Aaron Swartz: The Kid Who Helped Build the Open Web, and the Law That Broke Him
If you write Markdown, read feeds in an RSS reader, or have lost an afternoon on Reddit, you've used something Aaron Swartz helped build. Most developers I talk to know at least one of those projects. Far fewer know who he was, or that the United States government spent the last two years of his life trying to put him in federal prison for downloading academic articles.
He died on January 11, 2013. He was 26.
This post is about both halves of that story: the remarkable amount he gave to the web we all use, and the badly written, wildly over-applied law that was used against him, the Computer Fraud and Abuse Act (CFAA).
The work: a short list that's hard to believe
Aaron was born in 1986 in Highland Park, Illinois, and he was building things for the web before he was old enough to drive.
RSS 1.0 (age 14)
At 14 he joined the RSS-DEV Working Group and helped write the RSS 1.0 specification. RSS is the plumbing behind blog feeds and, today, nearly every podcast. Every time your podcast app pulls a new episode, it's reading a feed format that a teenager helped shape. RSS is also one of the best examples of what the open web is supposed to be: no company in the middle, no account, no algorithm. You subscribe to what you want and you get it.
Creative Commons (age 15)
He worked with Lawrence Lessig and the early Creative Commons team on the technical side of the licenses, the machine-readable metadata that lets software know what a CC license allows. Every "CC BY-SA" you've seen on Wikipedia or Flickr sits on that work.
Markdown (2004)
When John Gruber created Markdown in 2004, Aaron was his main collaborator on the syntax. Gruber credits him on the original Markdown page for his considerable help with the specification. Aaron also wrote html2text, a tool that converts HTML back into Markdown, which is still around in various forms.
I care about this one personally. Half the posts on this site are Markdown files, and I've written before about why I want my notes in plain Markdown files that will outlive any app. That idea, simple text that reads fine as-is and converts cleanly to HTML, is a big part of what Aaron helped design.
web.py and Reddit (2005)
Aaron was part of the very first Y Combinator batch in 2005 with a startup called Infogami. Infogami merged with another startup in the batch, Reddit, and Aaron became one of its co-founders. He wrote web.py, a minimal Python web framework, and Reddit was rewritten in Python on top of it after its original Lisp version. Condé Nast bought Reddit in 2006, and Aaron left not long after; office life at a big media company was not for him.
Open Library, Demand Progress, and SecureDrop
He didn't slow down after Reddit:
- Open Library (2007), with the Internet Archive: the goal of "one web page for every book ever published."
- Demand Progress (2010), an activist group that became one of the key organizers of the fight against SOPA and PIPA, the 2011–2012 bills that would have let the government and rights holders block whole websites. The internet blackout that killed those bills was partly his work.
- DeadDrop, a system for journalists to receive documents from anonymous sources. After his death it was taken over by the Freedom of the Press Foundation and renamed SecureDrop. Major news organizations use it today.
That's a lifetime of contributions, and he did it by 26.
What he actually did at MIT
In 2008 Aaron wrote the Guerilla Open Access Manifesto. His argument was simple: most scientific research is paid for by the public, but the papers are locked behind paywalls that charge readers, often while paying the authors nothing. Knowledge that belongs to everyone was being rented back to them.
In late 2010 he began downloading articles from JSTOR, a digital archive of academic journals, over MIT's network. Here are the facts that matter:
- MIT ran an open campus network. Visitors were allowed on it, and anyone on MIT's network had access to JSTOR through MIT's subscription. Aaron was a fellow at Harvard at the time, with his own legitimate JSTOR access there.
- He didn't break any encryption or steal any passwords. He used a laptop and a script to download articles quickly. When JSTOR and MIT blocked his laptop's address, he changed it and kept going. Eventually he plugged the laptop into a network closet in an MIT basement that wasn't locked.
- He downloaded about 4.8 million articles. Many were public domain or very old. He never published them, and there's no evidence he sold anything or made a cent.
- JSTOR got everything back and said it didn't want a prosecution. JSTOR settled with him in 2011 and said publicly that it had no interest in pursuing the matter further. In 2013, JSTOR opened up its early journal content to the public for free anyway.
At worst, this was a violation of JSTOR's terms of service and maybe a trespass in a basement. Those are the kinds of things that end with a banned account, a civil settlement, or a misdemeanor.
That isn't what happened.
The prosecution
Aaron was arrested in January 2011. The U.S. Attorney's office in Massachusetts, under Carmen Ortiz, with Assistant U.S. Attorney Stephen Heymann leading the case, charged him with wire fraud and multiple violations of the Computer Fraud and Abuse Act. In September 2012 they filed a superseding indictment with 13 felony counts, carrying a theoretical maximum of 35 years in prison and a $1 million fine.
For downloading journal articles. That the victim had already been compensated for. That the victim didn't want him prosecuted for.
Prosecutors offered a deal: plead guilty to all 13 felonies and serve about six months in federal prison. Aaron would not plead to being a felon for something he believed wasn't wrong. If he went to trial and lost, the government signaled it would push for a much longer sentence. Two years of legal fees had drained his money. MIT, which could have said "we don't want this either," stayed officially "neutral" (its own later internal review criticized how it handled the case).
On January 11, 2013, two months before his trial was due to start, Aaron took his own life in his Brooklyn apartment.
Aaron had written openly about struggling with depression, and no single thing explains a death like this. But his family was direct about what they believed. In their statement they said his death was "the product of a criminal justice system rife with intimidation and prosecutorial overreach." I agree with them. A brilliant young man was facing decades in prison and a felony record for life, and the people with the power to make it stop chose not to.
The problem with the CFAA
The Computer Fraud and Abuse Act was passed in 1986, the year Aaron was born. It was written partly in response to the movie WarGames, back when "hacking" meant breaking into military computers. It makes it a federal crime to access a computer "without authorization" or to "exceed authorized access."
The problem is that the law never really defined what those words mean. For decades, prosecutors argued they covered breaking a website's terms of service: use a fake name, share a password, run a script the site didn't like, and you might be a federal criminal. That's how you get from "downloaded too many articles too fast" to 35 years.
After Aaron's death, Rep. Zoe Lofgren and Sen. Ron Wyden introduced Aaron's Law to fix exactly this, by making it clear that breaking a terms-of-service agreement isn't a computer crime. It never passed.
Things have improved a little since, mostly in court, not Congress:
- In Van Buren v. United States (2021) the Supreme Court narrowed "exceeds authorized access." If you're allowed to access the data, using it for a purpose the owner doesn't like isn't a CFAA crime.
- In hiQ Labs v. LinkedIn the Ninth Circuit held that scraping publicly available data likely isn't "without authorization" under the CFAA.
- In 2022 the Justice Department said it would no longer charge good-faith security research under the CFAA.
All of that came too late for Aaron. And the core of the law, its vague wording and stacked-up felony penalties, is still on the books.
Today, everyone does what Aaron did
Here's the part that makes me angriest.
What Aaron did was write a script to download a large collection of documents he had legitimate access to, because he believed the knowledge in them should be free. Today, that is the business model of some of the most valuable companies on earth.
- Search engines have crawled and copied essentially the entire public web for 25 years. That's the whole product.
- AI companies have scraped the web at a scale that makes 4.8 million articles look like a rounding error. Training a large language model means downloading as much text as you can get: websites, forums, code, books, and yes, academic papers.
- According to court filings in Kadrey v. Meta, Meta downloaded tens of terabytes of books from shadow libraries like LibGen via BitTorrent to train its models.
- In Bartz v. Anthropic, a federal judge ruled in 2025 that training on books could be fair use, but that keeping a library of millions of pirated books was not. Anthropic settled for $1.5 billion.
- News publishers, authors, and artists have sued OpenAI, Microsoft, Stability AI and others over training on their work without permission.
Notice what those cases are: civil lawsuits, settled with money. Nobody at these companies was arrested by the Secret Service. Nobody was facing 35 years. Nobody was told to plead guilty to 13 felonies or go to trial. When a company with a legal department scrapes the internet, it's called "data collection" and it gets venture funding. When a 24-year-old activist did a far smaller version of it, to make research free rather than to make money, it was called a felony.
I'm not saying all of that scraping is fine. Authors and artists have real complaints, and those are worth sorting out. My point is that society has clearly decided that bulk downloading is a civil matter at most, something you argue about in court and settle with money. That's the right answer. It was the right answer in 2011, too. The difference is who was doing it.
What we owe him
Aaron is the reason I think the "open" in open web isn't just a nice word. RSS, Markdown, Creative Commons, SecureDrop, the SOPA blackout: they're all the same idea, that information and tools should be free to use and not owned by a gatekeeper.
If you want to honor that:
- Watch The Internet's Own Boy (2014), the documentary about his life. It's free to watch on the Internet Archive.
- Support the EFF, which has spent years fighting to reform the CFAA.
- Publish openly. If you write research, put a preprint on arXiv or your own site. If you write code, use an open license. If you write docs, write them in Markdown and let people copy them.
- Push for CFAA reform. Tell your representatives that breaking a terms-of-service agreement should never be a federal crime.
The web we have today is more locked down than the one Aaron helped build, and the law that was used against him is still sitting there. The least we can do is remember his name, and keep building the kind of web he believed in.
If you or someone you know is struggling, please reach out. In the US you can call or text 988 to reach the Suicide & Crisis Lifeline, any time. Outside the US, findahelpline.com lists free, confidential services by country.
Further reading
- The Internet's Own Boy: The Story of Aaron Swartz (Internet Archive)
- Markdown: Syntax by John Gruber
- Guerilla Open Access Manifesto (Internet Archive)
- EFF on the Computer Fraud and Abuse Act
- Notes That Outlive the App: Vimwiki, Org Mode, Logseq and Obsidian Compared

