FuckBigTech347

joined 3 years ago
[–] FuckBigTech347@lemmygrad.ml 2 points 17 hours ago

I don't use Gentoo but I still frequent the Gentoo Wiki and pick apart packages because it's such a great resource for OpenRC.

[–] FuckBigTech347@lemmygrad.ml 3 points 20 hours ago (1 children)

There is a project I've discovered recently which is similar to GPT4All, except you can throw multiple GPUs at the workloads (and yes it can use Vulkan): https://github.com/LostRuins/koboldcpp

I haven't messed much with it but it builds and works fine on Linux. The only thing I don't like is that the source tree has a bunch of Windows binaries in it.

But most importantly, it won’t work in the end. These scraping tech companies have much deeper pockets and can use specialized hardware that is much more efficient at solving these challenges than a normal web browser.

A lot of people don't seem to be able to comprehend this. Even the most basic Server Hardware that these companies have access to is many times more powerful than the best Gaming PC you can get right now. And if things get too slow they can always just spin up more nodes, which is trivial to them. If anything, they could use this as an excuse to justify higher production costs, which would make resulting datasets and models more valuable.

If this PoW crap becomes widespread it will only make the Internet more shitty and less usable for the average person in the long term. I despise the idea of running completely arbitrary computations just so some Web Admin somewhere can be relieved to know that the CPU spikes they see coming from their shitty NodeJS/Python Framework that generates all the HTML+CSS on-the-fly, does a couple of roundtrips and adds tens of lines of log on every single request, are maybe, hopefully caused by a real human and not a sophisticated web crawler.

My theory is people like to glaze Anubis because it's linked to the general "Anti-AI" sentiment now (thanks to tech journalism), and also probably because its mascot character is an anime girl and the Developer/CEO of Techaro is a streamer/vtuber.

[–] FuckBigTech347@lemmygrad.ml 12 points 2 days ago (1 children)

The Beautiful

The Good, The True

[–] FuckBigTech347@lemmygrad.ml 6 points 4 days ago

I have all my browsers configured where JS is disabled by default, no local data gets stored (cookies, cache, etc. gets deleted on exit) and I use containerized tabs, so I always get cucked by Anubis. There are several Websites I stopped visiting because of Anubis and I'm not just going to change my setup that I've been using for years across several PCs just to get to a piece of text. I really don't like this proof of work shit, especially not when it's done via bleeding edge JabbaScript.

I doubt this will affect scrapers as much as people like to think. Their crawlers likely run on enterprise grade Server Hardware on a massive uplink.

[–] FuckBigTech347@lemmygrad.ml 3 points 1 week ago

Exactly. All modern CPUs are so standardized that there is little reason to store all the data in ASCII text. It's so much faster and less complicated to just keep the raw binary on disk.

[–] FuckBigTech347@lemmygrad.ml 2 points 1 week ago

The newest PC I own is from 2014. The latest CPU µArch I know is Haswell. Zen Four? Sapphire Rapids? Funny fantasy names.

It annoys me though how newer video cards only come with DP and HDMI ports anymore. I'm still one of those people who actively use DVI and see no reason to get rid of my VGA/DVI Monitors, so I'm forced to deal with adapters. Fucking Capitalists.

[–] FuckBigTech347@lemmygrad.ml 4 points 1 week ago

What always gets me is that they can't admit and accept that something is overpriced, low quality shit when it is just overpriced, low quality shit. Instead they huff the copium and call anyone who does offer genuine criticism a "toxic hater" or whatever. Another argument I also keep seeing is "It's just not for you".

[–] FuckBigTech347@lemmygrad.ml 2 points 1 week ago

NVK doesn't support older cards though last time I checked. Pretty funny how I ended up with a stack of paperweights because NVidia dropped support and Nouveau/NVK can't get their shit together and instead of focusing on existing hardware they rather keep chasing the "latest and greatest".

[–] FuckBigTech347@lemmygrad.ml 3 points 2 weeks ago

AI? Look, I helped a friend fix a new install. It wasn’t Linux fault, it was a setting in the bios that needed to be changed. But the AI had them trying all sorts of things that were unrelated, and was never going to help. Use with a grain of salt.

I have the same experience but sometimes it was even worse; Sometimes the AI would confidently recommend doing things that might lead to breakage. Personally I recommend against using AI to learn Linux. It's just not worth it and will only give new users a false impression of how things work on Linux. People are much better off reading documentation (actual documentation, not SEO slop on random websites) or asking for help in forums.

[–] FuckBigTech347@lemmygrad.ml 10 points 2 weeks ago* (last edited 2 weeks ago)

It has a green lock icon with the word "Private" next to it so it's fine bro. western-journalist

[–] FuckBigTech347@lemmygrad.ml 1 points 2 weeks ago

You don't necessarily need a specific distro. What I always do is just use the distro I'm most comfortable with and do a minimal install that just boots to a TTY. From there I just try stuff out and see what works and what doesn't. The HP 2133 will require a dedicated distro though since it's a 32-Bit Machine. Consider something like https://archlinux32.org/

Right now I have a Toshiba Satellite C850D and an HP 635 and both are running relatively up-to-date Artix. I made sure both of them have an Atheros Wireless Card since those are well supported by Linux. I use them mostly to watch videos (720p works just fine especially with VAAPI) and Firefox works fine (although it's kind of slow, but modern browsers are massively bloated tbf.). If you can you probably want to swap out all the spinning rust with SSDs.

Besides the obvious casual use, you can always turn them into file servers or maybe even a Wireless to Ethernet Network bridge.

 

I figured I'd share my personal Notes on this particular Topic since Documentation on this is hard to find and most of it is outdated or scattered or leaves out certain details.

Maybe someone out there can find some of this useful.

How to setup Kerberos with NFSv4 on Linux in 2025:

  • This Guide assumes that:
    • You are familiar with Kerberos, NFS, DNS, Linux and UNIX-like Operating Systems in General.
    • All of the Machines involved can talk to each other.
    • All of the Machines involved have their System Date and Time synchronized.
    • All of the Machines involved have relatively up-to-date Software.
    • You have a running Server with:
      • A working default NFS Installation with Shares that can be mounted by Clients.
      • A working Kerberos V Installation where Clients can successfully authenticate and receive Tickets.
    • You are using the MIT implementation of Kerberos V. Heimdal and others may work, but haven't been tested.

1: Server-Side

1.1: Principals

kadmin into your Kerberos Server and add a Service Principal for NFS and a Machine Principal for each one of your Clients:

addprinc -nokey nfs/<Server's Hostname>
addprinc -nokey root/<Client-1 Hostname>
addprinc -nokey root/<Client-2 Hostname>
addprinc -nokey root/<Client-3 Hostname>
... etc ...

IMPORTANT: If you're using your own DNS Server and have multiple reverse (PTR) Records configured to resolve to your NFS Server's IP Address, then you MUST also add a Service Principal for each one of those Records.
For Example, if you have the Records my-nfs-server.lan and awesome-nfs-server.lan pointing at the same IP Address you do:

addprinc -nokey nfs/my-nfs-server.lan
addprinc -nokey nfs/awesome-nfs-server.lan

The reason for this is because Kerberos may try any Domain Name that is associated with your NFS Server's IP Address.

Next, make sure to add all the newly created Service Principal(s) to the Server's keytab:

ktadd nfs/<Server's Hostname>
ktadd nfs/my-nfs-server.lan
... etc ...

Then add each of the Machine Principals to their own keytab:

ktadd -k client-1.keytab root/<Client-1 Hostname>
ktadd -k client-2.keytab root/<Client-2 Hostname>
ktadd -k client-3.keytab root/<Client-3 Hostname>
... etc ...

Now quit out of the Admin Program and copy the newly created Kerberos keytab Files to all of your Clients!

1.2: rpc.svcgssd is dead

If your NFS Server Installation is mostly unmodified then there is a good chance that it's configured to use rpc.svcgssd for GSS authentication. But the reality of the Situation is that this Daemon is dead and in fact has been removed a long time ago.

gssproxy is its replacement, however it's a separate Component that needs to be installed manually.
It should be available in your Operating System's Package Manager. If not, make sure to look through the AUR and pkgsrc.
If for some reason none of them have it, then grab the source code and compile it manually:

$ git clone https://github.com/gssapi/gssproxy
$ cd gssproxy
$ vim BUILD.txt

gssproxy should also come with a couple of config files in /etc/gssproxy.
Specifically the Files 24-nfs-server.conf and 99-network-fs-clients.conf need to be present.
If for some reason you don't have them on your System, grab them from the gssproxy Git Repository. They are in the examples directory.

If everything looks OK start gssproxy (either via a Service File or manually).
Make sure that it stays running and doesn't print any errors. It should be treated as a part of NFS and needs to start with NFS Server Daemons.

1.3: exports

Any NFS Share that you want to use with Kerberos needs to have the sec mount option set.
For Example, if you have an /etc/exports File with an Entry like this:

/export/movies                          *(rw,async,no_subtree_check,no_root_squash)

Then you need to change it to look like this to "Kerberize" it:

/export/movies                          *(rw,async,no_subtree_check,no_root_squash,sec=krb5)

NOTE: NFS offers multiple ways to use Kerberos:

  • sec=krb5 is the most basic Option uses Kerberos for Authentication only.
  • sec=krb5i uses Kerberos for Authentication and verifies Data Integrity.
  • sec=krb5p uses Kerberos for Authentication, verifies Data Integrity and encrypts traffic.

If you want to use your NFS Shares in a Network with many different Users then krb5p is highly recommended.

Don't forget to run # exportfs -arv to reload your exports!

1.4: NFS idmapd.conf

Under [General] change Domain to your Kerberos Realm and make sure that [Translation] is using the nsswitch method.
Save the changes and restart the rpc.idmapd Daemon. Verify that your Domain is set correctly by running:

# nfsidmap -d

That should print your Kerberos Realm's Domain Name.

2: Client-Side

Make sure rpc.statd is always running.
IMPORTANT: Start rpc.gssd with the GSS_USE_PROXY Environment Variable set to 1 and make sure it is also always running!
Example:

# env GSS_USE_PROXY=1 rpc.gssd -f

If you have a Service File for rpc.gssd on your System, make sure it sets that Environment Variable! Edit the Service File if you must!
If for some reason rpc.gssd crashes and complains about some rpc_pipefs being empty then you need to mount that pseudo-Filesystem manually:

# mount sunrpc /var/lib/nfs/rpc_pipefs -t rpc_pipefs -o nodev

If rpc_pipefs doesn't exist then you need to load the sunrpc Kernel Module first:

# modprobe sunrpc

At this point you should be able to mount your NFS Shares on the Client:

# mount my-nfs-server.lan:/export/movies /mnt

You may need to be a bit more specific though:

# mount -t nfs -o sec=krb5p,vers=4.2 my-nfs-server.lan:/export/movies /mnt

If no errors occurred, confirm that the Shares are mounted correctly:

$ mount
...
my-nfs-server.lan:/export/movies on /mnt type nfs4 (rw,relatime,vers=4.2,sec=krb5p,etc.)
...

If everything looks good, you need to grab a Ticket for your User via your personal Principal:

$ kinit
Password for user@KERBEROSREALM.ORG: 

Now your User should be able to see and access the mounted NFS Shares:

$ touch /mnt/test
$ ls -l /mnt/test
-rw-r--r-- 1 user user 0 Jan 23 07:33 /mnt/test
$ klist
Ticket cache: FILE:/tmp/krb5cc_1000
Default principal: user@KERBEROSREALM.ORG

Valid starting     Expires            Service principal
01/23/25 03:28:00  01/23/25 13:28:00  krbtgt/KERBEROSREALM.ORG@KERBEROSREALM.ORG
01/23/25 03:28:00  01/23/25 13:28:00  nfs/my-nfs-server.lan@KERBEROSREALM.ORG

NOTE: Once your Ticket is expired your User will no longer be able to access the Shares until a new Ticket is acquired.

 

In 2021, senior officials from the Biden Administration, including the White House, repeatedly pressured our teams for months to censor certain COVID-19 content, including humor and satire, and expressed a lot of frustration with our teams when we didn't agree. Ultimately, it was our decision whether or not to take content down, and we own our decisions, including COVID-19-related changes we made to our enforcement in the wake of this pressure. I believe the government pressure was wrong, and I regret that we were not more outspoken about it. I also think we made some choices that, with the benefit of hindsight and new information, we wouldn't make today. Like I said to our teams at the time, I feel strongly that we should not compromise our content standards due to pressure from any Administration in either direction -- and we're ready to push back if something like this happens again.

Western media variants for the libs:

https://www.reuters.com/technology/zuckerberg-says-biden-administration-pressured-meta-censor-covid-19-content-2024-08-27/

https://www.washingtonpost.com/technology/2024/08/27/meta-zuckerberg-covid-misinformation-jordan-white-house/

 

Wish it was higher quality. Still an Interesting documentary from CGTN.

 

Website: https://www.openkylin.top/index-en.html

Git repositories: https://gitee.com/openkylin

I haven't tried it myself yet but I might give it a spin since it has a unique Desktop environment.

view more: next ›