82
Dessalines
PrivacyIE like Crypto AG:
In 2020, it was revealed that the Swiss company, Crypto AG, which provided secure communications services to ~120 governments throughout the 20th century, was secretly ran by the CIA and West German Intelligence. The CIA and later NSA were able to read encrypted communications for many countries such as Saudi Arabia, Iran, Italy, Indonesia, Iraq, Libya, Jordan and South Korea.
The Proton CEO did make suspicious US political statements despite being Swiss. That combined with their misleading marketing on social media.
This thread basically illustrates the challenges for a beginner, such as myself.
I've been locked into the Google ecosystem for nearly two decades and am now trying to free myself.
I'd like to migrate to a hybrid solution that involves self-hosted NextCloud synchronized with a cloud provider that I can trust more than Google.
However:
Proton apparently makes false, or at least misleading, marketing claims and doesn't fight a vast majority of its inbound government requests.
Tuta has been publicly accused by a member of the intelligence community of being a honeypot.
The rest of the email providers seem to implement even fewer protections, relative to these two.
So, what's a guy to do?
Now, to be clear, I'm not saying that either of these companies are bad or that I believe that they're actually honeypots. I'm just trying to illustrate the challenges faced by newcomers (and probably all of us).
While I'd prefer to absolutely maximize privacy and security on all fronts, given that my first goal is de-googling, I will probably start with Proton and NextCloud and re-evaluate from there, but I'm open to suggestions.
Thank you all -- I really appreciate this community.
Tbh for email I'd say don't bother with privacy as it wasn't meant to be private, as Dessalines said. If you care about data sovereignty (which is different to privacy, though often hand-in-hand), you can self-host email—it's not as hard as it's reputed to be. I've self-hosted my main email address for a couple years now and not had major hiccups. For the most part, after initial setup, it just runs. And if you're daunted by configuring it, there are out-of-the-box solutions like Mailcow you can use. I'd only really recommend it if you already have a VPS/home lab/etc where you already self-host things.
I intend to do that but basically wanted to have an off site copy, for both backup and deliverability purposes.
I don't have much in the way of privacy expectations for email, but I figure that Proton or Tuta are probably still safer than Google.
VPS/home lab
VPS is probably fine, hosting something this important on your own hardware sounds like a recipe for disaster though
No company is in a position to resist lawful orders from government (not good orders, lawful).
It’s why every company that sells security makes a big show about planning to leave some western country when they say they’re gonna do mass surveillance. It’s all they can do.
Email is not secure and cannot be made secure.
Do not ever send anything through email that you rely on being private.
I'm certainly not suggesting that email providers should resist lawful orders, but if Proton complies with 89% of requests while Tuta complies with 25%, it suggests a difference in methodology, no?
It could, of course, be the case that the Swiss are just much more skilled at sending lawful requests relative to the Germans, but that seems unlikely.
Email is a really tough one especially, because it wasn't designed with security in mind, and of course even if you're on a secure email service, 99% of the emails you send and receive are going to be with non-secure services hoovered up by google or AWS.
Anything is better than google at least.
the worse part is that; by the time security professionals' tribal knowledge is known to the general public; it's already outdated enough to keep you ensnared.
they say that you have to become your own lawyer to protect yourself and you have to become your own dentist/doctor to heal yourself; now you have to be your own secops to guard your information.
Signal for one.
Often ignored, online games. Non of the VPN which logs the history, TOR also isn't the panacea (network made by US secret service). Mandatory monitoring the traffic with Portmaster, PiHole or similar. FOSS from GitHub with a grain of salt. Good to have analytic tools in the bookmarks, eg Blacklight, Webbkoll, Exodus Privacy, Browserleaks, etc., preferable to use european alternatives. Using decentralized or /and selfhosted services. Common sense and always read TOS and PP before using the app or service.
Any VPN that isn't actively being sued by world gov/agencies to try and get their data is suspicious.
Alternatively any VPN company with the ability to store data is untrustworthy.
Also every cryptocurrency that exsts.
How do you feel about Tailscale?
they were talking about proxy VPNs, whereas tailscale is for building actual virtual networks to connect your devices, which is a completely different thing (besides sharing the same approval foundation).
If you were to distrust tailscale (and you're not simply self hosting headscale), an attacker might be able to access for otherwise non-public devices(' ports), reroute/MitM your traffic and monitor which device connects to which.
Not a privacy app, but you should definitely not think anything said on discord is private in any sense whatsoever
Tor comes to mind.
Technologically it's private, but if you're America and have the resources to create and control sufficiently many nodes you can undermine the protections.
The tor rabbit hole goes pretty deep, but ya based on the evidence I'd have to say its more a US developed counter-insurgency tool, rather than a privacy tool.
Wait 'til you hear who invented it...
That said, considering how many illegal services continue to run on it, I don't think it's as porous as some make it out. Definitely has well-documented weaknesses but the project maintainers tend to address them fairly straightforward.
Of course, you're also just as likely to be buying drugs off an Onion market that the FBI seized and kept running just to catch more bad guys, despite it also hosting illegal content itself.
Most likely all free vpns
Israeli actually, like express VPN
Maybe not a honeypot, but definitely too large for my taste by now: Proton. With Mail, VPN, password manager, file storage, AI and whatnot, it's one ginormous basket to put all of your eggs into, hopping it'll hold.
the owner is fine with fascism because fascism makes his product more lucrative
Did he say that? :o
https://web.archive.org/web/20250318235233/https://www.letemps.ch/articles/sur-les-reseaux-sociaux-l-entreprise-genevoise-proton-accusee-de-soutenir-donald-trump/gifts/ybn6gho5JTnuet2bcoda4jDMR573NaKyC9cKGuFh
not exactly. the more nuanced inspection of what he said was that donald trump's plans to deregulate the tech industry he expected to benefit his company. however, that deregulation is in service of allowing more surveillance capitalism, environmental degredation, and worker mistreatment. the wording i provided is what that ultimately means as an analysis of how and why proton would make more money in that type of environment
Be careful of accepting some of the criticism of Signal in this thread. For most of us, we have to make choices about secure comms from subject matter experts. Almost all the criticism I see of Signal comes from anonymous or otherwise random users online. If you believe in such a thing as expertise, please seek it out when evaluating something like this.
It is absolutely irrelevant who makes the criticism, what needs to be addressed is the criticism itself. If somebody gives you advice to simply trust people blindly then you should be very suspicious of their motivations.
Most issues are complex enough that we have to delegate trust. It's not feasible to verify every claim yourself. And trust vs "blind trust" is an arbitrary line.
The issues people bring up with Signal are very easy for anybody with a minimally functioning brain to understand, and none of these experts are able to provide a credible answer to them.
The key issues people point out over and over is that Signal is a central server hosted in the US that harvests people's phone numbers on sign up. The users are trusting server operators with their privacy at that point because there is no way to verify how this data is used. Since the server associates real identity with the account, it is in position to map out networks of people communicating. And if this data is shared with intelligence agencies, which they wouldn't be allowed to disclose, then those can trivially correlate the personally identifiable information with all the other data they have access to.
If there's a person of interest, and you map out whom that person wants to have private conversations with, that's very useful data. Once you know that, then you can start tracking all the activities of their associates, and map out a whole network of people. Say, people organizing unions, or coordinating labor strikes, and so on.
This is an obvious problem with Signal, one that doesn't take any significant expertise to understand, and one that has never been fully addressed. People talk about things like sealed sender, but that doesn't address the problem I just outlined.
The core issue is that you have to trust the physical infrastructure rather than just the cryptography. The protocol design for sealed sender assumes the server behaves exactly as the published open source code dictates. A malicious operator can simply run modified server software that entirely ignores those privacy protections. Even if the cryptographic payload lacks a sender ID, the server still receives the raw network request and all the metadata attached to it. Your client has to talk to the server and identify itself before any messages are even sent.
When your device connects to send that sealed message, it inevitably reveals your IP address and connection timing to the server. The server also knows your IP address from when you initially registered your phone number or when you requested those temporary rate limiting tokens. By logging the raw incoming requests at the network level, a malicious server can easily correlate the IP address sending the sealed message with the IP address tied to the phone number.
Since the server must know the destination to route the message, it just links your incoming IP address to the recipient ID. Over time this builds a complete social graph of who is talking to whom. The cryptographic token merely proves you are allowed to send a message without explicitly stating who you are inside the payload. It does absolutely nothing to hide the metadata of the network connection itself from the machine receiving the data.
This once again makes it very suspicious that Signal insists on running a single centralized server.
The fact that the US Government has adopted signal for it's own employees, suggests they have a backdoor.
good point
Who are the experts, and who pays their salaries? Crypto AG wasn’t lacking in experts.
You'll have to make your own determinations I guess, but be careful if you find yourself dismissing expertise in favor of opinion or motivated reasoning.
Probably various VPNs on the market
Especially Israeli owned VPNs. Which seems to be most of them lately.
Oh yeah definitely
Especially the ones aggressively marketed, or noted as independent when they cannot give concrete evidence for whence their finances and ownership come. Always question and investigate, and make sure trusted people know you do so.
I always assume the more popular it is, the more likely it is of being compromised.
I have no idea if it's the case, but I switched away from mullvad after seeing billboards and ads of it everywhere, even on city infrastructure like trains and buses.
Mullvad is very likely one of the few good ones. I'd suggest reevaluating it.
I have no reason to go back to it, and I switched away from it for the reasons mentioned: its grown very large, and has mainstream ads everywhere now.
My trust in them was definitely shaken after the recent news about fingerprinting exit IPs: https://tmctmt.com/posts/mullvad-exit-ips-as-a-fingerprinting-vector/
They were very responsive but this seemed like a huge fuck-up to me, to the extent that I question whether it was purposeful.
Not sure who else to trust because other providers like Proton seem even worse
if it makes you feel better i know an employee there and theyre a communist and say a lot of mullvad employees are lefties too, idk if they have a union or anything. nym vpn has chelsea manning backing it. not really a traditional vpn though its basically unfree tor that is not slow as balls, has the benefit of really good server coverage and few people blocking it. coolest thing is you can use a seedbox to route traffic to pay it down.
If the company is owned by "Kape" its ikely a Israeli honeypot:
https://medium.com/illumination/vpns-the-privacy-trap-4aef67f39634
Kape’s portfolio includes ExpressVPN, acquired in 2021 for $936 million; CyberGhost, purchased in 2017; Private Internet Access, bought in 2019 for $127 million; and ZenMate.
Together, these services account for three of the six most popular VPN products globally, serving approximately 7.4 million paying subscribers.
Kape also owns VPNMentor and Wizcase, review platforms that rank VPN services — including Kape’s own products — for consumers seeking expert guidance.
Most people only use vpn providers for streaming location hopping, torrenting, p*rn and on public networks. For day to day 24/7 use you are just trusting your VPN provider not to spy on your traffic instead of your ISP.
I know your example is the opposite, but any service that is run and hosted in the US.
It's one of the major issues with Signal.
Not to mention Graphite and Pegasus, Israeli spyware.
When parliaments have to inquire their own spy services, it's a sign that these spy services must be disbanded, as they are becoming a deep state of their own, intimidating and harassing politicians. After all, if you can't trust your own politicians, whom can you? And that's problematic.
Disbanding those services and prohibiting any secret services from ever forming, would also regain a great deal of trust of society in each other. And that trust in turn, can foster society to advance for mankind.
You got that right.
Signal I think. I don't mean that the end2end algorithm or messaging itself are itself unsafe, the algo has been shown to be secure. This is what people usually rebuke this with, with the reminder of Signal's OSS nature.
The issue the servers and the social networking data that can be harvested. The server code only partially exists in public and we just have to trust that that is actually what is running on whatever AWS server without tampering and self hosting is nearly impossible in practice if technically possible and nobody does it. The social network data (who talks to who) is more valuable than the actual messages logs, which give a massive, but mainly useless datasets. Until LLMs, like 10-15 years ago they were basically impossible to parse for any useful info without using large quantities of eye pairs. Basically if you are an organizer, criminal, government, part of a hunted opposition, you will leak the whole core group structure of your org with attached phone numbers. Whoever with that data can then target their devices and persons with other means. Plus it's literally built on top of CIA money. I think signal is totally safe and adequate for friends and family type of use, but not much else, but then all in all so is whatsapp, mostly since signal and Whattsapp share the same end to end algorithm.
Signal is def one, otherwise US government orgs like RFA and OTF wouldn't be defending and pushing for it so hard in western privacy spaces, nor fund it.
Have a look at Deltachat
Its starting to make headway: FOSS, Decentralised and anyone who is tech inclined can setup their own Relay.
It's funny how every poster who criticizes Signal inevitably makes a technical error. In your case, the claim that "Basically if you are an organizer, criminal, government, part of a hunted opposition, you will leak the whole core group structure of your org with attached phone numbers" entirely lacks basis. The Signal client - the OSS part we can and do control - does not divulge phone numbers.
You have this theory that Signal's servers are storing communication records. (While there is no evidence to support this, it's valuable to consider what they could do.) So the data that would be captured here is a network of hashed phone numbers and literally undecryptable messages. It's impossible for the adversary to determine any phone numbers they don't already know this way.
And since you can make a Signal account with a burner phone and create a "username", even a known phone number becomes useless against targets who don't want to be identified.
The phone numbers being hashed doesn't matter because of how small the input space is. A standard phone number is a country code plus 9 digits. If we assume that anybody looking at this information already knows what country the people they're targeting is from, that means there is 1000 000 000 possible phone numbers to check for any hash. Even if the hash is extremely slow, and takes 1 second to compute on a strong CPU, that still only takes 1000 000 000 / (60 * 60 * 24) = 11574 days, or 31 years to compute on a single core. For any large organization (like, say, any government or any large tech company), getting 1000 cores to run the hashes in parallel would be quite simple, reducing the time it takes to have a complete hash list down to 11 days to get a complete database of all possible hashes. Hashing phone numbers is literally just a mild inconvenience.
Edit: Actually looking it up phone number formats vary quite a lot by country, but the point still stands.
Signal doesn't run in a vacuum. It's main distribution platforms are app stores from Google and Apple. And most people are going to use stock smartphones from these two companies to sign up to Signal. But with them being under the same US jurisdiction, matching the two identities isn't that far-fetched.
The parent companies of both OS platforms are well known to funnel data and notifications to the US government. It too had no evidence to support it, until they admitted it. There's a setting for it now, but the person you're talking to might not be doing the same, so it's still out for profiling.
Other thing, they vehemently oppose F-Droid because "f-droid security flaws" bs, even though they can literally host their own repo for it without anyone else building their app. They would control every aspect of supply chain, but they didn't.
Besides that, they make it very inconvenient to get it from elsewhere, even though they did the bare minimum to provide a standalone installer, after an outcry. And with those stripped down installers, you have to deal with inconsistent notifications, because no apple/google. And they never ever gave unified push a look. I wonder why? Are they a small indie company with just a couple of devs?
Signal protocol may be "secure", but it's only a part of a bigger picture.
It's forced reliance on phone numbers, privacy averted platforms and unwillingness to work with opensource platforms and standards that lets it become decentralized and out of the hands of authoritarian government, leaves a lot to be desired.
Facebook's whatsapp also uses the signal protocol, but would you call it private or secure after all that zuck has shown to do? Signal creator literally helped them implement it too. I wouldn't touch a Facebook product with a 10 feet pole.
And now he's helping them again encrypt Meta AI, whatever that means. Why is he working with one of the worst offenders of privacy?
If that doesn't tell you these things are concerning, you do you.
https://lemmy.ml/post/48427945
All speculation. You gave them your phone number (which also means your real identity), so you should assume they have it. And because its a US-based company, it must adhere to US laws including key disclosure laws, which make it illegal for any signal employee to tell you that any US government agency has asked for this information.
https://en.wikipedia.org/wiki/National_security_letter
So the data that would be captured here is a network of hashed phone numbers and literally undecryptable messages
With this data you can build social networking graphs: who is talking to who, and when.
Also this is all the more suspect when you consider that US military / government agencies like OTF fund signal, and constantly try to push signal in privacy spaces.
The point is that they could. We are discussing honeypots here. They don't advertise the fact if they are.
Be the phone numbers hashed/encrypted or not they will still get your ip. They are not routing anybody's messages otherwise. Phone number is just more directly tied to a personal details, unless it's a burner, but with burners you lose the account if you need to log in. Also you can set your phone number public, so it probably can be seen by the signal servers at some point. And what about discovery through phone number and like the actual sending of the signal confirmation code? How is any that suppose to work if the servers don't know your actual phone number? And your anonymity trick only works if everybody you talk to does it, which they don't. If they want to profile you they can profile you directly or through the people you talk with. If the people you are trying to hide from don't care about getting message logs and just association with some group is punishable or can lead to punishment or death then tough luck.
And you miss the main point. practically speaking you cant self host a signal server, therefore you can't trust it fully (in a way 'fully' matters anyway). if you do it's unsupported and not recommended and you probably need a custom client to access it. That added with it being under American jurisdictions, and Signal starting as a spook project should really set off alarm bells.
Dating apps.
beyond the obvious ones? signal.
Sometimes I think that DNS providers could be, like NextDNS (I use them).
DNS providers can only see which sites you are visiting but not the actual content right?
I briefly used NextDNS but decided against using a DNS server tied to my email.
I wish there was a possible way to run an authoritative DNS yourself. The best I can do is a recursive server blah.
Yeah, that would be perfect. I thought some time ago about doing a DoT port -> nginx -> pihole -> unbound inside a cloud VM for the outside world , like this, but that would be too much work and maybe insecure.
You can tunnel your DNS requests via wireguard to your pihole server. If it has good bandwidth even the full traffic. Why would that be insecure?
Yeah, using a VPN would be good enough, but I want it to be open to the internet, without any port/config restriction, so I can access it from any device and anywhere, so the only remaining thing would be to host and open the port on a VM, only DoT and DoH, no :53 open (that would really be insecure, as DDoS insecure).
Bitcoin.
Hell, monero is the only crypto I think isn't a honeypot, since so many exchanges refuse to list it. That could just be how the government wants us to think though 🤔
It's not even that Bitcoin is a honeypot, it's that it isn't actually private at all, and through detective work a wallet can be connected to a person, as well as their inflows and outflows and what wallets they're sending or receiving money from.
yeah, the whole point of Bitcoin is literally everyone sees your transaction on there. not very cryptic if you ask me
Bluesky
Bluesky is like the furthest thing from a privacy app, which it doesn’t even claim to be.
Look what autocratic(ising) governments don't do shit against. And what their opposing governments tell about the autocrat(ising) ones.
Those are more likely to be honeypots imho, as they claim to be one thing but likely work for another.
COINTELPRO (or as I'd rather label it, SOWTERROR), showed that the US government heavily opposed mutual aid and class solidarity, and combatted the Black Panthers, by defaming them, inserting gun control laws once they started police watching, bombing their homes and so on.
That said... let's take an example. Correct me if I'm wrong, I'm working from my memory here.
DeepSeek is open-source, while ChatGPT isn't. DeepSeek does however, participate in censorship (which Muskrat's models also do, just in a different direction). DeepSeek is much more efficient, threatening Western models.
And imho, the latter is why many Western governments oppose using it. They say it's spyware, but how can we be sure ChatGPT isn't? It's proprietary shit.
But both models are honeypots, just in their own ways. Sure, DeepSeek is open-source as it claims. But it also censors itself to unnecessary extents. Then there's ChatGPT, which claims to be secure... but how can we know that when it's proprietary and we cannot crosscheck its source code?
DeepSeek the service censors, but you can run it yourself uncensored.
