This is a big weakness in FOSS communities, hell, in capitalist existence. People with resources can afford to spend their own time or hire someone else to focus on their contributions like a full time job while most honest contributers will be doing it during their free time because they need to pay bills and such.

Or an alternate implementation of the API that fetches it to flag any programs that call it.

You must be the most dramatic person in the universe, calling that a “meltdown”.

Lol I thought your link was “here’s a rocket designed by an LLM” rather than one designed by the non-LLM AI.

LLMs are a local minimum that tech bros are stuck trying to optimize to a generally useful point because its language abilities are able to fool so many (just like how a real person talking with confidence can fool so many).

This obsession with LLMs is making me question general human intelligence more lol. It’s looking more and more like we are just dumb apes but get lucky and every now and then a smart ape is born and teaches the other dumb apes how to bring their stupidity to whole new levels.

Yeah, it’s good enough that it even had me fooled, despite all my “it just correlates words” comments. It was getting to the desired result, so I was starting to think that the framework around the agentic coding AIs was able to give it enough useful context to make the correlations useful, even if it wasn’t really thinking.

But it’s really just a bunch of duct tape slapped over cracks in a leaky tank they want to put more water in. While it’s impressive how far it has come, the fundamental issues will always be there because it’s still accurate to call LLMs massive text predictors.

The people who believe LLMs have achieved AGI are either just lying to try to prolong the bubble in the hopes of actually getting it to the singularity before it pops or are revealing their own lack of expertise because they either haven’t noticed the fundamental issues or think they are minor things that can be solved because any instance can be patched.

But a) they can only be patched by people who know the correction (so the patches won’t happen in the bleeding edge until humans solve the problem they wanted AI to solve), and b) it will require an infinite number of these patches even to just cover all permutations of everything we do know.

Here’s an example I ran into, since work wants us to use AI to produce work stuff, whatever, they get to deal with the result.

But I had asked it to add some debug code to verify that a process was working by saving the in memory result of that process to a file, so I could ensure the next step was even possible to do based on the output of the first step (because the second step was failing). Get the file output and it looks fine, other than missing some whitespace, but that’s ok.

And then while debugging, it says the issue is the data for step 1 isn’t being passed to the function the calls if all. Wait, how can this be, the file looks fine? Oh when it added the debug code, it added a new code path that just calls the step 1 code (properly). Which does work for verifying step 1 on its own but not for verifying the actual code path.

The code for this task is full of examples like that, almost as if it is intelligent but it’s using the genie model of being helpful where it tries to technically follow directions while subverting expectations anywhere it isn’t specified.

Thinking about my overall task, I’m not sure using AI has saved time. It produces code that looks more like final code, but adds a lot of subtle unexpected issues on the way.

Ah it was just a reference to how that backdoor was found. I don’t actually monitor my boot time, though maybe I should at least have a script comparing it vs historic instead of just hoping someone else would find that kind of thing.

Ok, I figured the hardware support bit would be a longshot anyways.

Whatever is going on, it is missing real-time deadlines for some reason. It could be the configuration results in too much work, or the audio work itself isn’t that bad but the priority is low enough that other less time-critical work is getting in the way.

But yeah, even if it is solvable, there should be good defaults to prevent it from happening in the first place. Annoying thing is that there might even be profiles where one would work for you, but they might be hidden deep in the terminal.

A recent update added 104ms to my boot time and I am SEETHING and will get to the bottom of this and make those responsible pay dearly.

I haven’t gotten my hands dirty with this stuff specifically, but maybe you need to adjust buffer sizes to properly handle the different bit rates. Do you mainly see issues with higher combinations? The sample rate * bit depth is the important number, here. If you consider the problematic ones from that perspective, is there a threshold where anything under works fine but anything above has issues that get worse depending on how far above the threshold it is?

I’m not certain, but I believe the audio buffer is handled via a callback function that gets called when the audio buffer is some % close to empty, and then the program refills the buffer, plus some other overhead. That data left in the buffer sets the deadline for refilling the buffer; miss that deadline and the audio cuts out. Meet the deadline and audio is seemless.

A too small buffer will require the callback be called more often, and then the overhead can add up to missing deadlines. Alternatively, the % when it does the callback might need to be adjusted.

Another consideration is if your DAC doesn’t support the chosen sample rate and bits per sample, then there is probably another buffer of the supported size and a conversion from one to the other (and its own callback when that buffer gets low). That said, I don’t know if it’ll even list unsupported combinations because I’m having trouble thinking of a valid use case. But it’s technically possible, so maybe it is like that.

Anyways, those are what I’d be checking to debug this. If it is a setup problem, it won’t likely ever go away on its own, unless better defaults get set for those bitrates, but the ideal values depend on your system’s performance, so if yours is on the weaker side, it might never change.

So what’s the difference between this kernel and the Linux kernel? Are they both intended to be interface-equivalent (even if they aren’t in the same place on the implementation side)? Any fundamentally different design policies?

When I was in school, I wanted a Linux machine (since my school stuff was mostly linux and I wanted to be able to work locally instead of having to ssh in to school machines) but wasn’t comfortable doing it on my main PC, so I bought a cheap laptop and inatalled linux on that. Had the extra bonus of being smaller and lighter than my gaming laptop that was my main PC at the time, too.

Your options will probably be a bit more expensive (and apologies for suggesting a solution that involves throwing money at it if you aren’t in a position to get even a relatively cheap one) since it’s running windows and needs the hardware for that, including TPM if your school stuff requires win 11 (though if you can get away with win 10 or 7, you could probably get a cheaper machine). Though on the other hand, your tasks might not require a GPU, which can save a lot right there.

Then you can truly isolate your personal stuff from winsows, especially if you set your LAN up to never let the windows machine know that the linux machine even exists.

I also use this with consoles to play games I’d like to try but they have DRM or anticheat that I don’t want on my PC. Also kinda doing it with work, though the laptop belongs to them.

I didn’t have the patience to try out a few, but luckily Fedora has met my needs without annoying me as much as windows did. Though I am feeling a bit of an itch to try out some of the others, but am currently just chilling in this local minimum of effort and not regretting going from windows to Fedora one bit.

Yeah maybe. Could be trying to make it clear it’s not uploading to a cloud or something lol.

Or maybe it means “put it down” (as in record, not discard).

I’ve noticed libreoffice has changed their save symbol. It’s a grey rectangle (taller than it is wide) with a darker grey arrow pointing down.

Is that method different from using the hot keys to swap layouts? Like can I tell it to always use that mapping for that game or do I need to remember to run it each time I play the game and then set it back after I’m done (or automate that)?

Actually, there is one thing that is an annoyance that I haven’t been able to resolve. I use dvorak as my main layout.

Sometimes games get the keyboard right and keys are remapped to qwerty layout (and typing still uses dvorak). This case works better than on windows, since playing a game there either required the game itself to recognize keyboard layouts (best case), or remapping the controls (annoying case), or switching to qwerty (frustrating for typing because I’m stronger with dvorak now).

But sometimes instead it does the opposite and remaps the qwerty bindings to dvorak. As in, even if I swap layouts, wasd are all over the keyboard instead of all together. I need to exit the game, swap layouts to qwerty on the desktop, then relaunch for controls to work properly (and then I can sometimes swap back to dvorak in game and they continue to work). Often, the next time I launch the game, I’ll forget to switch it but it will just work this time.

And sometimes it behaves like windows did where I can swap the layout in game and keys change as you’d expect.

I have no idea why it’s inconsistent between these three options or where the “preserve key location despite the layout” feature is even coming from. Anyone have any idea about this?

I upgraded my gpu this weekend. Shut down, switched the psu off, swapped the old one out and new one in, booted into bios no issue (to check if I has left pcie on auto or needed to update it), then booted into the desktop (fedora cinnamon). Bam, after login only saw wallpaper, no mouse cursor or other UI.

Well, at least it’s kinda working. Time to figure out what’s going on. Terminal works. There’s some errors in the log but nothing to do with amdgpu or firmware failed to upload or anything. Software render just shows up as black screen. Reset my cinnamon session and boot back to the same thing. Fuck.

Then I try moving my mouse way over to the right and it shows up! Oh right. I have my TV plugged in for streaming to it sometimes and it ended up defaulted to the primary display, so my main desktop was only showing up there (and it was off). Right click, display properties, swap my monitor to primary, disable the TV until I turn it on.

This is about the magnitude of the average problem I need to deal with on Linux. Something isn’t working like I want it to, half the time it’s actually working but I misunderstood something or the default doesn’t match my intent and I need to adjust settings and then it’s perfect or close enough.

Or the other problem I had yesterday, tried monster hunter world for the first time and it wouldn’t launch. Played satisfactory for a bit instead (new gpu is noticeably smoother yay), then did a quick search, found that a specific version of proton works, switched to that version and it played. That’s the first game that has had such trouble for me.

It sounds like you might have some network places set up for windows to use but that are no longer reachable (or something along those lines) because that shouldn’t be taking so long so you might have things timing out in the background.

Or your internet is slow and it’s taking a long time to communicate with one drive or send its screenshots of your document to their creep department.

Or maybe a print driver that no longer exists still has an orphaned entry in the registry and it spends some time trying to locate it.

Or malware has set up hooks for any new window that pops up but the print to pdf dialog is set up in such a way that it churns very inefficiently on that window specifically.

I joke but any one of those might actually be what’s going on.

It will also give an error if you try to add a return value anyways.