skip to content
Profile picture of Pedro Canterini

Linux Rabbit Hole on My Desktop

/ 3 min read

I have been spending a lot of time lately going down a Linux rabbit hole on my mini-ITX desktop. What started as a simple idea to mess around with another OS turned into a lot of reflashing, reinstalling, and trying to figure out what actually feels good to live in every day.

Hardware

  • CPU: Intel Core i5-13600K
  • GPU: NVIDIA GeForce RTX 4080 Founders Edition, 16 GB GDDR6X
  • Motherboard: MSI MPG Z790I Edge WiFi
  • Memory: Corsair Vengeance 32 GB DDR5 5600
  • Storage: Seagate FireCuda 530 2 TB NVMe SSD

Distros I tried

  • Ubuntu was the easy starting point, mostly because it is familiar and simple to get going
  • Fedora ended up being the best out-of-the-box experience for this hardware and just worked really well
  • Arch Linux was fun to poke at, but it definitely sent me further down the tinkering path
  • Omarchy has probably been the most interesting one because the developer DX feels excellent and the opinionated defaults are actually good

Fedora impressed me the most when it came to supporting the hardware without much drama. Omarchy has been the one that made me want to stay longer, mostly because DHH has a very opinionated setup with strong defaults, shortcuts, tools, and apps that actually feel good to use as a developer. The part about shaping your own workflow is more what I like about Linux and open source in general.

What stuck

The hardest part for me is honestly how addicted I am to macOS. That muscle memory is real. Even when Linux is doing exactly what I want, my hands still expect Mac behavior half the time.

That said, I still really like how flexible Linux is once you lean into it. I use a Mac for work, but Omarchy is becoming the setup I want for everything else. A lot of that comes from being able to pile on custom scripts, personal dotfiles, and little quality-of-life tweaks until the machine feels like mine. That is still one of the best things about Linux.

AI tooling

Codex from OpenAI has been especially useful for helping me build and refine scripts, and OpenCode has been great for trying open models like Kimi, MiniMax, and GLM. Using AI for this kind of customization work has made the whole process faster and a lot more fun.

One of the nice side effects of all this desktop tinkering is that it keeps pushing me toward the next layer of projects. The more I get comfortable shaping my own Linux environment, the more I want to keep going with things like Proxmox and TrueNAS next.