We tested the end-to-end encrypted messenger’s new feature aimed at addressing critics’ most persistent complaint. Here’s how it works.
  • kbal@kbin.melroy.org
    48·
    9 months ago

    Whittaker says that, for better or worse, a phone number remains a necessary requisite

    Worse. It is for the worse. We sure did wait a long time for this half measure, Signal.

    • LWD@lemm.ee
      36·
      9 months ago

      Is there a reasonable alternative, though? Email addresses? Adding a cryptographic challenge to prevent somebody from generating tons of accounts?

      As far as phone numbers go, I’m not a big fan of Signal having them, but I definitely prefer not having to give them out! That change is a huge deal to me, as I can now communicate with people without handing them a phone number. And Signal has provided their client and server source code, along with evidence that their servers store absolutely nothing.

      Nowadays, the most likely way your Signal data will get leaked is if somebody screenshots it.

      • kixik@lemmy.ml
        6·
        9 months ago

        Jami doesn’t require a phone number, which is p2p. Xmpp (+ Omemo) doesn’t require a phone number and it’s federated… I mean, if a service is willing to rid of phone numbers, it’ll do totally without them.

      • RandoCalrandian@kbin.social
        4·
        9 months ago

        It’s the signal metadata that they want to keep associated with an identity

        They still can fulfill government requests for who is talking to who and how often

        • LWD@lemm.ee
          9·
          9 months ago

          Only the recipient number has been in the messages, so unless Signal servers have been compromised, and they’ve figured out how to associate sender IP addresses with phone numbers, and they’ve never been caught by the multiple government demands from them… I think it’s fair to say

          1. they probably don’t keep these logs, and
          2. they made it about as hard as possible to do
        • cjf@feddit.ukEnglish
          7·
          9 months ago

          Got proof for that last claim?

          I thought their sealed sender feature was meant to prevent exactly this scenario.

      • Որբունի@jlai.lu
        3·
        9 months ago

        The challenge of having your device solve a nasty PoW that takes minutes would not deter most people: a timer once is better than evil captchas, phone numbers, etc. I don’t understand why they use hCaptcha and not that.

        • LWD@lemm.ee
          10·
          9 months ago

          If one computer can create a single spam account every few minutes, imagine how many total spam accounts could be created by a small farm of computers, in a single day.

          • Որբունի@jlai.lu
            1·
            9 months ago

            A lot, but farming phone numbers from poor countries is also cheap and Signal sends them insanely expensive SMS. There is no perfect solution, spammers aren’t stupid. Since Signal is centralised they can enforce PoW incrementally if they get reports for spam, I still think it is way better than hCaptcha which is garbage.

      • Scolding0513@sh.itjust.works
        1·
        9 months ago

        phone numbers for spam prevention are a bandaid for a mediocre solution. the mediocre aspect being that it’s totally centralized when it should at least be federated like SimpleX. SimpleX is the ultimate solution to be honest with you, it’s federated, fast, extremely private and extremely secure.

        • LWD@lemm.ee
          1·
          9 months ago

          I don’t mind SimpleX. I use it myself, sometimes. But it is also currently a very obscure service, and it’s not exactly easy to find people to communicate with. Phone numbers, and now usernames, provide a jumping-off point for that.

          How do you find people on SimpleX and then make sure you’re talking to them in a group later on? Right now, that’s really hard.

          And right now, SimpleX is pretty small, so if it starts expanding in that first area, how would it prevent spam?