From GitHub Copilot CLI to OpenCode: A Pragmatic Take from a Pro+ User I’m a GitHub Copilot Pro+ subscriber, and I’ll start with this: I genuinely love the GitHub ecosystem . GitHub has shaped the way many of us work—code hosting, CI, issues, pull requests, and now AI-assisted development. Naturally, when GitHub released Copilot CLI , I decided to go all in. Betting on Copilot CLI as an Exclusive Coding Agent Despite its early preview status , I made a deliberate choice: Copilot CLI would be my exclusive coding agent . I wanted a CLI-first, first-party AI experience, tightly integrated with GitHub. I knowingly accepted the risks that come with preview software. Initially, the bet paid off. Contributing Back: When Feedback Turns into Features Early on, I proposed a small but important enhancement: a command to list available models . Issue: https://github.com/github/copilot-cli/issues/47 To my surprise (and appreciation), the idea was welcomed, discussed, and...
I’ve been a Linux user for a long time, and historically my comfort zone has been “hands-on” distributions: Slackware, Gentoo, Debian, Arch Linux. I also virtualize heavily. I started back in the day with OpenVZ, then moved to VMware Workstation… until I got fed up with the constant “you must patch your kernel” routine to keep things working smoothly. Recently I got an Asus ProArt P16 (AMD Ryzen 9 + NVIDIA GPU). Great machine—until I tried to install my usual favorite Linux distributions. To my surprise, none of them handled the hardware properly (or at least not enough to reliably use the laptop day-to-day). I then tried Fedora 43 , and it immediately handled the basics out of the box — the minimum needed to actually use the system. After a bit more digging, I realized I needed to explicitly enable the NVIDIA dGPU. The fix was simple once you know it: supergfxctl --mode Hybrid Then came the “random Google search” moment: I searched for a Fedora derivative that would handle the...