Comment on Head of the Signal app threatens to withdraw from Europe
vacuumflower@lemmy.sdf.org 3 weeks ago
About freedom, not freedom and various other things - might want to extend the common logic of gun laws to the remaining part of the human societies’ dynamics.
Signal is scary in the sense that it’s a system based on cryptography. Cryptography is a reinforcement, not a basis, if we are not discussing a file encryption tool. And it’s centralized as a service and as a project. It’s not a standard, it’s an application.
It can be compared to a gun - being able to own one is more free, but in the real world that freedom affects different people differently, and makes some freer than the other.
Again, Signal is a system based on cryptography most people don’t understand. Why would there not be a backdoor? Those things that its developers call a threat to rapid reaction to new vulnerabilities and practical threats - these things are to the same extent a threat against monoculture of implementations and algorithms, which allows backdoors in both.
It is a good tool for people whom its owners will never be interested to hurt - by using that backdoor in the open most people are not qualified to find, or by pushing a personalized update with a simpler backdoor, or by blocking their user account at the right moment in time.
It’s a bad tool even for them, if we account for false sense of security of people, who run Signal on their iOS and Android phones, or PCs under popular OSes, and also I distinctly remember how Signal was one of the applications that motivated me to get an Android device. Among weird people who didn’t have one then (around 2014) I might be even weirder, but if not, this seems to be a tool of soft pressure to turn to compromised suppliers.
Signal discourages alternative implementations, Signal doesn’t have a modular standard, and Signal doesn’t want federation. In my personal humble opinion this means that Signal has their own agenda which can only work in monoculture. Fuck that.
that’s a lot of words to say you generally accuse any programm that isn’t federated of having an agenda targeted at its userbase.
And lots of social woo-woo that doesn’t extend much further than “people don’t understand cryptography and think it’s therefore scary”.
A pretty weird post, and one which I don’t support any statement from because I think you’re wrong.
that’s a lot of words to say you generally accuse any programm that isn’t federated of having an agenda targeted at its userbase
No, that’s not what I’m saying. I used the word monoculture, it’s pretty good.
And lots of social woo-woo that doesn’t extend much further than “people don’t understand cryptography and think it’s therefore scary”.
Not that. Rather “people don’t understand cryptography, but still rely upon it when they shouldn’t”.
A pretty weird post, and one which I don’t support any statement from because I think you’re wrong.
I mean, you’ve misread those two you thought you understood.
Using mono ulture as a word doesn’t change the meaning here. If anything, its a pathway for the foal you ascribe.
I do give you credit about the second part - it would be better to have your own private key in chat apps, which isn’t handled by the app itself, at the very least to establish a shared key. I still think the existence of crypto is a massive boon to many, even in a “flawed” implementation with the “control” being on the side of corporations - tho if they are smart, they’d never store the keys themselves, not even hashes. Unless you’re part of the signal project, I doubt you know the exact implementation and storage of data they do.
Still, thanks for summarising your lengthy post, even if I had to bait you into it. Sometimes, brevity is key.
Using mono ulture as a word doesn’t change the meaning here. If anything, its a pathway for the foal you ascribe.
Of course it does. Federation can be a monoculture too (as it is with plants). A bunch of centralized (technically federated in IRC’s case, but united) services, like with IRC, can be not a monoculture.
Monoculture is important because one virus (of conspiratorial nature, like backdoors and architectures with planned life cycle, like what I suspect of the Internet, or of natural one, like Skype’s downfall due to its P2P model not functioning in the world of mobile devices, or of political and organizational one, like with XMPP’s standards chaos and sabotage by Google) can kill it. In the real world different organisms have sexual procreation, as one variant, recombining their genome parts into new combinations. That existed with e-mail when it worked over a few different networks and situations and protocols, and with Fidonet and Usenet, with gateways between these. That wasn’t a monoculture.
Old Skype unfortunately was a monoculture. Its clients for Linux (QT) and Windows and mobile things were different implementations technically, but with the same creators and one network and set of protocols in practice.
I still think the existence of crypto is a massive boon to many
That’s the problem, it’s not. You should factor psychology in. People write things over encrypted channels that they wouldn’t over plaintext channels. That means it’s not just comparison of encrypted versus plain, other things equal.
even in a “flawed” implementation with the “control” being on the side of corporations - tho if they are smart, they’d never store the keys themselves, not even hashes.
And that’s another problem, no. Crooks only steal your money, and they have adjusted for encryption anyway. They are also warning you of the danger, for that financial incentive. Like wolves killing sick animals. The state and the corporation - they don’t steal your money, they are fine with just collecting everything there is and predicting your every step, and there will be only one moment with no warning then you will regret. That moment will be one and the same for many people.
Unless you’re part of the signal project, I doubt you know the exact implementation and storage of data they do.
What matters is that the core of their system is a complex thing that is magic for most people. You don’t need to look any further.
I think you may need some sleep man. wtf are you talking about
Perhaps you need to get some sleep if you don’t understand what I’m talking about.
I get it messenger = gun wow i didnt know!
Holstering my phone now thanks
Unironically yes, communications (information and roads) were historically as important. Lenin’s call to “take post, telegraph, telephone stations, bridges and rail stations” kinda illustrates that.
What I meant is that abstractly having fully private and free communications is just as universally good as everyone having a drone army. In reality both have problems. The problems with weapons are obvious, the problems with communications in my analogy are not symmetric to that, but real still - it’s that people can be deceived and backdoors and traps exist. Signal is one service, application and cryptographic system, it shouldn’t be relied upon this easily.
It’s sometimes hard to to express things based only on someone with good experience telling them to me, making it an appeal to anonymous authority, but a person who participated in a project for a state security service once told me that in those services cryptography is never the basis of a system. It can only be a secondary part.
Also, other than backdoors and traps, imbalance exists. Security systems are tools for specific purposes, none are universal. 20 years ago anonymity and resilience and globalism (all those plethora of Kademlia-based and overlay routing applications, most of which are dead now) were more in fashion, and now privacy and political weight against legal bans (non-technical thing, like, say, the title of the article) are. The balance between these in popular systems determines which sides and powers lose and benefit from those being used by many people. In case of Signal the balance is such that we supposedly have absolute privacy and convenience (many devices, history), but anonymity, resilience and globalism are reduced to proverbial red buttons on Meredith Whittaker’s table.
lmmarsano@lemmynsfw.com 3 weeks ago
I don’t think you understand anything you wrote about. Signal is open source, is publicly audited by security researchers, and publishes its protocol, which has multiple implementations in other applications. Messages are encrypted end-to-end, so the only weaknesses are the endpoints: the sender or recipients.
Security researchers generally agree that backdoors introduce vulnerabilities that render security protocols unsound. Other than create opportunities for cybercriminals to exploit, they only serve to amplify the powers of the surveillance state to invade the privacy of individuals.
vacuumflower@lemmy.sdf.org 3 weeks ago
I don’t think you should comment on security if “open source” means anything to you in that regard. For finding backdoors binary disassembly is almost as easy or hard as looking in that “open source”. It’s very different for bugs introduced unintentionally, of course.
Also why the hell are you even saying this, have you looked at that source for long enough? If not, then what good it is for you? Magic?
I suppose you are an illustration to the joke about Raymond’s “enough eyeballs” quote, the joke is that people talking about “enough eyeballs” are not using their eyeballs for finding bugs\backdoors, they are using them and their hands for typing the “enough eyeballs” bullshit.
“Given enough good people with guns, all streets in a town are safe”. That’s how this reads for a sane person who has at least tried to question that idiotic narrative about “open source” being the magic pill.
Stallman’s ideology was completely different, sort of digital anarchism, and it has some good parts. But the “open source” thing - nah.
Exactly, and it’s not audited by you, because you for the life of you won’t understand WTF happens there.
Yes, it’s being audited by some security researchers out there, mostly American. If you don’t see the problem you are blind.
No, there are no multiple implementations of the same Signal thing. There are implementations of some mechanisms from Signal. Also have you considered that this is all fucking circus and having a steel gate in a flimsy wooden fence? Or fashion, if that’s easier to swallow.
Can you confidently describe what zero-knowledge means there, how is it achieved, why any specific part in the articles they’ve published matters? If you can’t, what’s the purpose of it being published, it’s like a schoolboy saying “but Linux is open, I can read the code and change it for my needs”, yeah lol.
Do security researches have to say anything on DARPA that funds many of them? That being an American military agency.
And on how that affects what they say and what they don’t say, what they highlight and what they pretend not to notice.
In particular, with a swarm of drones in the sky at some point, do you need to read someone’s messages, or is it enough to know that said someone connected to Signal servers 3 minutes ago from a very specific location and send one of those drones. Hypothetically.
Oh, the surveillance state will be fine in any case!
And cybercriminals we should all praise for showing us what the surveillance state would want to have hidden, to create the false notion of security and privacy. When cybercriminals didn’t yet lose the war to said surveillance state, every computer user knew not to store things too personal in digital form on a thing connected to the Internet. Now they expose everything, because they think if cybercriminals can no longer abuse them, neither can the surveillance state.
Do you use Facebook, with TLS till its services and nothing at all beyond that? Or Google - the same?
Now Signal gives you a feeling that at least what you say is hidden from the service. But can you verify that, maybe there’s a scientific work classified yet, possibly independently made in a few countries. This is a common thing with cryptography, scientific works on that are often state secret.
You are also using AES with NSA-provided s-boxes all the time.
I suggest you do some playing with cryptography in practice. Too few people do, while it’s very interesting and enlightening.
lmmarsano@lemmynsfw.com 3 weeks ago
Anyone can look at the source, brah, and security auditors do.
Are you in the dark ages? Beyond code review, there are all kinds of automations to catch vulnerabilities early in the development process, and static code analysis is one of the most powerful.
Analysts review the code & subject it to various security analyzers including those that inspect source code, analyze dependencies, check data flow, test dynamically at runtime.
Right, the protocol.
Stop right there: I don’t need to. It’s wide open for review by anyone in the public including independent security analysts who’ve reviewed the system & published their findings. That suffices.
They don’t. Again, anyone in the public including free agents can & do participate. The scholarly materials & training on this aren’t exactly secret.
Information security analysts aren’t exceptional people and analyzing that sort of system would be fairly unexceptional to them.
Even with state-level resources, it’s pretty well understood some mathematical problems underpinning cryptography are computationally beyond the reach of current hardware to solve in any reasonable amount of time. That cryptography is straightforward to implement by any competent programmer.
Legally obligating backdoors only limits true information security to criminals while compromising the security of everyone else.
vacuumflower@lemmy.sdf.org 3 weeks ago
In short - something “everyone being able to look upon” is not an argument. The real world analogies are landmines and drug dealers and snake oil.
You are not speaking from your own experience, because which problems are solved and which are not is not solely determined by hardware you have to do it by brute force. Obviously.
And nation states can and do pay researchers whose work is classified. And agencies like NSA do not, for example, provide reasoning for their recommended s-boxes formation process. For example.
Solving problems is sometimes done analytically, you know. Mostly that’s what’s called solving problems. If that yields some power benefits, that can be classified, you know. And kept as a state secret.
People putting those in are also not in the dark ages.
There are things which were wide open for review by anyone for thousands of years, yet we’ve gotten ICEs less than two centuries ago, and electricity, and so on. And in case of computers, you can make very sophisticated riddles.
So no, that doesn’t suffice.
Oh, denial.
There have been plenty of backdoors found in the open in big open source projects. I don’t see how this is different. I don’t see why you have to argue, is it some religion?
Have you been that free agent? Have you participated? How do you think, how many people check things they use? How often and how deeply?
Yes, but you seem to be claiming they have eagle eyes and owl wisdom to see and understand everything. As if all of mathematics were already invented.
It’s not about obligating someone. It’s about people not working for free, and those people working on free (for you) stuff might have put in backdoors which it’s very hard to find. Backdoors usually don’t have the “backdoor” writing on them.
Perhaps the reason they have so many resources is that they don’t miss opportunities, and they don’t miss opportunities because they have the resources.
0x0@lemmy.zip 3 weeks ago
You sound paranoid but it doesn’t mean you aren’t right, at least to some extent.
So what’s your solution for secure messaging?
vacuumflower@lemmy.sdf.org 3 weeks ago
Getting rid of monoculture via transports and cryptography being pluggable (meaning that the resulting system would be fit for sneakernet as well as for some kind of federated relays as well as something Kademlia-based, the point is that the common standard would describe the data structure, not transports and verification and protection).