Top
Best
New

Posted by LorenDB 11/20/2025

Verifying your Matrix devices is becoming mandatory(element.io)
210 points | 242 commentspage 3
g-b-r 11/20/2025|
One of the super confusing things is that even if you only use a single client it can be verified or not.

That's confusing even for very technical people; because, it simply doesn't make sense.

Saying "verified or primary client with recovery keys generated" seems too long, so they should just say something like "less secure" on the "unverified" sessions.

cowhax 11/20/2025||
The problem with Matrix adaptation has always been E2EE, or rather, the annoying implementation of it
adham-omran 11/20/2025||
Last time I used Matrix for our internal team notifications were beyond broken and we moved to Zulip, verification and authentication were also very funky at the time, I don't dare to try it again.
sschueller 11/20/2025||
I want to switch to SimpleX Chat[1] but at the moment there are issues with battery usage on android devices because of the way notifications are done. I hope this[2] or some other solution get merged soon even if there is a slight impact on anonymity.

[1] https://simplex.chat/

[2] https://github.com/simplex-chat/simplex-chat/pull/6205

irusensei 11/20/2025||
I had a more pleasant experience with SchildiChat hosted on a web server than the desktop Element clients.

I don't like the way groups/chatrooms are displayed to be honest. Its confusing. It feels like its trying to get away from the "server room/#somechat" model that works well with IRC and even with trendy current products like Discord.

bfkwlfkjf 11/20/2025||
Is this the ritual of getting together with a person and checking that their fingerprint match what you see on the app?

If this is that case what will happen is that people will start verifying everyone (because they might want to text to strangers that they can't bother verifying because the stakes are so low) and so verification will lose all meaning.

syntheticnature 11/20/2025||
It is not; I know we don't read articles here, but...
ranger_danger 11/20/2025||
Isn't this how TLS itself already works? "trust on first use"?
pavon 11/20/2025|||
Not in current practice. That is why you have to get a certificate from a trusted CA. If your CA isn't in the browser's cert database they will reject the connection even on the first time. If browsers allowed TOFU we would still be able to use self-issued certificates, without manually distributing certs to anyone that uses your service.

SSH is an example of TOFU.

majorchord 11/20/2025||
> we would still be able to use self-issued certificates

You still can... it just displays a warning message on first use, as does ssh.

treyd 11/20/2025|||
With PKI you're trusting a certificate chain up to a CA you already trust, by way of your OS or browser vendor.

A domain can layer on HSTS to that, which directs clients to additionally refuse to trust a new cert for a domain until the one you currently trust has expired.

scheub 11/20/2025||
That’s not what HSTS does. It asks the client to remember that you want to only use TLS for that domain and refuse to use unencrypted HTTP in the future.
olivia-banks 11/20/2025||
What exactly does this entail? I'm willing to be charitable in assuming that their use of "verify" isn't the modern usage of "give us your ID!" but I'm not enmeshed enough in the ecosystem anymore to know.
xethos 11/20/2025||
Respectfully, not even close. Verification is when I sign in from a new device, I use an existing device or second passphrase (either-or) to ensure that yes, it is me on both devices. I never have to reveal my ID, name, phone number, or email address to anyone. Not to Element, the Matrix Foundation, or the person running my home server where all my [encrypted] messages live.
ranger_danger 11/20/2025|||
My understanding is that there's two different types of verification.

Self-verification means that any new secondary devices you log into your account with will need to be verified by an existing login by way of an automatic popup that asks if you trust the device. It used to just be a Yes/No button but I think now they've added QR codes and/or emoji matching.

The other kind is verification between two different people, like when starting a direct message conversation, you might get the same emoji matching window to verify each other.

josephcsible 11/20/2025|||
Yeah, IMO "verify" was a poor choice of wording for what this is. It has nothing to do with remote attestation or any other form of Treacherous Computing, and it has nothing to do with your real-life identity. It's just "go on your old device and confirm that the new device is really yours."
goku12 11/20/2025||
If you don't mind reading an essay, here is mine from the same discussion: https://news.ycombinator.com/item?id=45989744
grishka 11/20/2025||
"Now the end-to-end encryption will leak into the UX even more and you better like it"

I'll say it again: E2EE will never become mainstream unless someone somehow manages to implement it such that it's completely transparent to the user while keeping all the features that people have come to expect from IM apps, like server-stored conversation history or support for multiple devices. By "completely transparent" I mean that the user doesn't have to do any extra actions whatsoever to make it work.

josephcsible 11/20/2025||
If that's true, then E2EE will never become mainstream. Consider this scenario: "My phone got lost/stolen/broken, so I just got a new one. I haven't logged in to this app since I got my last phone, so I forget my credentials for it. I'll reset them through my email. What do you mean my conversation history is gone?"

That's not really far-fetched. If you can get your conversation history back in that scenario, then so can the server operator so it's not real E2EE, and if you can't, then by your statement it won't become mainstream.

grishka 11/20/2025||
> If that's true, then E2EE will never become mainstream

Yes? :)

Given the choice, the vast majority of people would pick convenience over the kind of security that requires this much effort.

BrenBarn 11/20/2025|||
I more or less agree. And I also agree with the other commenter who says this may mean e2ee will never become mainstream. I think a lot of e2ee enthusiasts don't realize that the overwhelmingly most important feature for a messaging system is "when I log in, I can see all my messages". If there is a chance of that not happening, you're going to lose a lot of users.

I think there's the potential for a slight middle ground, but it would involve giving up a lot of the e2ee bells and whistles that privacy enthusiasts enthuse about (like perfect forward secrecy). You could image for instance a system where you have a single e2ee password and your data is encrypted on the server with that password. When you log in, you supply two passwords: your login password and your e2ee password. Then you have access to everything.

This tends to irritate people on both sides, since you can still lose your messages if you forget your e2ee password, and your privacy guarantees are also weaker, since the e2ee password can be a single point of failure that allows someone to read your messages. But people already rely on this level of security in other contexts. For instance, some cloud backup solutions encrypt your backup with a single passphrase. People are okay with having one password to unlock their entire hard drive's worth of data but not with one password to unlock their chat history?

I think it's worth exploring the space of e2ee solutions to find something that finds the balance between the levels of privacy and convenience that most users want. The thing is that existing apps that tout e2ee often do so to appeal to hardcore privacy advocates or people like dissidents in authoritarian states who are at risk of death if their messages are discovered. This level of security simply isn't a concern for the average person, and so they're not willing to take on the inconveniences that go along with it.

monerozcash 11/20/2025|||
> E2EE will never become mainstream

iMessage and Whatsapp are both mainstream.

grishka 11/20/2025||
Technically they are, but neither of them fits the strict definition of a E2EE messaging app, while also still hurting the UX.

Whatsapp is very insistent about backing up your messages to cloud services without encryption. To use it on desktop, you have to make everything go through your phone. And, afaik, you still can't transfer message backups between Android and iOS.

Even disregarding the extreme gatekeeping, iMessage relies on Apple managing your encryption keys so there are no confidentiality guarantees. Apple can, at any moment, give themselves a key to decrypt your messages.

Both Whatsapp and iMessage are proprietary, so it's also the case of "please trust us that we've implemented it the way we claim we did".

int_19h 11/20/2025|||
> iMessage relies on Apple managing your encryption keys so there are no confidentiality guarantees. Apple can, at any moment, give themselves a key to decrypt your messages.

It relies on Apple device managing your encryption keys, no? Which, yes, Apple can still access if it really wanted to simply by virtue of being able to push an iOS update that does that. But the same exact vulnerability applies to any app running on your iPhone.

monerozcash 11/20/2025|||
iMessage has worse UX than signal for key verification, but does support it. https://support.apple.com/en-us/118246

>Both Whatsapp and iMessage are proprietary, so it's also the case of "please trust us that we've implemented it the way we claim we did".

This is simply not true, any serious analysis of Signal would be performed on the binaries and not the source code. Having access to the source code does not make it any easier to discover well-hidden backdoors, but it is possible to exploit e.g. compiler behaviour in a way to create a backdoor that is essentially impossible to detect by reviewing source code.

Access to source code might very well make it easier to discover non-intentional bugs, but does not solve the problem of trust.

joecool1029 11/20/2025||
I mean we’re there for Signal. The parts that suck still are regarding access/retention of old messages which is an area Matrix is ironically slightly better about. But Signal we don’t need to think about verification, at worst it says this asshole has a new identity and then I have to tell them I’ve reset my iPhone for the 4th time this week…

Normal users do find retention important even if privacy/security minded users find value in ephemerality.

BrenBarn 11/20/2025||
Can you use Signal across multiple devices?
joecool1029 11/20/2025|||
Officially it supports linking other devices like their desktop app as a secondary. I currently use this to link into signal-mautrix on my matrix homeserver. This way I can access signal from multiple phones and multiple computers using a matrix client instead.
BrenBarn 11/20/2025||
But you still need one "primary" device and it has to be a phone, right? That's different from Matrix where you can have arbitrary devices that are all on an equal footing.
grishka 11/20/2025|||
Yes. And, annoyingly, when you only use Signal occasionally, these desktop sessions expire. And you have to link again. And when you do, you end up with a gap in your conversation history because "security".
int_19h 11/20/2025|||
Yes, and there's also a limit of max 5 linked devices.
jeroenhd 11/20/2025|||
You can use Molly to put Signal on multiple devices or you can bridge it into Matrix or XMPP, but you'll always need to run on one "main" device.
chrisjj 11/20/2025||
> This security update will give you assurance that when you receive a message from a contact, you can effortlessly assume it’s really from them.

Here's the thing. You can already! Whether you should or not.

ikkun 11/20/2025|
seems like it's just that element (the official, and most popular client) will ignore messages from unverified devices, but since it's part of the spec, other clients that want to be spec-compliant will implement this too. I don't think most other clients follow the spec that closely though.

I'm in favor of the change, the only downside I can think of is users with esoteric clients or simple bots that don't support verification won't be able to post to encrypted rooms with element users.

I feel like I'm alone in having good luck with matrix. I've been self hosting for nearly a decade to a handful of users, and it was a bit rough troubleshooting the encryption problems back when element was still called riot, but it's been a number of years since any of us have had a single encryption issue, and we added a new user recently with no trouble. we're still on 'element classic' though, the new 'element x' is a bit of a mess and loses the background sync feature, you need to set up a unified push server which I'm not looking forward to.

prurigro 11/20/2025||
For what it's worth, I've been using element x with unified push for a month or so now and I get notifications with message contents without any delay. Maybe they fixed the issue you're worrying about?

Self hosting the call/video feature became a lot more complicated though (and it's incompatible with the old system).

ikkun 11/20/2025||
my issues with element x are with the client itself, mostly missing features and bad UX. the main reason I don't want unified push is, it's just yet another thing for me to install and maintain, plus all my users need the client app installed. the ntfy server app even defaults to having a full web interface, fortunately it's possible to disable but it's just so much stuff to replace what used to be built-in to the app, to supposedly solve a battery life problem that I've never experienced.

I'm still going to get around to it, because element classic will be deprecated eventually. one of my users is on iOS and has a well-known bug with images not loading, which will probably never get fixed because they're focusing on the new client. and unfortunately I do have users that expect voice calls to work, so it sucks to hear that'll be annoying too.

eTomte 11/20/2025|||
I've had mostly good luck with Matrix too. Been self-hosting since 2022 and while there have been frustrations it has been pretty stable for basic chat.
tcfhgj 11/20/2025||
no, you are not alone, though I don't host
More comments...