Hacker News Reader

Comment by e1ven | Hacker News Reader

As I understand, Keybase chat is open-source? (https://github.com/keybase/client)

I don't have time to read through the code right now, but I'd love to hear how they implemented exploding messages with untrustworthy clients.

I've thought about it a few times before, and it seems one of the few places that closed software has an advantage - You can't easily force third-party clients to delete messages.

If they've solved that, I'm really interested to learn how it works!!

View on HN

It is impossible to implement this feature "safely" even with trusted clients -- worst case I take a screenshot or even a photograph of the device displaying the message before it explodes.

If you don't trust the person at the other end, this is never going to work. It's more useful for "we both agree that we don't want a paper trail" kind of thing.

A dead simple way to thwart the “screenshot” attack is to release a tool for accurately falsifying a screenshot. I’ve never seen this employed in practice though.

Photo, audio, and video evidence should already be dismissed until one is able to verify the integrity and source. All of these can already be believably faked - it's just a matter of educating people that a layperson can easily create fake things by using tools developed by research teams.

Fake text is the easiest to fake if you can identify the font used - any image editor will work. HN uses 9pt Verdana, even without using dev tools I could fake your post to say anything I wanted it to say since it would just be 9pt Verdana on a solid background set to text wrap every 1050px.

See: https://www.youtube.com/watch?v=ohmajJTcpNk & https://www.youtube.com/watch?v=AmUC4m6w1wo

Not even that much effort, just open the browser's dev tools and change the text in the post to say whatever you like.

I'm aware - but specifically excluded dev tools as faking text in scenarios where dev tools may not exist (eg: chat programs that aren't taking place within the browser) is still trivial.

That's an interesting way to combat certain kinds of attacks - those in which the screenshot 'reveals' embarrassing, illicit, or illegal communication: things you've done or intend to do that have no external evidence aside from the message itself.

But it does not combat attacks that are not embarrassing but rather a release of information you are known to have but which is intended to be kept secret, or which is easily verified: if someone captures your social security number, private key, or home address in a screenshot, you'd better be really good at bluffing.

If I release data that only I have, it’s obvious that I released it. If I release data that two people have, I can counter your screenshot with a faked screenshot of my own.

How is any bitmap editor not such a tool?

I think they're talking more along the lines of those online meme editors. Yea i could download the lion king and clip the frame with simba and then pull it into photo shop and then add text and then upload it to imgur. Or I could go one of those online meme editors in click the photo and enter my text and then get a link to the meme I made.

The whole point is to totally lower the bar for anyone to make a passable copy, thus removing all confidence that any screen shot is genuine.

If it’s made by the company making the messaging app, it will be pixel perfect, and will only take a few seconds to make instead of a few hours. And if you create one for a date in the past they can use the UI they had at that time.

Apps do screenshot detection but again nothing beat the old Polaroid.

Any DLP or DRM can be circumvented using analog means.

Absolutely - That's my understanding as well.

There are ways to mitigate (snapchat detects screenshots, etc) but no way to fully prevent - Someone could always use an external camera, etc.

I was just really hoping that they had come up with some sort of cool technical way to stop the ability to decode messages after XYZ time, even if they couldn't prevent it from being copied once decoded.

For example, imagine if a message were wrapped in two levels of encryption - Once from the user, and once with a key that you have to retrieve from keybase.io - If you weren't in the right time-window, you wouldn't be able to retrieve the second key.

There's lots of problems with that particular approach, which is why I was hoping they had come up with something awesome, not just asking the client nicely to delete it. It's a nice feature either way, though.

This is exactly how most DRM schemes work. You ping the keyserver a key, they check if you're still authorized, and if you are, they send you a key. There's even a bunch of tech that got built into commercial OSes [1][2] and hardware in an effort to try really, really hard not to expose decrypted content outside some protected path. But if a chunk of decrypted content is displayed on a screen, there's always the analog hole, e.g. taking pictures or a video of the display.

Nonetheless, it'd be interesting if all this effort, which got done because of the lobbying of major media rightsholders, was re-used for interpersonal communication.

[1] https://en.wikipedia.org/wiki/Protected_Media_Path [2] https://arstechnica.com/gadgets/2016/11/netflix-4k-streaming...

I think the solution is pretty simple: encrypt with a one-time pad and then store the pad in a box that will burn that pad at your desired time; make sure no one can ever see the pad and your secret will be pretty safe.

And how do you make sure the box will burn at the desired time? What if someone stomps on the fuse?

Integrated timer with small nuclear power source.

Reply on news.ycombinator.com