← All TWiT.tv Shows (Audio)
All TWiT.tv Shows (Audio)
Untitled Linux Show 234: Crescent Wrench AI
01:20:32
Mozilla CEO, The GPL Lawsuit, & The Return of Mainline
Loading summary
Transcript231 lines
- [00:00]
A
This week we celebrate Mozilla finally hiring a new CEO. And we hope he doesn't do dumb things. We talk about the Linux foundation and its finances. Then that GPL court case we covered, oh, quite a while ago finally may be getting closer to resolution. The mainline kernel packages are back on Ubuntu and a whole lot more. You don't want to miss it, so stay tuned.
- [00:26]
B
Podcasts you love from people you trust. This is Twit.
- [00:34]
A
This is the Untitled Linux Show. Episode 234, recorded Saturday, December 20th. Crescent Wrench AI. Hey folks, it is Saturday and you know what that means. It is time for the Untitled Linux Show. We're going to nerd out over the Linux kernel and software and some hardware stuff. It's going to be a lot of fun. I have my two normal holiday co hosts with us today. We've got Rob and Jeff. Welcome to the show guys.
- [01:03]
C
Good to be here. Yeah, we were just a day and non holiday.
- [01:07]
A
Yes. Yeah, well, Ken is. Ken is usually around non holiday, but he is working more hours than he wants to. Well, yes, Rob, yes, that was the implication. Anyway, we were comparing notes before we got started about the various side projects, like where we've done work for people in the past and was telling the guys, I'm actually trying to sell the little tiny business that I've got here locally to be able to work on some other things. We sort of decided that it's really nice though to have a server or a virtual server somewhere because we enjoy sshing into boxes and installing updates.
- [01:42]
B
Yep.
- [01:45]
A
Well, there is a business that is trying to do a CEO update. Rob, do you think, do you think this is the one or is this going to be another unintentional interim CEO over at Mozilla?
- [02:02]
B
I'm gonna say this is the one. It's the one big deal here. Yeah.
- [02:07]
C
So.
- [02:07]
B
So there is a lot going on at Mozilla this week, starting with a new CEO. So Anthony Enzer, Demayo de Mayo Demio, I don't know, has been named the new CEO of Mozilla Corporation replacing the interim CEO Laura Chambers. So he started out with a public message saying, quote, people want software that is fast, modern, but also honest about what it does. They want to understand what's happening and to have real choices. Mozilla and Firefox can be that choice. First, every product we build must give people agency in how to work or how it works. Privacy, data use and AI must be clear and understandable. Controls must be simple. AI should always be a choice. Something People can easily turn off. People should know why a future works the way it does and what value they get from it. Second, our business model must align with trust. We will grow through transparent monetization that people recognize and value, you know, like, like, like Google Money. Third, Firefox will grow from a browser into a broader ecosystem of trusted software. Firefox will remain our anchor. It will evolve into a modern AI browser and support a portfolio of new and trusted software additions. So lots of AI there, lots of trust, lots of other things. You know, I'm all for AI myself, but a lot of what I hear from, you know, many of the folks out there is that trust and privacy, at least for them, contradict with AI or in general, I guess they just don't trust AI. But so to me, you know, this message there, it's their way of trying to combine the very different paradigms, you know, how to make Firefox an AI browser built on trust and privacy. You know, the way that things have been going at Microsoft with Firefox is it is clear they have to do something. And maybe this is the right step. Sink or swap, I don't know. But you know, have no fear, it was, it was maybe implied in that statement, but if you didn't catch it, you know, if you, if you love Firefox and hate AI and Enzer Deo has confirmed there will be an AI kill switch to disable it all saying his, quote, rest assured. And this is what he said afterwards after some people called it out, apparently didn't catch that. He said there'd be choice and options and it wasn't exactly hard in there. Like maybe his choice between AIs, I don't know. But anyway, it says, rest assured, Firefox will always remain a browser built around user control that includes AI. You will have a clear way to turn AI features off. A real kill switch is coming in quarter one, 2026. Choice matters. And demonstrating our commitment to choice is how we build and maintain trust. And then also I think even before this on Mastodon, Firefox developer Jake Archibald said, quote, something hasn't been made clear. It wasn't clear, but anyway, it was kind of there. Firefox will have an option to completely disable all AI features. We've been calling it the AI kill switch internally, but I'm sure it'll ship with a less murderous name. But that's how serious and absolute we're taking this. So, yeah, I guess apparently the initial statement, people didn't quite read between the lines and it Wasn't necessarily clear that you're gonna be able to turn it off just because it's there. You don't have to use it, but it's there for those who want it. But, but, but, but. Let's go on. Firefox isn't the only Mozilla software with things in the works. Thunderbird has made some announcements for 2026 as well. Don't worry, it isn't AI, but it is Microsoft. Thunderbird 145 will see native Microsoft Exchange support. Okay, that's it for the Microsoft stuff. Not, not that. Not that horrible after all, is it? Anyway, they're also going to continue to work on their mobile offerings. And we should also see the awaited Thunder Mail and Thunderbird Pro released in 2026. And finally, we can look forward to what they say is going to be a refresh of the Thunderbird calendar user interface. User interface is something I've always thought Thunderbird needed. And then they made improvements recently. But keep the UI improvements going because I like them. The new CEO. Big things in the works with Firefox AI. And maybe not as big for Thunderbird. I don't know. It depends where you sit. Maybe they are as big. I don't know. Thunder Mail is pretty cool. So I'm looking forward to a lot of this stuff.
- [07:35]
A
You know, Thunderbird being able to talk with Microsoft Exchange is kind of the killer feature that it's missing right now. So if that's something that they can bring to the table, I think that's going to make Thunderbird suddenly a first class contender again.
- [07:51]
B
Yeah, it's definitely, definitely a big improvement for those who are users.
- [07:58]
C
I don't know how much people really.
- [07:59]
B
Hate AI as much as it's so I don't. Hold on. Jeff.
- [08:06]
A
Jeff, you're using the wrong microphone or something. You're 100 miles away. You got very small all of a sudden, not any better. That's hilarious. Yeah, I thought at first somebody, one of us was accidentally feeding audio back and I was hearing Rob like through speakers or something. Is that me making that noise? That was funny. Still not sure what's going on there. Yeah, the deal with Jeff, you just keep talking and when you're loud enough to interrupt us, we'll let you know. The deal with Mozilla and Firefox, it's interesting because, like, they want so badly to put AI into the browser. And the browser, at least the way I see it, the browser is the tool that can get you to the AI platforms, whether it be, you know, Copilot built into GitHub or any of the others that gets hosted on a web page. Except for a couple of really, really small use cases, I just don't see the point of putting AI in the browser. Now there are a couple that make sense and that's things like if a picture does not have an alt text, you can ask AI to generate an alt text. Like that is genuinely useful, but anything else they're wanting to do, it's like, why, why spend the, the, the work on it when people really don't want it?
- [09:45]
B
I mean, you could say, you know, I kind of hate really long form articles myself. I like to, I like give me something condensed. So I mean you could say summarize this page, but you can also do that with any other third party.
- [10:00]
A
Yeah. Do you, do you really want you, your web browser to be the thing that does that? I don't, I don't know.
- [10:06]
B
Oh, oh. Are we going back to the trust thing? Do you not trust the accuracy?
- [10:12]
A
It's, it's not even that. It's just like, it's the wrong, it's, it doesn't feel like the right tool for the job.
- [10:17]
B
Maybe. I mean, I, I, I compare it to like the text, to speech, visual aids, people, you know, it's like it's going a little beyond that, but I don't know.
- [10:28]
A
Yeah, yeah. How about now?
- [10:30]
C
Can you hear me?
- [10:31]
A
It's about the same. Jeff. I don't know what the deal is. Something is real weird there. Okay, I am going to go ahead and take it while Jeff works his audio still. And I am going to talk about the GPL and we may come back and talk about Mozilla stuff here in just a minute. But there is a big GPL case. I think we've mentioned it in the past. It's been a while. And there is now an update. Update. So this is the Vizio case. And what, what's going on here is Vizio makes TVs, they make smart TVs and they have taken GPL code and put it into their TVs. And this code, obviously it's GPL. And so that means everything that links to it is supposed to be covered by the gpl. And so the Software Freedom Conservancy sent a letter to Vizio and said, hey, you use the gpl, you should send us a copy of all of your source code. And Vizio said no. And actually what I think that is, they sent a partial copy of it. And SFC was like, nope, that's not Good enough. That is not how this works. They actually took them to court over it, which is a very rare thing to happen because court cases are generally expensive. So nobody really wants to do it. Well, this was quite a while ago. It's like over a year ago, I think, that this court case first happened and we now finally are getting some rulings out of it. And so we have a tentative ruling here that grants the Software Freedom Conservancy's motion on this issue. So here's the one specific issue that we have a tentative ruling on that a direct contract was made between SFC and Vizio when someone from the SFC requested the source code to a tv. Now this is a pretty big deal because one of the legal challenges to the GPL and to all of these copy left style open source licenses is that it's not a contract. And the people that you're trying, so like the company that you're trying to force to release the source code, they would say we're not a party to that contract, therefore we are not required to abide by it. And so that's one of the legal things that needed to be tested about the gpl. And it looks like it is going to win with the, you know, the clear intent of the GPL is going to win the day in this court case, which is a really good thing. Like going through a court case and being vindicated is going to, you know, in a legal sense, it's going to strengthen the GPL quite a bit. It's interesting, as you read through this, there's some almost bizarre legal arguments that Vizio was trying to make. Like if you're in a lawsuit related to the gpl, you automatically lose your right to ask for source code. I'm not sure how, you know, what hoops they jump through legally to try to get to that point, but seems like the judge was not having it, which is good. So this is not the end of this, but it seems to be a step in the right direction. I imagine within the next few weeks we will get a bit more news as this tentative ruling becomes sort of an official, an official ruling. Now that's not necessarily the end of the case, that's just this one particular, one particular bit that has been ruled upon. But it's an interesting update and like I said, it seems that things are going in the right direction. So. Welcome back, Jeff.
- [14:18]
B
Thank you.
- [14:19]
A
Can you hear me now? We can hear you now. Perfect.
- [14:21]
C
I blame Chromium.
- [14:24]
A
Sure, I, I do that too. I would blame Chromium. Too.
- [14:27]
B
Well, yeah, you're in a Firefox discussion. Of course Chromium's gonna be like, okay, I don't want to be a part of this.
- [14:35]
C
Yeah, we're just gonna be that. There's Gemini in there going, we're going to censor that. No, what I was trying to say before on that one was just going back to the Firefox. I don't think people hate AI that much. It's more of just how much it's forced on you all the time. And it's like, A.I. i think, is a wonderful tool, but I only want the tool when I need the tool. Not. It might be the best crescent wrench in the world, but I'm eating supper. I don't need a crescent wrench right now. Stop trying to force it on me.
- [15:07]
B
So since there's an off switch, you're good with this one. What firefighter is doing, Oh, I already.
- [15:12]
C
Have it turned off. I have AI because it pops up like I'm watching YouTube and a lot of times I'll just click on a video so it automatically jumps to 2x speed, and then it'll be like, oh, you want to summary this video and we'll do this for you. I'm like, no, just go away. When I want AI I will contact you. Until then, don't bother me.
- [15:33]
A
Yeah, yeah.
- [15:35]
C
And I don't like the automatic summaries. Automatic. We're going to do this for you. No, you'll do what I tell you to do. And until then, leave me alone.
- [15:44]
B
You know, though, you're actually doing the reverse of talking me out of AI it's making me want to actually try Firefox again just to see what this new stuff is.
- [15:55]
C
Well, I can't tell you much about it because it kind of got in the way. It turned me off. I turned.
- [15:58]
B
I know you can.
- [15:59]
C
I went in and killed it all. I actually put it in the discord. How to go through and filter all that out and how to shut it off. But no, I've done some wonderful things. I use it at work all the time. It'll do very complex tasks. You got to double check it. But I don't need it for just dumb little. Here's three paragraphs on a new motorcycle I'm looking at or the latest Nvidia driver or something. It's like, I don't need you to summarize this. It's already pretty short. It's fine. Get out of my way.
- [16:33]
A
I actually know how to read. Please don't do it for me. I don't need my reading shoot up before I get to it.
- [16:41]
C
As Jonathan's wife, just post in Discord, I am the boss here. Yes, I said stop. And did I stutter? You know, just.
- [16:49]
A
No. All right, well, here in just a moment, we're going to turn it over to Jeff and we're going to talk about a new kernel and how to get it. That's coming right after this.
- [17:02]
D
Hey, everybody. Leo laporte here with a little bit of an ask. Every year at this time, we'd like to survey our audience to find a little bit more about you. As you may know, we respect your privacy. We don't do anything, in fact, we can't do anything to learn about who you are. And that's fine with me. I like that. But it helps us with advertising, it helps us with programming to know a little bit about those of you who are willing to tell us your privacy is absolutely respected. We do get your email address, but that's just in case there's an issue. We don't share that with anybody. What we do share is the aggregate information that we get from these surveys. Things like 80% of our audience buy something they heard in an ad on our shows or 75% of our audience are it decision makers. Things like that are very helpful with us when we talk to advertisers. They're also very helpful to us to understand what operating systems you use, what content you're interested in. So, enough. Let me just ask you if you will go to TWiT TV Survey 26 and answer a few questions. It should only take you a few minutes of your time. We do this every year. It's very helpful to us. Your privacy is assured, I promise you. And of course, if you're uncomfortable with any question or you don't want to do it at all, that's fine too. But if you want to help us out a little bit. Twit TV survey 26, thank you so much. And now back to the show.
- [18:25]
C
This story caught my eye because if installing a later kernel in Ubuntu used to be. It used to be a script. I used to do it and there was a mainline tool and it did the work. And you just selected the kernel you wanted and installed it. It build the mod files, you know, against the header files of the kernel. It worked with the new kernel. It was pretty straightforward. It was. It was a graphical little interface and it worked great. Well, there was an issue a while back, you know, I. Maybe a year ago, two years ago, something like that. The mainline repository broke and the mainline repository is the Ubuntu repository where they build new kernels with the Ubuntu settings. So for whatever reason, something in there broke and they weren't building the new kernels now. Yes, I know people, you know, a lot of our, a lot of our users are going, you can just manually build it, pull the current settings of your existing kernel, use them to compile the new kernel. You'll be up and running, away you go. But that's more than a lot of people want to do. The reason I say pull the existing kernel settings is it just integrates better with your operating system if you match how the original OEM kernel was built with the new one. So it's, it just integrates a lot smoother so you don't have modules that are missing or anything like that. Now I was looking through various feeds for stories to talk about this week and you know, oh, what, what's going to catch my eye? So I'm flipping through, looking at, looking at stuff and I saw this and I opened it up because I was curious how they were recommending that the kernel get installed on Ubuntu and of course all subsequent flavors the Kubuntu lube onto, so on. Well, the story linked in the show notes gives you two ways you can do it, and both are using the mainline kernels that the Ubuntu team builds. Now, before we get too far, you know, and if you want to update your kernel, I just got to say this. These kernels are produced with no warranty. The Ubuntu kernel team will not offer support for these kernels in case you have issues, and they will not be held responsible for any damages these kernels may cause due to improper installation or use. Well, you might be asking yourself, why do they even make these? Well, the mainline kernels are for system maintainers and OEMs who will take the responsibilities on themselves. You know, maybe you're a Dell or some major company and you need hardware support or a feature newer that's in a newer kernel. And this way you can do it and you can have it match what Canonical normally builds. But because Canonical is not running through all the various quality checks that they do with kernels that get released and make it into the distributions, they don't know if they're going to be any odd corner case issues for any long term listeners of the show. They know that personally, I've run them in the past without issues. But I'm also not an Enterprise business that needs five nines of uptime. You know if I had a weird crash because of it, it's not the end of the world. I reboot and roll back to a previous kernel or whatever. So it's risk versus reward. But I never really had too many problems with them. They were pretty solid out of the box. Bottom line is I was happily surprised to see the mainline kernel tool is back. Now in the article linked in the show notes there are three commands to run. It'll install a kernel with the aid of that graphical tool. It is from a non official repository. Now it has been around quite a while and it's pretty well trusted. But if you are a person who really doesn't want to go that way because it's a non official repository. There is a second method in the article and it talks about how you can download all the kernel parts you need, which does require, you know, your CPU in your system. The example in the article is the files you need are for the x86AMD 64 kernel, which is what most people will need now I should say, if you're not hardware kernel savvy. Even if you have an Intel CPU, the AMD 64 files are the ones you're going to want. The 64 is the important part. I think AMD was tacked on there because I think they were one of the first ones or the first one to have 64 bit chips in the consumer sector. I don't know for sure. I'm swagging on that one.
- [22:55]
A
No, that's what it is.
- [22:57]
C
Okay. It then talks about how, you know, you go into the directory they're downloaded in and then you can use the sudo space d package space-I space *.deb to install them all. So still pretty easy. You know, you're not. You're not pulling header files, you're not manually massaging previous settings to apply to new ones and baking and all that kind of stuff. And you're not going through the options file to make sure your kernel is getting compiled with the proper options. It's because it comes from the mainline repository. It's all basically kind of been done for you. So easy to install just a few more steps and not as automated. The mainline tool lets you remove old kernels as well, so it was kind of an easy way to manage the kernels you had on your system. Now with that being said, if you do install and there are problems, you have your previous kernel still in there because most likely you didn't Uninstall it and almost all the bootloaders, basically all of them that I know of, I'm sure there's probably an exception or something, but it allows you to choose an older kernel if you so desire. So it's if you have a kernel that just really is bricking your system, you go into your bootloader and go, well, I'm just going back to this older kernel that worked and you can just jump in there and then you can remove the newer kernel or maybe there's some updates or reinstall because something didn't go good or whatever you need to do. So it's not an end of the road if you got a wrong kernel in there. Have a look at the article linked in the show notes for full details and you know, not sure exactly when it showed back up, but welcome back. Mainline.
- [24:43]
A
Yeah, you talked about some of the people that want to be able to use that and I will add another one and that is other maintainers that are maintaining packages. You want to be able to go and check stuff beforehand. You want some group of people running that new kernel to be able to find bugs before it becomes the official kernel. So them too.
- [25:04]
C
Yeah, that's a good call out I hadn't even thought of.
- [25:09]
A
All righty. Rob, what is up at the Linux Foundation?
- [25:14]
B
So the Linux foundation is a nonprofit that supports and funds major open source projects by providing neutral governance, legal and IP help, infrastructure and community programs so these projects can grow and stay sustainable. Sustainable and well, finances are looking pretty good for them over the past year and projected for the next year. Their newly published 2025 annual report shows an organization that is expanding its role into, you know, areas like AI and has the finances to suggest they have the capacity to keep scaling that work up. In the report, the Linux foundation forecast gross revenue of 311,343,021 for the 2025 year, crossing the 300 million mark for the first time. The breakdown shows over 133 million from memberships and donations, over 83 million from projects and services, and 29 million from training and certifications, and over 58 million from events, sponsorships and registrations, you know, plus other revenue in there too. On the expense side, the report forecasts almost 285 million across expense categories. So that, that leaves a good. Oh, I didn't jot down the math, but you're looking at over 30 million to, you know, throw, throw some money my way, I guess. I don't know. So out of that, the expenses, the Largest share is labeled project Support at almost 182 million alongside spending on training, projects, infrastructure, community tooling, event services, international operations and of course the Linux kernel project. The annual report is also a recap of where the Linux foundation put its energy this year, which highlights several notable launches and milestones across its portfolio. One of the biggest is the formation of the Agentic AI foundation, intended to provide neutral governance for open standards and tooling in the rapidly growing Agentic AI space. It's all about AI today, guys. The Linux foundation also points just like every day, just like every day, that's the life we live these days. So they also point to the contributions of Newton and Open Source robotics learning and simulation engine efforts supported by Disney Research, Google, DeepMind and Nvidia brought under the Linux foundation stewardship to accelerate open robot learning. On the networking side, you know, Linux Foundation Networking announced S M 1.0 this year positioning it as a modular AI native platform to help integrate AI into networking applications and operations. But you know, when you put this all together, this report shows a growing organization, you know, growing revenue, investing heavily in project support and expanding into new areas like AI governance is sustainability, funding models. And if this finance trajectory holds, it should strengthen the Linux Foundation's ability to provide even more funding, better tooling and more durable governance for Linux and open source projects that need the long term stability. You know, we're always talking about projects and you know, financing being one of the toughest, tough things that you know when, when somebody's just doing something solo and for, for just the love of it, they get burned out. So you know, hopefully the Linux foundation can help minimize that burnout by paying people to do the work.
- [29:24]
A
Yeah, so I've got a, I've got a wonder. Let's see. Can I full screen this real quick? Yep. Okay, let's see what we can do here because I am going to share my screen. Check this out. This is the screenshot from Pharonix article. And I've got to tell you, that gross revenue ring, I just, it's driving me nuts because it doesn't make any sense. It makes no sense at all. So that shows about 90% of everything being the membership of donations and membership of donations is not even 50%. And I looked at that and I went something, something is not right there. And then I went and actually pulled up the report and in the report itself they made an adjustment to those same, those same charts. It's like these make more sense. So I was sort of, while you were talking, I was Looking at that, going, did they let the AI make the ring charts? Maybe they should not have.
- [30:31]
C
I was going to say, boys and girls, that's why you always double check the work that's done when you ask AI to do something.
- [30:37]
A
That's maybe exactly what that was. Oh, goodness.
- [30:41]
C
Grace, you're on mute. Rob.
- [30:44]
A
Now it's you.
- [30:45]
C
It's your turn.
- [30:46]
A
It's your turn.
- [30:48]
C
Audio problems.
- [30:49]
B
I wasn't talking. I wasn't talking. I was just shaking my head and. And moving my lips to.
- [30:53]
C
To.
- [30:53]
B
Yeah, Exercise. Yeah, that's something, though. Had to be AI, you know, unless they, I don't know, added an extra zero, few zeros to their Excel spreadsheet on the. On the one, Something like that.
- [31:08]
A
Yeah. Yeah, maybe. Maybe the Excel spreadsheet. Because that's. That's one of the ways you would do that. You would create one of those. Maybe the Excel spreadsheet, they just selected the wrong row of numbers to get the graph from.
- [31:20]
B
To be fair, when I was reading my notes, I changed those numbers to, you know, I didn't read out the full number. I. I changed, you know, 82 million and 132 million or whatever it was in my own notes. One of them, I had like a K after it. So I don't know. When I was reading those numbers off, there was a time where you may have caught me, like, pause, like. No, these were all millions, right? I know these were all millions.
- [31:47]
A
They're all millions.
- [31:48]
B
So. So I don't know, maybe did something dumb like that.
- [31:51]
A
Yeah. Who knows?
- [31:52]
B
Probably AI though. It was AI.
- [31:55]
A
I bet it was AI. All right, I have got something that really surprised me in the. Well, in the Nvidia world. We're going to talk about here in just a second. Hello, everybody.
- [32:07]
D
Leo Laporte here. You know what a great gift would be, whether for the holidays or at just any time, a birthday, a membership in Club Twit. If you have a Twit listener in your family, somebody who enjoys our programming, and you want to give them a nice gift and support what we do, visit TWiT TV club TWiT. They'll really appreciate it and so will we. Thank you. Twit TV Club Twit.
- [32:34]
A
So the headline that caught my attention was that Nvidia driver packages are now using the open Kernel modules. And I went, what? Officially? Already thought that was going to be months away. Well, so the rest of the story is it's the Arch Linux Nvidia driver packages. This is still, I think, sort of a significant turning point. And it's tied to the fact that the stable Nvidia 590 drivers is dropping support for some old cards. And I think that is what led Arch to do this. But what Arch has done is that they are using the open source kernel modules, the. Oh, I forget what it actually calls it. They're replacing the Nvidia package with the Nvidia Open Nvidia DKMS with Nvidia Open dkms, et cetera, et cetera.
- [33:38]
C
And Nova driver.
- [33:39]
A
Yes, Nova, I believe that's what this is.
- [33:42]
C
Nvidia's solution is Nova, right now they're using a lot of no VU to load some of the.
- [33:48]
A
And this is not no vu. And so the, the article calls that out at the very end that you could also use Novoo. But that is not what this is. This is actually going, I believe, to Nova in Arch. So if you're running Arch and you go grab the Nvidia driver, by default, you're gonna be running Nova. And they apparently think it's ready, or maybe it's just Arch and previously Nova.
- [34:11]
C
Was doing some of the loading of some of the other code because Nova was not ready. Some of the latest benchmarking that had happened on Phoronix, they said it's still a few months away, but maybe they had some nice code go in that polished it a little more.
- [34:29]
B
That's a few months away for the normal slow distributions like Ubuntu, but for the big guys like Art already.
- [34:38]
A
Yeah, apparently. Apparently. Um, yeah, I don't run Nvidia, nor do I run Arch. So like, this is very much outside of my immediate knowledge, which is why it's useful to have you two here because one that runs Nvidia and the other has run Arch.
- [34:56]
B
Well, I got both, yeah.
- [34:58]
A
Oh, there you have.
- [34:59]
B
Kashi is based on Arch.
- [35:00]
C
So yeah.
- [35:02]
A
So I wonder if this means that like with your next update to cache, it's going to start trying to grab the open source kernel modules.
- [35:09]
C
I don't know, I'm on 590 right now. It just loaded that in. So I don't know if that's the open source modules or if it's the full proprietary driver.
- [35:20]
A
Now the other thing with this is it's not clear from the way the actual article is written up. Is this only for the Pascal drivers? Pascal, Maxwell and Older. And I'm not, I'm not 100% sure, but it looks like, from what I understand, it looks like they are moving to the Nvidia open in any way. I Think regardless, whichever way it goes, it's interesting to see, you know, this is becoming more and more of a thing and it's sort of getting to the point of being an option.
- [35:54]
B
Yeah. Pretty soon I'm not going to be able to rag on Nvidia for their proprietary ness.
- [36:02]
C
Well, in all fairness though, a lot of that just got moved to firmware and they have an open source, you know, oh look, we're all open load proprietary firmware. It's in firmware. So it's doesn't count.
- [36:16]
A
Yeah, well, I mean that's. Honestly, that's what AMD did as well. That's what these guys have done.
- [36:22]
B
I agree. It doesn't count.
- [36:24]
A
Does it not count? I don't know. It works on my Fedora machine. That kind of means it counts.
- [36:31]
B
Well, it works on your Fedora machine. That's what I mean. He's saying it doesn't count as open source because they just moved it to the firmware.
- [36:40]
A
Right. So you get into a whole other sort of discussion there. And I've had this discussion with some of the EFF guys. It's like where do you draw the line at what is and is not okay to run proprietary code on your machine? Like must every firmware blob be open? And that's the line that a lot of folks draw. The purists. And I have to ask. Well, no, no, no. There's more firmware on your machine than just the things that you have to load. Every chip has firmware running on it of some sort of.
- [37:17]
B
Exactly. Like if you, if it makes you run your. If it allows you to run your, your hardware on any system easier because it's. That piece is just in the firmware, then I know that's a good thing.
- [37:31]
C
Well, you start getting into what is firmware. If I have an NVME drive and it's kind of locked down, could you consider that it's stored on chip and it's not easily accessible if it's locked down?
- [37:44]
A
Your NVME drive absolutely has firmware on it. Hard drives have for a very long time.
- [37:49]
C
It's hard for that.
- [37:50]
B
I don't have to worry about too much.
- [37:53]
C
I mean. Yeah. How do you really get into the. You can get pretty pedantic in the software versus firmware, right? Yeah.
- [38:02]
A
So it's an embedded device. It's still running the Linux kernel. Is it software or is it firmware? Yeah.
- [38:09]
C
And if you say, well, it's stored on chip because that was one of the old definitions. NVMe all the data is stored on chip.
- [38:17]
A
Yep. Yeah, that's fair. I see what you mean. Indeed, indeed. All right, speaking of data stored on chips, there is an update for a file system. Looks like Jeff's going to talk about it. What is new in OpenZFS for all of you non Americans out there?
- [38:39]
C
Well, being American, it's going to be ZFS for me. But there is a new release of OpenZFS. It's version 2.4.0 and it's been set free into the wild. It's compatible with kernels 4.18 up to 6.18, so it won't do 19 yet or they haven't fully tested it to see if it's forward compatible. This release had 121 contributors, so this is a very supported and alive project. So there's, there's a lot of people putting time and effort into this. And that was 121 contributors to this specific release. Not just overall, in general, there's some nice key features with this release, such as now you're able to set a default a default quota for a user, a group or a project. For uncached IO you now have direct IO fallback to a lightweight uncached IO. When things are unaligned, there's unified allocation throttling. So it's a new algorithm designed to reduce vdev fragmentation, which means the reduce the fragmentation of the free space within a vdev or virtual device. It's not talking about the data itself being fragmented. There's better encryption performance using AVX2 for AES GCM, and that's advanced Encryption Standard Galoris counter mode. It's an authenticated encryption algorithm used by OpenZFS to provide both confidentiality and integrity for data at rest. It's the default and recommended encryption method in basically all the modern OpenZFS versions. Due to high performance and security, it'll now allow ZIL on special UDEVs or VDEVs when available. ZIL is a ZFS intent log and what it does is it logs a synchronous writes to disk, ensuring data integrity and consistently, especially after crashes or power loss. 240 also extends special small blocks to land ZVOL rights on special VDEVs and allows non power of two values and adds the ZFS rewrite P command to preserve logical birth time when possible. For minimizing incremental stream size, it adds ZFS space rewrite P which preserves logical birth time when possible. To minimize the I had that twice the stream size, it adds a dash a or a dash Dash all option, which scrubs, trims, or initializes all imported pools. There's a zpool scrub dash capital S, space dash capital E to scrub specific time ranges. So if you only want certain parts of your file system, changes scrubbed can do that. Add zpool prefetch t space brt to prefetch brt. That's the block cloning table, and it relaxes topology restrictions on special dedupe VDEVs. There's improvements to a shift handling multiple gang blocks, improvements and fixes, new deduplication optimizations and fixes, new block cloning optimizations and fixes. That's just a few of the improvements to the OpenZFS file system. Take a look at the article linked in the Show Notes for a lot more details. The article has a link to the GitHub release page, which includes all changes, and there's even removed module options. Happy reading.
- [42:33]
A
Yeah. So where does OpenZFS make sense? Who's the kind of the target audience to actually run it?
- [42:40]
C
I think it's more enterprise and you know, I mean home labs, I guess, if you want to really experiment with some of that.
- [42:48]
B
But replication, is that one of the.
- [42:52]
A
Things that really does well?
- [42:53]
B
Yeah. So a few examples where I've seen it in production is one Proxmox I don't use it on my Proxmox. I should have. I kind of wish I would have, because you have to have that for. For rep for replication and then also a backup appliance that we've used for work. Well, we do it virtual often. Often. But it uses zfs for, for replication and also just like Butterfs for. For the snapshots. So I I in my opinion, at least the two big features are replication and snapshots. Like you can, you can replicate and then you do like a Z send is the command and it will replicate over only the changes between your local drive and wherever you're replicating it to.
- [43:47]
C
Yeah, so it's way overkill for an average home user. Unless you want to play. Because part of it too is it can handle if you're writing data to a drive and you lose power and everything crashes because of the logging files and all that, it knows what was written and what wasn't and it can handle that kind of disruption.
- [44:10]
B
It also has like data scrubbing to protect from bitrot kind of things. And it's a lot like Butterfest, but with more to it, really.
- [44:22]
A
Is it just a coincidence that ZFS and XFS are just a couple of letters away? Or is one of those Intended to poke fun at the other.
- [44:31]
B
Oh, you know, I bet there's something there.
- [44:35]
C
OpenZFS was, my understanding, was designed to basically get around some of the copyright issues that ZFS has. Because otherwise zfs, or I shouldn't say. I shouldn't say copyright, I should say patent issues with.
- [44:51]
B
That's licensing.
- [44:53]
A
Yeah, yeah.
- [44:53]
C
Because there was always a gray area, could it be included in the kernel or not? And the decision was, well, let's just not include it because we're not 100% sure it's completely safe to include in the kernel. So Open was written to say, well, let's just do what ZFS does, only without the infringement issues or possible issues. That was never. It's never been truly sorted out.
- [45:19]
A
Yeah, it looks like OpenZFS is also not upstream in Linux, probably for similar reasons. Yeah, because you can run it on Linux. You can compile the kernel modules yourself.
- [45:33]
B
Yeah, Ubuntu is like one of the only ones I think that will offer it on install. I don't know really any others that will. And I think mainly because of the licensing.
- [45:44]
A
Yeah, makes sense. All right, let's see. We've got something very interesting coming next. Rob is going to tell us about a notable first for the Rust in the Linux project. And we're going to hear about it right after this.
- [46:02]
D
Hey, Everybody, it's Leo Laporte. You know about MacBreak Weekly, right? You don't. Oh, if you're a Macintosh fan or you just want to keep up what's going on with Apple, this is the show for you. Every Tuesday, Andy Inaco, Alex Lindsey, Jason Snell and I get together and talk about the week's Apple News.
- [46:19]
A
It's an easy subscription.
- [46:20]
D
Just go to your favorite podcast client and search for Mac Break Weekly or visit our website, Twitter, tv, mbw. You don't want to miss a week of Mac Break Weekly.
- [46:30]
B
Okay, folks, Rust is a great programming language and when done right, it can be more secure by protecting from memory access types of vulnerabilities. But remember that being developed in Rust doesn't automatically make it secure. You know, there are other types of vulnerabilities that could still be in the code. I actually had an argument with a. I don't know, fairly prominent. I don't know if that's quite. That might be a stretch, but a Linux social media moderator, I'll say on this recent. On this recently and you know, I don't know if it was, if, if I wasn't communicating clearly or what, but he did not seem to get What I was saying when I, when I kind of said being developed in Rust doesn't automatically make it secure, he's like, what do you mean it's more secure because it has this? I'm like, I didn't say it wasn't more secure. I'm saying there are other bugs. It's like. But that's most of the vulnerabilities out there. Anyway, it was, it was a fun conversation. I'm like, yeah, that's great. But anyway, here is another example. Oh yeah, one of the things I said, don't you listen to podcasts to keep up with this stuff? Anyway, here's another example. As the Linux kernel just hit a milestone, the first CVE has been assigned to Rust code in the mainline kernel. So Greg Cohartman announced that CVE2025-68260 applies to the Android Binder driver rewrite in Rust, which landed in the Linux 6.0.18 kernel. Binder is the core IPC mechanism Android uses to let different parts of the system communicate. The issue is a race condition, a timing bug where multiple code paths can access the same length data at the wrong moment. Because this code includes some unsafe rust, that race can corrupt the list, previous and next pointers and and trigger memory corruption that leads to a crash. The key detail here is that this CVE is a crash only bug. It's not described as a remote code execution or privilege escalation problem, but it does not. Oh, and it does not appear to enable attackers to take control of your system. But who knows, you know, sometimes they find a vulnerability, it bounces here and they get to there, but at this time it seems like it's just going to crash your system and annoy you. So that distinction is part of why this is being treated as a useful real world data point in the Rust. In kernel debate, Rust reduces a large class of memory safety problems, but it doesn't eliminate bugs, especially where unsafe code and concurrency meet. So Grego Hartman also used the moment to put the rust in context. The same day this Rust CVE was assigned, there were 159 other CVE is tied to fixes in the C portions of the code base, reinforcing the message that Rust is in a silver bullet. But it can reduce vulnerability, volume and severity as adoption grows. Though my injection here is, you know there is far more C code than Rust code in the kernel. So at a.
- [50:21]
A
159 to 1 ratio.
- [50:23]
B
Yeah, ratio, there we go.
- [50:25]
C
Statistical analysis.
- [50:26]
B
You know, 159 to 1 ratio is the C code to Rust probably a lot more than 159 to 1. So I don't know. When you only have one, it's really hard to do a real statistical analysis on it. But we'll see as we go forward what happens. But though I do think Rust will be the safer choice in the long run, as you know, Carl Hartman pointed out, it isn't a silver bullet. So just remember that. But for this bug, a fix is already available. If you're running Linux 6.18 update to 6.18.1 or newer if there's something else out by the time you hear this, and it's also present in the 6.19 RC1 the kernel CVE team teams General guidance applies here too. So update to the latest stable release or whatever is beyond these vulnerability ones and you should be good to go. So the inclusion Rust is good, but not a perfect silver bullet. It is still susceptible to developer error, just less susceptible. So you still gotta update when your security updates come.
- [51:43]
A
Yeah. So there's like three different points to be made here and you touched on most of them. But I'm going to reiterate, first one that we've talked about before is the the kernel has a CVE policy that says essentially any bug is a vulnerability because it's in the kernel. And that is maybe a little bit of an oversimplification and over exaggeration, but that is what they do. And so like if this were in any other program, it may not have even been considered a cve.
- [52:12]
B
I'm not sure though you could also then say those 159 may actually be a lot lower than those.
- [52:18]
A
A lot of those. Yes.
- [52:19]
B
That changes that ratio potentially.
- [52:21]
A
Yes. One of the other points to make and remake is Rust has the unsafe keyword specifically to be able to do kernel work because there are some things that you have to do with your kernel specifically talking to low level devices that is inherently unsafe. Rust can't do its fun compile time memory management magic because it's not actually managing memory, it is managing memory map devices. And that's quite a different thing. And then the last point I was going to make is essentially that there are plenty of other ways to screw up your code than just memory management. And we've discovered that too.
- [53:03]
C
Yeah, more is not total. So it's more secure. That doesn't mean it's totally secure, it's just a little better.
- [53:10]
A
And there's also kind of this like intersection that you're going to have because like as code has existed longer and people have banged on it. It's going to get more secure because you've fixed bugs. And so you've got all this Rust code that is very, very new in the kernel. And so just because of that, it's going to have a whole bunch of bugs in it. Maybe it has fewer bugs now than the C code did when it was first written. Yeah, but it's still. I think there's still a bit of catching up to do, so we'll see.
- [53:36]
B
In. In my argument, I had. I was trying to point at the Rust SU vulnerability and yeah, he kind of came back, was like, well, yeah, it's new. There's going to be bugs that. That doesn't make it not. I don't know. Secure is like. Well, yeah, that's why I said it's not secure.
- [53:55]
C
The Colonel's rusty. It's now a jalopy. Oh, yeah, I never said that.
- [54:00]
B
I mean, things to look out for.
- [54:05]
A
But it's not an experiment anymore.
- [54:07]
C
It is not.
- [54:09]
A
Rust is not experimental.
- [54:11]
B
It's here to stay.
- [54:12]
A
It's here to stay, probably. All right, let's talk about KDE for a minute. There's a couple of fun things in this week's KDE update. A few things that I am personally excited for. One is that KDE is offering is adding in 6.6. Most of this is coming for 6.6. They're adding support for ambient light sensors, which I know phones have that. I didn't know laptops had that, but apparently laptops like the new framework 13 has it. I'm very curious if the framework 16 has it because I've got one of those on the way. So support for that in KDE is going to be really interesting. I wonder if they're going to try to do dynamic brightness adjustment with that. That would be. That'd be interesting, like your phone does. There's also a fix in Plasma 66. This is something I have noticed, and that is in Plasma, trying to do a Windows HDR game sometimes looks terrible. And part of that is because the Plasma guys have some real strong ideas about the way HDR is supposed to work. And those ideas don't jive with the way that Windows decided HDR was supposed to work. And sometimes the games get caught in the middle. And so they're finally giving us a workaround of sorts that hopefully should automatically happen. Hopefully that'll fix some of the things that I've seen. They are, once again, flicks, fixing clipboard issues. Under Wayland, I've run into that one multiple times on kde. And then one of the other ones that I saw that I really liked here is that Plasma 6.6 will reduce the frame drop on monitors with extremely high refresh rates. Definitely don't want to drop those frames. And then there's some other portal stuff and it's all good. Like there's lots of goodness in here. But those were just some of the ones that really stuck out to me. We've got the link to the Pharonics article and then that has the link off to the Nate Graham blog post.
- [56:13]
C
My understanding was the sensor was so it could auto dim and take care of the.
- [56:19]
A
Well, I mean you figure that's why they put it in there. But I just don't know if KDE is actually going to use it for that or if the only thing that it's going to be is you can run a command line and it'll tell you how bright it is.
- [56:30]
C
I think they'll auto adjust.
- [56:31]
A
I imagine so.
- [56:33]
B
Well, as wizardling I think it is. In chat, I have Discord mode on, so I can't see his full name. But you know, he's kind of pointed out that auto brightness is often not that great. And I've had the same experience with my phones. You know, I've enabled on my phone and it just dims that poor moments.
- [56:56]
A
Most of the time. See, my experience with a lot of brightness is that most of the time you don't notice it because it's working well enough that it's doing what it's supposed to. It's only when it fails that you notice it.
- [57:07]
C
Yeah.
- [57:07]
B
And for me it was failing multiple times a day.
- [57:10]
A
Yeah, that's less than great.
- [57:14]
C
It's worked great for me.
- [57:16]
A
Yeah, I really enjoy it. Like when I walk in with my kids when my kids are trying to sleep, one of them wakes up or something and I've got my phone and it's not like, you know, the light of the sun. While I'm trying to browse discord or.
- [57:27]
B
Whatever, I just go in, go in there, turn down brightness. When I'm done, turn it up.
- [57:34]
A
You do it manually. What a plebian.
- [57:36]
B
Well, I, I would do it automatically if it worked nice.
- [57:40]
A
Yes.
- [57:41]
B
If it worked. What? Better I would use it. Awesome.
- [57:46]
A
All right.
- [57:47]
B
I guess a four year old, five year old phone, well, there's your problem. You didn't work well that either.
- [57:54]
A
But did anything work well four or five years ago it was a whole different world.
- [58:00]
B
Yeah, that timing there.
- [58:03]
A
All right, Jeff, let's Talk about video card performance. I think that's what we're talking about.
- [58:09]
C
What's new here, it's kind of a double header benchmark. But you know this. Earlier this month I covered the Nova driver and MESA drivers in relation to the Nvidia proprietary drivers and we even looked backwards at how the. From July until recently, the open source drivers, they've increased 20% in performance. But that was, that was a, that was another show. Well, today's story, we're doing that again but we're also looking at how the current drivers for AMD cards are looking and this is going to be more end of the day, what hardware should I get based on how well the drivers are stacking up and how well the hardware stacks up against each other because both Team Green and Red have been improving their drivers and since memory prices are so high, maybe a new computer is not in the works, but maybe a GPU upgrade might scratch the itch. So the Nvidia cards tested are using the 618 kernel with Mesa 26.0 develop NVK and Zinc for OpenGL graphics driver support. The proprietary driver is 580.95.05 and it's on the 617 kernel which is the default on Ubuntu 25.10. On the AMD Radeon side they're using their upstream and open source driver support using the same Linux 618 kernel plus Mesa 26 devel and combination. And that's, that was the. So it's, it matches a lot with the Novu testing with the MESA and the kernel. Now, as much as Michael Lerbl could, this is about comparing the different cards with the advancements made on the GPU drivers. The GPUs tested were the Nvidia 50 series and the AMD 9000 series. Well, what did we find out? Looking at all the benchmarks by geometric mean and just as a reminder, this is a mathematical mean that takes out spikes in the data. So it's basically it gives you a. A better average and one flyer doesn't totally wreck the data. So if, if anything's extremely outside of a normal distribution, it minimizes the effect of it. So the data where all the cards couldn't run was also removed to eliminate a shift in the data. So like in this case the Vulcan ray tracing benchmarks because the Novu driver doesn't support it currently. So they just. All results for the Vulcan ray tracing were not included in the final geometric mean. If you want a full open source stack then the Nvidia 9000 series cards, or I mean the AMD 9000 series cards are your best bet. The 9060 XT came in just above the 5060 Ti running the proprietary driver and just below the 580 running the Novu driver. The 9070 and 9070 XT were next to each other on the chart coming in above the 5070 running the Nvidia driver and just slightly behind the 5090 Novu and the 5080 Nvidia proprietary driver. So 5090 with the Novu 5080 with the proprietary, we're running just about the same speed on over all these benchmarks. Of course you know when you look at just proprietary, the 5090 was running way out in front, but it's kind of a whole separate card. It's kind of semi professional slash ultimate gaming, you know, Halo product for the average consumer. We, we again we have talked in the past how the Nobu driver still has a long ways to go but since July they've had a 20% improvement. So at their current rate of development it won't be too long. And there's not going to be much of a parity between the Nvidia proprietary driver and the Novoo Nova when it's ready or currently an arch is ready I guess driver. So open source is catching up. Take a look at the article in the show notes to see which benchmarks matter the most. To you. And depending on the specific benchmark, sometimes team red really does great against team green and it gets out of the normal stacking order. But because the geometric mean it's not all the time. But you know, maybe, maybe that's your use case. So take a look at the results and maybe it's time to treat yourself to a new GPU.
- [62:56]
A
Yeah, I can say the RX9070XT is quite nice. I did finally do the upgrade.
- [63:06]
C
Nice.
- [63:06]
A
Yeah, yeah, I'm enjoying it. It's nice to be able to run like unigine superposition. On my previous card I couldn't run the full 4K benchmark. It's like you don't have enough video card memory for this.
- [63:19]
B
Yeah, I think next, like you said, their memory is getting expensive. I think next time I upgrade my computer I'm going to need to find a new computer that supports old memory.
- [63:32]
C
Well the problem is a lot of people are doing that and the old memory is kind of expensive and especially the old CPUs.
- [63:38]
B
Well this way I could bring my memory with me. I just take it out of my old box and put in my new one.
- [63:42]
C
Oh, I see.
- [63:43]
B
Okay.
- [63:45]
A
That was a thing for a while. I'm trying to remember if that was the DDR3 to DDR4, the DDR4 to DDR5. You could get like the same CPU but do it with both kinds of memory. I think there were a few of those CPUs that could supported both.
- [63:59]
C
There were, but it was motherboard dependent so you couldn't just enter switch them. You had to have the either DDR3 or the DDR4 motherboard. And that was intel that was supporting that. AMD did not have that.
- [64:11]
A
That's been, that's been a few days since that last happened. I couldn't remember the exact details.
- [64:15]
C
Yeah, there's a few people rushing out and looking for used computers that people don't know what they've got. So they're like, oh, it's got 32 gigs of RAM. Okay, I'll buy it. You know, and people think, oh, I got it. And then they realize the memory and it's worth more than the computer computer. And, and the, the 50, was it 5007-00 and 57x3DS? They're all like sold out and going for twice what the modern CPU is because so many people wanted to just do a CPU upgrade to their old AM4 motherboard system.
- [64:49]
A
Yeah, yeah, makes sense. Crazy times, crazy times we live in.
- [64:56]
C
So you can get a hell of a deal on, you know, computer cases.
- [64:59]
A
Power supplies, all that other stuff.
- [65:02]
C
Yeah. And actually GPUs are not too bad right now. The availability is pretty good.
- [65:07]
B
Is that because people are taking their memory out and just selling that and like, well now I got all those other parts here.
- [65:13]
C
Well, a lot of people just aren't building anything and a lot of like storage, memory, even a lot of CPUs are going into enterprise servers. Now you can say, well wait a minute, the, you know, 9950 is not going into enterprise. No, but they're allocating that wafer instead of going to a consumer market. Now it's getting going to the EPICs and the enterprise market. So but cases, power supplies, they're just like, yeah, we're going to have some deals trying to move this stuff.
- [65:49]
A
It might not be a terrible idea to pick some of that stuff up now. And you know, at some point in the future the prices and everything else will come down. Hopefully at some point get the things.
- [66:01]
B
That buy the things now that will be usable in future products like power Supplies. Motherboard may not be a good idea.
- [66:11]
A
Because motherboard will be out of date. Well, I mean, that's interesting, though, because, like, you can run a. You can run a case from, like, 20 years ago. Those have not changed a whole lot.
- [66:19]
C
Power supply?
- [66:20]
A
Close to it, Just about.
- [66:22]
C
I'm still running a Silverstone fortress, too.
- [66:26]
B
Yeah. As long as your power supply has enough power to support whatever's coming, you should be good. Because maybe an adapter is needed. But they all pretty much, I think, the same voltages since. Pretty much forever.
- [66:37]
C
It's the same connector for a. Since the ATX standard came out.
- [66:40]
A
Yep, Yep. All right, well, here in a second, we're going to get into some command line tips. Before we do that, we're going to take our final break for the show.
- [66:51]
D
Hi there. Leo laporte here. I just wanted to let you know about some of the other shows we do on this network you probably already know about. This Week on Tech. Every Sunday, I bring together some of the top journalists in the tech field to talk about the tech stories. It's a wonderful chance for you to keep up on what's going on with tech, plus be entertained by some very bright and fun minds. I hope you'll tune in every Sunday for this Week in Tech. Just go to your favorite podcast client and subscribe. This Week in Tech from the Twit Network.
- [67:21]
A
Thank you. All right, Rob, you are up first. And the only thing I see here is sat. Sat. Why does this sound like a children's book or a very weird. It sounds like some insult that you would hear about from some other country like the Australians might call themselves. They would call their enemies this. You sot. Sot.
- [67:46]
B
Did I put SOT in there twice?
- [67:48]
A
You did, yeah.
- [67:49]
B
It's only one sot. I don't know, copy and paste error or something. When I put it in the notes, it's just sot. Just sot. I don't know what it means. I did not get into depth into that aspect of this command line. Tip.
- [68:07]
C
So, system observation tool.
- [68:10]
B
Hey, there you go. It's a system observation tool, and that's exactly what it is. Kind of like all the other top things, like, you know, top, etop, htop, whatever, other tops. It is kind of like a nice dashboard to kind of monitor various things for those who are looking. You can see my screen upper left has my cpu, has my cores, my threads, where it's at. Now, this is a vm, so not a whole lot is going on right now. It's purely for testing for the show you got right below, they have a box has the memory and various somewhat graphical representations with different stars. I don't know what they are. You got the disc with the read rights. You have a health score for CPU, memory, disk, processor, temperature. I have. Everything's at 100% health. Maybe it wouldn't be if I actually did some things on here. There's a bond box in the middle. This is sod actually. I don't know what that is. I didn't think about that. It has a bunch of stars in it.
- [69:21]
A
That might just be the logo.
- [69:23]
B
It might just be the logo. It may have nothing to do with anything.
- [69:26]
C
I installed this on mine and it says SOT and there's some kind of wavy little bit graphics that go by there. So I should say in a virtual machine.
- [69:36]
B
I am running this on a non graphical system. There's no X or Wayland or KDE or any of that. So I am running this on just the console. So that's why it's going to be different. You probably have a little more of a graphicalness to yours. But when I use command line it's often when I'm sshing into my servers and doing things. So that's my preferred use case for command line. If I'm on a graphical interface I might as well use a graphical tool. But anyway, I digress. You have a spot for networking connections details there. You have a spot for your processes and I'll move my head that's right behind me and then at the very bottom right you have a networking kind of showing your up and down, your bits per second, your max, total and IP and various details here. I'll get out of the way there. No, I. I can't get out of the way there. It's. It's there. You probably could see a little bit there anyway. So yeah, it's a nice little. You know, you need a one spot. Command tools just kind of monitor your system. There it is.
- [70:50]
A
It's SOT cool. It's easy to install with pipx.
- [70:56]
B
It is.
- [70:57]
A
That is the way to do it. That's the way I would tell people to do it. PIPX is great. I don't know if we've ever had that as a command line tip. Maybe we should. If we haven't. Pipex is really good. It's not the one I picked for today, but it's came up. It's what? Sorry.
- [71:10]
C
Yeah, well, I was gonna say it's what they. Because I installed it and I used their installation script and I Had to install pipx before it would work. It says, oh, we have this, we have this. Oh, you don't have pipx. So I had to install it. Then it said, oh, yeah, okay, here you go.
- [71:28]
A
Yeah, I saw that. And I did not want to run their script because I didn't want to have to take the time to look at it. And I'm not going to just run their script without looking at it. And then I saw, oh, we'd sell with pipx. Well, there you are. I'm sold. Looks like we've not covered pipx. I claimed that for a couple of weeks. We'll start the first, start the new year off and I'll talk about pipx. There we go. I have a real quick one actually, with thanks to my wife, the Mrs. Jonathan Bennett. Reminded me of something that we've not ever talked about before. And this is generally KDE specific in the way I'm going to tell you about it. Although I believe you can do this in GNOME and the other desktops as well. If you're on your desktop, like your desktop environment, and you're working with a window, you can right click in there and there is going to be an option to open in Terminal. KDE has this in Dolphin. I know that, like I said, GNOME has something similar as well. Open a terminal up with you already in the right directory to be able to look at it. That's cool. That's the one that she reminded me of. And then there's something else that I've been doing actually quite a bit while doing code compiles because you end up in these weird nested directories. You know, you're in source firmware, source bin output, PIO builds, you know, you're 15 levels deep. Suddenly it's like, I don't want to double click through all of that. Well, you can just go, you know, change directory in the terminal to wherever you want to go and then launch your file manager from there. So in the case of kde, it's just you run Dolphin CD to where you want to be and run Dolphin and it'll pull you up a Dolphin window that's right there. And then you can, you know, copy and paste it into Google Drive or whatever it is that you're doing. You've got your files right there to.
- [73:17]
C
Be able to look at.
- [73:18]
A
So not exactly a command line tip. Let's say about 50%, a command line tip to get to where you're wanting to work with right away and we'll.
- [73:28]
C
Round up for you.
- [73:29]
A
Yeah, There you go. And then tip of the hat, Mrs. Jonathan Bennett for reminding me of that one.
- [73:35]
B
Should have brought her on the show to give it to everyone.
- [73:39]
C
Yeah, that would have been good.
- [73:40]
A
I believe she has her hands full with the four children out there at the moment. Probably would not want to jump on the show. So anyway, Jeff Dog, as in what's up dog?
- [73:54]
C
Well, this actually started out as somewhat of a joke. So when we're all doing the show notes and we have a link to we all share a link to a same document and we're working on it. It isn't always filled out as we're still looking what we want to cover. And sometimes a command line tip, if a person has a blank line, someone might fill it in with something funny or silly little thing like maybe Jonathan will have an empty line and Rob will fill out something in there. Like as for Jonathan's command line tip is I should shave my head so I can be cool like Rob and Jeff or you know, one of many other silly goofy things. Well, I was going to do that to Rob a couple weeks ago and I was going to fill in Dog the Better cat replacement because my cat doesn't like me. But then I thought I should check to see if that was real. And to my surprise, Dog is a real command.
- [74:44]
B
I could have got a freebie.
- [74:48]
C
Dog is not a better cat command though. It is kind of like dig and it's a command line DNS client. It has colorful output, supports DNS over TLS and DNS over HTTPs protocols, and can admit JSON output. Now, the usage is pretty straightforward, at least for the simpler commands. For example, dogexample.net gives you a query of the a record of a domain using the default settings. Dog example dotnet MX looks up the MX records. If you add 1.1.1 it will use a specific name server instead. And if at the end of that you add a capital T dash capital T, it'll use TCP rather than U udp. So take a look at the band page in the show notes for the full usage of what goes well beyond a simple command line tip because it can do a lot of powerful things over a lot of different transport options with a lot of different outport options, and it can check a lot of different record types. So happy researching.
- [75:57]
A
Very cool Dog the better Dig.
- [76:01]
B
It looks like Dog is an application run by Alan Popey, apparently. Oh, maybe it was a snap by him. I did a snap install just now and it says Dog v.0.1.0 from Alan Popey Interesting. I don't know.
- [76:18]
A
I don't even know who Alan Popey is. Name means nothing to me.
- [76:23]
B
He used to work for Ubuntu. He's involved in a lot of projects.
- [76:28]
A
Okay.
- [76:29]
C
Right now it says DOG is maintained by Benjamin Ogham Sago. S A G O.
- [76:36]
B
He probably snapped it up. He snaps up a lot of stuff.
- [76:38]
A
Probably. All right, well, hey, that is the show and a little bit of behind the scenes news. This is the last show that we'll.
- [76:47]
B
Do live for the year.
- [76:48]
A
So all of you that watch live, you'll have to. You can watch the show when it comes out. I think it'll come out at about live time. We have one recorded and so you won't miss it. You'll get your, you'll get your fix. It won't be us live. We'll be off hanging out with our families and having Christmas. So I want to let each of the guys get the last word in on anything if they want to or plug their stuff. Rob, quit playing around with DOG and hit the buttons to plug your plug, plug your things.
- [77:18]
B
All right, so first I want to plug that few seconds that I've had with dog. I like that. That is a lot nicer than Dig or ns. Lookup already. That's, that's a nice program. Thanks though. I wish you would have caught that and I wish I could have found that for my own tip.
- [77:36]
C
Well, Merry Christmas. There's your present, Rob.
- [77:40]
B
But yeah, well, it would have been nice for me to be able to show people that. Instead I'm just going to show people my website because that is where you could come connect with me. You know, maybe I'll. Maybe I'll post some things during our Christmas break because I'm bored and, and I just need to do some Linuxy stuff. I don't know, maybe I will, maybe I won't. If you want to see if I do or don't come find me@robertp Campbell.com and on there is links to my LinkedIn, my Twitter, my blue sky mastodon. And if you really appreciate what I do, click the little coffee cup to donate a coffee to me. And if you don't, well, you can always just say hi in one of these other social media platforms and it'd be cool.
- [78:32]
A
All right. Yeah, very cool.
- [78:34]
C
And Jeff, two quick things. One is you can find me on LinkedIn through Rob, but I will tell you, if you send me an invite on LinkedIn, if you're not already connected to Rob, I probably won't accept your invite. Just so people know, I get a ton of random spam trolling, phishing type ones. So just a. Just FYI, if you want to connect, connect with Rob, then you'll find me. The other thing is Poetry Corner and I do have an end of the year special poem. So you. You will, you will finish out the year on a good one. But today's is. Roses are red Violence or blue unexpected air at line 32. Have a great week, everybody, and have a great year.
- [79:24]
A
All right, excellent. Appreciate you guys being here. Thanks for ending the year out strong with us. It's been a lot of fun today and every day. Of course. Thank you for being here. If you want more of me, you can go check out Hackaday. That's one where Floss Weekly lives these days. That is where, up until very recently, my security column lived. But as I explained in the last one that I wrote, I have had four careers and I now have four children. And it was time for something to give. And sadly, that security column is one of the things that's going to, at least for the time being, take a. Take a break. We're gonna step away from that one. So I'm down to three careers, which is still too many. So I'm still working on some stuff. But this one, this show's not going away.
- [80:08]
B
We.
- [80:08]
A
We like this one too much. We do appreciate everybody that's here, though, and those that watch, those that listen, where they get us live or on the download. And we'll be back. You get to see us in a week. We'll be back next year for another Untitled Linux show.