Live data from Hacker News

Signal says it won’t compromise on encryption

theverge.com

311–320 of 336 posts

Re: Signal says it won’t compromise on encryption

#311
post #217

Earlier quoted context omitted.

Signal has backup on Android, and fairly fast image management, but it's lacking on iOS. That points to a lack of developer resources. What you call WhatsApp history is other people being able to read your messages. WhatsApp is violently agressive in demanding you use your phone and give access to your contacts. You can't use it without feeling violated.

Signal is similarly coercive about contacts - only "Yes" and "Not now", and once they get them, there's no way to delete the data they've got. Desktop Signal still keeps showing me other people who also have Signal, but whom I've never even tried to contact via Signal, only because years ago I've mistakenly let them see my contacts.

You can deny the Signal app the address book permission and it still works.

Signal contact discovery is also the most privacy preserving approach available anywhere, using only the partial hash of their phone number. You don't have to share your contacts -- but yes, your contacts may share their truncated hash list, and hence discover that your number is tied to a Signal identity, but it is on their phone that the number maps to a name, not in any network traffic or on a Signal server.

We're really not getting any better solutions until secure MPC becomes more practical.

https://www.howtogeek.com/708651/can-you-use-signal-without-...

https://signal.org/blog/private-contact-discovery/

Re: Signal says it won’t compromise on encryption

#312
post #285

Earlier quoted context omitted.

> I see no reason to back-up my history ever. I don't use signal for anything that needs to be archived. Every single time this comes up, someone chimes in saying they don't need it. That does not invalidate the users who do need it. If I can't keep my messages from device to device, I'm not going to use Signal. With far more manual effort than should be required, I can move my messages from Android device to Android…

Signal on Android has backups and migrations. Maybe the IOS signal doesn't?

iOS signal has no backups and migrations routinely fail. I've lost years of message history multiple times attempting iOS -> iOS migrations. It's a shame it's so botched and has made me stop recommending Signal to people entirely.

Re: Signal says it won’t compromise on encryption

#313

Is there anything more secure than signal that is widely used? Maybe something that doesn’t leak metadata or require a phone number?

Uhh, Jitsi? Jitsi Meet? Free and open source of course. Jitsi has ZRTP encryption. Jitsi Meet uses WebRTC and its encryption isn't the same, but from the academic research papers I've read such as "Stegozoa: Enhancing WebRTC Covert Channels with Video Steganography for Internet Censorship Circumvention" from June 2022, it's convincing enough for me to have set up my own Jitsi Meet server on a Debian Linux Virtualbox…

> widely used

Re: Signal says it won’t compromise on encryption

#314

Earlier quoted context omitted.

Would it be possible for you to expand on Signal not being dependent on donations?

"we depend on donations from our community" + a $105 million 0% interest loan (until 2068)¹. ¹ https://en.wikipedia.org/wiki/Brian_Acton#Signal

Are you suggesting Signal should not repay their debts or the best time to do that is not as soon as possible? If not, regardless of if there’s no interest, that loan should be repaid as soon as possible in my opinion. In my opinion, 990s are public, that loan was publicly acknowledged, and see no issue with a non-profit raising donations or repaying their debts.

All that said, if it was up to me, Signal would have their 990s on their site, annual publicly shared audits, etc. - which to my knowledge they do not.

Re: Signal says it won’t compromise on encryption

#315
post #310

> Signal makes its code open-source. Is anyone aware of whether they are still obscuring spam related code? For the downvoters: - https://www.reddit.com/r/privacy/comments/qlw1ag/signal_is_a... - https://signal.org/blog/keeping-spam-off-signal/

Okay, but so what? Why is this a problem? You can run a functional signal server without their spam filtering code.

It's a matter of trust. If I can't trust that it's truly open source, then it erodes some of the confidence I have in them. I haven't personally tried running a functional Signal server without the spam filtering code, so I don't know if it truly works without it.

Re: Signal says it won’t compromise on encryption

#316

Earlier quoted context omitted.

"we depend on donations from our community" + a $105 million 0% interest loan (until 2068)¹. ¹ https://en.wikipedia.org/wiki/Brian_Acton#Signal

Are you suggesting Signal should not repay their debts or the best time to do that is not as soon as possible? If not, regardless of if there’s no interest, that loan should be repaid as soon as possible in my opinion. In my opinion, 990s are public, that loan was publicly acknowledged, and see no issue with a non-profit raising donations or repaying their debts. All that said, if it was up to me, Signal would have t…

> Are you suggesting Signal should not repay their debts or the best time to do that is not as soon as possible?

Cutting through the tug-at-the-heartstrings sales pitch: Signal does not currently "depend on donations", nor will they for at least 45 more years.

It is reasonable to expect the entire truth when asked for money, instead of getting Wikipedia'd.

Re: Signal says it won’t compromise on encryption

#317

Earlier quoted context omitted.

So does Signal: who when with whom Matrix does leak more meta data though, e.g. message relationships for instance, but what does that matter?

That's false - Signal has had sealed sender on by default since 2018. (the setting you can turn on is just to see an icon in the ui).

No, see. https://news.ycombinator.com/item?id=33355126

Re: Signal says it won’t compromise on encryption

#318

Earlier quoted context omitted.

Kazakstan mandated government issue man-in-the-middle TLS certificates: https://www.zdnet.com/article/kazakhstan-government-is-inter... The EU is following this govennment friendly move: https://www.bleepingcomputer.com/news/security/experts-urge-... It would not be difficult for India as well. I see itlikely Modi and BJP will make this move as part of some anti-terrorist legislation.

Personally I find it unbelievable that major governments are not already in possession of the private key for at least one of the 150+ root certificates pre-installed on my device.

Some govs officially operates their own root CA https://ccadb-public.secure.force.com/mozilla/IncludedCACert...

Re: Signal says it won’t compromise on encryption

#319

Earlier quoted context omitted.

Yeah, the Assistance and Access Act (2018) passed and is still the law of the land. It gave law enforcement, intelligence and the government the ability to compel Australian providers to backdoor their services. Can't tell you how often it's been invoked since it also allows for gag orders. Probably only 1 out of 10 Aussies would have even heard of it. The mainstream media here are about half a dozen ideologically al…

I guess now we share that plight.

Australian here. What Assistance and Access bill did is different, and clever in some ways. TL;DR: it doesn't attack encryption directly - it doesn't give the government power to direct anyone to hand over keys, for instance. In fact the bill specifically prohibits the government in asking anyone to introduce a "systemic weakness". A systemic weakness is something that would allow the government to spy on everyone - which is what India seemingly wants to do.

But as I said it's clever. It can specifically prohibit introducing new "systemic weaknesses" because they already have one that's more than good enough for their purposes. That would be silent security updates. They have given themselves the power to compel tech company (Apply/Google/Microsoft) to install a silent security update on a specific device. The "silent security update" would of course be a bug (spy intercept) of some sort. It doesn't bypass encryption because it doesn't need to - a human can only consume unencrypted data, so that's what the spy bug intercepts.

They've ensured that it will never be systemic to their own satisfaction by putting several hurdles in place, like independent judicial review of the bug and which devices will be targeted. The fundamental principle is the only acceptable reason for targeting someone is criminal activity. If those hurdles are respected (and it seems likely they would mostly be followed) it means the Chinese like surveillance society India seems to be trying to create would be very difficult in Australia, even with this law. Which I guess would make it a reasonable compromise between government privacy invasion and law and order concerns.

The flaw is it's impossible to know if they are being respected. All companies and people forced to inject these spy ware updates are automatically subject to a gag order. All that review I mentioned happens in secret, and they have specifically exempted themselves from publishing any meaningful information on who, what, why and how devices are targeted.

To finish the picture - if the Australian government was concerned about criminal behaviour happening over Signal, it's highly unlikely they would be approaching Signal as India has apparently done. (I can't for the life of me think of a reason why Signal would give a shit about what the India government thinks or wants. Ditto the Australian government.) Instead they would direct Google to inject keyboard and screen monitors into Android. Google makes a lot of money in Australia, so it's likely they would comply. Like I said - it's clever.

But not impossibly so. It only works if they can target a particular device. For a commercial products this is invariably easy - Apple, Google, Microsoft all want you to sign in with an identity so they can milk some profit out of it either by charging you or at least displaying advertising. But open source projects, like Fedora or Debian, go out of their way to not identity the users, and worse Debian creates audit trails like reproducible builds. So their users are largely immune to Australia's Assistance and Access Bill (2019). But they aren't immune to India's rubber hose approach.

Re: Signal says it won’t compromise on encryption

#320
post #309

Earlier quoted context omitted.

We're going a bit off topic here but I'll expand a bit. If we both (presumably) start from the premise of "this invasion is a bad thing", then following this up with "anyone who supports it is just a moron duped my propaganda" and/or "the Russian administration are simply crazy" isn't helpful. This did not happen in isolation, it happened against a backdrop of decades of western failures in diplomacy & deterrence. Th…

I said neither of those things

Apologies for strawmanning you
Post reply on HN