I agree with your general sentiment here (that such an arrangement is not trustworthy enough for me to feel completely private) but your delivery of said sentiment is really fucking rude, dude.

Even if it’s not secure enough for you or I to feel private, it likely exceeds the security necessary to satisfy most people’s threat models so they can not only feel private but objectively be more private than if they just used Google docs.

incremental or opportunistic privacy improvements are better than none, a fact that has seemed to be lost in elitist privacy circles these days.

I’m not sure what you’re talking about ? You’re not sending your private key to their server without first encrypting it first locally. Their servers are not doing the E2EE, your client is. The website front and apps are open source.

Yes they could send you a compromised front if you use it via their website, that’s a compromise you accept, otherwhise you could only use their apps which are open source.

I imagine it probably is inspected, just not by the public. They probably do it themselves.

And they may have contracts with certain companies specializing in this sort of security that also inspect it.

And there’s also the cybersecurity companies that test it whether they’re contracted or not. At some companies, their entire job revolves around finding bugs (especially security bugs) in other companies’ software.

Just because it’s not on GitHub doesn’t mean it’s not a good product that hasn’t been thoroughly tested.

Yeah because enterprises primarily use a ton of open source security tools…

ಠ_ಠ

Then it’s a good thing all of their products are fully functional and working as advertised, I guess.

Was that reply for me?

It is entirely possible to keep secure data on a server that only someone else with the password can access. They don’t store your password in plaintext, they don’t test whether what you typed is the same thing they keep on their servers. If the password works to decrypt your data then your client can read the emails. If not, your client gets gibberish and knows your password was wrong. With a secure system your password should never be sent to the server at all.

Now, that doesn’t mean it’s trustworthy. There could be holes in the security, and I certainly would feel better controlling my own server, but it’s not automatically insecure just because it’s hosted by them.

prove to me by pointing to the specific source code that would prevent Proton from your private key password

You want them to point to code (that’s not publicly available) that prevents Proton from capturing credentials?

What a dumb hoop to tell someone to jump through. Do you expect someone to actually post a block of code? Would you even be able to read it?

This simply isn’t really possible.

Even if they published open-source code for their backend, it wouldn’t prove that it’s actually what their systems are running.

And when you are storing your data on their servers, and decrypting it by sending over your password, there’s no way you can actually truly prevent them from accessing your data, if they were to modify how their systems function overall. (this is true for every company)

Even if they were using zero-knowledge proofs to verify and prove to you the computation done on the server matched what would be expected from published open-source code, then either their very own systems (and by extension, their administrators), or a different company’s proprietary TPM module, would be the root of trust for those ZK proofs, and would still have the same underlying trust assumptions of at least 1 company having the ability to potentially steal your information.

If you want to rail against Proton for this, you have to be against every single cloud-based instance of code that hosts encrypted data, by any company, for any user.

Saying Proton acts just like Microsoft is a laughable comparison to make in order to justify claiming a lack of privacy or security on Proton’s part.

Why? Is it because they’re both companies that offer online services? Guess what, loads of companies do that. But you know what Proton doesn’t do? Give away the contents of people’s files, like Microsoft states they do in their own transparency reports, that they conveniently stopped publishing in 2022. Microsoft handed over the content (not just IP, email, etc, but actual docs, communications, stored files, etc) of thousands of people’s accounts to law enforcement. Proton hasn’t given out content once.

And this doesn’t even consider the fact that Proton’s business model is privacy. For Microsoft, their users will keep using their services regardless of their privacy, but for Proton, if it comes out that their services are no longer private, nobody will use them anymore, because nobody who got them for privacy would need them at that point.

Such source code isn’t possible with the general audience service they offer, even if being open source were a requirement for credibility in any way.

You’re comparing them to a company with a long history of actively hostile behavior despite the fact that there’s never been a single hint of anything resembling hostile behavior from them, they operate from a country with meaningful privacy protections and only surrender data when compelled by their own courts (who only do so in circumstances that actually warrant it), and haven’t actually given up information that’s useful when required to because they don’t have it.

But access comes with office, so if you have excel you have at least a software that is intended to be used as a DB (efficacy aside)

Probably true for most companies but I worked at one that had plenty of DB servers and developers, even developed their own database tech. Still, Excelitis as we called it was rampant.