FedX

joined 4 months ago
 

A simple tool to figure out just how much LLM generated code has been used in a project. Obviously not perfect, but it automates the task of going through commits for "co-authored by" statements and other signs. In the off chance you don't want to use the CLI, there's also a web app!

[–] FedX@quokk.au 4 points 18 hours ago (4 children)

It's a straw man's argument. I don't plan on continuing to use Noctalia because of AI, therefore I shouldn't use Linux itself? These are two entirely different statements.

Is it problematic that the kernel is developed using AI? Yes, certainly, but that's not something I have the capacity to deal with at the moment. But why shouldn't I be allowed to switch to other projects that are stringently anti-AI like River and Quickshell? I will support any open source project that comes with blanket statement against AI usage on the project because I loath it with a vengeance to deep to put into words. But that's just me, these projects are freely available for anyone to use and if you don't mind LLM generated code, knock yourself out.

[–] FedX@quokk.au 24 points 22 hours ago (15 children)

I use it with Niri, please know, the community likes AI at least a few of the plugins are LLMs, doesn't default to giving you AI features, and it's fair to say a lot of commits are AI assisted. Whether or not that's a problem for you is your choice, but it is a problem for me and plan to not use it in the future. Regardless, it gets a tentative recommend from me. Its convenience factor of you just run one program to provide widgets, bar, notification daemon, clipboard manager, and such is really hard to pass up. Moreover, the common design language between all the components is quite nice with the color scheme system they provide working rather well.

[–] FedX@quokk.au 2 points 4 days ago

Got a tip recently to limit the search range to pre-2024. Won't work for everything, like current events, but will work for a lot, like heat transfer and normal shock wave equations.

[–] FedX@quokk.au 6 points 4 days ago

Similar(ish) boat, but I never got as into as you, really sorry to hear that and really proud you are trying to get away from it. I used the very first version of ChatGPT to get through a very annoying CS class (yeah, it was a really bad class). I messed around with some small local models in Ollama, but eventually realized I was too dependent on technology as a whole. I was using DuckDuckGo to get answers to questions I should be working through myself and using AI to take notes I would never reference. So I actually started sitting down and reading the textbooks, Wikipedia, and carrying a physical notebook (yeah, I know, way old-school). I started downloading information in forms I knew was AI-free (like Wikiepdia from 2024), and working offline. Even went so far as to ditch my phone this semester. It ended up working really quite well for me. I started getting much better grades in some really hard classes like aerodynamics, and annoying entry-level classes like thermo. So, really, my suggestion boils down to, read. Read encyclopedias, textbooks, and documentation. phone, bad book good.

[–] FedX@quokk.au 10 points 4 days ago (3 children)

Surprisingly only been mentioned once, but Chimera! It's weird and strange and brilliant!

If you want something slightly more normal (well, actually, it's not any more normal, just more common), then you can get PostmarketOS installed on anything if you go outside the documentation a little (it's really not hard, pmbootstrap is quite a good utility). If you take that route, I recommend OpenRC as not only is PMOS anti-AI, but so is OpenRC!

[–] FedX@quokk.au 13 points 4 days ago (1 children)

Genuinely some of my favorite projects are on there (before I knew they where anti-AI too)! Chimera, Dune3d, Krita, River, PostmarketOS, Zig, Quickshell, Alacritty, OBS, yt-dlp and MUSl all get my 100% recommend (really, check them out, they are all brilliant projects). There's also some sweet new finds in there, keep the good work going!

[–] FedX@quokk.au 5 points 6 days ago

Yeah, fair, I am very fortunately in a position to ditch having a phone altogether and never look back. Waydroid is a nightmare, but translation layers like Steams thing and the creatively named Android Translation layer do exist and kindof work.

Being required to use a phone to get snail mail is insane! The whole point of mail is that direct interaction with technology along with some pretty excellent security and privacy guarantees! I hate how much smartphones have become a piece of society.

Also worth noting, simply having a SIM tied to your name and taken with you is a privacy issue.

[–] FedX@quokk.au 12 points 6 days ago (4 children)

It's a conversation we have had before, but Android derivatives are really not a long-term solution. There are mountains of technical issues that myself and others take issue with that only ditching the Android ecosystem in its entirety can fix. Also, are we really sure phones are a necessity of modern life?

[–] FedX@quokk.au 3 points 1 week ago

I'm in this picture... and I don't like it.

[–] FedX@quokk.au 3 points 1 week ago

utI6Ra9GAMVhoE9.jpg

Apparently that line was supposed to be a joke... which I reject because the thing Spider Punk is referring to is a damn good metaphor for capitalism.

 

Updated version of https://quokk.au/c/onehundredninetysix@lemmy.blahaj.zone/p/1081324/spr-ule-ing with correctly labeled aircraft.

 

Based on a true story...

 

Based on a true story...

 

I am trying to mathematical model a passive Magnetorheological damper. Broadly speaking, this would work by having a magnet sit on the damper shaft near the damper body. As the shaft moves, it would induce an eddy current in the damper body, creating a magnetic field. Not only would this magnetic field oppose the magnet's motion, but it would increase the viscosity of the magnetorheological fluid inside the damper body. On paper, this would rapidly increase the damping forces with increasing shaft speeds, allowing for more controlled shaft speed.

Fortunately, I found a paper discussing a damper similar to my ideas. Unfortunately, I don't really understand the math, and I need the damping force equation in terms of velocity to validate the application I have in mind. Linked is the paper I am referring to.

For my purposes, I would also want to model separate traditional high and low speed compression and rebound circuits moving through traditional mechanical damping circuits, along with some other damping features. However, that's a problem for future FedX, right now I really just care about the broad differences in damping characteristics as a function of shaft velocity.

view more: next ›