Aaron Swartz's FBI File Released | Hacker News Reader

Aaron Swartz's FBI File Released

news.firedoglake.com

184 pointsrabble52 comments

Just a note about redaction -- my personal experience with FBI redaction is that it is completely subjective based upon the person who happens to be doing the redaction. There isn't some giant conspiracy to omit facts -- sometimes the person redacts too much, other times they screw up, etc. I've had names redacted on one page, only to appear non-redacted a few pages later. Even in cases where I've sent in death certificates for certain names (I know for a fact who the redacted person is), the name is still redacted in the output. I'm sure it is a very laborious, human process and all such processes involve a lot of errors.

Having had a great deal of personal experience with the FBI and the channels and management they're required to use and satisfy, I can verify that it's entirely accurate that everything is down to the individual. It's exactly the same kind of bureaucracy people complain about in large companies, except that the very bare possibility of change we hold on to in large corporate environments is not present.

Working with that organization in particular will teach you very quickly about the kind of immense roadblocks that exist to even the more "moderate" of conspiracy theories. Not only are they beholden to absolutely everyone, they don't have time to make things up or bury facts. Obviously there are exceptions, but this and most other government agencies are just loose conglomerations of individuals doing a never-ending amount of thankless work with the reward of being scrutinized and accused of every possible crime at every possible turn.

I hold government and its agencies in no high regard, because people are people and we all suck in our own way, but suggesting conspiracy at certain levels is just laughable.

"Having had a great deal of personal experience with the FBI [..] suggesting conspiracy at certain levels is just laughable"

http://en.wikipedia.org/wiki/COINTELPRO

Read that. Unless you can pinpoint how everything changed radically since then, I say heh.

Yes, this is an excellent point to consider. A conspiracy seems sexier, but it comes down to the human potential for error and the bureaucratic fear of torpedoing your own career (or worse, someone else's life) just because of a mistake in a menial task. If you were the redactor and the redactor's supervising manager, if something is a gray area and your lawyers aren't giving this their full attention, why not redact?

Of course, any one who is interested enough to file for documents may already know what he/she wants and will call you out (with a lawsuit) if it seems you've redacted things unnecessarily. But for other situations, it seems unlikely that the requester will either not go through the trouble of suing or not even know if what was redacted is worth suing over.

An example of a high-profile amusing snafu occurred during the Rod Blagojevich trial when someone using Adobe Acrobat forgot to complete the redaction process, allowing redacted text to be copy-pasted into another text editor: http://capitolfax.com/2010/04/22/blagojevich-hurls-allegatio...

Yes, that was a technology-error rather than one of judgment...but technology errors are much easier to double-check for and yet it's not an uncommon error in the digital age of public documents.

Overuse of classification labels is a huge problem. The incentives are skewed as you suggest, so there is no real benefit to spending the time required to make decisions on each and every individual fact or statement. Entire swathes of boring information are classified for no real reason and a large number of things that someone might use as useful data are sequestered in random agencies. The cost of complying with FOIA is not insignificant and over-classification jacks up the cost. I don't know if there is a really effective way to "fix" the situation without somehow changing the incentives.

They did not redact Steve Jobs's social security number, so it is indeed very arbitrary.

Plausible deniability..

(for those situations when a conspiracy might be necessary)

Amazon handed over his account details with very little prodding. No mention of a warrant. Making connections from AWS is probably more incriminating than making them from your home internet connection (your home connection has certain protection)

While we're talking about it... I sent in a FOIA request recently, and it came back a few days ago as 404 File Not Found. I previously sent in requests in 2009 and 2010 and never received replies. From talking with others, they claim that these requests should be on a record and that my 2013 request should have returned evidence of my 2009 and 2010 requests. Does anyone have a scoop on this?

I have always received responses to every FOIA request. You get an acknowledgement letter with a case number and then comes either documents, rejection, or a "not found" response.

And a bill. Both of mine incurred extra charges (~$80 each) for a one-page "non-response" brush off.

I'd love to see proof of this claim. The FBI covers the cost up to about 30 pages. Even the sample letter [0] the FBI gives you to use includes the following:

I am willing to pay up to [$____ ] for the processing of this request. Please inform me if the estimated fees will exceed this limit before processing my request.

[0] http://www.fbi.gov/foia/sample-fbi-foia-request-letter

Mine were from DARPA. The docs I was requesting should have been rather large (~100 pages x 15), so I was willing to pay up to $500 for 'em. Instead, they sent a curt reply (1 page or so) and a bill for $80. I'm not willing to post 'em publicly... but my email is in my profile.

So why would I not receive a bill/invoice for a brush-off?

Can you post these responses online? Especially the 404 one. Just curious.

I am traveling at the moment, but I could type them up when I get back. Shoot me an email reminder.

Here you go: http://diyhpl.us/~bryan/irc/fbi/foipa/response.html

It seems odd that his whole file would be an entire 23 pages in its entirety, and amount to 21 releasable pages without a single page relating in any way to anything other than publicizing the PACER records.

Surely they have more...

The FBI is part of the DoJ but I'm not sure if it is a fact that the FBI was involved with what went down recently. The DoJ has a separate FOIA process, so I am assuming that an FOIA can be made to them for all files related to their prosecution. I'd be shocked if some person or organization has not already started this process..

edit: DoJ FOIA http://www.justice.gov/oip/foia-request.html

FYI, Aaron himself posted his FBI file back in 2009:

http://www.aaronsw.com/weblog/fbifile

Has anything else been added since?

Looks like there are a few more pages, but most of it seems to overlap with what Aaron already posted.

Anyone else find it weird they included his social security number (page 5)? With the amount of detail they included it seems as though it would be fairly easy for someone to use any FBI file on a deceased person for fraudulent purposes, ID theft, or accessing accounts at companies who may not realize he is deceased yet.

Your social security number becomes public record once you die.

The situation with the Social Security Death Master File (SSDMF) / Social Security Death Index (SSDI) is very similar to the situation with PACER.

List from Nov 2011, obtained legally by http://ssdmf.info/

https://thepiratebay.se/torrent/7193029/Social_Security_Deat...

The SSA would love to insist that redistribution is illegal, because the price of the SSDMF is outrageous and a cash cow, but they don't have a legal leg to stand on. The SSDI is public record, and because it's a database with no creative element, it's not copyrightable even if it were compiled by someone other than the government.

Official sources for the SSDMF:

https://www.ssdmf.com/

http://www.ntis.gov/products/ssa-dmf.aspx

It would be really meta if Aaron had snarfed this one.

There is a database of deceased people with their SSNs available, it's called SSDI. All banks/credit agency should automatically flag these SSNs.

Social Security numbers are too widely used to really be regarded as secrets (how many government, bank, employer and call centre staff have access to yours) or to be used as authentication. They are a pretty good identifier of who is being referred to though.

It makes me sad the NYT story brought on extra heat. As a journalist, it stinks when collateral damage like such happens in the name of "truth" or some other subjective good. Good reminder to be careful what you say to the press.

Many of the pages were the same with different redactions, but the FBI withheld 2 claiming privacy, sources and methods, and the danger that physical harm would come to someone if they released them.

- Dan Wright, FireDogLake.com

Minor note: we don't do signatures at the end of comments on HN.

http://ycombinator.com/newsguidelines.html

"Please don't sign comments, especially with your url. They're already signed with your username. If other users want to learn more about you, they can click on it to see your profile."

Sorry about that. First comment.

Any non-scribd mirrors?

Just click the download button below the document. Here's a mirror if you can't access scribid for whatever reason: http://www.filedropper.com/126146785-aaron-h-swartz-fbi-file

The requires logging in, not exactly "pro privacy."

What about this: https://dl.dropbox.com/u/14707221/random/126146785-Aaron-H-S...

BugMeNot has a usable login http://www.bugmenot.com/view/scribd.com

It doesn't require logging in.

How much do you want to bet that someone manually goes through to scrub the personal information on every page...any other reasons the white out boxes would be so inconsistent?

This is correct. Redactions are typically handled manually. In the legal industry, for instance, this task is typically assigned to contract attorneys or paralegal staff.

There are of course auto-redacting solutions available, but due to the nature of the documents (scanned, inconsistent formats), and the potential consequences of either mis-applying a redaction or omitting one, manual verification is required.

> In the legal industry, for instance, this task is typically assigned to contract attorneys or paralegal staff.

Or junior associates :(

Hmmm--do I detect a bit of personal knowledge of this activity?

How else would it work? Do you trust a computer to do it? How would you handle challenges to the redaction?

Redaction is an A.I complete problem.

I don't understand. Was this file officially released?

Yes. Under the Freedom of Information Act, anyone can request US government documents (such as FBI files) and the relevant agency will send you the unclassified parts.

For any case or only for cases that were closed? And they don't even remove SSN, addresses and phone numbers?

You can request a file on any deceased person, and they'll give it to you in most cases.

Social security numbers of the deceased are also public information. Helps to prevent identity theft issues.

You can request any deceased individual's SSN application (which includes SSN, address, birth date, parents' names) directly from the SSA with proof of death: https://secure.ssa.gov/apps9/eFOIA-FEWeb/internet/main.jsp

Why are documents from legal proceedings worth millions of dollars? Aren't they public domain?

From the file:

"PACER normally carries an eight cents per page fee, however, by accessing from one of the seventeen libraries, users may search and download data for free."

And:

"The two accounts were responsible for downloading more than eighteen million pages with an approximate value of $1.5 million."

~18MM pages at $0.08/page ~= $1.5 million

So downloading a document is stealing printing fees from the courts, apparently.

My personal favorite part of the file so far:

"The Administrative Office of the U.S. Courts reported that the PACER system was being inundated with requests. One request was being made every three seconds."

I actually applied for a programming job with the Federal Courts in San Antonio about a decade ago when I was desperate for work. Cattle-call interview with ~100 easy questions, then they'd call you back that afternoon for a panel interview if you did well. I was criticized for dressing "too casual", despite wearing $400 worth of new clothes I couldn't afford. Apparently, suits are needed for programming, which also involves crawling around courtrooms installing equipment. Got the job offer a few days later for $55K/year. I mentioned that the newspaper ad said the position paid $78K, and politely attempted to negotiate. Five minutes later, I got an email saying they were rescinding the offer. Classy folks.

I hope you're happily employed now!

Posting your experience with the Federal Courts is helpful, thanks.

Thanks! Very happily employed now at a former startup that I also interviewed with back then. Decided not to go with them at the time because their 5-year plan was to sell the company within a year and move on to something else. They got bought out, I found another job, and when I got sick of that, got an interview here through a friend, and wished I'd started 6 years sooner.

Somewhat back on topic, since my current company was built on RSS initially, I doubt I'd even have this job without Aaron Swartz.

Sort of. They are not restricted by copyright, as I understand it, but, like uncopyrightable trade secrets inside companies, they are not freely available. Obtaining them from the original source (the courts) costs money.

Think of it like a company selling GNU software. There's nothing prohibiting that, even though once someone has it, they are free to redistribute it for free.

The claim that the docs were worth 1.5 million dollars is based on the arbitrary price set for access to PACER, not any underlying value of the documents.

Reply on news.ycombinator.com