5 ms·
Here's my problem: Let's suppose that I, or anyone else that reads Hacker News (like many of the YCombinator companies) creates an app (as many of them have),
by peter_d_sherman 7y ago
Here's my problem:
Let's suppose that I, or anyone else that reads Hacker News (like many of the YCombinator companies) creates an app (as many of them have), and this app becomes very popular in foreign countries, and goes on to get tons of users in those foreign countries...
OK, now let's suppose that a foreign newspaper, like the New York times, but a foreign version of that -- all of a sudden prints an article like this one, alleging, but not showing a comprehensive proof (like the ones in mathematics or logic) that there are all kinds of things unspeakably wrong with the app...
Well... maybe all of those things are true, but maybe they aren't...
How would I, as the end-user of such an app -- know for sure?
If the article isn't true... think of the gigantic swath of economic damage that would be caused by an article like this...
And if the article is true... well, it could be argued that you've done a great thing, and saved countless users' privacy rights...
That's quite the dichotomy of potential outcomes, based on what the truth of the matter actually is...
So here's my question to the NYT: In this era of Fake News (which you are no doubt aware of!), how do I prove to myself (as someone who is 50/50, that is, a skeptic who is willing to believe, if the proper set of facts is presented to them) that what you're saying is true?
?
Here are two things for the NYT to think about:
"Semper necessitas probandi incumbit ei qui agit"
The burden of proof lays on he who makes the claim
and
"Extraordinary claims require extraordinary evidence"
-Carl Sagan
https://en.wikipedia.org/wiki/Sagan_standard https://en.wikipedia.org/wiki/Sagan_standard
Saying that this stuff is true because the Intelligence Community says that it's true -- is not unlike making an "Argument to/from Authority":
https://en.wikipedia.org/wiki/Argument_from_authority https://en.wikipedia.org/wiki/Argument_from_authority
I don't know who the Intelligence Community is (much like I don't know who other abstract entities are like "China" or even the "U.S." are) I don't know who that is!
I do know who a guy named "John Smith" is, and I do know that if John Smith raises his right hand in a court of law and under oath promises to "tell the truth, the whole truth, and nothing but the truth, under penalty of perjury", then John Smith's claim, when he makes it under those circumstances, carries a whole lot more weight (to me) than the claims of any nameless, faceless, unaccountable organization -- by whatever name they go by...
Due Process (enshrined in our Constitution, which the media does not seem to give these days to all parties, despite the Constitution giving them their right of Free Speech) presumes innocence until proven guilty, in a court of law!
Now, all of that being said... I am willing to believe this article -- but I require a higher standard of proof...
- Wowfunhappy 7y ago> How would I, as the end-user of such an app -- know for sure? Monitor network traffic, or maybe even decompile the binary. What information is getting sent out of the app? Does it seem reasonable? If not, you have cause for concern. Unfortunately, modern platforms like iOS make this very difficult, frequently in the name of protecting security and privacy, ironically. This is a very big problem.
- peter_d_sherman 7y agoI agree. But to broaden your point, it could be argued that this problem affects every single application that uses the network... How do I know that Chrome, Firefox, IE, or any other U.S. made piece of software -- doesn't leak private information, one way or another? Intentionally or unintentionally? I've thought about this long and hard (as someone who wanted to publish an app), and the only two things I can come up with are: 1) Open the source code; that is, make it open source, and subject to peer review. 2) Inform users that due to the complexity inherent in computers and software/hardware/firmware stacks in this day and age, there cannot be a guarantee that your private information will stay private; thus, the simplest solution is don't put any private information on the app / don't assume that it will stay private.
- Wowfunhappy 7y agoInvestigate each layer one at a time. Not every end user will be able to independently verify every piece of software on their machine—going fully open source wouldn't do that either. But these types of investigations should be open to the general public so that peer review can take place.
- maqp 7y agoSecurity is about verifiability. If you can't verify it's secure, it's not trustworthy. I've been looking into secure communication stuff for almost ten years now, and there's a low bar that makes it easy to inspect whether the app developers are even trying. You want security for two things: Metadata and content. Content privacy is done with end-to-end encryption. You ask yourself 1) Can I find public key fingerprints from the app? If this is not available, the app is not end-to-end encrypted properly, and you should immediately abandon it. 2) If the app is end-to-end encrypted, can I inspect the source code to verify end-to-end encryption is implemented properly? If you can't do that, i.e. if the source code is not available, there's no way to trust the app. If you lack skills needed to do that, you can pitch into an audit by third party. 3) If the source is open, can I verify I'm using binary compiled from said source? I.e., does the client have reproducible builds. If all these three are true, then you can be sure it offers quite strong protection for your content. This is true for e.g. Signal. Regarding your Sagan Standard, the extraordinary claim is not "It has a backdoor because we can't check." The extraordinary claim is "It doesn't have a backdoor, it's secure, but no, you can't check that for yourself." At that point you need to have extraordinary proof that can verify the security. There's no pro-security argument in keeping an app proprietary, only monetary. --- Next, metadata: You ask yourself 1) Can I register the service anonymously, i.e. without giving any of my details 2) Can I connect (both register and use the app) via Tor? If not, your IP reveals your identity. 3) Who's hosting the service? Does a third party learn how often UID #4346455432 sends messages to other parties. I.e. is the app peer-to-peer. If the app is not peer-to-peer, your peers' bad OPSEC might deanonymize you if server-side data is cross-compared with an other social graph. 4) Can I misconfigure the system and lose anonymity (i.e. is the system Tor-routed by design or can you accidentally connect without Tor), and can you accidentally turn off end-to-end encryption and say something that might deanonymize you to the server? ToTok is not open source, it doesn't have public key fingerprints, and judging by the completely novice, non-technical name-dropping of a few primitives in their PR statement (https://totok.ai/news https://totok.ai/news): "Furthermore, we equipped ToTok with such high-security standards as AES256, TLS/SSL, RSA and SHA256, to diligently protect the user data. We also implemented a privacy framework that complies with the local and international legal requirements to safeguard our users at all times." It's trivial to implement backdoored system with said primitives: TLS is encryption that uses AES256, RSA, and SHA256, and it's only encryption between user and server, so server has access to plaintext data which can be given to the secret police etc. Even if they say they don't comply they might a) lie or b) be hacked. The only way to protect from that is to always use E2EE, something ToTok does not do. So no, there's no reason to trust them. They haven't even attempted to deploy E2EE or anonymity for their system. Even if they're not intentionally being evil, they're being ignorant about security and thus, ignorant about the human lives affected.