• 5 Posts
  • 160 Comments
Joined 2 years ago
cake
Cake day: December 27th, 2022

help-circle

  • I cannot access my homelab from my work network, so I cannot sync via Nextcloud. Syncthing would be better, but they just stopped supporting Android sync, which I need. Proton Drive doesn’t sync files on Android. On top of that, I don’t want to deal with sync issues because keepass isn’t designed for syncing like that. I’m not gonna go back to using Google, Microsoft, or Dropbox just for keepass. I’ve considered just keeping my db file on a flash drive, but all of the keepass Android apps I tried won’t automatically detect that the file exists when I plug in the drive.

    If someone has a better way for me to use it, please enlighten me.

    Bitwarden is slowly turning their stuff closed-source, and I hope they don’t turn to shit, but right now it’s what works.


  • Yeah, I’m talking about not just Nix, but NixOS. Nix (the package manager) can do a lot, but NixOS + disko + home-manager can literally be all of the configuration for your machine from drive partitioning through to dot files. Throw in nixos-anywhere and impermanence and you can have an insane amount of control over all of your computers.

    Ansible, Terraform, Chef, etc. do have some overlap, but the main difference is that those tools iterate through the system modifying it piece by piece and NixOS is declarative.

    If something fails in some of my bigger Ansible playbooks, it could mean 30 minutes of just running through all the steps again. I could probably break it into sections, but then I have to worry about making sure they all get run when things get updated. In my NixOS install, it’s way faster, I can roll back to a previous state, and troubleshooting is way easier in my opinion.







  • The main thing I have learned after switching to Linux full-time is that weird, proprietary hardware like this is almost never a good idea, for many reasons. It’s very easy to make labels for keys if you really want to, and if you need more functionality, having more buttons instead of layers is always going to be faster to learn and use. Especially if you are trying to use this as a home automation interface, it’s probably a better idea to have either a touch screen or a separate screen and keypad.

    Sure, this thing looks nice, but in a couple years (at most) it will be e-waste.




  • As you have in your post, Logical Increments is a good place to start.

    As others have said, AMD is your best bet currently, mostly because of raw performance compared to recent Intel offerings. If you have no limited budget or power requirements, here are my recommendations:

    If you have the paid version of Davinci Resolve, AMD does not have the best selection of hardware encode/decode options, but people have reported that Intel Arc GPUs work, so I would get and Intel A310 as a secondary GPU if that is something that you need.

    If you want the best of the best GPU, without going Nvidia, the AMD RX 7900XTX is it. Also, AMD has stated publicly that they are moving away from high-end GPUs, so there probably won’t be a better one coming out anytime soon.

    If you want to plan for more gaming than you stated in your post, the Ryzen 7800X3D is the best gaming CPU on the market, so I would get that. If you plan to focus on video editing, the 9950X is the best, but probably not worth the cost compared to cheaper 9000 or 7000 chips.

    If you go with a Ryzen 7000 or 9000 CPU, get DDR5-6000 CL30 memory.

    If you’re getting an air cooler for your CPU, don’t pay more than $50. There are a ton of great, cheap options these days.

    Get either the new Antec Flux Pro case (when it’s available, probably this month) or the Fractal Torrent if you care about best thermals and quiet operation. Everything else is a compromise.

    If you need HDMI 2.1, you’ll need a DP -> HDMI adapter on an AMD GPU because of a licensing squabble.

    Those are things I could think of off the top of my head. I don’t think I missed anything big.









  • I think about this a lot, and my take is that Linux is waaayyy better if you have perfect or close-to-perfect knowledge of how the operating system works and what software is available. Similarly, I think an argument can be made for Linux being better if all you need is a web browser and you’re not using really unusual hardware.

    Where things fall apart is for people who have very specific needs that are complex, even if they only need it 1% of the time, and they don’t have the technical knowledge to solve it with the power-user tools available. Microsoft has spent decades paying developers to handle these edge cases and ensuring GUI settings discoverability.

    At the same time, schools and workplaces have taught people the design language of Windows, and the network effect of having so much of the world’s end-user PCs running on Windows means that there are vast resources available targeted at people without technical knowledge. At this point, for better or worse, Microsoft’s design language is the global default for non-technical people.

    If a person never has to touch a setting because all they need is a browser, they don’t hit any friction and they are happy. If they need to do even one thing that requires them to dig into settings or touch the terminal, the difference from Microsoft’s design language is enough for that one frustrating experience to give them a bad taste in their mouth about Linux as a whole.