> I can believe them [WhatsApp] when they say they don't keep data.

If that's true, I'm in awe of their integrity. Skype on the other hand now has complete disregard for user privacy. Skype stores your voice mails and video messages forever[1]. This is something that they started doing 2-3 years ago and few people seem to be aware of it. It's amazing how low Skype fallen from its early days when it was considered a beacon of privacy and on the cutting edge of encryption and security.

[1] Details: Clicking on Preferences -> Privacy -> Delete history (OS X) or Options -> Privacy Settings -> Clear history (Windows) pretends to delete the voice/video messages but it merely hides them from your view. If you re-install Skype on the same computer or run Skype on a different computer, all those "deleted" voice mails and video messages re-appear. The delete and clear buttons are basically lies; there's no polite way to put it.

Skype was never a reliable option. They never published their protocols or security (just that one "review" iirc). They took many measures to prevent people from inspecting the client.

I thought Skype only stores the last 30 days of data on their servers, while it will store every piece of data received on your local computer indefinitely (unless you delete it, of course).

No. I personally have a video message that is almost 3 years old that I've tried to delete multiple times over the years and it still shows up if I do a fresh install of Skype on a new computer.

Very old voice messages are also accessible even if "deleted".

Text chats do seem to disappear, but at this point I don't believe anything Skype says. I figure they keep the chats forever as well.

Chats are definitely saved, but files aren't. You can't retrieve files after certain amount of time. I can also retrieve conversation from 3+ years ago. I don't remember if there is an option to forget conversation history though.

In all fairness, that's not say they're not being moved to deep physical storage. That's a practice in many organisations dealing with mass data.

storing data for 1 billion users and maintaining consistency is a pain the butt operationally. Whatsapp isn't storing the data. If they are, they're spending hundreds of millions on servers for data they can't monetize easily without alienating all of its users.

It is just text records, I doubt it will take hundreds of millions, even at whatsapp levels

No, a lot of pictures and audio messages as well go through Whatsapp

You used to could store your archive to Google Drive (at least with the Android client).

Edit: yes, the chat backup functionality still works. I have never tried to restore it though.

Go to settings > chats >> Chat backup to start backing up your messages to Google Drive. (Make sure you have 2FA on your Google account)

Telegram isn't end-to-end encrypted unless you're in a "secret chat", and that chat exists only between two different devices and can't be used by multiple devices on one account.

It's a "feature" of the normal chats/channels/groups/supergroups of Telegram that you are able to download them to other devices, or to restore them on a freshly wiped device, because (and I'm over-simplifying here, but the end result is the same) they are encrypted with a key known to the server, and which other devices signed into your account can then be authorized to use.

My point was more that if WhatsApp retained messages in some form they most likely would expose that functionality to its users (as Telegram does). So when people say 'apparently' they don't store messages, I'm inclined to believe WhatsApp/Facebook.

Digressing from the actual discussion, Whatsapp now allows you to keep a backup of your chats and related media on your Google Drive and then recover it when you move to a new device. Probably applicable only for newer versions of Android (not sure about iOS).

You can backup to (and restore from) iCloud, but then again WhatsApp (the company) does not have the data then.

Yes, I didn't mean to say they have the data. That's why I said I'm digressing from the topic (realizing the parent commenter might find it useful).

I wonder if they record the necessary analytics data from mining the text, and then delete the actual text. Depending on what data you're trying to mine, you may be able to get what you need on transmission, and then dump the source data in favor of the output of the analytics event.

The chats are end-to-end encrypted, so there is no plain text anyone could "mine"

That's true if you believe WhatsApp (and I do with some probability of certainty--not certain enough to trust it with data I wouldn't want the government to see, but enough that I'm comfortable hackers won't get my bank info).

But, the NSA approach to data collection is basically to vacuum it all up and, if possible, decrypt it later. This has two implications:

1. Metadata. "We kill people based on metadata" isn't a joke, it's a quote from Michael Hayden, ex-executive in both the NSA and CIA. End-to-end encryption doesn't hide who you're talking to or when.

2. It seems unlikely that it will be computationally possible for them to decrypt all traffic, but it would only surprise me a little if AES128 is breakable for high-value targets in the next 20 years: increased computing power, better multi-threading, better cryptanalysis algorithms, maybe quantum computing or some completely unexpected technology; it's hard to say what will come along.

In short: mining encrypted data still matters.

Unless they are covertly decrypting them on the fly using a master key.

Shouldn't that be impossible by definition?

I think it would be possible to include additional data in the E2E handshake that would allow a third party to extract the session key.