Right?
N.B. JSTOR explicitly settled with Swartz, suffered no financial harm, and begged the federal government to drop the case. The government refused and prosecuted Swartz anyway. Anthropic settled for $1.5 billion primarily because of the theft and storage of the files, but this was somehow a Civil issue...
If you're unaware of who he was, or what happened to him:
* He was one of 12 others who helped to spec out RSS' standard (when he was 14 years old).
* He is widely considered a co-founder of reddit, after merging with reddit, he rewrote the majority of its codebase from Lisp to Python.
* He was one of the reasons Reddit protested SOPA back in the day (the same day that good 'old Kim Dotcom was raided and arrested by the US, the irony of which being that they swore up and down that they needed SOPA in order to do what they did to Kim Dotcom).
* I didn't know this before: he also developed the markdown publishing format with John Gruber.
What happened to him?
He was charged with felonies for stealing PDFs for publicly funded research papers, he was tried as if he hacked these systems. He took his own life as a result of them throwing decades of prison time at him for what's hardly a serious hacking offense (seriously, Kevin Mitnick was way worse). Kevin Mitnick served five years, they were looking to throw up to 35 years at him, and a minimum sentence of seven years if he went to trial.
It is abhorrent to me that we have people doing more malicious hacks that get no repercussions. Someone at OpenAI / Anthropic should be 100% liable for criminal charges of what their AI models do. Need I remind everyone, that these models can run way faster than what we common folk can see, the speed at which they run these models internally can go so quickly that by the time the model is discovered to have hacked someone, it could very well be too late. Just try the "fast" feature of Claude Code, now imagine Claude launching thousands of sub-agents at top speed internally, the capacity for unprecedented destruction is the most concerning thing to me, I say that as someone who uses Claude every week.
He did hack these systems and MIT's network. They had video of him in the MIT wiring closet and a mountain of circumstantial evidence. I'm not saying it wasn't pursued aggressively out of political motivations, which seems supported if you dig enough, but the facts of the case were about more than PDFs.
If it was, impersonating home users en masse to scrape the whole web to train models would result in lifetime in prison. Now we call it fair use.
I mean, if you get into somewhere you're not supposed to be, and use that to get stuff you're not supposed to get, the general public is going to call that hacking. Doesn't mean it's true, but most people don't care about true!
None of the federal charges really had anything to do with copying the files [1].
> Kevin Mitnick served five years, they were looking to throw up to 35 years at him, and a minimum sentence of seven years if he went to trial.
You are comparing the ridiculous DOJ press release number [2] for Swartz to Mitnick's actual sentence. You should compare like to like, so should be comparing 35 years for Swartz to the ridiculous DOJ press release number for Mitnick which was 200 years.
[1] https://volokh.com/2013/01/14/aaron-swartz-charges/
[2] https://web.archive.org/web/20130211030902/https://www.popeh...
The Aaron Swartz case is important to know about and has catalyzed some good discussions about CFAA reform, but that specific case is a lot more complicated than your summary.
Before I begin: I agree that the punishment was too significant for the crime. I completely agree on that count. Discussing the facts of the case draws a lot of angry downvotes, so to be clear: I'm not saying he deserved any of these sentences nor that the laws are written well. I'm trying to explain the facts.
The PDFs were open to downloads on the MIT network. Straightforward downloading of the PDFs wouldn't have been a CFAA violation. The case became complicated because he started circumventing restrictions on the downloads. This escalated to breaking into a wiring closet and hiding a laptop in the closed-off infrastructure. Once someone breaks into physical infrastructure to circumvent access restrictions, it gets really hard to argue that they weren't doing a form of unauthorized access.
> He took his own life as a result of them throwing decades of prison time at him for what's hardly a serious hacking offense (seriously, Kevin Mitnick was way worse). Kevin Mitnick served five years, they were looking to throw up to 35 years at him, and a minimum sentence of seven years if he went to trial.
This is another part of the story that gets confused. Kevin Mitnick took a 5 year plea deal. The Swartz case plea deal was 6 months. The maximum sentencing numbers are repeated because they're shocking, but maximum sentencing was never on the table due to the low severity.
The plea deal (the correct comparison to Kevin Mitnick's case) would have been 6 months, not 35+ years.
I'm aware, though I will note, in the case of Kevin Mitnick the crimes were abhorrent, and I am sure Aaron was 100% in the wrong, and I don't think he should have gotten a slap on the wrist, but felonies + prison time for a few publicly funded PDFs is a bit drastic.
To be fair, I'm trying to remember things that happened over a decade ago, I could get some things wrong. I think me and you agree for the most part either way.
The reason I know these things is because I remembered it the same way you did up until a few years ago. Then I re-read the case and was surprised at how much it differed from the version that has been repeated everywhere for years.
Meanwhile, Claude does the same thing and it is deemed "progress" when the federal government profits from it.
It basically treats you guilty until proven innocent, which I hate. I know Aaron Swartz very likely did, it was an activist level "hack" but what if someone else "did it" and they threw the book at them, and they took the felonies despite not actually having done it? Lots of people have taken plea deals because its less costly than fighting the law in court.
Felonies, however, impact your rights as an American. You cannot own firearms, you cannot vote, you cannot hold public office, and you are a “high risk” economically and socially, making loans and employment difficult. State-level non-violent felonies can normally be expunged/sealed over time for otherwise lawful citizens, but federal felonies never go away.
He denied the plea deal because he, as an activist, did not want to suffer losing those rights. The prosecution wouldn’t budge and was adamant in making an example out of him because he was a figure of hacktivism. They also desperately wanted the CFAA (a poorly worded law written by tech illiterates in 1986) to maintain its broad power.
For further context, this was a response to Anonymous and other would-be hacktivists, coming right off the heels of Anonymous’s Arab Spring campaign and the hack of HBGary. The Obama administration targeted Swartz to make an example out of him and scare future hacktivists, coinciding with the arrest of Sabu (the co-founder of LulzSec) and him becoming an informant to round up the current group.
The Anthropic class action was already settled. It's what they mention in the opening of the article. Anthropic got in trouble for downloading the books (pirating) but it was determined to be fair use to use the content for training.
Aaron Swartz was not charged with piracy or downloading PDFs, even though that's how the story has been told. The charges were related to him breaking into a private wiring closet, hiding a laptop, and circumventing access controls on a system.
The Aaron Swartz charges were ridiculous. I'm not disputing that. However, there isn't an equivalent charge in this Anthropic case. Even if you or I get caught downloading pirated PDFs of eBooks from a torrent, the CFAA would not apply.
The main class action suit was already decided. The Stanford professors in this story opted out of the class action and are trying to get higher individual settlements with their own lawyers.
We shouldn't be using that as precedent; in fact, it should be the opposite.
No precedent was set. Nothing went to court. To the extent there is precedent it’s in the prosecutor’s career having been ruined by that stain. Using that failed prosecution to attack anyone risks it becoming actual precedent.
You're thinking of something unrelated.
We can totally take these civil crimes and turn them into a variety of felonies, with any luck we could put some software devs in federal prison too. Maybe after a few dozen of them get locked away for 10 years our profession might take ethics seriously.
Swartz "was prosecuted for multiple violations of the Computer Fraud and Abuse Act of 1986 (CFAA)" and wire fraud statute [1]. The CFAA angle specifically rested on a legal theory around escalating terms-of-service violations into felonies. That theory never found precedent because it literally never went to court. As a matter of prosecutorial discretion, it has never been brought up again because it literally tanked Ortiz's career.
> You're thinking of something unrelated
What do you think I'm thinking of?
> We can totally take these civil crimes and turn them into a variety of felonies
We can. It would mean reviving that legal theory. If it results in a conviction it becomes real legal precedent. That, in turn, means we get more prosecutions like Aaron Swartz's.
> Maybe after a few dozen of them get locked away for 10 years our profession might take ethics seriously
Cutting off the nose to spite the face.
Why not go after them directly with taxes and prosecution under non-harebrained schemes?
Who? Name names. You are making a bold, broad claim. Who are these specific people against whom you have specific evidence of their willingness to “distribute contaminated food … quite readily in service to their bottom line”?
A quote from that page:
> FDA’s leading hypothesis remains that this was likely an act of economically motivated adulteration.
Maybe the cinnamon case was an isolated incident, but the use of lead compounds to change a food's color is a trend. Per https://www.npr.org/2024/10/24/nx-s1-5119336/cinnamon-lead-f...
> We're aware of this practice in turmeric, where lead chromate has been added to enhance the color
Are you actually going to argue that people don't do things that are unsafe for others when money is on the line? We're living in an ocean of evidence that it happens.
Otherwise we're creating a system where it is unfair for some people.
Selective application of the law in favor of the rich and powerful is a perversion of justice, even if it does sometimes result in more just outcomes in individual cases.
> Similar to the claims in Bartz, the plaintiffs allege that Anthropic illegally downloaded, reproduced and distributed their copyrighted books from “shadow libraries” like LibGen and PiLiMi, and that Anthropic broke the law by downloading and storing the pirated files in a central library.
Say you have a network. You want to keep someone out of it. The tool for that is the CFAA.
If you have a book, and you want to stop someone from reproducing it, the tool for that is copyright.
They cover different domains. If you're not breaking into a network, you're unlikely to be charged under the CFAA.
(1) They admitted its Claude models had breached the systems of three companies since April; including Huggingface.
https://www.bloomberg.com/news/articles/2026-07-30/anthropic...
(2) More to the point, it 100% breached the CFAA by securing unauthorised access to a U.S. state government website hosting election-related data, pulling aggregate vote-count statistics from a gated government dashboard. The state government itself confirmed the access was entirely unauthorised.
https://seekingalpha.com/news/4657023-anthropics-claude-bypa...
If a human security researcher had deployed an agent that behaved this way, it would constitute a textbook felony under the CFAA for accessing those systems without authorisation.
Unfortunately different rules apply when you have the power to buy off Congress and the Appeals Courts.
Most recently a US appeals court ruled on 5 August that Amazon was unlikely to succeed on a claim that Perplexity's AI agents violated the Computer Fraud and Abuse Act by covertly accessing private Amazon customer accounts.
https://www.rte.ie/news/technology/2026/0807/1586941-artific...
RIP Aaron Swartz.
Criminal law, "beyond a reasonable doubt", doesn't mean "there exists, however implausible, an explanation".
He did break into the network - he was spoofing MAC addresses. You don't have a level of technical sophistication allowing you to do this, and then say "oh, I assumed the server closet with an "Authorized access only" sign, usually locked, seemed like a safe place, that's all, and since I was there, I figured I might as well just plug directly into a network switch in there."
Many judges are not technical. Most judges are not idiots.
For example, if the item was derived from a repository of stolen goods that was already illegal in the US on the basis of it being a repository of stolen goods.
https://arstechnica.com/tech-policy/2024/09/pirate-library-m...
Anthropic training their model outside of public sight without intent to redistribute, and had a large paper trail that expressed they believed they were acting within fair use.
A case which never went to trial and tanked the prosecutor’s career resulting in no more prosecutions of its kind to date.
Let’s not hanker for a return to that.
If he were still here, he would be 40 in only a couple of weeks.
In casual conversation in America English, it a derogaty term used to mean "poor and lacking in quality of life, freedom, and/or democratic governance."
What is the law besides "bizarre overlapping and bureaucratic rule making"? You can't charge people on vibes like "underlying values and beliefs"
Hence why bringing up a case where justice was not served proportionally presents an opportunity to discuss where potential improvements could be made or errors were committed.
I know it makes people who take comfort in institutions and rules uncomfortable to bring up the fact that none of this is real and all of it is a construct.
But we can in fact improve "the law" and have done so basically since the beginning of time, hence how the current law came to be. In the US specifically, this is literally the basis of how common law works (bottoms up vs. top down).
But the rich buy “justice”. Us plebes are meant to “face justice”.
> any coerced or poorly remunerated work. [1]
Someone who makes this statement this really want us to equate it with racist chattel slavery - the most heinous form of worker exploitation in modern memory. "Oh, I just meant: {the most common type of work in human history}" as a retreat is not very compelling.
Surprised to see someone casually admit that the most common type of human labor is slave labor.
One of the interesting things about copyright infringement is that if you weren't told it was happening, you'd never know.
And on the coercion side your definition includes jury duty.
Your definition doesn't work.
It is a recognized idiom in many English dialects:
> If people work very hard for long hours for very little money, you can refer to it as slave labor.
* https://www.collinsdictionary.com/dictionary/english/slave-l...
> very hard work for which people are paid very little
* https://dictionary.cambridge.org/dictionary/english/slave-la...
> (informal) work that is very hard and very badly paid
* https://www.oxfordlearnersdictionaries.com/definition/americ...
See also "working in the salt mines" where people do not literally take sodium chloride out of the ground.
* https://idioms.thefreedictionary.com/back+to+the+salt+mines
Your understanding of the word is not standard.
If you stop short of saying McDonalds workers are slaves, and that they are merely closer to being slaves, then you are agreeing with me.
Slavery is not being poorly compensated or coerced into work.
You could say minimum wage at $7.25 but very few people actually make that (1%).
Or maybe an entry level aid in an eldery home making $11/hr and cleaning the anus and genitals of old men that are hitting on them. Would you agree $11/hr is poor remuneration for that kind of job? Or would you argue there is no one poorly remunerated?
Seems strange. The question should be whether you legally have to do so or not. If you do whether it is impractical or not is not a reason to break the law. Perhaps it makes your business not feasible but so be it. If you ran a clothing store that required you steal from other brands and remix it for your customers, you couldn't say "it's impractical for me to get permission", no?
If this argument is extrapolated, why should streaming services get the rights to copyrighted works? It sounds like a lot of work.
Hell, why would Pirate Bay be any different than Anthropic in this case? Millions of pieces of copyrighted work, being used as intake rather than sold.
Anthropic trained on illegally acquired Stanford professors' books
The article does not say the training was illegal (although does discuss the professors push for a licensing model).
The noteworthy thing here, IMO, is the judge seems to have explicitly stated that using legally-acquired works to train AI models falls under fair use.
What the actual ruling found was that the professors should have been paid some tens of dollars for 1 copy of their book.
(Whether you agree with it or not is another matter)
But the car I bought with the money was totally legal.
So you come with this amazing idea regarding anything... how to build a chip, how to write software... and you write a book on how to...
Then a model eat it... spit it for free (they could get an advantage from being able to exercise that logic) and it's fair use...
Then... why torrent it, can't they sent someone to buy the book and scan it or buy it in pdf?
This is obviously hyperbolic, but it was a very poor choice of words.
It also relies on an assumption - that AI will become a substitute for buying books - that I think is likely untrue. You’ve been able to easily gain a surface-level understanding of anything for free online for decades, and for that matter, you’ve been able to easily pirate any book for decades! Hasn’t stopped people from buying books, or going to college, or doing any number of other expensive things that are technically unnecessary if the goal is just to gain information.
- Anthropic et al basically grabbed whatever content they could to scan and train the model.
- The vast majority of what they scanned were books they bought at wholesale in big lots. Stuff being weeded out from Goodwills or libraries and (mostly) bound for destruction anyway.
- They didn't copy the texts per-se, they studied the relationships between words to build a statistical/predictive model of natural language.
- Whether this a fair or illegal use of copyrighted text already seems like an adorably obsolete question. Regardless, the damage is basically done - they can neither prove one way or another how much any single text influenced their model and they cannot unscramble the egg without starting over.
- Rather than get stuck in court for decades splitting the hairs about copyright law, they wrote massive checks to the publishing industry and agreed to a bunch of restrictions around copyrights.
- For the foreseeable future, every independent copyright holder now has grounds to take grievances to Anthropic and get some form of compensation.
The court ruled that it was fair to train on books, but they had to acquire them legally. That's why these companies buy physical books and scan them now.
The Stanford professors in this article opted out of that last case and are trying to get higher payouts.
This is doing a lot of heavy lifting, especially when you consider the scale
It was how they got the books for training: they torrented collections of pirated ebooks. There are Slack messages from executives describing the process as "blatant violation of copyright." They were stealing books and they knew it.
They didn't start buying books legally and scanning them until much later.
No, the lawsuit was because they illegally acquired the books (e.g. downloading them from piracy websites)
These days they are sticking to more legal means like the one you described
> they cannot unscramble the egg without starting over
I would bet they do have copies on their hard-drives (which is probably legal for them to possess for their own use), and that they could simply omit them from their next training run that they probably do at major updates anyway.
I’d call that the opposite of “suffering”.
They've since bought vast numbers of second hand books and scanned those for training, which appears be legal.
Interesting phrasing from you, Simon. Is there any legal theory under which it should be illegal?
As for the training being fair use, this is still disputed:
https://www.courthousenews.com/ai-training-of-copyrighted-ma...
https://www.nytimes.com/2026/08/24/opinion/claude-pirated-bo...
Damn right too.
I had an issue recently where I asked claude code to log into a developer user on Chrome on my local host for a system which is fully local and part of the code base the agent was developing and it refused to do it, like flat out refused, I did it on multiple sessions and I could not get Claude to budge on entering the credentials - it required me to manually sign in.
Would be amazing if they used that same level of training for their intake exercises.
> [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.
After a 1.5 billion settlement, the professor/authors are saying it's not enough.
The judge had an interesting summary of AI:
"He compared AI training to how humans can read, internalize and remember a book’s themes and style. “Like any reader aspiring to be a writer, Anthropic’s LLMs trained upon works not to race ahead and replicate or supplant them — but to turn a hard corner and create something different,” he wrote."
“To put me to work that way, without remittance, is a form of slave labor,” said Wolff, professor emeritus of English and winner of the National Medal of Arts.
Yet again, Anthropic is the good guys. Though I think this is just the same story from a while ago but told through a Stanford lens, AFAICT?
... Which, if enforced vigorously, means new raw AI can only be trained by those who can afford the raw material.
The Court just handed Anthropic and other VC-backed AI projects a significant moat.
Very soon, all information will be free!
Or do you believe, all books should be free, authors shouldn’t be compensated ever? Why should we compensate programmers or any knowledge worker than if the fruits of their labor should thusly be freely distributed?
Supposedly the trillion dollar company isn’t in the set of “rich” here, but professors are.
I do not think training an AI model is Fair Use, because LLM "training" is being used as a metaphor for building a matrix with a computer. Only people and animals train and learn; computers run statistical problems over input, which in this case was illegal input acquired by humans.
But it does not matter whether training LLMs is Fair Use, "Fair Use" is a strange and arbitrary legal term that can be expanded to include AI if we wish. This was organized crime whether or not LLM training on copyrighted materials is fair use. It was organized crime whether or not they planned to use them for AI at all.
AI is irrelevant. They could have trained it on GPL'd or MIT code (as long as they then distributed according to the terms of the licenses), and works in the public domain. They could have gotten what they were doing explicitly declared to be fair use by the Library of Congress, and then purchased the materials, like they are belatedly doing now. It wasn't necessary to violate copyright law to prove the technology. They could have proved the technology, then asked for copyright law to be clarified, and licensed huge volumes of works from the very few publishers we have left. Becoming a criminal organization was an explicit choice they made, and they discussed its illegality, and even the risk that it would be discussed on HN, which might inspire prosecution.
The reason they are not being prosecuted is because the rich are beyond the law; prosecutors will not prosecute them. The RICO statute is clear, and explicitly mentions copyright violation.
We can sentence the executives harshly, but go easy on the lower level people, because they are intelligent and valuable to advance the technology. The executives aren't necessary for anything. They're currently running finance scams that are extremely dangerous for the economy. They belong in prison to protect the public.
This is an option that people (real people, with names) have, and are not taking. When they say that they care about law and order, and are worried about the behavior of these companies, they are lying. Trump could literally nationalize them for what they've done. They're vulnerable to any state's attorney who has any author in their state whose book was downloaded. Libgen/Annas-archive would have been the most important honeypot in history.