Hacker News Reader

Comment by codetrotter | Hacker News Reader

Off topic but it’d be funny if the censored text was literally just put in span elements with “background: black” styling.

Anyways, cool story and I like the phrasing of “borrowing the Lunik over night”.

Also, I think this document was probably digitized using OCR and then cleaned up by hand afterwards, considering the mistake of the word “film” having instead become “fihn”. That’s the type of error OCR would typically make, not a human:

> They filled one roll of fihn with close-ups of markings [...]

View on HN

This literally just happened in Canada a few days ago: a hugely significant data leak of TOP SECRET documents shipped with gray boxes under the most interesting parts of the document:

https://www.rebelnews.com/the_china_files

I'm sure it did, but Rebel News? That site is a cancer, not a news source.

Eh, It's the original source. I tried Googling around and couldn't find a better source.

Probably why I haven't heard about it. Rebel News reporting is a no-go zone for the Canadian media.

The Globe and Mail reporting on it was far more mild and less revealing, considering China has two Canadians locked up for years while they invited PLA to train with our troops. Nor did include the original source material: https://www.theglobeandmail.com/politics/article-global-affa...

Makes you wonder if we'd ever have found out about this otherwise.

The original source is actually kind of interesting. It's too bad this sort of thing isn't covered more regularly in the mainstream news.

That said, it's rather mis-characterized by Rebel News, I was only able to read the title of the Globe and Mail piece, but it seemed like a more accurate representation of the source document.

One thing which hit me when reading it, was the cache of email addresses, all to receive this 'Canadian Eyes Only' and 'Secret'.

For anyone, it creates a chain of people who are ... required? Important enough? Need to know? Said info...

There's like a few security implications here, such as, whom to apply pressure to. Some GoC department employees, on even LinkedIn for example, are restricted from indicating who precisely they work for.

For this very reason.

Even non-compliance-with-pressure tactics are a concern, for it helps to know whom precisely to attempt hacks against.

Some might say "Meh, they probably already know this and that", but things like this are precisely how they already know this and that.

This kind of attitude seems to be more of a "religious" approach than I'd like.

For me, personally, the more opposing opinions I can obtain and comprehend, the better my bounding limits on a topic can be.

Even if you really despise someone's personal or political beliefs, how does it benefit you to not fully understand their claims?

In Canada, we have an almost total blackout on interpretations of data that don't match the primary political party's desires and beliefs. They fund these "journalistic" organizations to the tune of billions of dollars, completely locking out any organizations with contrary opinions.

Is it your belief that this an ideal situation for Canadians wishing to discover if there exist alternative viewpoints consistent with the available data?

The actual link for anyone not wanting to visit such a site.

https://drive.google.com/viewerng/viewer?url=http://d3n8a8pr...

This has happened with PDFs where they just scaled a black square image over the redacted text.

XXXX XXXXX XXXXX XXX XX a Soviet upper-stage space vehicle.

No I meant like if the actual text was still there in the HTML.

I wonder if they've actually only replaced letters with X's. It seems like a poor decision to leave spaces in place, as people can attempt to guess the words based on their length. It's probably randomized, but why bother doing that instead of just "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"? Odd.

I wonder if GPT3 could be used to infer reacted text prompted by the surrounding article..?

But for something like this which is sort of a semi-serious report I kind of have the feeling maybe they're just doing some of these as spurious redactions for the laughs.

> I wonder if GPT3 could be used to infer reacted text prompted by the surrounding article..?

Generally, I would assume that the part of the text with the most entropy (redacted nouns and their adjectives) would be especially hard for a system like GPT3 to generate. There might be some clues in the surrounding texts, but perhaps someone would need to train it specifically on similar (un-redacted) documents?

Also, the other way around, marking parts as viable-for-redaction and then having GPT3 generate text to fill in those gaps would be a really insidious way of redacting content. I'm guessing a GAN-type network could even generate the visuals that are appropriate to the document that has been redacted..

If the redacted part was somebody's name or a project codename, there's basically no way to infer that unless the system has had source text with that info. But I guess in the case where they are redacting something like a "method or procedure" that has been written about unredacted in other contexts and then its application in this context is classified, I think that would be possible. I'm sure people are already thinking about how to guard against this.

Your second paragraph is interesting. It's basically an automatic disinformation generator. I think there's almost more security value in releasing a document like that with faked interpolated text standing in for redacted sections, and then a few blacked out sections to add to intrigue and credibility, than in doing a regular redacted document...but maybe there's benefits of the old way I don't understand. Not sure if I got what you were saying there, but I think I got your meaning correctly.

Please do not train GPT3 to think/act like government workers. Or politicians. Please. I beg you.

Copy and pasting the first blacked out section results in:

XXXX XXXXX XXXXX XXX XX a Soviet upper-stage space vehicle.

Reply on news.ycombinator.com