[ art / civ / cult / cyb / diy / drg / feels / layer / lit / λ / q / r / sci / sec / tech / w / zzz ] archive provided by lainchan.jp

lainchan archive - /sec/ - 1742



File: 1475568932035.png (1.31 MB, 300x214, _0033_Layer-17.png)

No.1742

What do you lains use for encrypted synchronous communication? I've looked into Telegram, Tox, and Wickr, but they all seem like soykaf. Jitsi is encrypted with OTR, but the UI/UX is so bad you're never going to get any unlains to use it. I really enjoy Discord, but building an OTR plugin seems like it would be difficult.

Is there fuarrrking anything out there that is both secure and stylish? Even the former would do in a bind I guess.

  No.1743

>>1742
WhatsApp without Facebook bullsoykaf (you can turn it off) is pretty fuarrrking good. UX rocks and everything is end-to-end encrypted. And the best thing is that literally everybody uses it.

  No.1745

https://coy.im/about/

CoyIM seems ok to me but I haven't used it since no-fuarrrking-body that I know gives two soykafs about having some modicum of privacy. I also have no idea on the UI if you want non-cli because they have, like, zero screenshots on their site or GitHub repo from what I can tell...

So, not much help but it's something to consider.

  No.1746

>>1745
Not OP, but that looks cool. Thanks for sharing

  No.1747

Conversations.im + OMEMO master race

  No.1751

whats wrong with tox?

  No.1757

ricochet!

metadata-free, real-time, onion-routed instant-messaging!

it's nice because there's no registration at all so you can just create as many disposable contacts as you want

i've got me and a dozen friends using ricochet as of now.

check it out at ricochet.im

  No.1758

>>1747
>>1745
i'll put my hat in for these two as well, conversations.im is the bee's knees!

but once matrix.org get's their OLM protocol specced out, i'm dropping xmpp like a ton of bricks and switching over to matrix fulltime.

and beware, coy is still in early beta, and afaik has not been audited and isn't in any distribution's repos afaik

ricochet has been audited, is packaged in the debian repos, and has installers for macos and windows.

  No.1761

>>1751
I'm not the OP, but it is dead.

  No.1763

File: 1475681784711.png (67.67 KB, 200x200, 1445722183259.jpg)


  No.1766

Signal is a very good messaging system. Pretty much idiot proof. When Whatsapp added crypto they (at least claim to) use the protocol designed for signal. The advantage of signal is that it is free (as in freedom) software.

Unfortunately you have to have a google-compromised phone or an iphone to use signal because of Moxie's autism. Once you have a phone set up then you can use the chromium app on a desktop.

  No.1767

>>1745
Yes, CoyIM is a GUI frontend built on https://github.com/agl/xmpp-client, same guy behind pond.
If you don't care for the clicky clicky, and need a golang based xmpp client which is no longer in alpha, with baked in tor support + apparmor pcf go with latter.

>>1757
FUARRK YES! use s/signal/ricochet, use tor

  No.1771

>>1742
I've used signal before. It worked well enough for me and is extremely easy to use. You can even set it to be used as default for text messaging on your phone.

  No.1773

Signal is solid as fuarrrk. I've been using Wickr also but i don't really like it.

  No.1776

While I haven't looked into it, just changing jitsi's UI should be rather simple, no?

  No.1778

The problem that I have with signal is that you need a phone number to register a device with it. I have a smartphone without cell service, and there seems to be no way at all to use signal with it.

  No.1779

Signal is best, imo.

  No.1781

Some friends and I started using signal once we began doing some shady soykaf. It's worked well and there's some records of what they've exposed when issued a subpoena: https://whispersystems.org/bigbrother/eastern-virginia-grand-jury/

  No.1786

File: 1475737101070.png (1.19 MB, 200x143, _0027_Layer-123.png)

A few half-decent solutions have been offered, but none of them seem to be what I'm seeking. I'm not looking for the perfect program; I'm just looking for something that is:

a) cryptographically sound and trusted
b) a UI that isn't vomit inducing (and non-lains can fuarrrking understand and use)
c) available on desktop environments

Given that this doesn't seem to exist, how difficult would it be to build Lainchat? Could we use existing encryption libraries?

I'm asking because I'm not a programmer (yet). Please don't taze me bro.

  No.1787

>>1778
There are plenty ways of doing it, from getting 5 euro sim without ID (if applicable to your coutry, if not you could always ask someone else who lives abroad), to using free web SMS receivers to register. I did not test it with Signal directly but worked multiple times with WhatsApp and Viber.

  No.1789

>>1787

Those numbers eventually get recycled and then you can lose access to the service if someone reregisters it. There are some SIP services that allow free incoming calls which can be an alternative method of activation vs sms.

Signal is very secure, but it's not built for anonymity and anyone who tries to use it that way will be disappointed.

As someone without a cellphone number or Gapps it's a non-starter.

  No.1792

>>1786
I don't want to sound like a broken record, but Signal checks all your marks man

  No.1798

>>1786
if you want to converse with nonlains, 1792 nailed it to a t.
and if you're not a programmer and are contemplating this as your first project, do not; getting all three of your requirements correct, especially the 1st, requires years of dedication. wet your feet with smth less huuuge and then gradually move up to riding the waves.

  No.1816

>>1792
>>1798
Actually, it doesn't. According to the lains, A is questionable and C is a definitive no so the search is still on.

As for programming, I definitely have other, less challenging projects I'm interested in doing first, but I was still curious about the the difficulty because I have friends that I might be able to convince to help with something like that.

  No.1821

>>1816
do you mind being more specific then, seeing as signal generally checks all the boxes/is the best currently available when you consider;
a) Moxie is well respected within the cryptography field and is one of the main heads responsible for the Signal protocol, now relied on and trusted by a handful of others (fb, ggl etc.).
b) UI is generally intuitive and clean.
c) Signal-desktop is a thing, yet requires pairing with another device running android/iOS.

project wise, if you pool efforts with a handful of others and focus on areas in which each of you excel at, this is a more realistic endeavor.
oh cool, mind saying what those other smaller projects may be?

  No.1822

Anything by OWS gives me the creeps, criticism is shot down by hiveminds and it has at least one full time wage slave.

https://reddit.com/u/uph

  No.1841

>>1822
I don't know how what I'm looking for doesn't exist in 2016. I seriously can't wrap my mind around it. How is this a novel idea?

  No.1843

>>1821
I'm not sure if enough lains would be interested in the project to see its completion. Do you think there is a chance?

As for my other projects, I don't really want to discuss them here. If you're on the IRC, perhaps we can talk more there.

  No.1853

>>1821
Its not out of the realm of possibility, but really why do you want to create just another thin client for IRC?
After all thats what lainchat is, within the browser aw.
Btw djb's nacl is the golden standard as far it goes for crypto libs, so you should get comfortable with that.
Understood, I'll swing by somepoint in the week.

  No.1916

Tox will be the messenger. It isn't now, but it will eventually.
But that takes work, now I can't contribute and neither can you as it seems. But design-wise (as in what they're aiming for) Tox blows the competition out of the water in my eyes. Also IIRC Ricochet uses some of Tox's code. Ricochet added integrated anonymity, but at the loss of file sharing and group chat. But Tox can be anonymized with OSes like Whonix so I'll let you weigh that situation. Maybe they'll merge or one will be absorbed in the future.

Upcoming features they (Tox) are working on are upgrades of present functionality, better documentation, synchronization, slimming of toxcore & lighter resource-use as they're entering the mobile platform and eventually a code-audit.

In the meantime I think I'd recommend OMEMO'd XMPP and/or Matrix protocol with Olm (any day now...), routed through Tor of course.

  No.1918

>>1853
What about a desktop client that uses a secure XMPP server and forces the use of OTR? Would that be the easiest means of achieving something that is lightweight and secure? Existing clients that have these are absolutely hideous and clunky and abandoned long-ago.

>>1916
Didn't the original developer of Tox leave the project because the development started heading in a direction they believe is inherently insecure? I remember reading something about that before.

  No.1919

Signal depends on Google Play Services since it uses Google Cloud Messaging to talk to Open Whisper System's central server (which you need to personally trust). It's de facto lead developer, Moxie Marlinspike, is bemusingly hostile to any attempt to remove this unwanted dependency, closing a pull request for WebSockets support for little or no justification, attacking those who are pushing for this, and threatening the LibreSignal project with a ban from Open Whisper System's service with the additional mention of legal proceedings. I can't personally recommend Signal to anyone because I can't use it. No matter to me because i've got no friends haha XD
>>1841
There's a reason there are so many secure messaging efforts, because somebody thinks "hey that system is like kicking dead whales down the beach" and instead of trying to change it or fix it they go and roll their own. Sometimes, though, this is required since the architecture and approach is different, say between distribution using DHT (Tox) and centralisation with a third party server you need to trust (Signal).
>>1853
>djb's NaCl is the gold standard
libsodium is the new hotness with API compatibility with NaCl as well as an improved build system and portability. https://github.com/jedisct1/libsodium
>>1916
Quote from the Antox developer on IRC (paraphrasing slightly): "Antox is a really well designed app. I don't use it though I like battery life LOL"
This is Tox's biggest failure in my opinion; you can't have a joined up identity on your phone and your laptop, for example, and the distributed model is just a plain bad idea for running on a mobile phone, since you need to be listening and jostling for peers all the time, which has it's impact on battery life and data usage.
>>1918
Tox had a major soykafstorm when one of the leaders involved in the project nicked the Tox Foundation charity's dosh and used it to fund his own little jolly. Their split amounted to a new website being registered.

  No.1924

File: 1476540087319.png (1.77 KB, 150x80, zom1.png)

There is a new mobile XMPP Client called Zom:
https://zom.im/

It is a merge of Conversations and ChatSecure and focuses on easy usage:
https://chatsecure.org/blog/chatsecure-conversations-zom/

It features encrypted group chats, voice messages and file sharing.

  No.1927

It increasingly looks like the solution is going to require the use of a phone, which I absolutely abhor. God, why?

  No.1944

>>1927
>It increasingly looks like the solution is going to require the use of a phone, which I absolutely abhor. God, why?
Majority of the population doesn't cares for laptops/desktops when you're mobile does it all.

  No.1945

>>1927
XMPP clients are getting a following. Like Conversations and that Zom.im thing. Those XMPP servers can also be used on the desktop with Pidgin, Empathy or another XMPP client.

  No.1949

XMPP should be your choice. Signal is good but it require your mobile phone number, which probably require your ID card in some countries (at least in my country)

  No.2018

>>1919
The best way to fix tox would be to allow it to work in a client-server model. I should be able to have my server handle all the in and out while my mobile just receives push and pushes up to it.

UX is a different story.

  No.2029

Matrix.org was brought up in the other related thread about encrypted group chat. Do you lains trust it? Does anyone use it?

  No.2406

>>1758
matrix is looking really amazing and can pretty much keep up with stuff like discord it seems like.

  No.2409

>>1743
Telegram is a better option overall, but Whatsapp is owned by Marky Mark, so I don't trust them at all, still if you're looking for mobile options, the truth is that probably there are none, since most at least require a phone number.

  No.2415

>>1787
> to using free web SMS receivers to register.

This is extremely bad idea, as anyone can mitm you. In case of signal it will show your contacts the key has changed, in case of whatsapp or tg, anyone can silently hijack your account - those dont do 2fa, anyone can reset your password via that public gateway SMS.

That said, apps using SMS as sole authentication factor is no-go privacy regardless, but it is convenient for most normies so not much you can do about it.

As odd as it sounds, fb is still best out there - most normies are there, e2e (albeit you need to enter that mode explicitly), works with seven proxies, no need to send your dox and authentication is simple password.

  No.2430


  No.2443

>>1743
they'll still track and then sell your metadata.

  No.2497

>>1766
Silence is an alternative to Signal built from the same codebase. It works over GSM so no mobile data/Wi-Fi required, though it will cost you text rates if that is a concern.
It's available on F-Droid if you would rather not have Botnet-Play Services on your phone.

  No.2515

>>2497

Silence is not an alternative to Signal. It only works over SMS, which Signal doesn't support anymore. They're completely different and non-interoperable.

  No.2605

So the final word is that you can't have crypto without doxxing yourself (e.g., Signal)?

  No.2607

>>2605
Pretty much.

  No.2608

conversations for android
xmpp-client by adam langley for pc

I don't get why more people dont use xmpp, but then again I guess they do since thats what facebook chat is based off of

  No.2980

Why?

  No.3084

RIP crypto.

  No.3087

cryptocat seems good
https://crypto.cat/
or you might like this
https://crypto.dog/

  No.3088

>>3087
I deleted it !!!

  No.3099

>>1761
I know I'm replying to a two-month old post, but it seems like very few people know that Tox development has moved from irungen2's repository. Most people are contributing to the TokTok repo, which just had a stable release 11 days ago. Tox is pretty far from dead.

  No.3126

>>3099
Link lain?

  No.3159

Wire is what I use. It's not perfect and it's centralized, but it's open source, uses end-to-end encryption, has many features, and is swiss-based. It also looks really nice and has a modern design. The only other privacy-focused messenger I've found with a nice design is Telegram. A program's look is important if you want family and friends to switch from their programs to something like Wire. Most people unfortunately care more about a program's features and design than it's security and encryption.

  No.3205

File: 1483228654616.png (149.24 KB, 200x143, lainchat.jpg)

strong e2e crypto open source lainchat when?

  No.3210

>>3205
soon :)

  No.3242

I see the argument against Signal based on requiring Google Play Services, but is it still a good tradeoff? Is a libre, end-to-end, cross-device system worth the information that Google Play Services can send to Google?

If would say yes, but even if you don't think the tradeoff is worth it now, keep in mind Moxie has said he *wants* to remove dependencies on Google, as soon as the crash analytics functionality can be replicated. It might be worth supporting the project now, hedging your bets that the Play Services dependency will be removed in the future.

  No.3243

>>3242
fyi moxie has renewed efforts to go full websocket :D

https://github.com/WhisperSystems/Signal-Android/pull/5962


s/o to eightbitkid!

  No.3339

>>3159
Telegram is unaudited.
Wire I don't know the current state, but they used variable bitrate for voip so that's a no go, but i think their text encryption was fine.

Personally pretty much solely using https://matrix.org currently. Irc bridges are pretty great and even a telegram matrix is in the works.

  No.3340

>>3242
There is still fake gapps. And the copperhead devs have noise, which is a signal fork without gapps. Though i think it's not completely free.

  No.3447

File: 1483925901287.png (282.16 KB, 200x151, cultural_appropriation.png)

>>1742
It seeems like the only thing with potential beyond meme/fad is XMPP+OTR.

What does everyone think of this?

  No.3454

>>3447
IRC+OTR and email+PGP is also good.

  No.3510

>>3447

I use VPN+TOR+XMPP+OTR. For text there's no major speed issues.

For email, hidden services + PGP.

Send files on free hosting or as attachments encrypted with GPG.

For idle chitchat I'll use whatever other people want to use, just with a fake identity and adequate access obfuscation, usually just a VPN.

  No.3511

https://medium.com/@RiotChat/exciting-new-riot-release-get-ready-for-chatting-securely-acc93ecfe0a#.x00vuajf3
Riot is GREAT, and you can host your own server. In my experience it works as well as slack without some of the integrations.

  No.3563

OTR is obsolete.
Embrace OMEMO

  No.3593

>>3563

If anything except a single Android app had stable and complete OMEMO support, I would do so. But every OMEMO-capable desktop client I've tried is horribly broken.

Especially Gajim on a lot of systems. Instant kernel panic because it's trying to read /proc and a whole bunch of other files and directories to gather entropy (https://bugs.launchpad.net/ubuntu/+source/gajim/+bug/1452006) but it's got a really fuarrrked-up way of doing so, and it's a known bug that's almost 2 years old and remains unfixed. Gajim's developers blame broken drivers, but that's pass-the-buck bullsoykaf.

  No.3594

>>3511
>Riot is GREAT

Except that it slaughters battery life on android.

  No.3618

I've been using XMPP + OTR for quite a while now. When I need to send someone an encrypted e-mail, I usually send them an email. Same with files.

>>3563
I would do that If I actually knew someone who used it.
Wanna try it out? swedish [] creep.im

  No.3683

Signal is incredible. Not only does it work fantastic and let you edit most options you want and run smooth on a phone, but you can easily convince friends to use it as well. At least 15 of my contact list use it now, and numbers went up after Nov 8th

  No.3685

>>3683
I love Signal and wish everyone would use it. The only problem is that it can't send SMS/MMS on iPhone, which means people would have to use iMessage for normal text AND Signal for Signal users. Which is too incoveniant for basically everyone.

  No.3708

>>2430
resetthenet recommends cryptocat.

cryptocat has been found to be using cryptographically insecure methods and there are programs out that can decrypt cryptocat communications.

Do not trust that website since it is likely a honeypot.

  No.3709

If there is a centralized server being used for communications do not use it. I've noticed most of the mentioned apps here do. You should all be ashamed.

DHT, self made VPN or nothing.

  No.3770

>resetthenet

Just because they're an activist organization doesn't mean they know soykaf about data security

  No.3775

>>3709
thing is, it's nearly impossible to actually set up a network without at least one server to bootstrap off of.

  No.3817


  No.3833

>>1781

wrong app for that. Signal is for non-shady stuff, just in case something sensitive leaks out over day to day communications. Which sensitive things generally do.

Signal uses a centralized third party server owned by someone else, so there is no guarantee it will be safe or unmonitored.

  No.3834

>>3775

No it is possible, because any singular node can populate any other. And it can happen assymetrically. That is if node A has peers[B, C, D] and node E has nodes [F, G, H], and node I needs to boot strap, Strapping into A or E will bootstrap you into the entire 'Graph' so long as atleast any one of the nodes knows of any other. You just need a common enough nexus.

In the context of security, it would be relatively easy to write a wrapper script in python that uses Tor to get a list of peers that can segway you into any other part of the graph. So you're idea that DHT requires a centralized server is only partially true in the sense that theoretically if all nodes are not aware of any single other node, then the protocol by itself could fail. But that would assume complete control of all network segments.

Not to mention dht can work over private subnets (ipfs at least), which brings VPN's or any other tunneling/virtual network scheme.

I would say it's nearly impossible not to bootstrap off of a centralized server in the case of dht.

  No.3838

>>3834
but where do new nodes go to get a peer list? There has to be a centralized list somewhere, hosted on some server(ultimatetely).

  No.3845

>>1743
whatsapp end to end encryption is useless.

  No.3850

>>3838
There's a litany of seed servers, but even barring that, all you have to do insert one peer manually (or programmatically) and you've populated the rest of the swarm.

The bootstrap list is large and you can export the list using standard shell commands such as

ipfs bootstrap list >save

Then save that as a backup or add new peers to this list. [1]

Here reference this: https://w1r3.net/zGdUdT.txt

it's the output of my swarm. you can use this to bootstrap later. If anyone single node is up when you need it. You'll repopulate your list.

The entire list is /not/ centralized. Again Graph theory. As long as one node has a path to any other node in any capacity (route) you can reach it or learn of it. The bootstrap servers *might* have the entire peerlist. But not necessarily. Especially clients that use their own bootstrap servers (specifically different from the default). But even then should they request a resource (hash) from a node they dont have from the list, they'll learn of that new node through the working of the protocol (and the general nature of how dht works).

Really what it comes down to is you reading up on dht. [2] In particular read about how magnet links work.

REF:
[1]: https://github.com/ipfs/examples/tree/master/examples/bootstrap
[2]: https://www.neowin.net/forum/topic/846692-how-does-dht-work-in-laymens-terms/

  No.3851

>>3850

I though up of a better way to explain it. Think about degrees of separation in terms of human contact. You're connected to any human on earth by virtue of any other human (at least the probability is exceedingly likely). So lets take phonebooks on cellphones and limit ourselves to people that own cellphones. DHT works like this. You (Alice) contact Bob, and he gives you his phone book. Then you look up tha tphone book and can look up even more phone books from those contacts. Eventually if you continue this process long enough you'll come to know every single other person that's connected to atleast one other individual other than themselves.

  No.3854

>>3594
>>Riot is GREAT

>Except that it slaughters battery life on android.

You can adjust the sync delays, helps alot

  No.3855

>>3845
Source? At least give a reason.

  No.3860

>>3855
A story broke a little bit ago how facebook put a backdoor in.

Basically you should never, ever trust closed-source crypto. It will not protect you when soykaf hits the fan.

  No.3861

>>3860
That article is based on a post from early 2016 and I don't know if I'd call it a backdoor. Whatsapp just favors usability more. Not that i think it's good how they do stuff calling it a backdoor is just a clickbait imo.

  No.3992

>>3511
>>3339
I am working on a matrix client atm, gitla.in/f0x52/neo. It will use most of the ui design ideas of Telegram, and use PGP for encryption, since it is easy and secure.
It's still in progress (not even functional yet) but I thought it might be a good idea. I'm currently figuring out the matrix api as I'm going, because it isn't that well documented

  No.4123

File: 1486912436185.png (23.64 KB, 200x143, ClipboardImage.png)

>>3992
L A I N C H A T I S C O M I N G !

  No.4124

>>3992
Why not use matrix encryption ?

  No.4128

>>4124
The matrix encryption is in like alpha, and hasn't had an audit yet (I think). So I would rather use trusted and proven algorithms.

  No.4129

>>3992
I also decided that my current code was horrible (python with webkit-gtk, using javasoykaf for most things). So I am starting this project over with C++. I did the first commits today at https://gitla.in/f0x52/neo-qt

  No.4130

Why not make it cli ? People can develop frontends if they want.

  No.4134

>>4130
That is actually a very good idea. I'll start working on it soon!

  No.4149

>>1767
>>1771
>>1773
>>1779
>>1781
>>1789
>>1766
>>1792
>>1821
>>2605
>>3683
>>3685
>Signal
Please no. All lainons should know much better than to trade off their own phone number while running google services for the placebo effect of privacy. And yea, its essentially placebo because it runs on third party severs, where literally anything can be seen by an intruder. If you even remotley care about privacy you should not hand it over.
>inb4 but Snowden!
Snowden is a relativley smart guy who made a stupid comment. I don't understand why he is vouching for them when it has properties that he advocate(s/d) against continuously (mass 3rd party collection of data and forking over personal info).

  No.4157

>>4149
> literally anything can be seen by an intruder.

Care to elaborate on this? Extraordinary claims require extraordinary evidence.

  No.4158

>>4157
Some moron literally walks in to their server room and walks out with a copy of hundreds of thousands, if not millions, of chat logs. This is under the presumption that they are storing the chats unencrypted. Even if they were storing them while encrypted, most people in the position of the intruder are either all ready high up enough internally to know their own encryption's flaws and how to decrypt or they are smart enough or have a powerful group backing them that will figure it out.

  No.4159

>>4158

We're all wearing tinfoil hats in here, save some for the rest of us will you? You've gone through almost the entire box.

  No.4164

The biggest issue with actual secure chat clients is that (ironically) normals don't trust them. I don't know enough people that would ever use any secure client, as much as I'd like it.

  No.4171

>>4149
>here literally anything can be seen by an intruder.

it's end-to-end encrypted with state-of-the-art key exchange, dumbass.

  No.4172

>>4171
Nice job regurgitating whisper system's marketing.

Signal is of a centralized design and the server end is closed source. To make matters worse the initial key exchange passes both of your keys right through their servers. That other lain has a right to be suspicious.

  No.4207

File: 1487349307820.png (74.81 KB, 200x68, 1280px-Ring_VOIP_logo.svg.png)

Has anyone looked into GNU Ring?
https://ring.cx/
https://ring.cx/en/about/technical
https://wiki.archlinux.org/index.php/Ring
>an official GNU package
>based on the DHT, distributed network technology, avoiding the use of servers
>peer-to-peer discovery and connection
>A global account identification mechanism based on blockchain implementation on an Ethereum technology.
>Identity managed by X.509 certificates
>use of well defined protocols (SIP, DHT Kademlia, TLS, etc.), methods (continuous integration, code reviews, etc.), portable languages (C/C ++ 11)
>multi-device support
>GNU/Linux, macOS, Windows, and Android support
>IPv6 support

  No.4209

>>4207
Looks effective. Will ride it if my mom manages to proceed the installation&setup procedure.

  No.4213

>>4172
Nigga you can't have encrypted communications without exchanging public keys. Are you complaining about the fundamental laws of mathematics now?

  No.4214

>>4213
if you send keys through a centralized server then.... they server can replace them with their own. Which is why key transmission is the most dangerous part of a PKI protocol. Also why keysigning parties are big IRL.

  No.4219

>>4214
And...? Signal has a means to verify out-of-band too. In the meantime, all your secentralized services using Kademlia DHT still rely on a bootstrapping node, which could MITM you just as easily as OWS's servers could.

  No.4228

>>4219
>decentralized services using Kademlia DHT still rely on a bootstrapping node, which could MITM you just as easily as OWS's servers could.
Initial key exchange does not occur through the DHT.

  No.4252

>>4228
>be bootstrapping DHT node
>target connects to the network
>they ask for the nearest nodes
>give them a list of compromised nodes
>if they ask for a particular node, do both a real search and also return a node which MITM the connection
>success

  No.4267

>>4252
Malicious nodes could harm availability, but how would they compromise the confidentiality or authenticity of a message?

  No.4540

For mobile communication I mostly use Signal (which I can't recommend enough) or WhatsApp (since my family is using it).

On the desktop I usally use XMPP with OTR if I want to communicate encrypted.

I don't think writing another piece of software and a new chat system/client etc would be much of help right now. It's more important to spread the word about secure, existing tools & try to improve them/support them.

  No.4541

>>4540
I totally agree. XMPP with OTR or OMEMO is the best option. Creating new, non interoperable, protocolls that bring no groundbreaking benefits will only harm the goal of reaching the general public.

  No.4748

>>4149

it uses FCM to deliver push messages because google has infinite money and it's one of the only decent infrastructures available.

If you've ever tried to use xmpp with otr you (especially across two servers or intercontinental) you know what i'm talking about.

not to mention the battery drain from doing websockets on a phone

The key exchange isnt just for one key either. axolotl uses DH with KDF to maintain short term session keys and signal comes equipped with safety numbers for verification purposes since It's built on back of otr; and it has even better built in deniability than otr because of the handshaking method with KDF.

You can build it yourself so you don't have to verify. You can also verify with any valid voip number so once again your arguments hold no clout.

  No.4749

>>3860
you mean the sensationalist guardian story that was a bunch of bullsoykaf ?

It wasn't a back door. So many infosec pros have commented on that.

  No.4753

>>4749
It might not be a backdoor, but it can be used as one if wanted.

  No.4763

>>1742
Threema. i know, not open source, but they have features services other messengers don't have and they got reviewed.

  No.4765

>>4753
just check the security keys youy baka

  No.4783

>>3992
Update on neo, I am almost done with basic functions, you can login and receive raw json when an update happens. Basic send/receive will probably be done later this week