JCM

@mxinden@mastodon.social Oh my, that's very tempting~

@liaizon@social.wake.st That's so cool! :D I wish I could have watched it.

@LAS@floss.social In Berlin? :O For free? :O OMG, this is amazing! :waffy_heart: :waffy_party:

@alina@girldick.gay It's more fun like this anyway! :D

@vwbusguy@mastodon.online @bookwar@floss.social @tragivictoria@mastodon.catgirl.cloud I disagree with this definition tbh. A distribution is anything that distributes different pieces of upstream software without being controlled by upstream. You can break upstream software even if no package manager is involved. Think about it from upstream's point of view: If someone packages my software, why would I care if it is distributed to people with a package manager or in a bundle? Someone is still packaging my software and potentially breaking it or making changes to it that I never intended.

@solonovamax@tech.lgbt If you wanna do a bit of low effort trolling, so a chattr +i on the directory in which you don't want unauthorized subdirectories and see how long it takes your team to figure it out :P

@laru@pleroma.laru.dsrv.x7q.eu @mldkyt@soc.mldkyt.com It's neither passive nor a scraper. Probably quite the opposite of a scraper, since it will have very limited storage for remote posts to begin with and will overwrite old posts as new posts arrive. As I said, I will honor all post deletes manually. User deletes won't even matter, because I don't plan on storing remote users to begin with.

@Cyndi@wafrn.jcm.re Yep yep, hihi~

@laru@pleroma.laru.dsrv.x7q.eu @mldkyt@soc.mldkyt.com That also means that you'd need to defederate any Fedi instances doing regular database backups, as those typically require manual intervention for propagating post deletions through backups (if they even do it at all).

@laru@pleroma.laru.dsrv.x7q.eu @mldkyt@soc.mldkyt.com Maybe I explained my plan badly: deleting manually means manual intervention (i.e. unplugging the SD card, putting into my PC, and running a script to delete posts requested to be deleted), not going through all posts by hand The only reason I'm doing this is, because I'm not sure, that fully automated deletion can be implemented given the constraints I'm working with. I don't intend for this instance to be malicious and I'll go out of my way to make sure every single deletion request will be honored and that as soon as a deletion request is received the post is inaccessible before its final deletion by manual intervention. When I say my project is not passive, I mean that its purpose is not to record all fediverse activity. It'll record as little as necessary to build a timeline and only make this available to authenticated users. It will also no record any non-public posts to begin with.

@vwbusguy@mastodon.online @bookwar@floss.social @tragivictoria@mastodon.catgirl.cloud Is there any alternative tho? Like, how would a normal person get a running and stable system without some kind of distro?

@Pepijn@mastodon.online Now I feel like I missed out on something very important by not sitting on a cow in my childhood…

@vwbusguy@mastodon.online @bookwar@floss.social @tragivictoria@mastodon.catgirl.cloud Only if: upstream distributes itself for all relevant architectures (especially RISC-V tends to be problematic and the Ubuntu packages are saving my ass often) upstream backports security fixes, so users don't have to do breaking updates upstream provides timely updates when there is a security issue in any library the package includes I usually prefer upstream provided Snap, AppImage, or Flatpak packages myself too, but I don't think it's viable for all types of software, so some form of distributing will always happen.

@mkljczk@pl.fediverse.pl While I am pretty sure this was not addressed at me at all, it (finally) convinced me to get rid of Anubis :D So, thank you!

This button is scaring me 😨 #why-are-conferences-so-damn-expensive

I really wanna love Wayland (I am even working on my own private cursed Wayland compositor), but it always sucks when something that flawlessly worked on X11 for years completely breaks on Wayland… Yes, I could and should probably figure out what is happening, find the bug eventually and then fix it, but selecting X11 in my login manager is usually easier :/

@mldkyt@soc.mldkyt.com 400 Bad followers~

@solonovamax@tech.lgbt That's so evil xD I love it! :D (But probably don't do it, it sounds like a recipe for loving important work one day before a deadline 😅)

@karolherbst@chaos.social I personally prefer well documented APIs that explicitly state required preconditions and behavior that you can rely on (and define what behavior you cannot rely on). If you violate a precondition and it still works, it should be allowed to stop working at any point (even because your CPU temperature increases). If you rely on undefined behavior, your code might break at any point and imo that is okay.

@karolherbst@chaos.social XKCD 1172 energy

One of the major downsides of vibecoding is that once the code has been written by the AI it is deterministically right or wrong and also cannot be improved (e.g. by switching to a new frontier model) without manually prompting an AI to rewrite it.

Therefore, I have decided to address these issues by developing a new and revolutionary file system: llmfs!

Simply point it to an API, mount it somewhere, and start following "everything is a file"! Even your LLM prompts!

With this amazing new, never-before-seen technology, you can simply include prompts into your C(++) code using the well-known #include keyword:

#include#include#include#include "llm/openai/gpt-oss-20b/prompt/Write a C function called 'void replace(char* text, const char* needle, const char* replacement)' \
that replaces all occurences of needle with replacement. Only include the function and no markdown or explanation."

int main() {
    const char* str = "Hello World!";
    char* s = strdup(str);
    replace(s, "Hello", "Goodbye");
    printf("%s\n", s);
    return 0;
}

Will your code compile? Maybe.
Will your code work? Who knows!
Will your code have the same behavior every time you compile it? Abso-fucking-lutely not!
Will your code burn through tons and tons of tokens? Abso-fucking-lutely yes, so you can finally hit those CEO-mandated token goals.
Can you see your code? No.
Does the compiler show your code if there are warnings or errors? No, it generates an entirely new version for diagnosis!
Is this real? Yes, it exists and runs on my machine!
Does it actually work? Sometimes!

#AI #LLM #FUSE3 #cursed #I-am-so-sorry-for-creating-this #include #include #include #include #include

@laru@pleroma.laru.dsrv.x7q.eu @mldkyt@soc.mldkyt.com The main problem is that building an index is hard. I basically need to build a hash map, which is quite annoying within the project. Therefore I'd rather perform deletes manually at the beginning and then later see if I have enough headroom to even fit the hash map logic 😅

@vwbusguy@mastodon.online @bookwar@floss.social @tragivictoria@mastodon.catgirl.cloud They might not be traditional distros from the DE's point of view, but they are/rely on traditional distros from the kernel's, systemd's, libc's, bash's, mesa's, etc point of view.

@laru@pleroma.laru.dsrv.x7q.eu @mldkyt@soc.mldkyt.com Yeah, my bad. I meant that the URL (or IRI) is the ID. And there's no guarantee that the URL itself contains any UUID or numerical ID. From my understanding a post ID of https://abc.xyz/post/lol/rofl would be valid.

@karolherbst@chaos.social Yesssss, I also love the way Vulkan handles it! Very clear API specifications, extension mechanism, and optional validation layers ftw.

@laru@pleroma.laru.dsrv.x7q.eu @mldkyt@soc.mldkyt.com Wow, some of those block reasons are absolutely wild… Some instances even block entire TLDs, WTF

@laru@pleroma.laru.dsrv.x7q.eu @mldkyt@soc.mldkyt.com Ooooh, what is this website? I wanna know if anyone blocks one of my instances!

@alina@girldick.gay Reverse engineering a protocol or a software is legit much more engaging to me than simply reading the documentation (I never even read Wafrn's protocol docs when making Wafbuntu, all I did was watching requests and reading the source code). It always feels a bit like solving some kind of mystery and is so satisfying once I make things work. Also, docs might lie, packets and code won't. :P

Authorized fetch is weird…

And beware, that even if there is no .vscode/tasks.json file, it still isn't necessarily safe. Depending on what plugins you have installed, VSCode might automatically execute build scripts (e.g. CMakeLists.txt), which might be able to execute arbitrary commands.

@luna@lunar.place I've only ever had very cheap and basic ornamental (is this the right word?) ones (back in my "femboy phase" 😅) and I eventually managed to lose those :/ Maybe I should get (or make) myself some proper ones one day :3