Reasons are usually just newest kernel/mesa/etc. Most of the time the difference is very small, and often inconsequential. However, every now and again there is a major development that might make it worth it (IE: The graphics pipeline that all but made dxvk-async obsolete)

I see you all over this thread and I want to share something you might find interesting.

You keep mentioning the server can’t handle the anti cheat because it needs to trust client data. Here’s an interesting thought: how is client anti cheat supposed to work when it needs to trust input data?

Look up direct memory access cheats. TL;DR Two computers are hooked up such that PC 1 runs the game, PC 2 reads memory from PC 1, and can then output keyboard/mouse inputs, as well as wallhacks/esp. How is the client side anti cheat supposed to know that the keyboard and mouse inputs are legitimate? How is the client side anti cheat to know wallhacks are being used when they are being rendered on an entirely different machine?

As a C# developer on Linux, I wish this was more true than it is. Working on a multi project dotnet solution in VSCode is still far behind Visual Studio / Jetbrains Rider.

Its also worth pointing out that the more you add to VSCode, the slower it becomes. If you add the toolkits to make it compete with Jetbrains products, it isn’t nearly the same lightweight editor anymore.

Won’t speak to Webstorm, but hard disagree when it comes to Rider. VSCode/Zed really fit into an entirely different category from Jetbrains IDE’s. Lightweight editors vs full fat development environments. There are use cases for each.

The truth about abs workout and diet is the same order tonight and tomorrow is fine but most importantly I will send you the best way to get the latest Flash player to play with my family 😁🐱

deleted by creator

The plugins would almost certainly work in a VM, but I imagine that latency would become a big headache. For my purposes, I picked up a Beelink mini pc and called it a day.

So in terms of DAW (Digital Audio Workstation), Linux already has Bitwig, Reaper, Arour, LMMS, and possibly others. Personally, I find the bigger issue comes from plugin developers (the DAW is your main program, and you add your sounds/effects through plugins). Most companies are not delivering anything Linux native. Many of these plugins can be bridged with compatibility software, and will work fine that way. However, most of these plugins now are also using their own install/activation software center, and they are often a nightmare in Linux.

Music production is the one thing I currently keep a windows mini PC around for these days. It’s not impossible to make the transition to Linux, but the last thing I want when pursuing a creative endeavor is technical software challenges holding me up.

You can say that speaks volumes about the character of the author (though you are the one assigning said “shame”). You were asking why this report deserves credence. The points raised in the report have citations such that you can decide where you fall on the presented issues.

It looks pretty well cited to me. The fact that it was written anonymously doesn’t really take away from that.

90% sure wireguard (the VPN server) is going to need an open port if you want to connect from the outside.

It is and it isn’t. It’s super dependant on use case. They bill on operations, not bandwidth. Obviously if you are hosting video/audio to be streamed, that could mean massive savings.

Lots of people, often unknowingly. If you run apt install firefox on Ubuntu, you’re getting the snap version.

I’m a software dev with quite a lot of experience in server admin. I’m also a full time Linux user, and run a lot of services both at home and on a rented VPS. I had oddly enough never used Ansible before, but the instructions on that GitHub page should make it pretty simple.

Yeahhhh…

Obviously it can all depend on your requirements, but this N95 system has been pretty eye opening on how much people are over-speccing their builds for home server use. It has 8Gb of memory in it, but I seldom see it use more than 2. The box is doing DNS, Jellyfin, torrenting, VPN, private git, etc.

I used the Lemmy Ansible method to deploy. At the time that I first installed it, it was the recommended method vs a docker compose. It is a little bit of setup, but is pretty simple to get going. Just follow the instructions and it should just work.

It would also result in a metric shit-ton of traffic and data storage.

Really depends how many instances they want to federate with. I run a single user instance for all of my personal Lemmy use. Looks like it is using 20Gb of bandwidth per week, and the VM it runs on only has 32Gb of storage (and it runs other services, too)

Same, but even lower (Beelink N95). My whole stack of two NAS units, mini PC, switch, router, and modem average a load of 50 watts.

Murphy!

FWIW: I’m running jellyfin and a whole host of other services on a Beelink with an Intel n95 and 8gb of ram. Runs like a champ.