• 0 Posts
  • 14 Comments
Joined 3 years ago
cake
Cake day: July 8th, 2023

help-circle

  • cyrl@lemmy.worldtoLinux@lemmy.ml•Kitty 0.50
    link
    fedilink
    arrow-up
    4
    ·
    4 days ago

    Ooh, custom shaders will let me properly dim inactive splits/windows - much better than my current solution which changes the alpha channel in a not-wholly-intended fashion.

    Personally, I landed on using Kitty after…

    • being a long time <whatever-terminal>+tmux user and getting sick of problems Ctrl+<char> or delete keys etc.
      • I addressed most of these but there was always gremlins and the whole stack felt hacky with tmux emulating essentially an entire terminal, inside another terminal
    • I tried looking at tmux alternatives, but none seemed to entirely support fancier terminal features like graphics and have vim/tmux-like keybinds - looking at you ZelliJ
    • I then started looking for a replacement terminal which would give me the necessary bits of tmux - splits, tabs, scrollback search & copy, vim keys. I don’t need shells to persist between logins/SSH sessions
    • I landed on wezterm, easily configured in Lua like my Neovim, but it seems its not really got the attention of the original developer and gave the impression of falling behind on features, raising questions of security issues
    • kitty was the only other terminal I could see largely offering the tmux-like layer out of the box, needing just a bit of config for vim/tmux-like keybindings and using Neovim for fancy scrollback features.

    I’d love to know of alternatives to Kitty - the main dev’s attitude does leave me with some concerns, but he is a long time open-source developer with multiple successful and maintained projects, for now that keeps me onboard. It seems like he’s pushing forwards with new terminal features like multi-font and image support, which is interesting if nothing else.

    Perhaps I should go look at herdr as a tmux replacement and use ghostty or Alarcitty, but I’m not so interested in the AI features.


  • My answer for this has been to use Gadgetbridge - its an open source app which can replace the e.g. Garmin app, connecting to the device and gathering activity and sensor data plus forwarding notifications etc as you’d expect the first party app to do. The data remains exclusively on your phone, no servers involved.

    Its not perfect, the Garmin UI is slicker and some minor features may not be supported but overall it’s really impressed me for a community effort.

    I’ve been using this since March 2025 with a Garmin Forerunner 265 without any issues.

    There are a wide range of supported devices, so you can buy the hardware you like/find a good deal and not worry about privacy based on the vendor.

    Gadgetbridge Site Supported Watches

    Hah, as of today there at 666 supported devices!










  • I think theres two aspects to it

    • social licensing - e.g. Google tries and gets away with X, so the others now perceive they’re less likely to receive excessive pushback and are emboldened to go for it. This is why I feel there needs to be a pervasive and continuous push towards consumer rights - smaller cases can snowball quickly in the wrong direction, never so in the reverse.
    • share price driven margin pressure, a figurative ideal business that balanced perfectly its price/margins/costs against consumer demands/buying power would still be pushed to make graph go up and right - the usual enshittification argument.

    The first feeds into the second, once your competitor moves against consumer interests, C suites are/perceive they are then under pressure to match peers, else fall behind.

    Screwing over customers is baked into Capitalism, even more so with the current scale and concentration of a handful of business operating in a weak regulatory environment.


  • Anything that reduces the footprint of LLM’s is welcome, however…

    • making LLM compute cheaper in datacentres won’t mean lower total power/cooling/space/water consumption, like adding lanes and traffic, it will just mean more usage as it gets cheaper (and a short-term bump in margins for the LLM owners)
    • these are still highly dedicated chips that are always going to be bound up in the mega-scale datacentre deployments
      • what happens if there is a paradigm shift in the exact compute architecture? Loads of junk servers and no applications able to make use of such a glut
      • these do nothing to push LLMs out of the datacenter and into non-corporate hands, which is the only spot where we might see fewer privacy concerns, less corporate control etc

    If we’re stuck with the current compute/corporate paradigms, at least alternatives nibling at the unhealthy dominance of nVidia and the cloud giants is some small benefit.