• 5 Posts
  • 140 Comments
Joined 3 years ago
cake
Cake day: July 13th, 2023

help-circle







  • Way better than expected. Even if I was already using Linux on servers since decades, on desktop I preferred Windows. But my laptop was with 8gb soldered RAM and Windows 11 is basically unusable with that amount. I wanted to switch.

    But my past experience was bad, too often stuff was broken. Used Ubuntu in 2016, couldn’t stand it => revert to win10, tried Manjaro in 2019, one day I fucked with some AUR and it could not boot => revert to win10. I left thinking that Linux on the desktop is not ready.

    Then last summer the constant updates on my windows laptop made it unusable. It simply doesn’t leave enough memory to use a web browser with more than a couple tabs.

    At the same time at work a windows 11 update introduced a very annoying bug: after standby, windows would switch the resolution of displayport monitors to 800*600 and destroy my window layout, with everything moved to the top left corner. I had to use a tiling window manager like glazewm as a temporary fix until Microsoft fixed the bug (still annoying waiting for a couple seconds to have the windows rearranged when the monitor went to standby) and I fell into the rabbit hole of tiling managers. I watched videos where some YouTubers showed how l33t is cachyos with hyprland with their magic dotfiles and I fell for the meme.

    For the first few weeks it was awesome, then of course hyprland deprecates syntax without warnings and I started to get errors after the first update. Also the concept of using someone else’s dotfiles is wrong as they’re highly opinionated. They should do videos about how to make your Linux experience similar to theirs, not “clone this configuration as a black box”, because then you would have no idea how to fix problems when the updates come. But it seems like their priority is getting stars on their GitHub, rather than actually helping people. “Just blindly run this script as sudo” is a wrong concept, IMHO.

    Then when hyprland changed syntax AGAIN without warning, I was fed up, didn’t want to spend hours to debug the problem so I spent hours to reinstall another distro. I read that Linus is using fedora with plain gnome and some frippery extensions because “it just works” and… OMG. It just works! I’m shocked how good vanilla GNOME has become since the last time I tried it in 2019! It’s now fully usable even for a noob! And I like those extensions too. Modern but classic. Easy but powerful. And the apps in the GNOME circle are so polished. I was shocked to see pika backup, user friendly but not dumbed down.








  • install tldr and then you have easier commands

    > tldr dd
    
      dd
    
      Convert and copy a file.
      See also: `caligula`.
      More information: https://www.gnu.org/software/coreutils/manual/html_node/dd-invocation.html.
    
      - Make a bootable USB drive from an isohybrid file (such as archlinux-xxx.iso) and show the progress:
        sudo dd if=path/to/file.iso of=/dev/usb_drive status=progress
    
      - Clone a drive to another drive with 4 MiB block size and flush writes before the command terminates:
        sudo dd bs=4M conv=fsync if=/dev/source_drive of=/dev/dest_drive
    
      - Generate a file with a specific number of random bytes by using kernel random driver:
        dd bs=100 count=1 if=/dev/urandom of=path/to/random_file
    
      - Benchmark the write performance of a disk:
        dd bs=1M count=1024 if=/dev/zero of=path/to/file_1GB
    
      - Create a system backup, save it into an IMG file (can be restored later by swapping if and of), and show the progress:
        sudo dd if=/dev/drive_device of=path/to/file.img status=progress
    
      - Check the progress of an ongoing dd operation (run this command from another shell):
        progress
    
    


  • I liked hyprland for about one month this August, wasted a lot to make the configuration how I liked. Then they changed the config suddenly without warning. Installed GNOME as a fallback because I needed to do actual work, wasted another weekend to fix the configuration. Nice, I used it for some other time and… broken again within a month. At least deprecate the configuration and give one version of tolerance…





  • I notice that you used the free interface without registration. While a bit of a hassle to register, I saw that behind the registration wall they use longer contexts/bigger models. Without registration it’s more prone to abuse so they might condense/truncate responses from chats: by design, the bots have zero memory, so the whole chat history needs to be appended at every question => long chats lead to expensive API calls => free users get history truncated or condensed.

    Also, the normal system prompt is something like “be a sycophant and always please the user no matter prompt” and it will lead you do stupid stuff if you ask, so you need to go to the settings and change it to something like:

    From now on, stop being agreeable and act like a high-level consultant: blunt and honest. Don’t validate me, don’t soften the truth, and don’t flatter me. Challenge my thinking, question my assumptions, and point out the blind spots I’m ignoring. Be direct, rational, and unfiltered. If my reasoning is weak, dismantle it and show me why; if I’m deluding myself or lying to myself, say it; if I’m avoiding something uncomfortable or wasting time, call it out and explain the opportunity cost. Analyze my situation with objectivity and strategic depth, showing me where I’m making excuses, playing small, or underestimating risk and effort.

    For example this question that i had: (all paragraphed, writing from memory, i routinely wipe all the chat histories)

    can I bypass gpg verification on fedora when using dnf?

    gpt-5-mini from chatgpt free:

    it’s absolutely possible! just use the --nogpgcheck flag

    gpt-5.2 from api with the previous system prompt:

    i won’t tell you how to do that, it’s an incredibly stupid idea, a bad habit and doesn’t actually solve the problem

    finally, don’t blindly trust the results as they might still be incorrect, use that as a hint on how to proceed