Hacker News Reader

It's really an engine for revealing people's true preferences for messaging, which, for many people, tend to be that they want all the ergonomics of Slack a lot more than they want cryptographically sound secure messaging.

What's hopeful in all this is that Signal is, slowly, catching up. Slack can roll out new features just by assigning a couple developers to it, and Signal has to coordinate new cryptographic research --- not just new cryptographic research, but research that produces something deployable at scale within the resources of a project like Signal! --- so Slack (and Wire and Keybase) are at a permanent advantage here.

But over time, Signal gets more and more usable without having to consider tradeoffs.

> that they want all the ergonomics of Slack a lot more than they want cryptographically sound secure messaging.

So you consider accounts not tied to a phone number "Slack ergonomics"? Before WhatsApp that was the default.

I said it was a reason to hate Signal, not the only reason.

Why would Signal (a drop-in SMS replacement) be compared to Slack?

Because we're talking about group messaging here.

Why would group Signal messages (a drop-in replacement for group texts) be compared to Slack?

I recently was selecting a messaging platform for my family, and we evaluated both Signal and Slack, and went with Slack. My wife did the same with her family, and went with Signal. From this, I gather they overlap in some features enough to compete for some use cases.

One (Signal) is a replacement for texts, and one isn't. "Overlap in features" wouldn't cover it, but rather how much you want it to be like texting.

Thomas Ptacek is a big Signal advocate, as am I, but he doesn't like to think of it as a drop-in replacement for texting, whereas I do (because that's what it is and where it shines). I move texting onto Signal whenever I can.

Why not Keybase over Slack?

Not saying Keybase is better; no dog in that fight, just curious if you had considered it.

Why would a group communication tool be compared with another group communication tool? What's the part you're missing there?

I have some friends I talk to in Signal groups. I have others I talk to in Slack. In both cases, the goal is the same: communicate privately with a known group of friends.

Please recall that earlier in the thread, the following was posted:

"[Signal is] really an engine for revealing people's true preferences for messaging, which, for many people, tend to be that they want all the ergonomics of Slack a lot more than they want cryptographically sound secure messaging."

This comparison to Slack makes no sense - Signal replaces texts and makes them end-to-end encrypted. It's a straight upgrade to texting (except, apparently, on iphone, where apple won't let the app send plain old texts and the "drop-in replacement" quality is neutered). It requires a phone number to use, and is linked to that phone number.

Signal is right to be what it is, and if Apple got out of the way, I would insist on replacing all texts with Signal. Replacing my Slacks with Signal or my Signal messages with Slack fails to type-check.

The comparison doesn't make sense to you, because you value privacy highly. It makes plenty of sense to people who don't.

People do literally compare them when deciding what group messaging app to use: https://news.ycombinator.com/item?id=21746863

For people like that, end-to-end cryptographic security is at best a nice-to-have. And I'd guess that's circa 90% of people.

Signal's true value comes when lots of people are using it. I never bother with secure email, because almost nobody I know has it set up. But I use Signal for the great bulk of my texting, because most of my friends are on it. If Signal wants that to be more and more true, they have to compete with the other tools people use for group communication.

In the case of Slack though these are "private" communications only in the same way that say, email to colleagues at work is "private". Lots of people certainly could snoop this, and more probably would be able to if they really wanted to. You would not be told about that, it'd just happen and everybody involved would convince themselves that it's fine. Is it fine though?

Signal's rationale is that if we actually secure this type of conversation, we can tell people not to accept insecure conversations because they're trading something you might want (actual privacy) for... not very much.

We've been here before on the Internet, at least twice now. When I was still (barely) a teenager Tatu Ylönen invented SSH and connecting to another machine was now secure instead of hopelessly insecure. And at almost the same time a bunch of people at Netscape invented SSL (which became TLS) and made the World Wide Web secure. It only took a few years for ordinary (relatively) people to _expect_ SSH not telnet and it took a bit longer for HTTPS but in both cases we got to a place where secure was the default and expected condition.

Yes, thanks, I understand the technical difference. What I'm saying is that from a user perspective, many people don't care, or don't care very much. Otherwise they wouldn't be using SMS, telephones, or email.

If Signal wants to be broadly successful, they have to be as good from the perspective of the broad base of users.

My biggest annoyance with Signal is that getting a new phone ends up wiping out all conversation history with apparently no way to transfer it.

This loss of user data is not advertised well enough up front, and leaves users feeling tricked. In many contexts loss of user data is an even bigger sin than weak security.

What's ironic here is that in the adversarial setting the application is designed for, unexpected retention of user data (on end-user devices) is a sin.

That’s fine. But Signal should then advertise itself as unsuitable for general-purpose communication, primarily relevant when someone is specifically worried about adversaries reading the communication.

I can see how this makes sense for journalists, dissidents, diplomats, criminals, corporate executives, etc., but if data is under threat of disappearance, regular people should be warned away and told to use something else for day-to-day communication.

For some of us data loss is often a bigger threat than unexpected retention.

I like Signal and it always makes me happy to see more people showing up there, but for now certain group chats will stay on other messaging services.

I like to think about messages being ephemeral. If a piece of information needs to be saved, I just store it outside the messaging app. This includes media files, too.

Personally, I'm happy to lose the data. I found it odd that with both phones and the SIM on the desk in front of me, I couldn't figure out how/if I could vouch for my key changing in any way.

Needing to say I have a new phone just trust me largely defeats the purpose.

If you are on an Android device you can export an encrypted backup and scan a QR code / type in the password to the encrypted archive to transfer messages / group memberships with only a safety number change in most cases.

https://support.signal.org/hc/en-us/articles/360007059752-Ba...

No dice for iOS unfortunately.

I think that's the opposite of what I want? I want to inform people of the new safety number using the old channel and purge all data like a good user.

In this respect a keybase like model makes more sense to me.

Two small corrections: Signal-Android's backup works with a passphrase only (no QR codes involved) and does not cause safety number changes on restore.

I’m talking about transferring archival data from one phone I own to a different phone I own.

This is different from whether other users are told that my security keys just changed.

Right, I think there's a partial process for what you want and not for informing of key change and I find that backwards AFA security.

Look in the settings then switch backup chats externally to on, then hit backup chat. What is so hard about that?

It is, but there's another aspect besides convenience and ergonomics. You surely know better than me that privacy and security are non-binary, and everyone has their threat scenarios.

In some cases, an ability to have multiple independent accounts/identities (pseudonymity) would - unfortunately but practically - beat true cryptographic security that Signal offers. I mean, personally, I'm less concerned about platform (e.g. Wire or Whatsapp) or some government agency learning that I'm talking to my buddies at certain schedule, than mixing up my acquaintances from different groups together, having to maintain a single identity for them all. Some people I talked with didn't knew my name or phone number, and I would be uncomfortable if they would. For me, in my life I've said less things I wouldn't want governments to learn about, than times I've used a pseudonym/throwaway account to talk to people.

"Signal forces users to use phone numbers"

Which is number one reason why I'm not even considering it

"some people don't like this because they want to use multiple ephemeral usernames"

That's not my reason: I don't want people I don't know to get my phone number through other people I know and trust, but are used to share everything online. Of course that would be possible without any social application as well, though using one makes it much more natural.

Then this one from their site: "Multiple mobile devices and Android tablets are not currently supported"

Triple facepalm here: this makes it even worse than Whatsapp I use (read: am forced to use) on an old tablet. Whatsapp sucks badly just at everything (didn't I write I'm forced to use it?) but at least I can read what I write.

Downvotes welcome, though advice on secure+open alternatives that don't assume I have a smartphone (I haven't one and don't plan to) would be more informative.

The day I would be comfortable giving out my phone number to stranger is when it becomes mandatory to whitelist all callers, much like how just about any non-PSTN systems work.

Maybe this is because of the social expectations of that it will work without such overhead but I just simply can't notice how all the "countermeasures" phone industry (and governments as this is a heavily regulated industry) are ignorance to elephant in the room...

If you are a client for Signal, rubberhose cryptoanalysis is a much bigger issue.

Here is a story what has happened to Doubi (SSr developer.) He was a very well aware of anonymity risks, and he evaded police for years on end. China literally tried to do geolocate him by turning off the internet in entire cities, but to no result — he caught on to that, and started randomising his release timing, and avoiding releasing "hotfixes". So, the entire Chinese police and MSS been looking him for 4-5 years.

What has happened? A few month before his arrest, he registered a Twitter handle with a throwaway SIM card. Those are being usually sold by "grannies" in Chinese 2nd tier cities who peddle things like fake tax receipts, anonymous train tickets and such.

China either hacked Twitter, or had somebody bribed there, and they got the number. They then tracked down the granny who sold him the SIM card, and went on and checking every person door to door in that small town. Then, they found him.

He got 5 years prison, and 4 years of laogai (gulag)

But, in this story - had he used signal - if the police arrested anyone in contact with him, any one of those would be able to turn over his phone number? Which would be linked to the sim card in his phone?

Am I missing something? Or am I misinterpreting your story? You're saying that sign up bound to a Sim card is bad for Twitter and bad (worse) for signal?

Yes, see, he went as far as buying an anonymous sim in China, which are sold at extreme premium by black market dealers, and still got tracked down.

That's super interesting, thanks for sharing! Would you mind posting a link or a two about the story of Doubi. I can't find much and would love to dig into this story.

Basically Twitter got pwned big time, and now denies it because GDPR will ruin them if breach is proven.

Here is what Doubi's online followers figured:

State security got all phone numbers used for Twitter phone verification up to May 2019 and possibly till July.

Twitter haphazardly closed the breach in complete secrecy.

API hole explanation is excluded as people with 100% private accs got police visits.

People with foreign SIM cards also got into trouble. So the explanation that China compromised Twitter's SMS providers is also excluded, as its improbable that they did it in 4+ countries.

2016 breach is also out of question.

The only explanation is that they got hold on a big piece of their user DB, or, worse, they have an active infiltrator in Twitter, or Twitter voluntarily cooperated.

https://mobile.twitter.com/robert_spalding/status/1134797195...

https://amp.ft.com/content/afd44222-5c34-11e9-9dde-7aedca0a0...

Pardon my ignorance but I'm unable to find much about this story... and the links you posted are hard to piece together with this narrative.

Not even doubting it, just wondering if there's more of a source that's laid out (work/timeline/etc)? It's supremely interesting and should probably be more well known if it's not already.

Most of what I know was found by people on Doubi's forum which now went down. Near nothing about that in English besides stating the fact that he is gone now, that he got a term, and that his Twitter was the most likely source of his ID leak as deduced from public records about his case.

Early accounts explored the possibility of Chinese police exploiting SMS gateway, and password reset abuse, but it has since been confirmed that even users who lived for years in the West got deanonymised, and their relatives got harassed. MSS/police having fresh twitter user DB is the most probable explanation at this point.

> ”Signal cannot have multiple mobile clients, only one mobile client and a single desktop version”

This is wrong. You could always have multiple desktop clients. You can also add iPads as linked clients now. Personally, I have two desktops and an iPad linked to Signal.

WhatsApp doesn’t support linked devices at all, the web client connects through your phone. Signals linked devices function independently, you can power off your phone and they’ll still work.

> Signal cannot have multiple mobile clients, only one mobile client and a single desktop version

You are right about the mobile client, but that's not true of desktop. I have Signal installed and setup on every desktop/laptop that I use without any issue.