Beginner questions thread
hendrik @ hendrik @palaver.p3x.de Posts 8Comments 1,844Joined 4 yr. ago
hendrik @ hendrik @palaver.p3x.de
Posts
8
Comments
1,844
Joined
4 yr. ago
KDE is the correct answer. I guess you could also learn coding, and then any piece of open-source software would become "customizable" to you...
I don't think the internet gave particularly good advice here. Sure, there are use-cases for both, and that's why we have both approaches available. But you can't say VMs are better than containers. They're a different thing. They might even be worse in your case. But I mean in the end, all "simple thruths" are wrong.
I'm not sure what OP meant but I think none of the comments here are nailing it. If we say, we can't do anything unless it's 100% certain that all instances comply with the protocol, we might scrap the whole Fediverse idea altogether. Any post or comment or vote or deletion could be tempered with in one way or another. I mean we're clearly making an effort here to do federation. I don't really agree with the why and how of the whole discussion. My own point is, the software seems to have some bugs. Rarely, some comments and posts don't federate to me correctly, and more often than that, deletions don't federate correctly. Which seems to be one of OP's problems, but also while dealing with spam or malicious activities.
On the other hand, everyone who thinks it's super easy to just delete everything has never had a look at the consequences. Moderators and Admins sometimes need to deal with bad people, there are technical reasons involved. And bad people also misuse features. It's complicated for several reasons, difficult to get it right and it's always a balance between opposing legitimate interests.
But with that said, this doesn't apply to bugs. Lemmy should at least iron out the software bugs to federate activities to other instances properly.
Hehe, I don't know if they don't want to understand it, or if it's a lack of technical knowledge... But yes. In the digital realm, a copy and the original are identical in every way, no matter how you twist it. And you can't even properly transfer any item it in the same sense as it applies to physical items. (Unless we're talking about quantum computers or something like that...)
Sorry, this just isn't correct. Yes, you can ask for almost anything and it'll be alright and merely asking a question is completely legal.
The issue is, you then proceed to do a second step. And that is transferring the data. And that is a separate thing. You then initiate the actual transfer. Your computer actively does that. It keeps the transfer going and recieves the network packets. It literally copies them into RAM and then copies them again onto your harddrive. To make your local copy. The uploader merely reads it from their harddisk and hands it out, they do one copy operation less. Though they're still the distributor.
I think any expert witness would testify in court, that your computer as the downloader does two copy operations, at least in the technical sense of the term. And that you've ultimately also initiated the transfer as the downloader due to how TCP/IP works.
The thumbdrive example is a bit construed. I think you might get away with that, though. Unless you plug it into your computer. Because then all the copying to RAM and harddrive etc starts again. But I think just pocketing it is posession (which doesn't seem to be wrong), and not necessarily copying.
But like: how do other laws work where you live? Can you instruct someone to do something illegal and you're fine? I can't come up with anything normal, let's say I hire someone to kidnap my child/wife to teach them a lesson. Or I hire a hitman to kill my arch enemy. Am I fine dong that? It's a bit over the top. But where I live I can certainly get into trouble if I make people do something on my behalf. Which I'd argue doesn't exactly happen here. It's a bit more complicated... But your concept of law doesn't seem to make much sense to me.
Fair enough. I guess we can skip the other options then, at least for your case. The replacement isn't implemented in a very thoughtful way, I agree. For technical reasons, you can't have it your way either. A platform with tree-style comments or replies can't have a comment in-between deleted entirely, or the rest of the tree will collapse. So there needs to be some empty placeholder. Or you just can't use platforms which allow users to reply to each other, but that's more a you-problem. I agree though, if you delete it, it can't have your username or content left behind. Thanks for raising this concern. I'm not sure if anyone ever put it on the agenda for Lemmy.
Well, did the uploader push it onto my computer, or was it me who clicked on something and initiated the transfer? I'd say it's the latter. So the downloader initiated the copying process... I mean if I steal an orange in the supermarket, we also don't say it fell into my hands and somehow they're guilty...
And additionally you might find other local laws like this: https://www.nysenate.gov/legislation/laws/PEN/156.35
and generally, this is related to wire fraud, computer fraud, unlawfully obtaining information from a protected computer... Which all seem to be crimes and something people get charged with if someone wants them convicted of a crime. See Aaron Swartz for example.
In Germany we definitely have this:
and we have Störerhaftung which can also get you in trouble even if you didn't do it yourself.
In the USA it'd probably be something like https://www.law.cornell.edu/uscode/text/17/106
which says the copyright holder has exclusive rights (1) to reproduce the copyrighted work in copies [...]
Which is kind of what you do when downloading something. There will be a copy on your harddisk... And (1) does NOT limit this to redistribution, like (3)... I'm not sure how it turns out in practice. I don't follow American court rulings that closely.
Damn leechers. And doubly so. First they steal the books, and then they don't even give back to the pirates. And it's not like Anna's Archive or Libgen weren't struggling already. So Meta is just harming everyone involved.
Sure. Question is: How can we improve? Is this a symptom for another missing feature? Or do we want to not address it and just provide one nuclear option? Would you for example like a feature for ephermeral comments, which auto-destruct after a week or so? PieFed has something like that for posts. Or the ability to categorize comments so you can find them later on? Or an option to (regularly) wipe your history, so you don't have to delete the whole account...
That's why I ask for exact reasons, and not just a vague feeling about how the platform is bad... I mean it is for some edge-cases like this. And I don't see how Lemmy would improve on this in the near future. Seems some of the groundworks still don't work properly. But this doesn't have to apply to other Fediverse software.
And sometimes I struggle to relate. I for example don't post anything on social media that's very private in nature. So I don't really have the use-case where I post someting publicly on social media, but then I want it gone at the same time. I suppose we just post different things, because I can see how you wouldn't have your daily state of mind available forever.
Well, I think 1) is correct, while 2) 3) and 4) are mostly misleading. The Reasoning Models are still LLMs. Just that they emulate some inner dialogue before writing the final text. I wouldn't use the word "accuracy" in context with AI. It's far from accurate these days, no matter what we do. And in 2), it uses intermediate reasoning. But the Mechanism is still predicting tokens based on statistical likelihood. And regarding 4) also the patterns still come from data, so do the intermediary "logical" steps. And if we look for example at the "thinking" steps of Deepseek R1, it's often unalike human reasoning. I had it reason very illogical steps and outright false things, and despite that, it sometimes arrives at a correct conclusion... Make of this what you will, but I don't think "logical" or "accuracy" are the right words here. Table rows 2, 3 and 4 are oversimplified to the point where it's not correct any more. IMO
Exactly. And I sometimes find myself in the position where internet enshittification and content vanishing harms me more than it helps. So I'd like to balance this with the other side of the medal, where people might have legitimate interest to do so. But so far the argument has been "just because". And for me, that argument doesn't tip the scale to their direction. I still have tangible arguments not to over-delete. While the other side seems to be very theoretical.
Tl;dr: Yes, it's complicated.
Hmmh. I think 1) just means it has to be implemented properly. But you're right. That sounds exactly like something a developer would do. Unlink the information and at the same time add a timestamp that immediately links it again 😅
And I'm not sure about 3) I'd have to read the GDPR again. Afaik it just mandates the user is provided with the ability to do so. Not that it needs to be the default.
And 2) is kind of my question. I suppose a user who is about to delete their account, might not be super relaxed and ready to deal with the intricate details. I mean they could be pissed and want out asap. Or something happened and they need to get it over with, quickly. Either way, it's probably not the right time to bother them with 500 questions and make them learn about the consequences. Though... They need to do the right thing. Once their account is gone, and it turns out they would have liked to delete more (or less), that's not really possible any more (without manual admin intervention). So maybe it's down to: delete everything in any case, and accept that it has a negative effect on the content on the platform.
It also has to be balanced with handling abuse etc since malicious actors use the same features to cover their tracks.
But I'm probably getting way ahead of where we are. OP said deletion doesn't even propagate through the federated network correctly. So realistically, we probably don't need to bother with the details several steps down the line.
Uh, that's a complicated question. I don't know whether BLAS or Vulkan or SyCL are faster on an iGPU. I think I read many different takes on that. And I suppose it probably changed since I last tested it. People are optimizing the code all the time and it probably also depends on the processor generation and things like that. All I can say setting up SyCL is a hassle and requires like 10GB of development libraries. And I didn't see any noticeable improvement in speed. Either I did something wrong or it's not worth it on my computer. And Vulkan made everything slower on my 8th generation laptop's iGPU. But I'm not sure if that applies generally. But I'm currently sticking to the default backend, I believe that's BLAS. But again on KoboldCPP they replaced OpenBLAS with NoBLAS(?) recently and I haven't kept up to date and it's just too many options... 😅 I don't have any good advice. Maybe try all the options and see which is the fastest... Seems to me using the iGPU likely makes it slower, not faster.
Heheh. Sure. If I'm not at home, my sex-toys are up another stranger's bum. It made no sense for us to own anything anymore. My brain implant keeps me indifferent and drugged. Consuming and obedient... I'm happy.
Good use-case. Would it suffice to "unlink" the information in that case, instead of deleting it? I think that'd solve both problems. The posts and comments would stay in place for everyone to keep using them, but it'd say "by [deleted user]", so it's forgotten that you (or someone) wrote it.
I'm not sure. And we somehow need to present that to the user without overwhelming them with several options, delete account without data, delete account and unlink content, delete account and content...
Hmmh. I mean sadly we don't have an abundance of free software developers, let alone kernel developers. So in reality we just can't take them from anywhere. More often than not, it's hard enough to find one person. So I don't see how we'd get a second one on standby. But I agree. hypothetically, it'd be nice to have more than enough people working on important software projects, and some leeway.
Nice, that seems to be very usable. Now what does it cost compared to an M4 Pro or MacBook?
Thx. Seems I have some knowledge-gaps in pop-culture 😅
I suppose that line means llama.cpp tried to allocate another chunk of memory, roughly 2GB and that failed because there wasn't any memory left. I'm not sure about the details, maybe it's the KV cache and additional stuff that is required for the computation aside from the model itself? Have you tried lowering the number of layers to offload to the iGPU and see if that works? Like lowering the value to
-ngl 20might leave additional space for other important things.