Hacker News new | past | comments | ask | show | jobs | submit login
Darpa wants to make a better, more secure, version of WhatsApp (trustedreviews.com)
106 points by hhs on April 15, 2019 | hide | past | favorite | 96 comments



Aside from the typical encryption worries here are some features I would love to see being offered on a alternative messenger app

-ability to give out your alias or username to people to reach you on whatsapp instead of a phone number, something you can change later after banning the contact.

-transcribe voice messages to enable search among them.

-ability to hide the fact that i have listened to their voice message.

-ability to hide the fact that i am on whatsapp network and receive 'add to contact' requests from people where i can accept or reject without them knowing i actually rejected.

-hide the 'online' in conversation window when im reading the message.

-save my chat history in the cloud so that i dont lose everything when i move from iOS to Android or vise versa.

-ability to save incoming and outgoing media attachments to the cloud for access later

-full desktop support for video and voice calls

-ability to record the calls, and have their transcribed texts easily searchable.

-ability to ignore certain kind of messages from certain people such as 'allow text messages only, no calls, no pictures, no videos.' etc.


- user accessible backups. WhatsApp backs up to your Google Drive account but you can't access it directly. Only WhatsApp may do that (f_ck you Google for contributing to WhatsApp's lock-in). Local backups are encrypted so you can't read them. Exports are incomplete (for example, on one chat it would give me the last month or so when I could scroll back more than a year). WhatsApp discards old messages at its discretion without letting you configure otherwise. I lost a ton of very important texts from the last months of my now-deceased mother. I f_cking had to scrape what was left from WhatsApp Web.

- open protocol and open source clients and servers, with network working independent of any particular provider to be able to ditch any misbehaving one.

- no expiration on software versions. Why are we being forced to update versions to view old messages?

Damn is software becoming user hostile.


>save my chat history in the cloud so that i dont lose everything when i move from iOS to Android or vise versa.

Implementing this could make key management either less secure, or more difficult for users. If the user still had access to both devices, it wouldn’t be too bad. But if you want the users to be able to ‘recover’ their message history on any device, then you’d need to use something like the mnemonic seed phrase, which isn’t an improvement in UX, or security.

The audio transcribing sounds difficult too, as you’d have to do it all client side if you wanted to preserve the security model.


Is there any issue with the Automatic Key Backup feature in Riot?

>Once enabled, your device will maintain a secure copy of its keys on your server. To ensure those keys can only ever be accessed by you, they are encrypted on your device, with a key that you either store yourself or secure with a passphrase and upload to your server. It is important to understand that to protect your privacy your keys will never touch the servers unencrypted. https://medium.com/@RiotChat/the-big-1-0-68fa7c6050be


This is probably the most elegant solution to achieve that functionality, but it still has a few security downgrades.

To back up the keys, you either need to manage a server yourself, or store them with a 3rd party that you trust.

To protect the backed up keys, you either need to manage another key, which just kicks the can down the road, or use a passphrase. If you give your users the option to protect their keys with a passphrase or with a mnemonic word list, a lot of them will pick the passphrase because it’s easier. It’s also less secure, especially if your users choose a weak passphrase, which a lot of them will.

I really like the whatsapp solution, because it has all the security under the hood and doesn’t let the users revert to less secure behaviour. It has an obvious UX tradeoff (with lost message history), but the alternative is to downgrade the key management UX to that of managing a cryptocurrency wallet.


> -full desktop support for video and voice calls

Any desktop support would be nice. WhatsApp web is not a real solution.


It kind of is, at least it is 80% there. What is your problem eith it? Having to handshake almost everytime?


For me, the problem is that it requires internet access. When I found that WhatsApp was deleting important messages and that the backups and exports that I was counting on were useless, as I mentioned in another comment[1], I found that the easiest way to save those messages was to scrape them from the web client. However, connecting the phone to the internet meant that it would receive more messages and hence delete more messages. I took screenshots manually of several days of the oldest available messages to avoid losing them, but it wouldn't have been necessary if the desktop client didn't require internet access.

[1] https://news.ycombinator.com/item?id=19669740


What if you left your phone at home and it's not connected to the Internet?.

I usually leave my phone at random places in my house and have to find it and 'wake it up' so that it connects to the wifi and I can use Whatsapp web.


no gif upload support



I think most if not all of that is possible with matrix.org.


Matrix should have improved XMPP rather than added yet more fragmentation.


> -transcribe voice messages to enable search among them.

> -[..] video [..] calls

> -ability to record the calls, and have their transcribed texts easily searchable.

> -ability to ignore certain kind of messages from certain people such as 'allow text messages only, no calls, no pictures, no videos.' etc.

Apart from these I believe Telegram offers already the other features (voice desktop calls are a bit unreliable often)


>-save my chat history in the cloud so that i dont lose everything when i move from iOS to Android or vise versa.

the key is that the data has to be encrypted. currently i think whatsapp uploads your chat history as plaintext. this basically allows the corps. to get your data.


> -save my chat history in the cloud so that i dont lose everything when i move from iOS to Android or vise versa.

Is there something in the existing WhatsApp cloud backup feature that prevents you from bringing those backups between iPhone and Android?


I did a lot of research and apparently its impossible because if different database formats the messages are saved on two platforms. At least that's what I understood on some random forum.


It is possible to do so, using third party applications. I've done it with the help of a friend, although it required using paid Windows apps that looked like there was no way they could possibly be legitimate and not a scam - they worked like a charm. the process involved an application downgrade, forcing the iPhone to make local backup instead of backing up to the cloud and a handful other things, all of which where automated.


yeah I saw those and to your point I thought there is no way these are legit and I thought it's definitely a scam


- ability to add custom emoticons to your library, and ability to share them


Isn't that called Signal? Maybe just put the funding into extending that?


I tried signal, much more inferior to WhatsApp in terms of functionality and usability


I use signal on the reg, and the short comings are both superficial and unrelated to the security. In my opinion.

What are your specific concerns? I could list things like: lack of bold and italic , poor voice message UI, lack of video conferencing , lack of group names, ...

Note that WhatsApp was wildly popular for over half a decade before it got most of those features.

In other words: I don’t get the point of talking about these cosmetic blemishes. * Signal today feels like WhatsApp at its inception. I.e. a messaging app with the potential to be used by everyone and their dog.

Or is there some fundamental problem you’re encountering that I’m blind / oblivious to?

* edit: let me clarify: “... talking about these cosmetic blemishes as if they’re fundamental blockers to adoption.


they are not just cosmetic..

* no search (big one for me)

* no web client ( another big one)

* app gets stuck in "retrieving message..." if there is a poor signal like travelling under a tunnel

and there are many cosmetic ones like the ones you pointed out.


Text search across full message history has been part of stable production releases (on Android at least) for a decent long while, just fyi. (Desktop client has also gotten better (but is obviously far from perfect); those messages end up retrieved later on, fwiw).


Not sure if it's because I run the beta version but I have the search feature in individual conversations and use it often. There is a desktop app and I agree it's not the best. The desktop app is the only place I've had the retrieving messages issue. So really the only gripe I see you having is no decent non-phone client. For me that's not a show stopper at this point in exchange for a secure, easy to use application.


> . So really the only gripe I see you having

you ignored all my complaints saying you don't experience them. wtf.


Note that the horse a buggy was wildly popular before the car came along, so as this follows Tesla would be find starting a horse and buggy company. I'm a bit exaggerated in my sarcasm here to really illustrate how arrogant it is to assume that features you don't value are 'cosmetic blemishes' when they really do fundamentally hinder the adoption of the product. While you may value security as a priority that overcomes all others, but most by network effect influence, do not. And if you really care about security, it follows that the 'cosmetics' matter as well.


Yes, although pushing a firmware update to turn a horse carriage into a model Y is, perhaps, slightly less practical than adding Markdown support to a chat client.

The original comment was a blanket "much more inferior in terms of functionality and usability." I hope to elucidate that, in the context of a DARPA grant, it might be worth the effort to look to the future, and differentiate between fundamental problems and fixable issues.

I wouldn't say the same for an XMPP client, for example. But WhatsApp and signal are very, very close.

Perhaps even closer than a horse and a model Y. ;)


A few points that I hope dont come off snarky

1) It's open source, so if you can code and have time, you can help.

2) Some small friction adders seem like a small price to pay to not have a facebook product installed on my phone and suck all my metadata down.


You’re right and I agree with you, but as to point 1, they don’t really act like a proper open source project and it seems more like a “source dump” than anything. It’s hard to compile on your own and they don’t like forks connecting to official servers either.


That, and "just go contribute" is really just a cop out answer. Users can desire features without it requiring that they learn to develop software. And really, does anyone want brand new, fresh developers contributing to a privacy-oriented product? Of course not. We need experienced and skilled developers to do that, because mistakes can be costly.


To be fair he was talking about usability and not encryption/privacy features. I agree mostly though, even capable developers don't have time. Feature requests make software better. If you want people to use your app, you should build for the user. It seems they want people to actually use it and didn't open source just to share knowledge of the code. They probably open sourced it to inspire trust.


"and they don’t like forks connecting to official servers either"

Don't like? Well, nice words for activ blocking.


> It’s hard to compile on your own

It was pretty easy to compile your own desktop version. I had to change some colors because white was hurting my eyes and i was able to do it easily.


Unfortunately, the Android app requires Play Services to compile even if not to run, which means F-Droid can’t include it by their own free software requirements. See the forum thread: https://forum.f-droid.org/t/signal-discussion-about-google-p...

Building it reproducibly and with 100% FOSS is very hard and you can’t fork it, otherwise they’ll stop you from using the official servers.


A 500 billion dollar corporation will always be able to beat nonprofits on usability. The question you should ask yourself is whether the marginal difference in usability makes up for all the negative consequences of being a Facebook user.


> A 500 billion dollar corporation will always be able to beat nonprofits on usability.

You'd think. And this is often true, but not always. I am frequently surprised by how bad applications from major companies can be when it comes to usability.


Currently the answer is a definite no, but that may change in future.


There are real tradeoffs between functionality and usability that any "more secure" chat service would have to deal with.


In this Telegram made many nice choices. It can be end-to-end secure but it is also a viable Whatsapp alternative.

Sticker, channels, bots and supergroups are essential to attract communities.


WhatsApp is default end-to-end secure, and Telegram is not. Recommending Telegram as a more secure alternative to WhatsApp is malpractice.


Yes, because there is an inverse relationship to usability and security. Signal won't add a feature until they can call it entirely vetted and secure.


>inverse relationship to usability and security

Not necessarily. You can easily have all kinds of shitty usability without it having anything to do with security.


An inverse relationship between better security and usability perhaps? More security == less usability. Look at PGP encrypted email as a shining example of security sucking for usability.


How is Signal inferior to WhatsApp "in terms of functionality and usability"?


Just a few things:

- Signal is a lot less "polished" - FB has spent tons more resources on making a slick streamlined easy to use interface. Even though signal looks similar its missing all the little things.

- there is no easy to use backup. Whatsapp can backup to google drive or icloud and restore it on another phone. Signal you have to figure out how to deal with and transfer your backup file[1]. This isn't easy for regular folks, Signal needs to be able to connect to cloud storage (preferably of your choice).

- Whatsapp deals with and displays media better/faster

- calling works better. last time I checked whatsapp has easy group video and audio calls, Signal had neither.

There is actually lots more. I prefer the idea of using Signal over whatsapp, but I found it not close enough yet. Especially for non-techy ppl.

[1]: https://support.signal.org/hc/en-us/articles/360007059752-Ba...


> Whatsapp can backup to google drive or icloud and restore it on another phone.

As long as it's Android to Android or iPhone to iPhone. Oh, and Windows phone backup transfers never worked, not even to another Windows phone.


Personally, I don't want backups. I don't want to worry whether some joke I made 10 years ago would look bad in a future data dump. If messages could self destruct after three years that would be lovely


I personally will never user a personal messaging app without long term backups. In twenty years there I will have many fond memories of my past conversations. I see it (almost) no different from keeping old letters.


It feels much worse to use than both Telegram or WhatsApp, both on Android and on OSX. I had installed it on iOS and Android and uninstalled within 10 minutes of trying to use it. That is how bad my user experience was.


My main complain is that whenever I start the desktop client after few days, it takes forever to load the old messages. I'm not sure, but I think it tries to download all messages and attachments before showing the UI, which can be potentially made faster by first trying to get the messages from person I'm trying to talk to and downloading attachments after text messages (Not sure about implication to security protocol here).


I had issues with lag and dropped packets when I was calling from Europe to South America with Signal about 6 months ago. I was really cheering for Signal (still am) and had suggested we try it, but it failed quite quickly. We moved the call back to WhatsApp within a minute or two, and had no problems whatsoever. So at least in my mind WhatsApp is still much more reliable.


It has almost none of my contacts on them. Pretty useless as a messenger as such.


Invite users and explain why? I went from zero contacts to every single person i need to reach at any frequency is now on Signal.


Put effort into improving usability of Signal then? I don't understand.


A lot of the issues can't be solved without relying on an authoritative 3rd party (which is what facebook, whatsapp, skype, etc do). For example, because all messages are p2p, you can't view conversation history that isn't stored locally or sent by the other party upon request (which may not include deleted messages, etc).


Well, on that regard, neither can you on Whatsapp itself.


Signal relies on an authoritative third party to exactly the same extent that WhatsApp does as far as I know.


Isn't signal peer to peer? How can there be an authoritative 3rd party, unless they get rid of all the advantages of using p2p?

I don't know anything about signal, I just assumed it was exclusively p2p based on 30 seconds of googling. But assuming it is all p2p, my above statement should be correct.


Its end to end encrypted and ciphertexts aren't stored long term on Signal's servers, but Signal isn't truly p2p. To send a message a user encrypts the message using the keys for the user, sends them to Signal's servers to be queued, Signal delivers the encrypted message to the receiving user, the receiving user decrypts it. Phone calls would probably be the closest thing to p2p, but it still uses Signal's servers as a broker for connecting the two parties instead of some other form of discovery.

Signal also does not allow for third-party clients to connect to their hosted Signal servers. This further makes them an authoritative 3rd party, as you need to run the code approved by Signal to connect to Signal's servers to send messages to Signal users. Not that I am complaining about this structure, just sharing knowledge.


Interesting, but wouldn't this method would still have its own set unsolvable of problems? For example I don't know if it would be possible to search on the server, since all messages are encrypted.


I use it nearly every day since nearly two years and I have had not many problems. The voice calls are actually higher quality than via Skype (when calling the same person at roughly the same time).


I like Keybase.io - it seems better than Signal and is nice to use.


Really? I fail to see _any_ difference between the two in terms of features.


WhatsApp is Signal under the hood. And as I recall, Moxie helped them set it up.


Not exactly. WhatsApp uses Signal's communication protocol, but has some flaws in the way they implemented it, by design.

https://www.cyberscoop.com/whatsapp-signal-group-chat-vulner...


WhatsApp is Signal Protocol under the hood. That does not make the two equivalent from a security perspective; Signal goes through a lot of extra trouble to protect metadata (note, for instance, how it didn't even really have user profiles until relatively recently).


That's exactly my thoughts.

Interestingly, WhatsApp uses the same e2e encryption protocol under the hood... or at least it used to I know Facebook is trying to merge it with messenger which certainly means removing the e2e encryption.

Good news is this means Signal could probably be on parity with WhatsApp with a little government grant money love.


This makes it sound like DARPA wants to take some commercially developed tech and adapt it to the military, but just for the record that grossly misstates the relationship and the general history of Silicon Valley.

Ever wonder how Siri got her name? From SRI International, a research org largely funded by the DoD.[1] DARPA is the origin of much of the tech, including secure Internet tech. Just one recent example: Tor[2]. It's developed at taxpayer expense and either given away to private industry or "transferred" for a pittance.

[1] https://en.wikipedia.org/wiki/SRI_International#Employees_an...

[2] https://www.onion-router.net/Sponsors.html


I don't think there's an official origin for Siri's name. Someone said it was named after a child. I suspect (without evidence) it was named after the female robot in the pilot episode of the Logan's Run who ran Mountain City, a paradise city.


From : https://www.sri.com/engage/ventures/siri

https://www.sri.com/work/timeline-innovation/timeline.php?ti...

"Siri, the first virtual personal assistant, arose from decades of SRI research in artificial intelligence (AI). The technology was developed through the SRI-led Cognitive Assistant that Learns and Organizes (CALO) project within DARPA's Personalized Assistant that Learns (PAL) program, the largest-known AI project in U.S. History, and joint work with EPFL, the Swiss institute of technology.

SRI spun off Siri, Inc. in 2007 to bring the technology to consumers, raising $24 million in two rounds of financing.

In April 2010, Apple acquired Siri, and in October 2011, Siri was unveiled as an integrated feature of the Apple iPhone 4S."

I'm not verifying that - I just found it via google search "sri Siri"

I like the link to Logan's Run :-), it could be also true.


It's mildly entertaining that we have calls for a more secure and trust-able version of an encrypted chat service, but other agencies in the executive branch are calling for loopholes and vulnerabilities for the same encrypted channels.


Do we really need yet another chat app? I think there are far too many of them. Can't we go back to XMPP again and make everything compatible like it was ten years ago?


From the description, it seems like it's less about the chat app itself and more about developing and designing a new secure protocol and infrastructure that will influence future products by commercial or other entities. They want people to take the tech they develop and make their own stuff more secure, which sounds awesome.


Amen.

If anyone wants to be on XMPP with minimal fuss, try https://quicksy.im/ . This is an XMPP client, with registration based on your phone number.


Seems like Keybase's key model would work well here: https://keybase.io/blog/keybase-new-key-model


> Exist completely within a network

I bet this requirements implies working seemlessly between online and offline networks. In which case this project is not just building another XMPP/WhatApps/Signal/etc. They are building another Cabal (https://cabal-club.github.io/) :-)


The title makes it sound like they want to use whatsapp, not that they want encrypted chat.


Agreed, i think the title is a bit misleading.


What's wrong with the current encryption? I'm told that Signal's encryption technique if very superior and probably the best in the industry.


Just wait till Telegram open sources their code?


Original solicitation is from June 2018, not clear that this story is anything but a reheating of that announcement, found here: https://www.fbo.gov/index?s=opportunity&mode=form&id=c244cde...

(As an aside, it is always a bit unsettling when a federal function is represented by something that looks like a link farm. I mean, I know you elected Trump and all, but "FedBizOpps" probably predates him by a decade)


Besides signal, there is also

1. Threema. Swiss based but not Open Source AFAI

2. Frozen Chat. Android client based on Jabber XMMP/OTR


Frozen Chat seems to be removed from the Play Store, and I can't find many references to it online. Never heard of it before.


Oh, wow. It is on my phone but not in the play store anymore. Strange.

https://github.com/Frozenbox/frozenchat


Anybody using Wire? I've recently been looking for a skype replacement myself and seems like wire is a pretty good fit.


I always wonder why few people here acknowledge Telegram, is it little know or is it not considered secure enough?


Unencrypted chats by default.

Self-written encryption.


There's also Conversations, an Android XMPP client that supports OTR/OMEMO.


Don't they have one already!


Trust us, this totally isn't another Dual_EC_DRBG


Darpa had absolutely nothing to do with Dual_EC_DRBG.


DARPA's not doing this for shits and giggles; they want to use it (or later iterations of it) for military communication. They also want to open source it.


It's exactly how Tor came about.


They should invest in the BEAM virtual machine. IIRC, Whatapps is built on Erlang and its VM.

There are many benefits in helping an existing project with active developers. If dealing with the politics of managing existing project is annoying, just fork it and make their own branch. The original owner can deal with the merging if they want the contribution.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: