I used to buy them cheap in AliExpress, but since the new european tariff it's ridiculously expensive, you'll pay dozens of times the prize of the product in tariffs.

Amazon and other big general retailers are ridiculously expensive as well.

Local buying is not an option because in my city there is not a single electronics shop.

Buying from small specialized online shops, the best option if it were not for the lack of selection, meaning that to order all you need for a project need to hit several shops and be buried in shipping cost.

What do you hobbyists do since ali is not an option anymore within the EU?

I'm looking for something that just gives good results with their own engine, not proxying google or yahoo, without LLMs being involved.

I've used startpage and have my own searxng instance self hosted. But both use mostly google results with just a nicer frontend.

I was looking Kagi as an alternative. But they seemed to move to become another AI company, bundling their subscription to the search engine with a subscription to an LLM, so I'm not paying for that.

Imagine you would see a fact that contradict your deepest believes, the core of who you are. This fact is uncontestable, is evidence in its purest form, is something that happened and you know with 100% certainty that it happened. But this fact also is against what define you as a person, your most profound thoughts. Either political, religious , philosophical or moral believes, whatever sits deepetin you.

Do you think this evidence would make you change your mind, or not?

I you've got an old piece of hardware with nvidia hardware you know the pain. Nvidia stopped supporting 390 drivers back in 2022. Meaning that a few years later, as new kernels rolled in people will start being unable to install those drivers. Now in 2026 AFAIK there isn't any maintained distro that keeps supporting a kernel old enough to work with those drivers. So good bye old nvidia gpus.

People here had several paths:

1.- If you got iGPU you could use that, less performance, but usually still works. 2.- Use nouveau drivers. Nouveau it's not very performant now, with old hardware is even worse. This should be just the last resource as the experience would be miserable, my gpu even struggled with some desktop environments using nouveau, while igpu worked just fine. 3. Switch to windows. Windows usually keeps better retrocompatibility, and while the drivers are still unsuported, you can install them no problem in windows.

I decided to move to windows 10 ioT LTSC, with support until 2032. Because I had an old laptop that I still liked to use for some lightweight gaming. Nouveau was completely unable to perform good enough for any game. iGPU worked better, I was able to play some games, but still I was missing the extra performance the dGPU gave to me. So I, sadly, moved to Microsoft OS.

But lately I've been thinking, what if there were a solution? And keep looking deeper. And I found it. Te DKMS for the drivers could be patched to work with modern Linux Kernels. But it was DIY. There used to be some maintained projects and repos, but most are now archived. The only thing I found still maintained is this AUR repo:

https://aur.archlinux.org/packages/nvidia-390xx-utils

Great, some people managed to get it working! My day was saved! But.. it was AUR, which means Arch or derivatives. And I really didn't want to go that route. This laptop is just for media and playing, all through GUI from my couch, I don't want to be messing with the keyboard. And I knew that with Arch, specially using AUR patches that was unlikely to go smooth.

I had always use linux mint on this laptop and I was very happy, so I though. why no replicate whatever AUR is doing in linux mint? And I got to it. I installed mint, and lock the kernel to 6.8 (with support until 2029) and starting following the AUR to patch the nvidia drivers manually. But the kernel expected in arch is not exactly the same as is it's in Mint (which uses a Ubuntu kernel with backported security patches). So I hit a wall, it didn't work. Then is when I found the second most useful resource in this journey:

https://github.com/earldodd/nvidia-390-kernel68-patches

A guide on how to manually patch the DKMS, specifically made for linux Mint. It didn't include any ready to go binaries or scripts, but it was very well explained. Following that I managed to make what I thought it was impossible. And I finally got to see my old laptop running the nvidia proprietary drivers again (I almost cried here).

So yes, it is possible. On linux forum most people said it was not possible. That it could not be done. But it is. And if you want to do it I found those two resources the better path to achieve it.

Now we arrive to the bitter end. I think while possible I won't finally keep this config, and I probably just keep windows 10 ltsc and after that loses support switch back to linux mint and just disable the dGPU. why? Because the process to patch the driver was painful, hard to replicate, fragile, and led to a lot of missconfigurations and broken things. I spent two full days on this. I don't want to do it again each time an update messes something up. Maybe in the future I would try to create some installation script once I figure it out who to fix all the secondary issues that emerged with this Frankenstein, but by that time probably kernel 6.8 would be discontinue and I would need to start again for the next Ubuntu LTS kernel, a little discouraging, but who knows.

I just wanted to share this penguin journey with all of you. And share those resources for anyone in need of them.

UPDATE: Finally I decided just ditch the gpu all together. Installed my old good reliable mint and blacklisted nouveau. A little lost in performance in some games, but I think it will be worth it.

Let's call the noisy coworker NoiseGuy for short.

Situation, I work in a open office, in the same space there are some rooms with actual doors. NoiseGuy works in one of these rooms but he always keeps the door open.

NoiseGuy is very noisy about everything. He is constantly stapling papers at full power, whistling, yawning out loud, singing, and sometimes he just shout up some words out loud like "WEEELL!!!".

I'm quite sensible to noise so I thought it may be just me. But I've asked other coworkers and everyone is sick of it. When he shouts specially loud everyone look at each other.

I did some research and he doesn't even work in my department. He apparently got demoted here for having issues with his old department and being sent to my office was, apparently, his punishment.

He knows he is noisy, and I think he is doing it on purpose to bother everyone around him, that's why I also think he keeps his door open. Everyone in office have decided just not to acknowledge him very much in a "do not look at the monsters" way. But he is just not stopping, it's been years like this.

I don't want to create a bad ambient in my workspace, but I'm starting to have a bad time each day when I hear him. I don't want to switch jobs as I have otherwise a great job.

Options I have thought:

1. Talking with my boss. My boss is not his boss so they have no power over him.

2. Talking with his boss or human resources. They were the ones who sent him here, I don't think they would be doing more as he is impossible to fire or have any other work-related punishment.

3. Talking with NoiseGuy. He has apparent personality issues. I don't want to mess with that. Most likely he would be even more noisy knowing he is bothering someone.

4. Having a meltdown and going to his office and slam the door close. I fantasize about this, but he will just open the door again and I would be the crazy one in the office.

This is not about any specific case. It's just a theoretical scenario that popped into my mind.

For context, in many places is required to label AI generated content as such, in other places is not required but it is considered good etiquette.

But imagine the following, an artist is going to make an image. Normal first step is search for references online, and then do the drawing taking reference from those. But this artists cannot found proper references online or maybe the artist want to experiment, and the artist decide to use a diffusion model to generate a bunch of AI images for reference. Then the artist procedes to draw the image taking the AI images as references.

The picture is 100% handmade, each line was manually drawn. But AI was used in the process of making this image. Should it have some kind of "AI warning label"?

What do you think?

Reminder: This post is from the Community Actual Discussion. You’re encouraged to use voting for elevating constructive, or lowering unproductive, posts and comments here. When disagreeing, replies detailing your views are appreciated. For other rules, please see this pinned thread. Thanks!

I think one of the issues with online arguing, from most takes on it, is that the main reason people have to argue is to spread an idea. Whether it's by convincing the opposing part of the argument and making them change their mind, or by changing or reinforcing the mind of anonymous readers of the argument.

Most of the time this leads to one of two conclusions: If someone tries to change the other person's mind they will, most likely, find themselves hitting a wall, which will lead to frustration, disinterest, or personal attacks once it's seen that the other person will not change their mind. If they do not care about changing the other person's mind and just want to make clear that their own position is the right one to have, then the argument becomes a game of winning and losing. This could be achieved by many ways, depending on the context, it could lead to insulting and trying to put group pressure (via downvotes for instance) to make the other person's opinion seem as the "bad" one. Or via creating a game of rules, and play that game better to become a winner. Please excuse the small attack I'm about to make on this very space, but part of this second approach is the rules of debate, as in consider arguments without sources, emotional responses, or fallacies as losing points in the game of arguing. And often when the other part falls into one of this issues the goal quickly becomes to point out all this "faults" the other person made, so they are clearly shown as the loser. Don't get me wrong, it is important to argue without fallacies, and to be able to prove any statements that one's make. But I don't think anyone gains anything when the argument becomes a match on who is able to ask for more sources, link more articles and identify more fallacies.

That being said I'm going to just link some literature that support the basis of this statements. Can Arguments Change Minds? . This article goes into great lengths to show something that's easily seen when arguing online: People don't change their minds from an argument. The process of changing someone's mind is very complex. The article explains some study cases where people from extremist backgrounds changed their minds over time, in a context of discussion, but it's stated that this change had a lot more going out that just a proper intellectual discussion.

Why bother then? In my opinion, the best thing we can get when arguing with someone whose opinion differs to our own is to understand them. To find out their way of seeing things, their motives, their reasoning. That's a great value. And to get this often we need to let them talk the way they want to talk, this tend to lead to some undesirable things, like mentioned fallacies, unsupported claims or straight up bigotry and name-calling. But I think that it is still valuable knowing if that's their only reasoning, or trying to push past those to see if there's something more in depth about why they don't agree with us. But, ultimately, focusing the discussion in getting a win, will often make us miss a lot of valuable information that we could have gotten if we just saw the argument as a way to understand the other person, and of course, to understand ourselves. And not only for us to understand them, but to them to understand us. Explaining our point of view in the clearer way possible, and focussing not on winning when we talk about our opinions, but on showing why we have those opinions. To be able to reach a point of "I don't agree with you but I understand you".

Of course the big elephant in the room here is that taking this approach to it's logical conclusion would mean letting some people express ideas that we don't want to be expressed. The obvious example here is hate speech. Should hate speech, or extremist arguments be allowed, and discussed? If allowed, what's our goal when engaging into an argument with them, to convince, or to understand and make the other part to also understand us? This is where I'm more torn apart, as the logic of this reasoning leads me to believe that the best is the later, but it confronts with everything I've learn about how to deal with hate speech and dangerous ideologies until now. Thus why the (OPEN-ENDED) tag, and why I hope for anyone to jump and give their opinion on this.

This is not a question about if you think it is possible, or not.

This is a question about your own will and desires. If there was a vote and you had a ballot in your hand, what will you vote? Do you want Artificial Intelligence to exist, do you not, maybe do you not care?

Here I define Artificial Intelligence as something created by humans that is capable of rational thinking, that is creative, that it's self aware and have consciousness. All that with the processing power of computers behind it.

As for the important question that would arise of "Who is creating this AI?", I'm not that focused on the first AI created, as it's supposed that with time multiple AI will be created by multiple entities. The question would be if you want this process to start or not.

I cannot stand google news any more, too much spam, clickbait and advertisement. So I decided to try to selfhost an RSS aggregator to make myself a news feed that I would be comfortable with. Being RSS such an "ancient" thing I thought there will be many mature systems, but I'm not sure that's the case..

As far as my investigation goes there are two main options out there** TT-RSS (tiny tiny RSS) and FreshRSS**. There seems to also be miniflux but it supposedly have very few features.

So I tried the both main ones and I ended up kind of disappointed, I hope that I'm missing something. My requirements are:

1-Have a nice interface, card view, phone friendly. Basically being able to look the same as google news looked. So both have a pretty dated interface. And terrible responsive UI for phones. I was kind of able to make a "card view" with TT-RSS but looked hideous and didn't really work on phone screen, also applying themes broke TT-RSS, this will be recurring theme but it looks like TT-RSS is constantly breaking a rolling release system makes it very unstable and many plugins, themes and third party apps don't work right now because some new update broke everything. So native theming wasn't going to be a thing, so I tried third party apps. I found many that worked with FreshRSS and settled on Feedme, it looked exactly as I wanted, great. One point for FreshRSS. Feedme was supposedly compatible with TTRSS but I could not login, I have the suspicion that one update broke integration. I'm not even try to attempt to ask in their forums as I see that some time ago somebody asked the same question and got banned from their forums.

2-Being able to filter or prioritize feeds The problem is that I would love to suscribe to very diverse feeds, some would post maybe over a 100 post per day and others maybe one post every week or even month. So if let everything by default the former would flood the feed and I would never see the post from the little feeds. Here both offer categories that I could use but ideally I would love to have a curated main page. FreshRSS supposedly have a priority system but it seems quite simple and not effective for my needs, AFAIK you can put some feeds in "important feeds" but it only would show those feeds in that category then. TTRSS does have an advance filter system that is complex enough and with some fiddling I think I could make a set of rules that satisfy my needs. One point for TTRSS.

3-Being able to suscribe to any feed or even scrape webs that doesn't provide feeds. Here FreshRSS wins, I have zero issues subscribing to everything I wanted. With TTRSS I couldn't even subscribe with some pages that did provide with a feed, even if it was in an unconventional way. TTRSS devs say that is the webpage problem (even if FreshRSS had no problem with it). Here another point to FreshRSS.

And that is it, I do not exige that much. But I wasn't able to find a system that ticks those three checkboxes. FreshRSS was so close. But unless I am missing something you can't really create a curate feed that prioritizes and sorts feeds and posts in the way you can do with TTRSS sorting, if there is a way please let me know. And without that the whole thing becomes useless from the flooding feeds. And while I'm in love with TTRSS filters and sorting system, the whole app seems to unstable and with so many bugs to be usable, at least in my desired usercase (and I've seem many people complaining about TTRSS updates breaking things all the time).

My two main questions are:

-Am I missing some other self-hosted app that could do all I wanted?

-Am I missing some FreshRSS feature or extension that could curate a main feed with my own rules?

Any thoughts?