Autocorrect hates me, I am sorry.
I run lemonade on my strix halo box, I change the models I use, but the qwens are pretty good.
I access it via Tailscale on my phone, then I have a page I can use. The Lemonade interface forces you into the app, so you’ll want some other interface for your LLMs, as the app is junk.
I used to run llama.cpp and llama swap, but lemonade combines them and makes managing different ai backends easier.
Archwiki is pretty good, it has a lot of breadth, but it’s not a manual and it’s not easy.
Sorry, that guide was written for systems running systemd and the install manual you followed was a from an opinionated Arch user for SysVinit.
Oh, this guide is 11 years old written by someone who hated Wayland, so no, modern KDE won’t work with it.
Just use Arch Wiki, it’s the best for beginners, but also don’t install Arch unless you know what you’re doing already because it’s the worst for beginners.
I mostly use big picture mode now, so it’s not a huge deal. I just think it’s tacky.
Didn’t know you could turn it off on settings.
Why do these need 5-10 seconds to start at all?
I know steam does it so they can jamb a big pop up ad in your face, but this seems asinine.
Is it though?
I try it but it doesn’t feel good. It feels like a clone of Office 2004.
—
I just installed it, open an xlsx file, it opens to the document explorer panel not my document.
Okay, weird, where’s my file? Open it again.
It’s your first time here, customize the app! Here are 5 different UI paradigms to choose from. Okay, I guess compact?
Don’t forget the theme! Sure, okay, use automatic theme and default 2004 style icons.
LibreOffice needs to restart to apply settings. I watched it dynamically update in the background, but okay.
Restarts Tip of the day: you can change your appearance settings! Run this action now? No. Why? Close.
Okay, I spent all that time customizing my tab bar, and yet, Calc doesn’t use a tab bar, so there’s no change at all after all that? No idea if that’s a macOS thing or not because it puts the toolbar values in the menu bar.
Okay, well let’s make a pivot table. Seems decent, it’s a clean Excel clone here, decent functionality. Needs a count distinct and some custom aggregations.
What’s this “Data” field though? Drag it to first slot: LibreOffice hard crashes. At least it saved my document.
Reopen file, let’s zoom out. Click and drag zoom slider, it creates a glitched rectangle around the active cell that tears vs rest of screen that doesn’t always update. It looks like the full page only zooms on certain values while deagging, but on release that sticks.
Let’s make a chart. Select 3 items plus total in my pivot. Click chart button. 3 bars in the chart, where my total bar? Data field is blocked out on pivot take charts. Change chart type, wait 3 seconds for a re render. Try to move the chart but have to click off first.
Okay, let’s delete the chart now. Press delete button: “This function cannot be completed on the selected objects”. Huh? Oh I have to click off, then select the chart again, then delete it.
—
That was my 10 minute experiment. It feels just as bad in a web browser. The Skia renderer is so much better than the old system, but like, I found all this crap in 10 minutes.
I really want this to be good, but it needs a critical editor to go through it with a fine toothed comb. Honestly, they should crib Google Sheets, since it’s got a better interface than Excel now.
Okay but surely people get that BSD and Linux aren’t remotely the same thing, even though both are Unix-like and POSIX compatible, right?
Just a heads up, I pulled the drive stats from Backblaze’s drive survey, for most brands failure rates accelerate at 4-5 years.
At that point Hitachi, Toshiba, and Seagate all increased in failures. Western Digital didn’t, except for their 2014 batch that immediately dropped off.
Western Digital post 2020 are the most reliable drives.
Hitachi, Toshiba and Seagate seem to hit a cliff around 2-5 years now, but that only kills about 10% of their drives, so with replication you’re probably fine, just make sure to mix drives so you don’t get correlated failures.
The agency of just being able to fix something yourself is amazing.
Most of my open source contributions are bug fixes and reports. Can’t do that with closed source!
The short version is it makes your whole system deterministic and auditable, and that does some automatic stuff that is very nice.
Basically you write a configuration for your computer, then it generates atomic images from your config and manages your system for you. You update by generating then switching to the new snapshot. If anything goes wrong, you roll back, and your system is exactly as it was.
The whole system lives as code, it’s tied to git, and up to hardware specific things, I can checkout the exact same system across machines.
This is really great for people like me who dislike having to manage long term state like a grub config change or managing firewall rules or fiddling with committing systemd changes and managing them directly.
The packages are very well thought out.
For example if you add a systemd service it has rules for starting and stopping and the system is smart enough to know when to use them. That makes things like updating a server really nice, it’s just one command and everything is done for you, anything not affected by an update doesn’t restart, so services run smoothly (I have one config that manages all my services, their URL, firewall rules and healthcheck, and automatically adds them to a system monitoring dashboard for example).
As another example I can manage accounts on my system very easily, and I can clean them up by just removing that user from the nix files, and the system will do all the Linux user management stuff for me. This is great on a server where you want limited permissions users running specific applications — I just define what they should have and it works.
That has costs. Everything has to be thought of as either a system package or a user package, updating system packages is always a snapshot change with lock file update. You’re also locked into using nix packages which can lag releases (though this is most package managers).
Your average Arch Linux user probably doesn’t care or like this, they like the full control and flexibility Arch provides (which is just different tastes/philosophies).
You can install user local applications too outside of your main nix configuration (with something like homebrew, which works great on nix), or you can use home-manager to add more nix and manage your apps and generate user account dot files files.
You can also use nix-env to create reproducible software builds without containers, but I find it a little heavy to manage and haven’t learned it.
Ohh I have an old i9 intel MacBook I could migrate to do this…
Only had 512gb though, it’s an old dev machine.
How’s the Linux support on it? Since it’s all thunderbolt ports I’m wondering how well it works.
I really like nixos, which will not even tell you to reboot if the kernel is unchanged. Plus it will restart my services correctly when I update or pull new container images on my server.
It’s very very nice.
I compiled Linux last week and I don’t even know how it works
That’s fine because those vms are configured by and run on nix.
Yeah, venvs are the answer to “why is that folder 8gb?”
Containers are the answer to “why is my hard drive completely full”
I’m surprised the cooling unit doesn’t have a system that’s just state (on/off) and desired temperature, and that the enshittification subsystem controls the cooling at all.
Docker
That’s so stupid.
downloads and runs entire Linux distro just to run a python script without managing dependencies
I pointed lemonade at my llama.cpp model folder but it uses json files to configure them. I think it was fine for importing them or you could have an agent write them for you. After that you can manage models inside the app
Overall it wasn’t a huge deal, it’s easier than managing llama.cpp directly with rocm updates.