TOS on my website that says no training freely with their scraper so …
You forgot to connect your petty grievance to the point of the conversation re. the state of accounts
Why is the grievance petty?
The company is completely free to void the contract and stop selling their services to anyone they don’t want to?
How is that relevant to anything I've said?
Since what you said applies to the company as well. If they don’t like customers breaking their possibly (or certainly at least in some jurisdictions) largely legally unenforceable ToC conditions they are entirely free to stop doing business with them.
Morally equating both sides seems distasteful since the relationship is mostly dominated by the companies. In a free competitive market it would be different but since were are talking about oligopolies/monopolies it obvious doesn’t work that way since there is only an illusion of choice.