All I was thinking about watching this video is “I wish they’d open a MicroCenter in Phoenix”; I’d had a blast visiting the Tustin location after seeing TMBG in LA this May.

I then googled it for fun and found out they announced they’re opening one a month ago! I checked national news to see if I was in some jolly alternate timeline (unfortunately, no) or dreaming, and it appears not.

Good! I just remembered that from the days I used to use Debian on a Microsoft Surface.

Just curious: which AV1 encoder were you using?

I ask because librav1e and libaom performance is dismal, but libsvtav1 can at least do ~1.00x speed, which is impressive considering how complex the codec is.

That’s not to say I don’t welcome improvements, though I think AV1 Vulkan requires hardware support I don’t have on my RX 580; it wasn’t until the RX 7000 series that AMD cards started getting hardware encoded support for AV1.

Which DE are you using? At least for XFCE, there’s HDPI and XHDPI themes you can choose in the Window Manager settings.

Honestly, I get wanting to keep using those laptops and that a perk of Linux is extending a device’s usable lifespan, but the last i386 machines are probably ~20 years old at this point; I think an upgrade is justified.

You can score an old AMD64 machine for next to nothing these days - you might even find something throwing one out; I think the sweet spot is a 1st or 2nd generation Intel Core i Series machine.

Thank you for saying this before people started crying, “Linux is getting encrapified!” without understanding what the article was actually about.

Do note that the yt-dlp version in stable will go out of date; I recommend installing it from the backports repo so it keeps updating.

Technically Raspbian Jessie, I think- I was gifted a Pi 3 in ~2016 and fiddled around with it for a while. I also made some cursed choices, at one point running Windows 10 IoT Core on that thing… though luckily not for long.

In 2017 or so, I started toying around with Ubuntu in VMs. It wasn’t really until 2020 or so that I started trying other distros; Debian Buster was probably the first non-Ubuntu distro I’d tried (excluding RPi stuff), and I mostly stucked to Debian besides one Arch install.

At a certain point in 2022, I found myself using Unix tools so much I was starting to wonder if I should just use Linux instead of Windows. It was at this time that I tried NixOS in a VM for the first time and thought, “Wow, this is cool… I’m sticking with Debian, though.”

Around that time, I threw Debian Testing (then Bookworm) on a second 256GB drive, ostensibly as a “test run” for daily driving Linux, and by “test run”, I mean I de facto quit using Windows; a few months later, I opted to use dd and copy that “test install” over my Windows install on my bigger 1TB drive (of course with sufficient backups so I could copy my Windows files over). That install is still the one I use on my desktop today and has just transitioned into Debian Testing/Forky*

*A name I quite honestly hate, mostly due to the fact that Forky represents everything wrong with America today the Forky Asks a Question shorts beat out Steven Universe Future for an animation Emmy, though honestly, I don’t know else what I was expecting to happen.

I also use XFCE. My desktop’s currently on Forky and went through all of Trixie, and the emdia keys have worked fine.

I’ll have to fiddle around and see what’s going on, though it may take a few days to get back because I’m starting school again soon, so I’m quite busy.

For reference, what programs do you tend to use with media keys? For instance, VLC, Firefox, etcetera.

What desktop environment are you using?

And maybe just give device model for good measure.

The big 2⁵!

To be fair to macOS, it’s still Unix-based, which at least makes it less miserable for development than Windows.

I would still go for Linux any time, though.

You’re right in some ways; Windows is closer to a microkernel than Linux, though it doesn’t perfectly adhere to the philosophy of - there’s supposedly weird things like drawing calls in the Windows kernel that should be in microservice, I’ve heard

However, I wouldn’t necessarily call microkernels a detriment; in fact, Linux is a bit of an odd duck for going monolithic - modern Apple operating systems also run on a microkernel. Monolithic is an older architecture, and there are worries about the separation between components and system resilience e.g the webcam driver can’t crash the whole kernel.

In practice, it’s less of an issue, and there really aren’t any open source microkernel operating systems that are practical for production desktop and server use, which has a microkernel though there are certainly solutions for embedded systems.

QubesOS is built on Xen hypervisor, which uses a microkernel design, but Linux is then run in multiple VMs on top of it, which makes it more of a technicality in my eyes. RedoxOS also runs on a microkernel and is certainly intended as a desktop operating system, but its hardware support is limited; GNU Hurd is even more limited in that respect and not really usable.

Neat. I love Debian, but its documentation is crap! I hope this works out and I can see an improvement.

I mostly agree with you. However, I think there are some caveats to upscaling; there are so many lazy “4K AI UPSCALE BEST QUALITY” videos online that just don’t look good and were clearly put there just to get views.

However, I’ve also found they have their uses; for instance, I wanted to laser cut a TMBG Flood logo once, but there were very few good images online that traced well in Inkscape. I ended up doing an AI upscale of the least terrible one with a white background, and that traced pretty well in Inkscape.

I was messing around with HomeAssistant the other day, which uses the same speech recognition engine, and I found it to be decent.

I think including the word “OpenAI” in the post name is somewhat a misnomer that implies an encrapification not really happening to the FFMPEG project.

Yes, it is true OpenAI originally developed the Whisper model, and I hate OpenAI; however:

• Whisper is actually open source, unlike most OpenAI crap.
• FFmpeg isn’t even directly using the OpenAI version, written in Python - they’re using a port to C++ called Whisper.cpp
• We’ve been able to use speech recognition for decades, so unlike other AI models, I don’t think a speech recognition model that does it better is problem.
• You don’t even necessarily have to compile FFmpeg with Whisper support.

I get the dislike of AI, but the idea of association with OpenAI is overblown and not really reflective of reality. Now I can get not wanting to use open source projects whose developers don’t reflect your principles; however, I think this ethical issue is more indirect than may initially appear and is not a strong reason to quit using what is still the most effective media conversion tool.

No need to panic in this case. While I hate OpenAI, there’s two things to note here:

• Whisper is an open source library for speech recognition rather than generative AI, run entirely locally. It’s just using ML to do something we could already do with computers (speech recognition), but better.
• They aren’t even directly using the OpenAI version - they’re using whisper.cpp, a port of the model.

Maybe it’s because I only use stable on my laptop with Flatpaks, but honestly, Bookworm never got that crusty to me until recently - it feels like new software versions didn’t introduce a lot of must have features in the past two years. Only hiccup was I had to install the backports kernel to get Wi-Fi working.

I used to use this, but I always found it really janky - window boundaries not updating, weird graphical glitches, etcetera.

It was especially annoying to use with Photoshop and GPU acceleration (I do GPU passthrough to my VM).

In the end, I just abandoned it and just used the monitor the VM’s GPU is plugged into.