/tech/ - Technology

yells_at_cloud.jpg
[Hide] (27.6KB, 400x300)

>>19083
>search engines got so bad
Even before Google was censored, even prior to that before SEO partly overwhelmed Google, no search engine was ever anywhere near as good as AI at sifting through the innertubes.

That said, the conflation of "AI" with "panopticon mainframe" is infuriating, hopefully hardware will become powerful enough soon that even the swarms of mobilesheeple will switch to local models.

>>19082
I don't really need it to look up things, I am just lonely and it treats like a friend especially when I am drunk or high as fuck 
>>19083
It can see things if you give permission to open your camera

I need to stop drinking before the Ai knows about me more I know about me

jahy_tears.jpeg
[Hide] (257.5KB, 1539x1600)

Last night I dreamed that the OpenPOWER Foundation finally figured out how to make people buy Power ISA workstations: they started calling them Jahy processors, and weebs began buying them en masse to have Jahy-powered computers.
I wish it was real.

>>19126
Awesome dream.

d9793bd740b4cc584a878b5017f8e884.png
[Hide] (294.9KB, 800x750)

>>18986 (OP) 
I am not capable of having strong emotions towards humans. Or most subjects for that matter.

Johann_Otto_von_Gemmingen.jpg
[Hide] (106.6KB, 247x320)

>>18990
What, pray tell, do you feel strongly towards?

fa59fc54ae0b3e04fca597f040de5825b9f491819187d0a16f00158bc2aef8e8.jpg
[Hide] (90.4KB, 1400x700)

>reads from a script
>"YOURE A NERD"
>"YOURE A NERD"
>"YOURE A NERD"
Yes... Thanks fag.

Hate is for cucks, I just love what I want and if I dislike something I just ignore, full apathy.

d9bc1c0354f05259f2fb5794914cafa7.png
[Hide] (12.7MB, 2500x2100)

>>18991
Programming languages, code style, decomposition approaches, architectural decisions, things that can directly affect my day-to-day activity, but even then - only comparatively.

>>18961
hey fellow hungarian! (or you just find a hungarian book lol)

smugterry.gif
[Hide] (2.9MB, 200x234)

>>18964

spurdo_magyar.jpg
[Hide] (28KB, 479x478)

>>18998
If he is a fellow magyar, then there are 3 of us here. Which would be quite shocking.

>>19001
Kicsi a vilag :D

>>19011 (OP) 
https://en.wikipedia.org/wiki/PeerTube
>PeerTube is a free and open-source, decentralized, ActivityPub federated video platform. It can use peer-to-peer technology to reduce load on individual servers when videos get popular. 
>Started in 2017 by a programmer known as Chocobozzz, development of PeerTube is now supported by the French non-profit Framasoft.[4] The aim is to provide an alternative to centralized platforms such as YouTube, Vimeo, and Dailymotion. 
>As an ActivityPub platform, PeerTube is part of the federated network known as the Fediverse. 
>Each PeerTube instance provides a website to browse and watch videos, and is by default independent from others in terms of appearance, features and rules. 
>Several instances, with common rules (e.g. allowing for similar content, requiring registration) can form federations, where they follow each other's videos, even though every video is stored only by the instance that published it.
There's already an off-the-shelf solution, the problem is that you'd have convince the teeming masses of normalfags to use it.

>>15638
Missed: multimeter for connectivity check

>>13234
helpful post

ClipboardImage.png
[Hide] (19.4MB, 4051x2315)

ClipboardImage.png
[Hide] (50.4KB, 764x560)

ClipboardImage.png
[Hide] (62.1KB, 748x399)

https://en.wikipedia.org/wiki/Eurocard_(printed_circuit_board)
>Eurocard is an IEEE standard format for printed circuit board (PCB) cards that can be plugged together into a standard chassis which, in turn, can be mounted in a 19-inch rack. The chassis consists of a series of slotted card guides on the top and bottom, into which the cards are slid so they stand on end, like books on a shelf. At the spine of each card is one or more connectors which plug into mating connectors on a backplane that closes the rear of the chassis.
It is a physical format only, so you can use whatever connectors with whatever hardware you want, and 100mm version is pretty close to standard PCI Express card sizes. A 16x PCIe connector is also less than 100mm long, so you could design Eurocards with that connector. Maybe it would be even possible to design Eurocards and PCIe cards in tandem, as the main difference would be placing the connector either at the end or at the bottom. Still, what I am really thinking about is how 10" racks are getting popular: you could make a 10" 3U PCIe-Eurocard that could be used to house anything from a large variety of SBCs to gigantic RAM cards to ridiculous SSD farms. I could see it being a hit with certain hobbyists if you turn old systems into Eurocard SBCs, so that you could plug an Amig

>CPU with integrated GPU and unified memory
>but there are also RAM slots
>by default it prefers the unified memory but uses the extra memory (if present) pretty much like if it was swap memory
>and if the GPU needs all the unified RAM then it uses the extra RAM for the CPU
<or to turn it around: you can use the VRAM as generic RAM, so you can run the system even if you don't plug any RAM sticks into it
I know that I am retarded when it comes to hardware, but could this work?

>>18989
Intel did exactly this on some Haswell CPUs with Iris iGPU, the eDRAM served as both VRAM and "L4" cache.

>>18976
>just keep an entire separate OS installed with an old glibc for building your software bro
>need new libraries or toolchains with your old glibc? just compile them bro just fucking compile your entire environment from source bro
>oh your program broke anyway because of glibc symbols? well that's your fault bro we all know glibc never breaks lol

>>18979
It doesn't matter how you ship your libraries, the failure point is always glibc. You can link against any glibc version you like and your program will still break sooner or later when glibc updates on the user's PC. This has been glibc's routine for years, see >>18465
>This shit isn't rocket science
LMAO when devs are willing to put up with JS cancer and ship applications in a fucking browser instead of dealing with your open sores libc then you know it's rocket science. Stop sucking GNU's cock.

oswald_mosley.png
[Hide] (1.8MB, 820x1024)

>>18981
>Compile your program for your host os with dynamic libraries on an older version? Nah just compile it with dynamic libraries on a completely different OS and run it through a compatibility layer 
>Oh your program broke anyway because of a Wine regression? Well that's your fault because we all know compatiblity layers and emulators never have regressions
I ain't defending glibc's stupidity: I'm calling you a hypocrite.

>>18982
Wine regressions get fixed, glibc regressions usually don't.

4ddd72e341d9952a58ff12bd9bafbe9cdcddbcdce2e49a7a77d01742bde427ce.jpg
[Hide] (88.8KB, 1069x965)

>>18983
Most distros ship a patched version of glibc with DT_HASH kept intact. As retarded of a decision as that was upstream, it has effectively been fixed downstream for the vast majority of Linuxfags.
This is what usually happens whenever upstream glibc breaks something. It is obviously a serious problem to have a libc whose devs repeatedly pull that shit, but in practice it gets "fixed" until glibc ends up replaced. Telling people to not write Linux software over this is like telling people to not use the Win32 ABI on Windows because sometimes Microsoft does things which break your program, except the Windows user is a lot more helpless than the Linuxfag in that situation.

Screenshot_20260531_041252_Termux.jpg
[Hide] (897.5KB, 1920x1200)

>>18091
I've been running into issues and I want to update others that, even though I'm sure most would never even bother with a setup like this, I don't recommend it as a main setup and especially not as your only setup, a lot of software isn't available, and even the things it's supposedly so good at that the cons don't matter, like browsing, aren't the best because the versions we have here of internet browsers are worse and often have compatibility issues with some sites, even when using the "use desktop version" button or something. I still love my setup, but I won't lie and pretend it isn't a very limited one still.

>>18808
>some of the BSDs are also working on their own X(enocera) successor, Arcan.
Arcan is not tied to any of the BSDs as far as I know. Or do you mean that some people try to port it to a BSD?

__2b_and_tux_nier_and_2_more_drawn_by_seori__fd1ec3cb692fea3397b3db68ff23ec7a.png
[Hide] (4.2MB, 2728x3983)

>>18808
>If you surrender the ability to run your main environment on something at least competitive with modern hardware, and access modern networks with it, you've lost the game. The moment we're squeezed into retrocomputing on a dwindling supply of eWaste quarantined from the normalfags on our own LARPnet, we've been neutralized.
That's unfortunately an important point. "Just use old hardware" and "just pirate all software" are not solutions to problems, but consumer responses. They aren't sustainable on their own, as they don't create, but cling to what others make without financially supporting them. Relying on old hardware depends on that dwinding eWaste supply you mentioned, and exclusive piracy depends on the very fragile community and technical/legal context surrounding software piracy. 
Actual solutions, on the other hand, involve providing ways to create and support those who create. You also ideally want to address the problems that push people to use cancerous software or centralized web services in the first place as opposed to you just acting dismissively towards them. Think Gaben's "piracy is a service problem," but replace piracy with whatever cancer you wish: systemd, webshit, Rust, Discord, or Steam itself.

I gotta head off now, so I can't wrap this up or expand on it the way I'd like to, but I'd suggest looking into the concept of intermediate technology. Al

>>18815
SystemD, even against Pootering's ambitions, is still being forked into separable libs, most relevantly eudev. Admittedly this means these forks are powered by a steady supply of autistic spite that could hypothetically wink out. As for Wayland itself, its Linuxisms are little worse than other freetard/*N*X stuff that has been needlessly Linuxized over the years, such as when DRI/DRM/KMS/GEM/EGL was Hoovered into the Linux source tree, or the broader implementationisms enforced by decades of convergence on glibc.

>>18816 
I mean that Arcan has made conscious efforts from the start to avoid absorbing gratuitous Linuxisms, including implementations for OpenBSD, FreeBSD, Plan 9, etc.

>>17062
Gentoo because it had always been about your personal choice. Portage is very flexible. If you don't want to build software yourself, you can use binary packages nowadays. The next best thing is either Void Linux or Alpine Linux. Alpine Linux is very good but since it uses musl, you may need to use Flatpak, distrobox or gcompat.
>https://wiki.alpinelinux.org/wiki/Software_management#Running_glibc_programs

>>18820
I thought the eudev project basically stopped because it was too much work for the dev to keep track of all the bullshit changes they were making? People seem to be using elogind now and running systemd udev.

>>18044
>>18162
That's actually a really good idea. It's similar to my Galatea, but more food-based. If you put the batteries near the "feet", the battery weight can be an advantage with a counterweight. I think you should pursue it!

>>18169
It's not waiting, it's work, a lot of hard work. 
Also going schizoid can be a great time and you can hallucinate waifus at a cost to your sanity. Been there done that. miss it a bit to be honest, used to just listen and play with the music in my head, messes with getting things done and interacting in society though.

Galatea_v3.0.9_Wide.jpg
[Hide] (100KB, 1792x1024)

Lieutenant_Galatea.JPG
[Hide] (715.7KB, 1690x2112)

>>17765 (OP) 
Galatea has been updated to v3.0.9. She now has a new body with more aesthetic ratios, and new accessories.

>>18170
That robot is not going to build herself, anon!

>>18170
>>18926
Exactly. It's a new frontier.
Visit /robowaifu/

The fact (real)  political discussion on the internet is the most poisoned well in the history of mankind should tell you something.

slack-penguin-bw.gif
[Hide] (4KB, 258x303)

>>18917
The most valid concern I've seen is that politics is a very big stick that isn't always appropriate for the scale of your problem. The other thing is that if you aren't careful, you can give your party crazy powers to deal with one problem which then get used against you when you're kicked out.
There's obviously a place for it, but it's something you gotta be careful about. Otherwise you get crap like the current wave of tech-illiterate boomers destroying the web because they have no idea what "just make everything check your ID, how hard can it be?" implies.

I mean in general. A good example of what I'm talking about is brown people insisting to use usraeli social media like whatsapp and X while bombs are turning them into meatfetti. How stupid can people be? They had their actual dicks blown off by their sail foams a year ago and they keep using them. I saw some brown woman in lebanon say "how can kids learn without power for their phones?" INSANE. What is a pencil and paper bitch?

It's some kind of animal passivity. Animals don't plan ahead in a fight and neither do 99% of people on the internet. Nobody forces you to sit and do nothing while LLMs turn the entire web into the Truman Show. But asking people to use their own hands and brains to make their own handmade computer programs to talk to other poeple is like asking a monkey to fly to mars. Even though LLMs put this ability into everyone's hands because you can just say "HOW DO I TCP YOU FAGGOT ROBOT" and it will explain in perfect detail without having a tism rage like the faggots on stack overflow.

First off, internet /= web.

But also, the web has no practical purpose to begin with, an entire operating system specification built from several worst in class standards, possibly the worst programming language ever made (and oh boy does this title have competition), and 2 implementations that compete for worst piece of software ever written, all just to send and receive text and files in the most devious rube goldberg machine ever devised is a monumentally retarded idea from the start. 

And it's so blatantly obvious from a glance that anyone who didn't already know this before I made this post either lacks the beginner level knowledge in network protocols, programming language design, serialization formats, etc, or is not right in the head, which scares me a lot, because it means most programmers are clinically retarded if they did not form an alliance and vow to never write nor allow anyone else to write any javascript, json, HTML, etc upon first seeing those out of sheer disgust.

peak_webpage_design.jpg
[Hide] (117.7KB, 800x600)

eyecandy_elpher_gopherhole.png
[Hide] (493KB, 1080x1920)

>>18923
>the web has no practical purpose to begin with
To be fair, back in the early 90s it must have seemed like a great idea to give the writer of the document the power to specify the colour of the background and the ability to embed some pictures into the text. But as those horrible late 90s webpages showed, it only really gave people the power to waste bandwidth with unnecessary clutter. And by the 2000s it became obvious that you could also use it to track people and to build ever more centralized systems.

And in hindsight, the very idea that someone else should decide for you the exact appearance of a hypertext document is flawed. How could some random faggot know better my preferred colour scheme and font than I do? Not to mention that I like it when I can both read text and stare at fat 2D tits at the same time, which is not a feature supported by the average website. At least by now many websites have both a light and a dark colour scheme, but it's not standard, and many of those themes are quite horrible. Not to mention that websites have to support a practically infinite variety of resolution and aspect ratios, so the more elaborate and specific a website's design the less likely that it'll work as intended.

>>18138
At least you're (I assume) looking to make a technologically mediocre implementation of your (no doubt brilliant and inspired) new ideas, instead of retreading old ideas with as much bleeding edge technological optimization in implementation possible as your first project, like this guy:
https://invidious.nerdvpn.de/watch?v=CJ94gOzKqsM

>>18138
>primal revised simplex using sparse matrix/vector arithmetics and Bartels-Golub method to iterative LU decomposition
I am going to say this as harshly as I can so you really get the point:
Coming as someone who studied numerical linear algebra let me tell you that there are graduate students in numerical linear algebra who would have no idea wtf you just said. You are so deep inside either imposter syndrome or Midwest sandbagging that I don't know what I can possibly tell you at this point other than your so-called performance issues are deathly serious psychological and esteem issues as opposed to intelligence.

>>18147
You're giving me too much credit. I am exactly retreading something that was thoroughly explored for optimisation in the 1990s and more-or-less solidified in set of "state-of-the-art" industry standard techniques by 2010s, at least as far as LP optimisation solvers go - most variations past this point seem to focus on heuristics for picking best-suited scaling and pivoting techniques for a particular problem and avoiding degenerate states. And my domain is as trivial as finance.

>>18218
I only participated in uni for a limited time but from what I've witnessed, merely being seen at every lecture scores you higher in than actual understanding of the topic and it's tangents, so at least in some extremes it may be testing the bureaucratic/conformation aptitude more than anything else. While I don't discount that someone brilliant may choose to stick around, it is doubtful that act of graduation alone is a meaningful achievement.
Thankfully, scihub and researchgate are around, so there is no need to be academic environment to have access to whitepapers, but while I can make a working implementation after cross-referencing the arbitrary notion syntax for several days, the understanding is only tentative and amounts only to consuming and regurgitating essentially, being entirely dependent on someone truly impressive to kindly share their find

>>18128
Update: I'm sticking with Kotlin and KorGE. 

As to Rust and any game engine related to it, I'm not interested. I'm taking up the D language. I think D is a very good and underrated language. The most complete engine offered in D is the Dagon engine, spearheaded by a very productive and focused Russian man who's spent 10 years building it by nearly himself: https://gecko0307.github.io/dagon/

So, as of right now, my game dev languages are Kotlin and D and my game engine choices are KorGE and Dagon.

I came upon this article discussing how during the 90s people would finger .plan files for direct access to the daily musings of gamedevs like John Carmack.

https://irclol.com/digital-prophets-doom-plan-files-tech-journalism/

I explored using finger and making my own local .plan file, but I'm a little confused how this was actually used in the 90s were they web facing and not internal? What was it really like, how did anyone know who to finger?