The admin of Blahaj is openly interested in exposing trans people’s alt accounts and outing them on their mains. And somehow it’s the biggest trans instance. We need a community and admin reaction in favour of defederating people who do that.
But why? Is there a compromise taken on privacy in favour of visibility and mass adoption of whatever fediverse client they’re using? I don’t understand this, especially since I also find the strongest advocates for privacy right here.
A lot of Lemmy adopters joined with rose tinted glasses, and came with a lot of good ideas, like getting data out of the hands of big companies, making it easy to access it (as Reddit locked down APIs), etc. Which is all good, but a subset of them believe “not officially belonging to one company” is good enough. As for how your data is handled online, a subset of them believe nothing can be improved, and a subset believes it shouldn’t be improved because your data shouldn’t belong to you at all.
And Lemmy is made up of all sorts, so there’s overlap between Reddit refugees and diehard fans. That interaction is a lot more implicit here, but the friction is a lot more visible on sites like Mastodon where similar privacy discussions have been happening.
I’ve not seen any of these arguments. Though it may be all downvoted to hell and back.
My main gripe with adding privacy features to Lemmy is that the whole point of Lemmy is that all data is already publicly available and for Lemmy to continue working the way it does it’ll need to remain that way. And because of that there’s nothing that can be done to stop bad actors setting up an instance and selling all the data they collect.
At least in the EU (and UK to a lesser extent) no major corporation would be able to get away with selling that data, so the spent man hours on allowing privacy settings would be wasted time.
It doesn’t necessarily need to remain that way. For example,we should have the option to make our profiles private. We should also be able to create pseudonyms for content we submit. The content will still be federated, but not necessarily linked to one user ID
The only privacy setting I can encourage on any social media site is don’t share private stuff about yourself and never link to your account from other accounts
That is part of the problem though. Proper privacy allows you to express what you want to, without self censorship. The issue is not: don’t speak about x, but rather: speak about it and feel comfortable that you can do it in a safe environment. I fully agree with the account linking though
Well, how about having a local API and have no calls at all to your cloud infrastructure? Probably too easy and you cannot lock people into your ecosystem.
From any practical standpoint, this makes so much sense.
Sometimes my Tesla fails to unlock for some reason and I have to disable my VPN and then stand next to it like a God damn idiot for 10 seconds while it calls it’s servers in fucking California to ask it to unlock my car.
As if I needed yet another reason to never ever own a Tesla.
My car has this crazy technology in it: You can stick the key in the door and twist and it’ll unlock. Even if the network is down or the battery is dead. Arcane, right?
Hell yes! My sister-in-law has your same year but the diesel version and that thing is a champ. It’s rated at 45 mpg on the highway but she typically gets 50+, even with nearly 200k miles on it.
I had a 2004 1.8t Jetta for 12 years but I swapped it for a Prius. I love the Prius features and fuel economy but I miss how damn quick that my Jetta was, plus I loved the interior color scheme.
Haha yeah there are other, more reliable methods but the “phone as a key” is also super convenient when it works properly, which is most of the time. It just would be a lot smarter if it worked locally.
…Or if there were an alternative option that didn’t rely on software and electronics is my point.
Cars have had electronic remote keyless entry for decades. It’s not new. Some of them even have phone apps that duplicate that functionality. No one but Tesla has been stupid enough to remove the keyhole, though.
I understood your point. My point is those electronics make it more convenient to use. Would I appreciate ALSO having a physical unlock mechanism? Sure. It also increases the attack surface.
Cars have had electronic remote keyless entry for decades.
I think it could definitely be possible to do locally, and I wouldn’t want a car where I have to connect to servers to connect to it. But I am also not sure I want a car that can be opened with a command on the car itself. The code to access your CAR being stored locally on the car itself, with no server side validation, does seem kinda scary. It’s one thing for someone to manage to get into your online login where you can change the password, it’s another for someone to literally be able to steal your car because they found a vulnerability. It being stored locally would mean people would reverse engineer it, they could potentially install a virus on your car to be able to gain access. Honestly, as a tech guy, I don’t trust computers enough to have it control my car.
The issue you are experiencing likely has nothing to do with the VPN. Network connectivity is not needed to unlock the car. I have been in places with no cell phone signal and it still works.
I do sometimes experience the same issue you are. If I wake up my phone, then it works. So it may be working for you not because you disabled the VPN, but because you woke up your phone and it then sent out the bluetooth signal to let the car know you were nearby.
It’s a bit of both! Certain commands to the car can be done locally via Bluetooth OR via Tesla servers. The tricky bit is that status always comes from the server. If you are on a VPN that is blocked (like I use NordVPN and it is often blocked) then the app can’t get status and as long as it can’t get status it may not even try a local command. It’s unclear to me under what circumstances it does local vs cloud commands, and it may have to do with a Bluetooth LE connection that you can’t really control.
When you don’t have service, or you’re on VPN, it may be worthwhile to try disabling and reenabling Bluetooth. I have had success with this before. If you’re using android, it seems like the widget also uses Bluetooth, so you could try adding the widget to your home screen and using that. You can also try setting the Tesla app to not be power controlled, so it never gets closed.
Either way, there’s a definite engineering problem here that feels like it should be fixed by Tesla. But I can at least confirm that, even in situations with zero connectivity, you should be able to perform basic commands like unlock and open trunk without data service.
I’m glad the people with this device are getting traction on using it with their HA, but holy hell this is a complete non-starter for me and I cannot understand why they got it in the first place. There’s no climate automation I would ever want that is worth a spying device connected to the internet and a spying app installed on my phone.
Probably more. Your app can use the local API then as well. And AWS is insanely expensive, especially if you forget to block log ingestion to Cloudwatch (ask me how I know).
I’m cynical so I assume they are turning a profit selling user data. So the lost money is not from AWS expenses but from not having installed apps to steal more data.
What you’re describing is an issue with all of social media. While your concerns are valid, I don’t see your arguments as privacy issue. I honestly prefer post and comment history being transparent and accessible. It’s much like Reddit and this format fits much better with an open forum style of platform.
Don’t post private information and it’s a non-issue.
Also, can’t you just delete posts and comments like on Reddit?
Also, can’t you just delete posts and comments like on Reddit?
Nothing ever dies on the Internet. With the federated nature of Lemmy, it’s possible for deletes to not sync across instances, especially if there’s defederation that happens.
If you’re not running your own server privacy policies are not even worth the pixels they’re presented on.
Literally, you’re just taking a random person’s word for it (whoever the admin is). A website is a black box, you have no idea what’s going on on the back-end.
The only way to be in complete control of your user data is to run your own server and be literally the only user on it.
Even then, any public comments you make are, you know… public.
Ask me no questions and I’ll tell you no lies. It asks much less of my instance admins if it’s understood that my information was never private to begin with.
All your data will pass over other hardware owned by other people. The only real online privacy is not connecting to the internet to begin with.
And now we’re entering into the realm of encryption, especially end-to-end. Generally speaking, just because you’re sending information that touches other people’s hardware, doesn’t mean it’s public and readable.
Even then, AMD, Intel and now Apple CPU chips are suspected to be backdored. NIST has been slow to adapt a standard post-quantun E2EE algorithm, with some rumours of self-sabotage mandated by NSA (like they have already done in the past). The Tor network is extremely vulnerable to traffic correlation by big parties.
Encryption theoretically gives you what you describe, but in reality you still need to put a lot of thrust in things like your own hardware.
I think that’s worth considering: an open-source volunteer project requires and leaks way more data than a private corporation it’s mimicking.
It couldn’t be that one has had loads of VC funding for *checks notes… 15 years. Whereas one has been barely funded for five years and has more people complaining than adding code.
Actually, it makes perfect sense that an open source project that doesn’t have a big organization behind it isn’t going to have the same capability anywhere near as quickly. Reddit also makes money from advertising. The money for Lemmy is from donations and an abysmally small set of grants.
Hell, Matrix, an actual open source communications protocol is 9 years old and they still haven’t gotten encrypted video group chats working properly and if I recall correctly still offload a lot of that to JitsiMeet. I was using Matrix/Riot.IM (now Element) in 2016 and it was garbage that barely worked, and updates constantly broke what previously worked, etc. It took time to become better and Matrix does have a whole ass organization backing it.
For comparison, Lemmy has been around for about five years and they’ve had far less financial backing and developers contributing to the project. Matrix has governments like France and Germany lining up for services for private communications, which means they’ve literally got people paying them for the service of helping manage their Matrix servers. Lemmy doesn’t have the same advantages. They don’t have a service or ads to sell (no ads is part of the appeal.).
For what its worth, Veilid exists, if you’re looking for a better framework to start with than ActivityPub.
I remember a little while ago a thread with someone from kbin gloating that they could see what everyone was voting, and accusing the people upvoting comments they disagreed with of being bigots in a vaguely threatening way obviously intended to produce a chilling effect, and people found this surprising because that information is not public on most instances.
I basically agree with the people saying open info is just the nature of posting on a public forum and of federation, but there could be improvements, even just in awareness of what is and isn’t private.
This is a great point because in the Lemmy UI, this information isn’t shown, and you can’t even list out all posts you’ve upvoted. As most of us coming from Reddit, we’re used to upvotes being private, and probably assume it’s the same. I understand the technical reasons for having the information public, but it is not clear from a user perspective that it’s public.
What’s extra confusing is that I’ve seen people asking about how to get this information from the API, with the answer being that you can’t (I guess to protect privacy?). It’s only accessible to federated servers, but then those can do what they want with it including publishing it to everyone.
I don’t know about you but I want the companies to take self hosted and Foss solutions seriously. The fact that they are wanting to work with him is a major step in the right direction. It would be dumb to discourage companies from supporting foss.
Are they supporting FOSS, or looking to buy out the project to make it a closed in-house solution and avoid the bad publicity they created this last week?
I have a feeling that you might be misunderstanding what the actual purpose of lemmy is. lemmy has taken quite a few design decisions from Reddit which is exactly the same way. Both platforms are public places where all content is shared. Anyone using them needs to be aware of that fact. Mastodon might be a better fit for you as it is more focused on individuals rather than public communities.
What irritates me many times when I enter Lemmy is that instead of my Nick at the top right, someone else’s Nickname appears for a moment, before changing it to mine. This is a sign of an open account sharing channel, which is quite serious and should be fixed quickly. Security at Lemmy is apparently non-existent.
It occurres sometimes, I see a random nick from strangers. It means that my account obviously is públic and even shared. I will be attentive and I will try to take a screenshot, before the nickname changes to mine while Lemmy loads.
It’s not easy to catch, because it’s only a moment when Lemmy loads and just sometimes. For now I always have my eyes to the top right corner when I enter Lemmy.
It gets weird fast, because before privacy controls in the Lemmy source code mean anything, we need trusted third party verification of a server’s patch level, and security controls.
That can be done, and I think Lemmy has a shot at getting to that point, but it’ll be awhile.
In the meantime, I suspect the Lemmy developers are hesitant to add and advertise features that you can’t be sure are actually correctly enabled on your instance.
But yeah, let’s not let perfect be the enemy of moving toward better.
Edit: Assuming you completely trust your instance admin, we could start adding some basic privacy to actions taken on your home instance.
But as soon as the user starts interacting via federation, all bets are off - because the federated instance may he malicious.
I think we might see one or more “trusted fediverse” groups emerge in the next few years, with instance admins making commitments to security controls, moderation, code of conduct, etc.
So, in theory, the lemmy software could start implementing privacy controls that allow users to limit their visibility to whichever part of the fediverse their instance admin has marked as highly trusted.
But even then, there’s risks from bad actors on highly trusted instances that still allow open signups.
Anyway, I totally agree with you. It’s just a genuinely complex problem.
If all the people complaining would just contribute to the codebase this wouldn’t even be an issue.
Often, you even see the devs coming into threads like this and making suggestions, like “make a pull request.” They want more people contributing.
It’s tons of people whining, very few people contributing. Guess what? While at a certain point, adding developers stops increasing productivity, there’s a small window where adding developers does increase productivity.
If I am correct, Lemmy only has four main developers. That’s well within the range to add more developers and increase the productivity, making new features and security come faster.
So I get it, but things take time, and are complicated, which you thankfully can see.
People whinging about it in threads does nothing to change it. Donating to Lemmy’s development costs or contributing code does.
So much of it sounds like it sounds like its from less-technically-inclined people (some of its valid critique from experts, but they generally… write bug reports and do pull requests…) who just want it to be better but the only way they know how is to “bring awareness.” Well, all that “awareness-bringing” just amounts to spreading FUD.
I think we might see one or more “trusted fediverse” groups emerge in the next few years, with instance admins making commitments to security controls, moderation, code of conduct, etc.
There is now at least one system in place for admins to vouch for other instances being non-malicious, and to report suspected instances. It is called the fediseer: gui.fediseer.com
If Lemmy cared about privacy, contributing source code & opening tickets would not require opening accounts with a for-profit, US-based, closed, prorietary service owned by a publicly-traded megacorporation that has shareholders to appease & a history (as well as current) record of EEE (embrace, extend, extinguish).
I mean it took the code production of from workers for the Commons, packaged it up, & sold it back to the workers—often in violation of the license if not the spirit of free, ethical, or similar software. All AI generations should be CC0 / 0BSD licensed.
Choosing proprietary tools and services for your free software project ultimately sends a message to downstream developers and users of your project that freedom of all users—developers included—is not a priority.
Lemmy has many privacy problems that have nothing to do with public comments you make. For example, the “hide posts that you have already read” option requires that the server track what posts you have read. There is no public activity involved in reading a post. So the Lemmy server should not track that info. If that feature is to exist at all, it should be implemented purely on the client. The same can be said about subscriptions, and for that matter about voting (server should discard voting info after a brief interval for abuse detection). The Lemmy software in many ways naive about this stuff.
I don’t disagree on those points, but I think it’s the nature of Lemmy being decentralized that makes all those things necessary.
server should discard voting info after a brief interval for abuse detection
What if the server has not federated out the votes yet? Some of that stuff can get backed up in a queue. There’s definitely a possibility that votes could get “lost” on the way. Hell, that already happens, and that’s with a system that tracks them.
Servers have to keep a lot of this info to pass to other servers. If I upvote something on Lemmy.blahaj.zone, it doesn’t mean that upvote has been federated outward to hundreds of other servers yet. I would assume this is part of how Lemmy is able to keep things “organized” between all servers.
In other words, a lot of the privacy complaints come from technical limitations of how Lemmy works. Lemmy, by it’s decentralized nature, has to transfer tons of data back and forth between all Lemmy instances.
However, there are technologies that are trying to work around this kind of technical limitation. You might be interested in something like Veilid. I’m not sure about the details of putting together a Veilid-based social-network, but I’m willing to believe it’s possible.
I don’t see anything in your post that indicates any reason to track what posts a person has read. That should not be tracked at all. Reading posts should be completely anonymous.
I don’t see why voting necessarily has to track who casts the votes. But, because untracked voting can be abused so easily, I can understand deciding to retain the info for let’s say 24 hours. Hopefully that is also enough to handle those propagation issues.
Really, imho, server instances shouldn’t have a web interface at all, just an API. Web apps would make API calls to the server and reformat the response for use by the browser. The API call to read a post should not require any identifying info or require the user to be logged in. Read tracking and subscriptions should be handled by the client, and in the case of a public client (web app shared by many users), the private user info should be encrypted in case of a server breakin or seizure. The encryption key would be based on the user password and transformed to a browser cookie when the user logs in, so it is never stored by the web app. With most people using mobile clients these days, alternatively, the info can be kept completely on the client device and maintained by the mobile app.
github.com
Active