all 96 comments

sorted by: hot top controversial new old
[–] 128 points 1 year ago* (last edited 1 year ago) (22 children)

TL;dr of the article :

  1. They keep your private key on their servers.
  2. Their implementation allows for AITM attacks.
  3. It's closed source.
  4. There's no perfect forward secrecy.

This secret stays between you, me, and Elon.

I hope politicians use the hell out of it, so we can see what they really think when it gets (inevitably) hacked in a few weeks.

  • source
  • hideshow 22 child comments
  • [–] 23 points 1 year ago (13 children)
  • [–] 26 points 1 year ago (1 child)
  • [–] 4 points 1 year ago

    Are you sure that site is trustworthy? It kinda reads like an LLM being told to explain the difference between two names for the same thing and basically rephrasing the same thing. I'd imagine it might just be a different name to get rid of a male-coded word.

  • source
  • parent
  • [–] 0 points 1 year ago (4 children)

    They are stupid, but not that stupid.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 15 points 1 year ago (3 children)

    Never attribute to malice what can be attributed to incompetence.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 12 points 1 year ago* (last edited 1 year ago) (2 children)

    I used to give the benefit of the doubt but when there are bad incentives in play and shit keeps happening.. then perhaps that is naïve sometimes, unfortunately.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 100 points 1 year ago* (12 children)

    offering me end-to-end encrypted chat

    No one - not even X - can access or read your messages

    This key is then stored on X’s servers

    So...they're just blatantly lying?

  • source
  • hideshow 12 child comments
  • [–] 16 points 1 year ago* (3 children)

    It's encrypted with a 4 digit pin so they'll have to spend at least 316.8809e-10 years on brute-forcing it.

  • source
  • parent
  • hideshow 3 child comments
  • [+] 4 points 1 year ago (6 children)
  • [–] 2 points 1 year ago* (5 children)

    There are hardware for that called hardware security modules, but yeah I definitely wouldn't trust Twitter's implementation - especially because they probably just need the auth team to tell the HSM that the user logged in when they didn't to get that key

    A proper implementation would use multiple security measures and require a reset (delete) of certain private account data before the account access can be reset, otherwise the user's password would be needed (for key derivation) or some other secret held by the user's devices (in the TPM chip or equivalent)

  • source
  • parent
  • hideshow 5 child comments
  • [+] -3 points 1 year ago (5 children)
  • [–] 2 points 1 year ago* (4 children)

    I've run a cryptography forum for 10 years. I can tell snake oil from the real deal.

    Musk's Twitter doesn't know how to do key distribution. The only major company using HSMs the way Musk intends to is Apple, and they have far more and much more experienced cryptographers than X does.

  • source
  • parent
  • hideshow 4 child comments
  • [+] -1 points 1 year ago (4 children)
  • [–] 1 point 1 year ago (3 children)

    You sound like an antivaxxer defending a crank

  • source
  • parent
  • hideshow 3 child comments
  • [+] 0 points 1 year ago (3 children)
  • [–] 1 point 1 year ago (2 children)

    If you can't demonstrate that you know more about cryptography then me, it's time for you to admit you're wrong

  • source
  • parent
  • hideshow 2 child comments
  • [+] 0 points 1 year ago (2 children)
  • [–] 1 point 1 year ago* (last edited 1 year ago) (1 child)

    This is incoherent bullshit.

    You're choosing to pretend it's nothing so you can dismiss legitimate criticism.

    An engineer hearing about some novice trying to build a plane using difficult methods that only one or two companies with immense expertise has succeeded at would be correct to assume that plane would be unsafe.

    A doctor hearing about a tiny clinic attempting treatments that only big medical research facilities have pulled off are correct to assume they're charlatans.

    A cryptographer hearing about somebody attempting to build E2EE using methods that very few are capable of implementing correctly and without having the expertise on hand are correct to call that snakeoil.

    Cryptography is INFAMOUSLY complex. E2EE is infamously difficult to make easy ("Johnny still can't encrypt"). The worst part is that cryptographic failures are almost always 100% silent!

    There's a reason almost everybody copies Signal's protocol, and that everybody else who does it in-house keeps having vulnerabilities.

    Multi user key management (PKI) specifically is wildly complex.

    They're doing cryptography in the browser - famously difficult to make it work decently because there's no reliable code pinning solution, no reliable protected key storage (no TPM protected keystore) and absolutely no auditability. And that's on top of the risk of getting served malicious Javascript via XSS attacks, or by the host getting hacked, or by a maliciously issued certificate (there's 800+ certificate authorities, FYI, no cert pinning = easy for a state level actor to MITM)

    They're not doing transparency logs of user keys. Even whatsapp has started doing that.

    I haven't seen evidence of them attempting user key verification

    Twitter/X has only displayed signs of LACKING the necessary expertise.

    To pretend that's wishful thinking from me just reveals how little you care about expertise.

  • source
  • parent
  • hideshow 1 child comment
  • [+] 0 points 1 year ago (1 child)
  • load more comments (-1 replies)
  • load more comments (-1 replies)
  • load more comments (-1 replies)
  • load more comments (-1 replies)
  • load more comments (-1 replies)
  • load more comments (-1 replies)
    [–] 44 points 1 year ago (1 child)

    ...yet? How bout just not trusting it at all?

  • source
  • hideshow 1 child comment
  • [–] 30 points 1 year ago (5 children)

    That "yet" is the narrative hook to trick us into feeling like it will soon be trustworthy, and that our assumed suspicions refer to a temporary state of untrustworthiness. Clever girls!

  • source
  • hideshow 5 child comments
  • [–] 7 points 1 year ago (3 children)

    Feels like Bluesky’s federation promise.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 1 point 1 year ago (2 children)

    I think you can install your own bluseky instance and federated with others.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 1 year ago* (1 child)

    I don’t consider the PDS stuff to be fully federated. That’s just keeping your data on a different server, as far as I understand it. To be federated it needs to be a full interoperable server like mastodon, or lemmy.

    You should also be able to host a non federated instance, or one with limited federation.

    If they have moved past that, and I can open a server and have people sign up for accounts, then I stand corrected.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 1 year ago*

    Bluesky federates across different layers, it's modular, it doesn't have a comparable same-layer federation. It is fully interoperable, just not by the method you're used to.

    You can host your own partial appview now (caching and indexing your and your friends' comment), and multiple people have managed to run their own relays for cheap (caching most of the posts in the network), and you can pull the rest of data you need to browse from the other relays and use the service as usual. You can run your own moderation labeler, use your own app, just your own account, etc...

    Just look at the interoperable blacksky project by a bunch of black devs making their own infrastructure for accounts and moderation, etc.

    To be non federated, all you have to do is not announce your server and not accept arbitrary connections

    Due to content addressing, limited federation isn't really a thing by the usual definition. You can filter content from any PDS you don't like, but can't really control who can see already public posts

  • source
  • parent
  • How about: "You probably should trust or use X at all... ever."

  • source
  • [–] 20 points 1 year ago

    Hey y'all. Reminder not to trust a platform owned and operated by a Nazi manchild.

  • source
  • [–] 19 points 1 year ago
    [–] 19 points 1 year ago

    Yet? What kind of idiot would imagine that X would or could provide actual secure communication?

  • source
  • [–] 18 points 1 year ago

    Brain damaged people trust x again.

  • source
  • [–] 18 points 1 year ago

    Shouldn't trust it yet.

    Or ever.

  • source
  • [–] 14 points 1 year ago

    XChat, has some red flags.

    With a white circle and a swastika inside?

  • source
  • [–] 14 points 1 year ago (2 children)

    It's proprietary, how could you possibly trust it?

  • source
  • hideshow 2 child comments
  • [–] 12 points 1 year ago (5 children)

    Why are people evening using this site anymore? It’s been severely compromised.

  • source
  • hideshow 5 child comments
  • [–] -4 points 1 year ago (2 children)

    Cause it has an audience unlike mastodon or bluesky. All the other alternatives are dead.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 11 points 1 year ago

    "xchat" sounds like one of those porn chat rooms

  • source
  • If you trust ANYTHING Musk has for you well then have I got a bridge to sell you.

  • source
  • [–] 10 points 1 year ago

    Quick everyone, install this just so that if Pete Hegseth invites people to the next airstrikes chat group, your satirical JD Vance account will be next to the real JD Vance's account and he'll probably add you both and figure it out later.

  • source
  • [–] 9 points 1 year ago (1 child)

    I don't trust anything coming out of Elon's fascisthole. Deleted the app when he bought it and never looked back.

  • source
  • hideshow 1 child comment
  • [–] 9 points 1 year ago (1 child)

    Never trust any social media sites "private" chat.

    Especially not one of the big ones run by weirdo fascists. You know Elmo is going to snoop on anyone relatively famous, or that just say something he doesn't like.

    In all honesty, there's zero reason to even have accounts on them

  • source
  • hideshow 1 child comment
  • [–] 6 points 1 year ago

    Even if the server had zero knowledge of your private keys (which is doubtful), I'm sure the client code won't have any backdoors. It's only the social media "platform" owned by the world's most thin-skinned billionaire.

    if (message.contains("elon") || message.contains("musk")) {
        upload(chat.privateKey)
    }
    
  • source
  • parent
  • [–] 9 points 1 year ago

    Yet? More like never.

  • source
  • [–] 9 points 1 year ago

    probably??? try definitely and ever

  • source
  • [–] 8 points 1 year ago
    [–] 8 points 1 year ago

    I wouldnt trust X with a picture of my shoes

  • source
  • [–] 7 points 1 year ago (1 child)

    Signal and encrypted email only.

  • source
  • hideshow 1 child comment
  • [–] 4 points 1 year ago

    Friends and I swapped our group chat to Signal the day Trump was inaugurated...the first time.

    If things keep going the way they are, no one should be communicating on anything but encrypted messaging apps.

  • source
  • parent
  • "The guy who helped install Donald Trump, did a Nazi Salute at Trump's victory parade on live TV, supports authoritarians, and who has declared war on transgender people to the point you're not allowed to say "Cis" or "Cisgender" on his platform, has created an end to end encrypted chat."

    All of this has the same vibes as the time Brigham Young University amended their code of conduct to allow people to come out as queer, let some students come out, and then changed the CoC back and expelled the students.

  • source
  • [–] 12 points 1 year ago (1 child)

    Our good friend Elon cannot be trusted? I don't believe you, this must be propaganda to discredit his good manners.

  • source
  • hideshow 1 child comment
  • [–] 6 points 1 year ago

    It’s like a regular encrypted chat but with peepholes and racism.

  • source
  • [–] 5 points 1 year ago

    shouldn't trust it yet shouldn't trust it ever

  • source
  • [–] 5 points 1 year ago

    Don't trust it at all.

  • source
  • [–] 4 points 1 year ago

    Cool, and I bet it will be just as trustworthy as WhatsApp (i.e. not at all).

  • source
  • [–] 3 points 1 year ago* (last edited 1 year ago)

    LOL nope. I'd use anything else. As in NO X.

    It was eyerolling back when that dickhead decided to sell blue checkmarks instead of issuing them only to verified celebrities.

  • source
  • [–] 2 points 1 year ago (1 child)

    I do really like E2EE but why do I need it in everything?

    If I want to talk to someone I would rather them message me on Signal or something that I trust more.

  • source
  • hideshow 1 child comment
  • [–] 2 points 1 year ago*

    Probably shouldn't? How 'bout definitely shouldn't, ditto for Twitter in general. Give ATproto shit all you want but at least you can move to an independent PDS with it.

    Granted ActivityPub is still ideal over ATproto, but both are better than a centralized black box.

  • source
  • [–] 1 point 1 year ago

    Ah, new ways for Kegsbreath to expose his idiocy.

  • source
  • [–] -2 points 1 year ago (3 children)

    I trust it but there is a major misunderstanding of end to end encryption. Some implementations the platform holder does not have a key to decrypt data but it is far from a requirement. All end to end means is there's a blocker preventing the network from seeing what you send not twitter who im assuming has a copy of the key.

  • source
  • hideshow 3 child comments
  • [–] 6 points 1 year ago (2 children)

    That is NOT end to end encryption. That is transport layer encryption. So basically SSL

    End to end is from sender to recipient. No one in the middle should be able to read anything

  • source
  • parent
  • hideshow 2 child comments
  • [–] -3 points 1 year ago (1 child)

    It's like ssl but done at the application layer. Nobody in the middle can read it except it's nobody in the middle of you and twitter and twitter and the recipient. If you put something on a platform and they have the key they will always be able to read it if they want to.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 2 points 1 year ago

    are you being dense on purpose? sender and recipient are both users. never twitter.

    i'm not saying that twitter's messaging is secure. the opposite. what i'm saying is that you statement of "All end to end means is there’s a blocker preventing the network from seeing what you send not twitter" is 100% objectively wrong.

  • source
  • parent