Untitled Linux Show 189 Transcript

Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.
 

00:00 - Jonathan Bennett (Host)
Hey folks, this week we're talking about problems at Canonical and you can run Linux inside a PDF, oh yes. And then there's Rust in the Linux kernel and the new RISC-V mainboard for Framework that maybe you shouldn't buy. Hey, it's all interesting. You want to hear about it, so stay tuned. Podcasts you love From people you trust. This is Twit. This is the Untitled Linux Show, episode 189, recorded Saturday, february the 8th, rustrated by Frust. Hey folks, it is Saturday and you know what that means it is pregame for the big game. No, it's time for the untitled linux show it's it's uh, it's not just me.

00:48
We've got rob, we've got jeff and we've got ken and uh, apparently there's something else going on this weekend, but we don't care about that. We are here to talk about linux and uh, all of the crazy things that's going on in the linux world right now. And there are some crazy, and we're going to talk about some of them, and we're going to start with Rob, who has a bit of a warning for everyone.

01:13 - Rob Campbell (Co-host)
Be careful. So you know, from time to time we do bring up job postings from like open source and Linux companies when there's something interesting that catches their eye, and I'm not sure. But I think we have brought up openings at Canonical before. Well, this week I found a rant on Reddit from someone that applied at Canonical and although you know I'm not suggesting you don't apply at Canonical, suggesting you don't apply a canonical, I am giving this as a word of warning to be prepared for a four-month grueling process that may not necessarily end well. But you know, and I haven't been able to obviously verify every detail of this rant, but when comparing against some other posts and reviews on Glassdoor, things seem to line up pretty well with what this person experienced. So this rant on Reddit states he applied for a software engineer job and then he continues on saying that first they waste tens of hours of his time first with endless forms that that took him 10 plus him or her, I them, I don't know 10 plus hours to complete, and then an IQ test, which that's not too crazy, or an IQ style test. I've taken them before for employment. I mean, it's been a long time I've done mechanical aptitude tests for jobs way, way in the past I even took a reading test once, which everyone up for that job kind of thought was weird because it was actually an internal job and they made everyone take a reading and, I think, some other tests. But anyway, I digress. And this writer rant. He says he also had to take a language test. I think he means spoken language, not a programming language, even though it's a software engineer, because he does go on to say to talk about programming language stuff later on. Because after that, after all this tests and paperwork he had to go through, I counted seven interviews First, a technical interview on Python coding.

03:40
Second, a manager interview with the hiring team. Third, another tech interview. This was on system architecture and general tech questions. Probably could have combined it all into one. But fourth, an HR interview. Fifth, another manager interview, but this was from people who were not on the hiring team. Six, the hiring lead did an interview. And then seventh, there was a VP, a vice president interview and he says the hiring lead and the VP both gave really positive feedback.

04:18
He basically checks all the box and, like I said, him, her, them I don't know what they are, I'm just going to say he, because that's what keeps slipping out of my mouth. Um, so so this seems overkill, uh, to me, and a lot, a lot of management hours spent to interview one person, uh, but if that gets them the best then, uh, from an end user perspective I suppose that may be a good thing. But here's the rub and that stuff is like okay, that's grueling and crazier than any other job I've ever been involved in, other job I've ever been involved in. But the rub and the real part that's really kind of sad is, after four months of going through all of this, he was given a job offer that's not the sad part yet and he accepted the job offer. So. So after that and they did say they were in Europe, so in context here After that he gave his four-week notice at his current employer. So he's already geared up to put in his notice. He's done in four weeks they decided to cancel his application. They withdrew his job offer after all this time and then offering and accepting it After four crazy months scoping out every part of this person's life, what could they possibly have discovered in one more week that they missed all these other weeks?

06:16
In my experience, the hiring process takes two to maybe four weeks before an offer is given, and then in the US here the standard is kind of a two-week notice. So in that short amount of time I could see something being missed that wasn't discovered. But in four months that's the same. So what kind of due diligence is Canonical missing in their four-month process? How much management time is being wasted doing nothing?

06:56
In the comments there some of the things I've heard is that they often do not get the best person because in four months a lot can happen. They've already taken other jobs. So then they go to the next person. Oh no, they've already taken another job too. You know, you got to be a little quicker if you want the best people. Also due diligence, I guess you're going to get the people who really, really want to be with you. But there's a funny story in the comments too. Somebody said that their wife was going for some government job and the interview is like three months later they had a schedule shows up and, uh and um, hr person took him to the office. No one was there. It turns out the hiring person died two months previous. So that's just a side note that things can happen in three months and you know to go on and provide a little support to this.

07:51
On glass door uh, canonical gets a 3.3 stars. Now, maybe to be a little fair, people aren't. Normally people are more likely to go online and complain than say good things. But when I compared it against any place I've worked, 3.3 stars is less than any place I have worked, and with a rating of 53% of employees saying they would recommend to a friend and only 47% approve of the CEO, mark Shuttleworth. And a lot of the reviews on there are pointing to poor management. Others also list the hiring process is a con of Canonical. So again, I'm not saying you shouldn't apply there to work there, but if you do be prepared and maybe hold off on giving your notice to your current employer until the last minute, maybe somehow try to set yourself up for success if something doesn't go as you hope.

08:55 - Jonathan Bennett (Host)
Yeah, so the terrible employment practice seems to be just pretty typical in the tech world. I've heard horror stories like this before, where you know there's interview after interview and it takes forever to get to, you know, and weird interview practices, weird tests, asking people to do weird things for their interviews. That is unfortunate, but kind of normal. I do wonder and looking through the Reddit thread, other people were asking this too. I'm actually really curious if it's even legal to make someone an offer, they accept it and then renege on that offer before employment starts. I think this individual probably does have a pretty strong case of basically a breach of contract, a pretty strong case of, uh, you know, basically a breach of contract. Um, so it's not the, it's not the four month thing that really bothers me as much as it is the right. Essentially, they they failed to follow through on the, on the obligation they had.

09:55 - Rob Campbell (Co-host)
Um, yeah, and I could see if that, if that offer was contingent upon uh, background checks or I know in the old days they don't seem to do it much anymore, it seems a lot of places but contingent upon a drug test or something else like that. But I've got to imagine after four months that stuff was sorted out.

10:19 - Jonathan Bennett (Host)
One would hope. We don't have copies of the emails. We don't have copies of any contract that was signed. We actually better look at them. So you know, just in full fairness to canonical, it is possible that there's more to the story than what we are seeing, but on its face it does sound a little. Sounds a little sketchy.

10:39 - Jeff Massie (Host)
Yeah, it is a Reddit post after all, I mean, but we get, we got to always take that with a grain of salt, yes, and kind of a. You know. It's interesting, though, too, being on the other side trying to hire, you know, like, hire a programmer where we, you know we would have. It would be interesting. We'd have tests where we'd say, okay, write this simple thing, you know, maybe a bubble sword or whatever, and you can use any language you want, you can use pseudocode, you can whatever you want, and you'd be amazed at the number of people. Go, well, can I have access to google, can I? I need, you know. Basically, I'm a script kitty and I need to have, uh, have, google so I can look, all, look up all the answers but you know canonical I, it's a big company at thousands of employees.

11:32 - Rob Campbell (Co-host)
I think, um and for for a software engineer, you're bringing your vice president in. I don't know if that's maybe it's vice president of a section or a division or something or some part of it. Maybe not like shovel horse number two, but that that seems rather overkill too yeah, the big corporations.

11:57 - Jeff Massie (Host)
I haven't been involved with that. You don't go that high a ceo or president.

12:02 - Rob Campbell (Co-host)
They hire a good they a good team under them and they let them do their job. You know, a vice president hires good management directors. Whatever under them, let them do their job.

12:13 - Jeff Massie (Host)
You know, don't try to micromanage every hiring offer unless it's the vp of hr yeah, they don't have the vps at most like a big company like that, you you probably get like a director or something, because after you're about three steps away, they don't really care because they're never going to talk to you, they're never going to deal with you. It's, it's more on, okay, who, who are the? You know their direct leadership and maybe one step, maybe two steps above them, that'll have some interaction. Beyond that it you'll never see the person.

12:49 - Jonathan Bennett (Host)
Yeah, I don't know how Canonical does it.

12:50
I know I've read about some businesses here in the States big, rather large They'll do things like they'll have skip meetings.

13:01
This is something that SpaceX does, something Musk does that seems to work out well for them, and it's the idea that, no matter who you are in the business, you can say well, my manager's not listening to me, but I really believe that I'm onto something here and you can jump up a couple of ranks in the hierarchy and that's pretty interesting. This is totally unrelated to Canonical, it just came to mind and so you know it's not that unheard of for even a business the size of canonical, for one of the if there's multiple VPs, for one of the VPs to be involved in hiring decisions, and it also depends upon what level you're going to hire someone on at right. So, like this could have been just another engineer that was going to be, you know, taking care of mitigating CVEs or or whatever, or maybe they were hiring an engineer to be the lead of some new feature. So I mean there's there's a lot of, there's a lot of wiggle room here.

13:54 - Jeff Massie (Host)
Um, that that is, that is possible to at least imagine yeah, the company I'm at, we do skip levels, we we have it every so often yeah, no I it's a great idea.

14:06 - Jonathan Bennett (Host)
personally, I like it a lot.

14:08 - Rob Campbell (Co-host)
I have been familiar with companies where the VP does get involved in hiring decisions, but they work with the directors and the managers. I had this interview with them and they were great at this, not so great at this. Blah, blah, blah. What do we all think? And they offer their input, but getting it out of interview, that's just. I don't know. But yeah, I don't know what kind of software engineer job he was going for.

14:40 - Jonathan Bennett (Host)
Yeah, so I think that's really the takeaway here is that there's a lot of information we don't have about this Um, but I think it is also worth warning people, not just about canonical, but on the job hunt in general to, you know, be careful and sometimes companies aren't necessarily. Well, I think you can just generically say companies are not looking out for the interests of their potential employees. Like you're two different. You're two levels away from what the company is actually looking out for. So a good company will look out for its actual employees in addition to looking out for, you know, its customers and the thing that it's trying to do uh, even even less of a requirement, I suppose, to look out for the interests of its potential employees. So just keep that in mind when you go and have an interview.

15:35 - Ken McDonald (Co-host)
It's really easy to become a customer of that company.

15:40 - Jeff Massie (Host)
What's that, Jeff? I was going to say I wouldn't resign from my current job till I had the the next one, you know nailed down and I had an actual employment contract yeah, just really get let the ink dry a little bit.

15:54 - Jonathan Bennett (Host)
He had an offer and yeah, so that's right, like that's the question. He had an offer. He said he accepted the offer, but was there an employment contract? Was there ever that final yes, you have it quit your job? We don't know.

16:09 - Rob Campbell (Co-host)
I've never gotten an employment contract before the day I started working. I've never had an offer and then they said here's a contract, You're all signed up, here's a contract, come in this day, and then you fill out all the uh onboarding stuff I've. I've never worked anywhere that gave a contract ahead of time. So that's not a thing.

16:31 - Jeff Massie (Host)
anywhere I've been all right let's, but it wasn't a full contract. It was like we are hiring you, we are, so there's they were. They were legally saying we're engaging with you to hire you and yeah, a letter of intent.

16:46 - Ken McDonald (Co-host)
In some cases that gets called do we need to do a youtube video showing how to get hired?

16:54 - Jonathan Bennett (Host)
uh, that was kind of clunky, ken, but we are going to hand it to jeff to talk about youtube channels. What's it? What's this about, jeff?

17:03 - Jeff Massie (Host)
well as linux. We all have to start somewhere, and this first story, or my story, is going to help beginners, intermediates and possibly even advanced users as well. If you take a look at the link in the show notes to a HowToGeeks article, you'll find a list of 10 YouTube channels that the author says made them a better Linux user. Personally, I've learned about several new channels because of this article. I mean some I knew of, but many I was unaware of. But going through the list, number 10 is the Linux experiment, and the author comments that this is their favorite channel, which I guess maybe 10 is the best instead of one. I don't know, this is in their order so I'll have to let them talk to that, but they said the channel covers a lot of content, such as news updates, exploring different distributions, going over new tools, guides and a lot more, and the author calls out some of their favorite videos. They do is where they answer Linux misconceptions. Number nine is DistroTube, and personally I'm familiar with that one. Like the last channel, it covers a wide range of distributions, guides, tools and answers questions. The creator especially loves the DistroTube DT, loves tiling window managers and has written some scripts that users can use if they want to copy his setup. Number eight is Learn Linux TV, and the channel focuses on command line tools, how-to guides and covers bash programming. The channel is especially suited for new users who have just switched or are considering switching to Linux, making the content very beginner-friendly. Number seven is Switch to Linux, and this is also geared towards the very beginning Linux user, and there's also a dedicated website for the Linux news and tutorials linked in the channel, so they have a kind of a sister website to go with the YouTube channel.

18:53
Number six is Michael Horn. He's a content creator from Australia and, like many channels, he covers a wide set of topics in the open source arena. Something that's a little different, though, is that, in addition to covering common Linux questions, he also covers desktop environments and is pretty passionate about Linux gaming, and I agree with the author here that not many Linux channels cover Linux gaming beyond a passing comment. Number five is TechHut. Now, this one's a little different because it covers Linux, windows and Mac systems, and in this channel you'll see a lot of comparisons between the different operating systems side by side, so you can see how Linux stacks up. They also cover Linux servers and delve into hardware along with their normal guides.

19:37
Number four is not a stranger to this channel, as we've talked about this YouTube channel in the past Brody Robertson. His channel is a little different that he doesn't make a lot of guides, reviews or tutorials like, unlike you know, like other channels do, most of his videos are news, opinions and discussions. He can be a little controversial and polarizing depending on if you agree with him or not. Number three is is Infinity Galactic. The schedule for video releases is not as rapid as other content creators, but he is noted for having interesting, unique and alternative takes on Linux that you don't normally find in other venues. The author calls him out on his video for productivity apps, for example, which helped the author find a lot of hidden gems that he's come to love. Number two is Linux text. The author notes that the videos are more like documentary style videos and when reviewing different distributions, the content creator doesn't go into reaction videos but rather goes into in-depth research and gives all the good and bad about distributions. So this channel is for those who really want a deep dive when comparing distributions.

20:48
Number one is a bit of surprise to me, as it's nine to five Linux. The author talks about how the channel covers a wide range of topics and mentions that they also have a website. I think listeners to this channel would know they have a website and maybe not realize they actually have a YouTube channel. That was me. I only know the website, not the channel. One bonus that I personally was going to throw in here and you get this for free, I would like to add if you enjoy level one techs, they also have a level one Linux channel. They do a lot of hardware on Linux content, but they do hit software as well and it's a little more enterprise geared. But I really enjoy Wendell over there. So take a look at the article linked in the show notes and, whether you're a beginner or to the advanced user, there's probably something in there that you can find helpful and enjoyable. So happy watching.

21:45 - Jonathan Bennett (Host)
I'll add one more to that list, and that is another friend of ours, jeff Geerling, and he is sort of famous for the let's run full size GPUs on the Raspberry Pi, among other things. But Jeff is also another one to tune into if you're trying to keep sort of your finger on the pulse of the single board computer world of Linux.

22:05 - Rob Campbell (Co-host)
So I'm not a huge consumer of YouTube content. I kind of go there when I'm looking for something specifically. But I'm going to call out one of those on your list Learn Linux TV. I have gone there for some Proxmox stuff when I was first setting it up and I know he used to be a listener. I don't know if he's a listener anymore, but he was the one who ran with one of my great distro ideas Nice, yeah. So that is one that I am familiar with really only because when he was in our Linux chat I kind of checked him out after that.

22:47 - Jonathan Bennett (Host)
Yeah, pretty cool. All right, so let's talk Calibre. This is the ebook reader, right Ken?

22:55 - Ken McDonald (Co-host)
Yes, it's my favorite choice for managing ebook and we can thank Marius Nestor, since he wrote about the latest update to my favorite e-book management software. Here we're talking about Caliber 7.25. The highlights, according to Marius, include support for importing KFX files from 2024 Kindle devices that use the MTP protocol and a new option to adjust the size of the link and note icons in book details. You can now set custom icons for items in the tab browser by right clicking on them and choosing manage icon for this value. Calibre 7.25 also fixes a few bugs, including an issue in the Nook driver to make the Glowlight Floor eBook reader work on Windows systems, and an issue in the MTP driver where internal storage and SD card were swapped on some devices that have buggy firmware that assigns a lower ID than the internal storage to the SD card.

24:06
This release adds several new sources, including Alternatives, economics, afrique, exit 21, orient 21, and Football League World, if you're into those sorts of uses. I was looking to see if there was anything new as far as the Linux side, for new sources available through. There. Didn't see anything recent.

24:37 - Jonathan Bennett (Host)
Yeah, you know I don't find myself using an e-book reader, especially on the desktop hardly ever at all, but it is nice that it's out there. I boy, I really I do audiobooks these days, either audiobooks or like paper.

24:50 - Ken McDonald (Co-host)
I will still occasionally sit down with a paper book uh, one of the plugins for uh caliber is uh for overdrive Overdrive so you can link to different Overdrive websites or libraries, and it's actually been upgraded so you can use it to link to your audio audible books. Oh, cool.

25:15 - Rob Campbell (Co-host)
And also, ken, you can use that to manage your ebooks while reading that on like a tablet device too, right?

25:22 - Ken McDonald (Co-host)
Yeah, in fact, when I first started with it, I was using it to manage the eBooks that I had bought through. I can't think of the name of the site now it's been so long ago, but they had eBooks that were designed for the Palm PDA.

25:42 - Rob Campbell (Co-host)
So, Jonathan, you don't have to be a reader of ebooks on your desktop to utilize uh caliber.

25:48 - Jonathan Bennett (Host)
You just need to just manage it there yeah, I've got, I've got all of one device that sort of makes sense, and that's actually.

25:55 - Rob Campbell (Co-host)
I've got an ipad that somebody got me, um, but it's still not my favorite thing to read on you could get the um, the, the, the pine eink I don't remember what they call it, but there's that e-ink Pine tablet and then read these yeah.

26:10 - Jonathan Bennett (Host)
I don't know, that could work.

26:12 - Ken McDonald (Co-host)
But I went from my Palm PDA to the Nook. I think that was the second version. I still got it around here somewhere. It's just so old that I can't get updates for it anymore.

26:26 - Rob Campbell (Co-host)
And I want to apologize to Jonathan's wife for implying that he should buy anything I didn't yeah for real, it's the Pine Note.

26:38 - Jonathan Bennett (Host)
There it is. I'm trying to see. Yeah, there it is $400. Oh, my goodness, it's kind of expensive. It's going to take two. See, yeah, there it is $400. For $400. Oh my goodness, it's kind of expensive. It's going to take two weeks.

26:48 - Ken McDonald (Co-host)
Yeah, and I've actually got a recommendation for e-book reader to put on there.

26:52 - Jonathan Bennett (Host)
Oh.

26:53 - Ken McDonald (Co-host)
FB Reader.

26:54 - Jonathan Bennett (Host)
FB Reader. Is that for what? For Android?

26:58 - Ken McDonald (Co-host)
It's available for Android. There's even a Linux version.

27:01 - Jonathan Bennett (Host)
Nice, all version, nice, all right. Well, let's steam along. And speaking of running on linux, rob, where else can we run?

27:13 - Rob Campbell (Co-host)
linux. So over the years, many people have shared fascinating places that linux can run, you know. But you know you don't see a lot of that these days because Linux runs on almost anything you know. But with this story, you know, I have to bring back the trend, as this is just too weird not to share. You know you've heard of Linux running on ATMs, linux running on toasters, coffee makers, heard of Linux running on ATMs. Linux running on toasters, coffee makers, refrigerators probably. But would you believe that Linux can run on a PDF file?

27:56
Pdf, the portable document format, the thing used to, you know, print documents and sometimes fill out forms. Well, vk6, also known as Adding 202010 or Allen, a high school student with a obvious passion for programming and, I guess, web development, cybersecurity, has managed to run Linux inside a PDF file on the browser. I checked it out. You can even download it and run it locally on your browser. You can't run it online, at least not in the Mac OS PDF viewer. That's what I was using it on at the moment. I haven't tried on actual Adobe, but I'm guessing it's just a web only. So this work that he did builds upon an earlier project of his where he got Doom running within a PDF, the other software that runs everywhere. He now released what he calls Linux PDF, powered with a modified version of the tiny EMU RISC-V emulator that has been compiled to ASMjs using an older version of em scripten so users can use the virtual keyboard to interact with linux pdf run commands.

29:30
I checked out briefly. It didn't have an internet connection, but I was able to do an ip space a and see that the local ip was 127.0. I don't remember what the last was, but maybe one or you could send keystrokes to a virtual machine using the text box. So you know, it's pretty amazing that Alan has been able to programmatically running within a PDF, get Linux running within a PDF. But it kind of begs the question do PDFs have too much ability to run whatever? You know this is supposed to be for documents. I think PDFs power may. Maybe you should get dialed back a bit. If you can run Linux on it, you can run anything on it you can. You know any malware can be ran on a PDF. So you know, don't be clicking any of those PDFs in your email. Guys because guys, girls, folks because that. But besides that, this is another fine example of a high schooler that has an amazing career ahead of him. Compare this to to what I was doing at that age.

30:47 - Jonathan Bennett (Host)
Well, let's so I've got it. I've got another link. Before before, this guy made uh linux work. He made doom work and we covered it on hackaday, and so there is a doom pdf that you can go and run in your browser if you want to. It works in Chrome. I was playing Doom. I will not lie, I was playing Doom while Rob was talking, and the way that this works is that, yeah, there's a little script engine inside of PDF renderersers because the pdf spec calls for it. Um, and yeah, they've got. They've got javascript running in there, and once you have javascript running, you're sort of in a position to bootstrap the whole world yeah sounds very safe and we do have.

31:37 - Jeff Massie (Host)
We do have a question, ant Pruitt on a PDF. Why?

31:42 - Jonathan Bennett (Host)
Because you can.

31:44 - Jeff Massie (Host)
Yeah, it's just a dumb trick, you know. It's just like hey, wouldn't this be wild? It's the only reason, but it could be too, because Adobe said you know, we got rid of Flash. How else are we going to be the main leader in security vulnerabilities? It, it does yeah we got to do something to get that title back I thought you were going to say something else, I don't know.

32:10 - Rob Campbell (Co-host)
I thought you were going to say it's going to replace flash in some other quality way, and I was going to say that the video of it is not the best. But no, you, you got it right, uh, with your no, I mean it is.

32:23 - Jonathan Bennett (Host)
It is really. It is sort of amazing. And there's this. There's this long tradition of once once you sort of have control over a device, the next thing that you just have to do as a hacker is run doom on it. We've, we've covered doom on so many different things over on hackaday. Um, you know people, people will do things like, um, I think it was the the nintendo entertainment system, which does not have enough horsepower to run doom. Someone hacked a cartridge together to where they could. They could, um, I think they did it wirelessly, but they could flip individual frames over into the cartridge and then display the frames on the tv so that you could run doom on an NES. I mean, just the crazy thing. I've covered people running Doom on SIP phones everywhere they can, possibly. It's just, it's sort of a rite of passage.

33:11 - Rob Campbell (Co-host)
ILag says first JavaScript, then the world.

33:17 - Jonathan Bennett (Host)
Yep, yep, very much so. All right, let's move on to something else that'll run doom, and that is that fancy nvidia overpriced 5090 graphics card oh yeah, well, this, this is a two-in-one uh segment here.

33:36 - Jeff Massie (Host)
So first, something we haven't covered yet is NVIDIA has released the 570 series beta Linux driver with support for the 5000 series of GPUs. But in addition to supporting the 5000 series cards, there are a number of new features and bug fixes included in this release. The NVIDIA settings control panel will now use NVML rather than NVControl to control the GPU clocks and fan speed. This allows related functionality to work when using Wayland where the NVControl X extension is not available. So they do make a note, as some operations which were previously available to unprivileged users due to the privileges of the X server may now require elevated privileges. So basically, because of the greater security in Wayland, you might have to use your sudo to get things kicked off correctly. Now. They have added support for variable refresh rate on systems with multiple displays and they've added an application profile to improve performance on Indiana Jones and the Great Circle. Along with that, they've also fixed some with a profile where there was some corruptions, and it should fix Assassin's Creed Valhalla and Assassin's Creed Mirage. There was also a bug fix that would cause some Vulkan applications to crash when responding to window resize events, and they've also updated GPU overclocking to be available by default in GPUs that support programmable clock control. Now we had this before, but previously that was only available when bit three was set in the cool bits X config option. So now it should just be standard and you don't have to go through that. They have enabled 32-bit compatibility support for the NVIDIA GBM backend and if you have an HDMI monitor that would not work with some DVI outputs, that's been fixed and the bug they found was actually introduced in the 555.58 driver. Take a look at the article in the show notes, as it has all the information to the driver updates and a link to the actual NVIDIA page where you can see all the additional changes they made in this beta driver. There's a lot more improvements and fixes and whatnot.

35:54
I just kind of hit the high points, because now let's talk about some hardware. The second article linked in the show notes is where Michael Larable over at Pharonix reviewed the RTX 5080 and the RTX 5090. Michael received both the cards and he benchmarks them in gaming versus several 2000, 3000, 4000 series cards, along with the AMD 7000 series cards, so a big pool of graphics cards that you can compare what you're running to what these new cards will do. Now Michael is using the 570 driver that we just talked about and he does mention that it's working good and hasn't found any instability or performance issues. Since the cards have been out a little bit. People interested probably have already looked at a lot of the Windows benchmarking and the Linux benchmarking falls right in the line where the cards are roughly 30% faster for the 5090 versus the 4090. And the 5080 is roughly 14% faster than the 4080 Super. It should be noted that the 4080 Super, the 4080 and the AMD RX 7900 XTX they're almost the same performance, so you're not getting a lot there.

37:07
How did the 5000 series cards do with heat? Again, you probably already know, but pretty well. They tend to shed heat quite efficiently, which they had to, because when you look at the power numbers, they are power pigs. The 5000 series takes a lot of power and the 5090, for example, pulled a peak of 578 watts in Michael Erbel's testing. Basically it has a 37% increase in power for the 30% increase in performance. The 5080 had a much smaller increase in power, but it also had a much smaller increase in performance.

37:42
The real bottom line comes to should you purchase the cards If you're on a 4000 series card. Don't upgrade. It's not worth it If you're on a much older generation than possibly. But the problem right now is the 5090 and the 5080 have had such a small volume ship. It's almost a paper launch and some people have said it basically is because so few numbers have gone out the door to retailers.

38:11
If you want to get one, I would plan on playing the long game.

38:15
Don't expect to get one for several months, but if you watch all the websites and check in every once in a while, maybe you'll get lucky and come across one that's actually priced reasonably.

38:26
Almost all of them you'll find out there right now are either double the retail suggested price or more, and it's all because scalpers have them. I would not expect any relief in price or supply until AMD releases their cards in March or April. Personally, I think it's because NVIDIA was trying to get AMD to release hardware before they were ready, and I don't think NVIDIA was ready either, but they had their drivers in a better place. So I think they were trying to make AMD look bad and when AMD then pushed out, they said we're not ready, we're not going to play NVIDIA's game, we're going to take those few months, make our driver and polish everything up better. I think that took NVIDIA off guard a little bit and they just didn't have any of the hardware I don't think they were quite ready to truly release. But that's just my personal opinion. Take a look at the article linked in the show notes if you want a deeper dive into the individualized performance results. But until then, happy gaming.

39:34 - Jonathan Bennett (Host)
Yeah, it's interesting. I've been doing a little looking on what AMD is offering, rumored to offer, and you know they first came out and said, oh yeah, we're going to release the 9070 in March. And during their earning call, Lisa Su said we're going to release it in early March and as of right now it looks like March 6th is their release date. So I think they have kind of pushed things forward a little bit in response to NVIDIA. But to your question of should people go out and buy an nvidia card, the new one, no, you should wait a month, you should buy an amd card.

40:10 - Rob Campbell (Co-host)
I've I've found one for about three thousand dollars.

40:13 - Ken McDonald (Co-host)
Is that too much or well, let's see three thousand dollars for $3,000 for one NVIDIA card or for PS5 and an Xbox.

40:23 - Jeff Massie (Host)
Yeah, they're about three grand is overpriced.

40:29 - Jonathan Bennett (Host)
They're getting scalped is what's happening. People are buying them and then selling them for ridiculously high amounts.

40:37 - Jeff Massie (Host)
And for here we have a big retailer called Micro Center, and there's a lot of them that are, I mean, monstrously big computer stores, and they had some that were getting, you know, some didn't get any and several got single digits, you know when they could sell hundreds of them.

40:56 - Rob Campbell (Co-host)
Yeah, so there's just, they just did not have supply yeah, quippy's reminding us that's a lot of coffee, so that is you guys really want. Yeah, if anyone wants me to get a 4090.

41:11 - Jeff Massie (Host)
Uh, do the math, that's a lot of coffees 50, 90, you can't, you can't get a 40 90 oh, that was a 40 90, I looked up oh, how many bitcoin does that equate to?

41:22 - Jonathan Bennett (Host)
less than one by far. Yeah, I, I will accept bit. I'll accept bitcoin donations. Uh, word me so I can make a go, make a new um address that I actually have access to yeah, I I can't find a 5090.

41:38 - Jeff Massie (Host)
I could find a 4090, no 5090s yeah, the 50s are all scalped and they're running. There's some of them almost $6,000.

41:48 - Ken McDonald (Co-host)
But the thing is, people are buying them. That's the sad part. And now they're just sitting there waiting for somebody in the AI community to buy it.

41:58 - Jeff Massie (Host)
Yeah, to buy it. Yeah. Well, now that you know that could be, there is a legitimate case for the 5090 because it's got 32 gigs of ram in it. If you compare, people say, wow, it's really expensive, even at, say, five thousand dollars. When you look at the higher end uh, nvidia, professional level hardware, that's still pretty cheap.

42:24 - Jonathan Bennett (Host)
Yeah, yeah, that's that's always. The interesting thing is that it's a lot more reasonably priced than the, the pro, the workstation cards and the, the server cards for doing this stuff yeah, and when you look at the power, people go, oh my gosh, 600, why that crazy?

42:39 - Jeff Massie (Host)
Not at the enterprise level.

42:44 - Ken McDonald (Co-host)
That's like whoa, we're going to save money. Isn't that why Microsoft bought a nuclear reactor?

42:50 - Rob Campbell (Co-host)
I got to call this one out from Eric here he said a big store called Micro Center.

42:58 - Jeff Massie (Host)
Yeah.

42:59 - Rob Campbell (Co-host)
Yeah.

43:00 - Jonathan Bennett (Host)
Yeah, all yeah, all right, let's move on. Let's talk about what's new with FWAPD. Again has the story.

43:16 - Ken McDonald (Co-host)
Yes, I do. D2.0.5 introduces support for more Elan fingerprint readers, as well as support for emulating devices reading EFI keys, support for skipping device tests by CPU architecture and support for the Starlight magnetic keyboard from Star Labs. This release also fixes several bugs to clean up Dell Kestrel devices when disconnected, To ignore the boot entries that do not exist when checking the DVX, as well as ignoring EFI binaries that are zero-sized or not well-formed. If you do want to find out of anything that you're operating hardware-wise is being supported in this latest one, I do recommend checking Marius's article I've got it linked in the show notes because there is quite a bit that was added there as well.

44:17 - Jonathan Bennett (Host)
Yeah, fwapd just sort of runs as part of your update process these days on a lot of distros, right, like it's very possibly there on your Linux install and doing its thing and you're just not even aware of it, right.

44:29 - Ken McDonald (Co-host)
I know in KDE Discover it's actually pulls it up sometimes, but I always run it under top grade because top grade lets me hit the system. Updates the flat pack, updates snap updates, homebrew as well as flop deep.

44:55 - Rob Campbell (Co-host)
Yeah. It's not like the old days where they'd say don't update your firmware unless it's broken, because you might break it.

45:04 - Jonathan Bennett (Host)
I mean, let's be honest, that's still kind of true.

45:09 - Jeff Massie (Host)
But with everybody pushing everything out so fast now it probably is broken half the time.

45:15 - Ken McDonald (Co-host)
Well, it's broken as soon as you install it. Until you get that update. It's not quite like it used to be. It's not quite like it used. Install it Until you get that update.

45:21 - Jonathan Bennett (Host)
It's not quite like it used to be. It's not quite like it used to be. There's your show title. All right, speaking of things that are different, let's talk about the mess. Rob's going to tell the story this time instead of me. What's the new Linux soap opera drama thing that we get to talk about today?

45:44 - Rob Campbell (Co-host)
so people who listen like to know that I often like to share all the awesome things that asahi linux and hector martin are doing to get linux running on the apple Silicon computers. You know the M1, m2, max I don't know if they've got M3 or M4s yet, but you know and the amazing work that they have been doing has always been upstreamed to the mainline Linux kernel to where even Void Linux, a week or two ago, was recently able to provide their own Apple Silicon version of their distro. Well, recently there has been some arguing in the kernel mailing list against having Rust in the Linux kernel and apparently this was too much for the Asahi lead developer, hector Martin, as he has removed himself from being an upstream maintainer of the ARM Apple code. So the patch he put in just yesterday if you're watching live Friday states, and I quote I no longer have any faith left in the kernel development process or community management approach. Apple slash ARM platform development will continue downstream.

47:02
If I feel like sending some patches upstream in the future myself, for whatever subtree I may or may not, anyone who feels like fighting the upstream fight themselves is welcome to do so. So at this point, asahi, you know they're going to continue to move forward with their code, but they may no longer upstream it into the Linux kernel. So other people's like void can use it. I mean I guess they can. It's just not going to be in the kernel right out of the right out of the box. Others, you know they're free to wrangle the patch and upstream them as they wish, but it won't be Hector, and you know this. For all the, for all the good things that have been going on, it's kind of not really a good thing for Linux on Apple Silicon. Who knows what the future may bring.

48:00
But there is a small potential light in the dark sky that may keep things going upstream, as co-maintainer sven peter um commented just shortly after. He says give me a few days to figure out what we'll do. I think we can keep tree going forward. So hector isn't gonna do it, but maybe the co-maintainer will or maybe he's going to bring Hector back in. I don't know that statement's open to interpretation how they're going to do that. And you know I haven't given a chance to dig into the actual argument. And Asahi Linux itself isn't. It's not random rust, so I'm not quite sure how it directly affects them. I guess you know he kind of just lost faith in general in the maintaining process of the kernel. So you know, hopefully that's not a bad sign for the Linux kernel itself, but I guess maybe there's too many passionate views. But Jeff has a scoop and more details on that argument and what was going on there.

49:14 - Jonathan Bennett (Host)
Yeah, we're going to move right through and let Jeff pick it up at this point, because he's got the other half of the story.

49:19 - Jeff Massie (Host)
Yeah, we've got to say it was kind of a co-story here. So you know, as a longtime listener of the show, you've heard us talking about Rust becoming more and more common inside the kernel and it could be said the friction really started to surface about Rust being in the kernel last September when Microsoft engineer Wedson Emedia Filho stepped away from the Rust for Linux project because he was frustrated with the pushback he was receiving. Last month things boiled over again when a proposed abstraction to allow Rust-written device drivers to call the primary C-based kernel core DMA API, met with resistance from kernel maintainer Christoph Hellwig. Hellwig is quoted as saying no Rust in kernel. Slash DMA, please, slash DMA please. He further went on to say keep the wrappers in your code instead of making life painful for others. Interfaces to the DMA API should stay in readable C code and not in weird bindings, so that it stays greppable and maintainable. Hellwig was then challenged by Red Hat software engineer Dan Denlio Krumich, and Helwig was not happy about that and said don't force me to deal with your shiny language. Shiny is what he said Of the day. Maintaining multi-language projects is a pain and I have no interest in dealing with that. I have no interest in dealing with. If you want to use something that is not c, be that assembly or rust, you write the interface and deal with the impedance mismatch yourself. As far as I'm concerned, hellwig wants all the rust abstraction maintained at the in the driver's code and he doesn't even want it in a central section of Linux because which would then require another maintainer. So he went on to say if you want to make Linux impossible to maintain due to a cross-language code base, do that in your driver, so you have to do it instead of spreading this cancer to core subsystems.

51:28
Now the article in the show notes goes on to talk about how Rust is actually safer language because of how it handles memory, and the argument against it continues on about having multiple code bases complicate support. They do also mention there are initiatives and projects like TrapC or MiniC and SafeC++ that actually tried to make the C language, c++ language, less vulnerable to memory bugs. Linus Torvalds and Greg Kohartman have not chimed in on this debate and the author actually tried to get a comment or, but no, nothing was no, no replies. So now, reading the comments on this article, it seems like things fall into one of three buckets the. We should use rust to make things safer. Group the C is perfectly safe as long as you know what you're doing. And having two different languages in the kernel will make things a nightmare. Group. And a minority group saying everybody should use fill in the blank of some other language you probably haven't heard of because it's so much better.

52:38
Now there is anecdotal talk from people who've had to maintain projects with two different languages and felt it was a nightmare. There's also people talking about how Rust cannot fully replace C also people talking about how Rust cannot fully replace C. Now there is also a comment that while compile time checking makes a lot of things safe, it's never truly safe. So take a look at the article in the show notes for more details. I'm sure by the time you hear this there's many more comments that will have been made to that article. But you know I'm honestly curious about what our panel here. You know what they think. You know they program more than I do, and what do you think about? Whether Rust and two languages in a project overall are safer, or will it be too much of a nightmare to maintain? And you know the safety of Rust that rust brings is not worth it. So thoughts uh.

53:34 - Jonathan Bennett (Host)
so there is another um and I've linked to it. It's a message on the colonel mailing list. That is actually pretty uh. It's pretty important to realize that this is part of the context of what's going on here. And uh, there's there's quite a tale on this, but Hector Martin made the statement if shaming on social media does not work, then tell me what does, because I'm out of ideas.

53:58
And then Torvalds and this is on February 6th, so right before Hector stepped down, torvalds responded with how about you accept the fact that maybe the problem is you? You think you know better, but the current process works. It has problems, but problems are a fact of life. There is no perfect. However, I will say that the social media brigading just makes me not want to have anything at all to do with your approach, because if we have issues in the kernel development model, then social media sure isn't the solution. Technical patches and discussions matter. Social media brigading, no, thank you.

54:34
So there was pushback on more than just technical things. There were more than just technical problems. Hector Martin has also deleted his Mastodon account, so, like there's just, it's a multi-pronged story. There are multiple problems. It is unfortunate to see sort of the players not getting along and not being able to make the progress that they wanted to. But at the same time, sometimes you just you just don't come to conclusions and you have to go separate ways. So I don't know, other than that I don't have, I don't have real strong feelings. I understand why the maintainers are not interested in having Rust code in what they maintain. That makes sense. If I don't speak Rust at all, I really don't want Rust code in the section of code that I'm responsible for. That's a pretty reasonable take.

55:23
At the same time, if I'm a Rust developer, I don't want some old C developer to stop me from being able to get my code in the kernel. That's a pretty reasonable take too, and so we're just kind of caught between these two viewpoints. And every once in a while somebody like Hector he's not the first one to do this, he's not the last one to do it, but every once in a while they get tired of the fight and they just decide to pack up their toys and go home, and that's sort of OK too.

55:52 - Jeff Massie (Host)
Yeah, and that's sort of okay too. Yeah, but what? I jeff, jeff and then ron, oh okay. Well, the one of the things I can see is you know, it's one thing to say, okay, I can learn rust, but as the languages evolve over the years, you know you can have things change and then that interaction might not be as compatible. You know it's going to require more heavy lifting because maybe c adds something or rust adds something that used to work, but now it doesn't, and it's. It's the long term support.

56:18 - Rob Campbell (Co-host)
I could see where it could be problematic, but I don't know yeah, I think they both, like like jonathan pretty much said, I think both sides there make good arguments and you know I can understand both perspectives and really thinking about you know I've I've always thought, ah cool, adding rust. But really thinking about it, I'm probably more on the side of not having rust in it. Just because you know it's a complicated project already and you know this is a c project, maybe it's more maintainable to keep it as c. There is another project out there I can't remember what it is right now, but where the kernel is rust base I. I think there is.

57:09
Is it Redux, redux OS, that's right, I believe, yeah, redux, and I don't know. Maybe for people who want Rust to be the future, maybe going to that and making that a better kernel is a better path and maybe someday that could take over and replace Linux. Yeah, yeah, yeah.

57:31 - Jeff Massie (Host)
They call out in the article. I think it's. I'm trying to look. It's like Tractor, I think is what they called it. Yeah, and it automatically converts C code to Rust. So that would help you know that kernel you know kernel and Rust project.

57:49 - Rob Campbell (Co-host)
Even though there is something that can convert C code to Rust, I bet it's not going to convert the entire Linux kernel perfectly, and if you ran it, I bet there's a lot of tweaks and fixes that would have to be made.

58:04 - Jeff Massie (Host)
Yeah, my guess is from how converters work.

58:08 - Ken McDonald (Co-host)
But it would do a lot of the heavy lifting.

58:11 - Rob Campbell (Co-host)
Yeah, I've used them to upgrade I don't know old versions of PHP to new ones and that seems like a pretty simple upgrade and I still had to go through and fix a bunch of bugs afterwards.

58:28 - Jonathan Bennett (Host)
Yeah, the kernel is not exactly normal C code.

58:31 - Ken McDonald (Co-host)
Does anybody remember upgrading from Python 2 to Python 3?

58:36 - Jeff Massie (Host)
Yep.

58:37 - Jonathan Bennett (Host)
Yep, I have those scars.

58:39 - Rob Campbell (Co-host)
I've never migrated any code.

58:42 - Jeff Massie (Host)
Well, in defense of keeping the kernel C, they've been working hard to get rid of the assembler in it too. Colonel C, they've been working hard to get rid of the assembler in it too. Now there's a few little bits they don't think they can probably replace, but they've been trying to get as much assembly code out of it as well, just so that it's all in C.

59:00 - Rob Campbell (Co-host)
As awesome as I think Rust is, I'm going to vote for leaving it. C. Maybe that's an unpopular opinion and a vote that doesn't matter anyway.

59:11 - Ken McDonald (Co-host)
I get a feeling what we'll see is, as time moves forward, a co-project for translating all of the C code to Rust.

59:26 - Jonathan Bennett (Host)
Maybe I will tell you what I think we're going to see. We're going to see things continue as they are. So there are going to be some developers that are still wanting to do rust in the kernel and so long as there are developers that are willing to work within the kernel process, more and more rust code will land in the kernel. We'll start getting more and more rust drivers and so long as there are rust developers that want to make that happen, that will continue forwards, with some pushback from some maintainers, obviously, because that's what's already happened. What may happen is that enough of the Rust developers get enough pushback that they get tired of it and stop wanting to work on it, at which point Torvalds will come along and say, ah, nobody's using this Rust stuff anymore. We're going to yank it out of the kernel and it's all going to go away. Those are pretty much the two different directions that this can go. Um, and I, I don't know. I see a middle go.

01:00:17 - Jeff Massie (Host)
Yeah, I see a middle ground in there. I bet you there is. Time goes on. There's going to be more rust code and we will have that central abstraction section of the code that they will have a separate maintainer. You know it's going to be enough in there going. Okay, let's make it more efficient. Let's just have one centralized translator between Rust and C and the Rust people will have to take care of the translator and the C people still have their C code.

01:00:48 - Jonathan Bennett (Host)
Yeah, there's challenges though, because when you then go to change something on the C side, you've also got to make the updates to the rust bindings to be able to make it continue to work, and part of the problem is the people working on the C code don't want to mess with those rust bindings, and that's part of the central tension here. There's a video, rather famous at this point, of one of the kernel maintainers going you can't make me learn Rust, you're quite passionate about it, and I think that's sort of. The central conflict here is that there's just developers that want to continue writing in C and don't want to mess with Rust. Yeah, it's just not clear exactly how that's going to get fixed.

01:01:36 - Jeff Massie (Host)
Yeah, when I was saying the central repository I don't know what you call it, the branch or whatever that does the translation. That would be up to the Rust maintainers to maintain the C. People are not going to do that.

01:01:48 - Jonathan Bennett (Host)
Yeah, but so here's the problem. So say, you change the way the internal, because, okay, torvalds makes the statement over and over again we don't break user space. That does not extend to the kernel API, the internal kernel API. They change that every other Tuesday if they want to. They have no compulsion to keep that the same. In fact, sometimes I think they change things on purpose just to break out of tree code, but that's totally beside the point.

01:02:18
Well, so you've got. In that case your Rust code essentially becomes the same as that out of tree code. And so when you've got a maintainer that's going to redo something in the API, that's going to break those Rust bindings. And then when they push their patches to redo something in the API and it breaks the Rust bindings, that's going to break kernel compilation for people. And so there's either right. So like either you've got somebody writes C code and Rust code makes the changes and makes it in both sides, or you work together, the C people and the Rust people work together to make the changes on both of them at the same time, or they just mess each other's stuff up and it doesn't work. Right, like there's not a whole lot of room because it's in the kernel and everything is so tightly wound together there's not a whole lot of room for just ignoring each other and going on If you have both kinds of code in there.

01:03:14 - Rob Campbell (Co-host)
Yeah, you're telling me on it.

01:03:18 - Jonathan Bennett (Host)
Keeping rust out of there. But I mean, there's a few places where it makes a whole lot of sense, right, like you want to write a, you want to write a driver for some USB program or some some USB device, and you don't necessarily trust what's on the other side of the USB device, and so on the other side of USB cable, and so you want to make sure that you're going to handle all that incoming data in a memory safe way. Well then, hey, let's write that driver in rust, and if you've got the bindings to be able to do that, then great. Um and so there's, there's a few places like that where it makes a lot of sense and it's a really neat idea, but I just don't know if it's going to last. I don't know if they're going to be able to pull it out and fix all of the interpersonal things that go with it.

01:04:01 - Ken McDonald (Co-host)
Well you know? I think you don't. So you're basically saying that once you've got this driver for this set of devices written, no matter what language it is, you need to lock down the API in the kernel to never change as far as working with those drivers, I can guarantee you that that is entirely a non-starter.

01:04:28 - Jonathan Bennett (Host)
That is one of those things that has been suggested in the past. In summary, you kicked out the door as no, we are not going to lock down our API.

01:04:36 - Rob Campbell (Co-host)
I think, no matter whichever way this goes, one side or the other is going to be frustrated. They're going to be frustrated, frustrated, frustrated, f-rustuss-tated Frustrated.

01:04:55 - Ken McDonald (Co-host)
So we need to hire a counselor.

01:04:58 - Jonathan Bennett (Host)
Yeah, I guess I will say one thing that the whole idea of going out to social media and complaining when people inside the mailing list don't agree with you, it's not great, like it's just kind of in bad taste. Don't, don't, don't do that.

01:05:11 - Rob Campbell (Co-host)
Don't make the problem, don't, don't put the problem in a bigger space than it already was hey, I know, whenever I get burned by a company and they they won't like give me a refund or something, you go to social media and they get it to you pretty fast.

01:05:26 - Jonathan Bennett (Host)
They listen much quicker that way yeah, but if you're an employee of the company and you try to do that, that's sort of a different ballgame. Yeah, that's.

01:05:35 - Ken McDonald (Co-host)
That's a good way to become a customer of the company.

01:05:40 - Jonathan Bennett (Host)
All right, let's move on. And Ken has one more update for us. And then I've got some news too, ken what is new in LibreOffice?

01:05:51 - Ken McDonald (Co-host)
Well, my third article is coming from Marius Nestor again, and he informs us about the Document Foundation's release of LibreOffice 25.2. Now this release introduces a new privacy feature that removes all personal information associated with any document, such as author names and timestamps, editing time, printer name and configuration, document template, author and date for comments, as well as track changes. Now LibreOffice 22.5 also brings support for customizing the color of non-printing characters and the background color of comments and for customizing the theme, independent of the system or desktop environment theme, such as provisional support for font relative, first line before text and after text, paragraph indentation in LibreOffice Writer and a new macro manager dialogue that combines the functions of five existing dialogues used for basic macro and scripting framework macro language management into one dialogue. There are also some new accessibility features in LibreOffice 22.5, such as improving warning and error levels in the accessibility sidebar with an option to ignore warnings. Support for user interface elements to report an accessible identifier that can be used for assistive technologies Are we talking about AI here?

01:07:37
And the ability to correctly report positions of user interface elements in Linux systems, including on Wayland, on the accessibility level. Now I've just touched on some of the highlights. If you want all the gritty details. I do recommend reading Marius's article, where he does have a link to the changelog as well.

01:08:02 - Jonathan Bennett (Host)
Yeah, so that's interesting. You mentioned AI and in the context of LibreOffice, I had over on Floss Weekly I had an interview I had the interview last week it's going to go live this upcoming week with the product manager from Thunderbird and we talked about both their potential merger with LibreOffice several years ago and them looking at what to do with AI inside of Thunderbird and that was a lot of fun. We sort of concluded that there are some very limited places where AI might make sense in Thunderbird, in the same way that there are very limited places where it makes sense inside of something like LibreOffice.

01:08:41 - Ken McDonald (Co-host)
You mean assistive technology?

01:08:51 - Jonathan Bennett (Host)
usually refers to things like a screen reader, um, or, you know, even even text, text to speech, or speech to text, uh, or magnifying glass. Even those sorts of things are usually what you think of with assistive technologies, but sometimes they can be, uh, ai driven. You could probably make the argument that anytime you're doing text to speech, you're sort of doing ai, uh, I don't know, maybe I'm not, I'm not an ai expert. Don't at me, bro, um anyway.

01:09:14 - Jeff Massie (Host)
Well, I, I, you know I use it in uh for work, I use ai and my office products and, yes, I use microsoft because I have to at work. But you know it's, it's pretty handy, it. You know I use it for a lot of it's like grammar and spell check on steroids. So I was showing somebody I work with. I wrote just three sentences and they were very outline-ish you know, check with this person about this thing and it was three sentences, kind of like that. I had it in word. I said make this a nice outline format. It spit it out in a nice format, made it a little, you know, cleaned it up. Then I fed it into powerpoint and said give me a presentation on this. And it generated slides with images based on what I'd said. Now there was stuff I would have to fill in, but it shaved off a lot of what I would have to do as far as, oh, let me go find a picture and let me put it here and make it look nice, and it just just did it.

01:10:20 - Rob Campbell (Co-host)
I haven't used any of that in Office. But I could see it being. I was going to disagree with Jonathan and say I could see it being I was going to disagree with Jonathan and say I could see it being useful in LibreOffice. I did not know that the Office suite was that far along with AI. Yeah, so I'm humored.

01:10:40 - Jonathan Bennett (Host)
I'm humored that the only way that Jeff can get the indenting and the bullet pointing it right inside of Microsoft Word is to get ai to do it for him.

01:10:50 - Jeff Massie (Host)
Although that checks out, it is a pain yeah, but you know it's also it does good, I find, for search for like it's kind of a fuzzy search and it picks up a lot of things, so I can see it you being useful to you know, like I said, spell check on steroids better search, definitely it's.

01:11:13 - Jonathan Bennett (Host)
It's not, yeah, yeah, it's not revolutionizing anything, but it's just making things a little faster, it's, you know so I mean in firefox, one of the things that they're looking at doing or already doing, I don't remember which is, when you come across an image and it does not have the alt text filled in, they're asking they have the ability now to ask an AI what is this image showing, and it be able to generate alt text, which you know. That's, that's pretty interesting and there's, you know, there's a few other places where, like, it really makes sense of translation from one language to another. Is another one where it really makes sense translation from one language to another is another one where it really makes sense.

01:11:51 - Jeff Massie (Host)
Yeah, and and I will to reply to uh, david crosland commented. He said, yay, clippy everywhere, but it's not there until you actually say help me now or it's, it stays back and you don't have to use it at all. It's not dancing around the screen or anything goofy like that.

01:12:10 - Jonathan Bennett (Host)
Yeah, all right guys. Have we talked about the Framework RISC-V main board? Have we talked about the fact that it's actually now available?

01:12:20 - Rob Campbell (Co-host)
We talked about it being announced. I don't think we mentioned it being available.

01:12:25 - Jonathan Bennett (Host)
Yeah, okay, it is now available. The Deep Computing RISC-V main board for the Framework Laptop 13 is now available, and you can get the main board for just $199, which is really interesting. They also have I don't have the link to it here, but with this they've also now started selling a kind of bare bones Framework 13 that does not have a main board in it, and that way, you can get both of these things and put it together yourself. It ends up being I'm trying to remember it was several hundred dollars under $1,000. I don't remember exactly what it ended up being. It was enough that I was looking at it, going, man, should I buy this? And it was also enough that I decided, no, I can't spend that much money right now. So that is now available, which is really interesting.

01:13:18
But I've got a second link here off to an xcom user the social media network previously known as Twitter network, previously known as Twitter Laurie Wired, security researcher and otherwise Linux geek, I believe who says don't buy the framework RISC-V board. No hate, but the U74 cores it's based off of are really old and don't support vector instructions. If you're looking for a RISC-V dev board in 2025, you want something like the Spacemit X60 or the Zanti C908 cores, and then she gives a couple of options here, like the Banana Pi BPIF3 and the Canon VK230 are good boards to play with that actually have this newer spec. And so this is kind of like the, the x86 64 levels that we've talked about in the past. You know where some of them have an extension and others don't.

01:14:19
Uh, I'm trying to remember what. What was it? What was it? Wasn't it the extension that, like the original pinium and then the pinium 2, had? Um, I've mmc is what's coming to mind, and that's totally not right. I don't remember what it was actually called. Um, you know the, the, the fancier I'm drawing a blank the fancier math.

01:14:41
Oh well, I can't remember what it is yeah not not the math coprocessor, it's the, it's instructions built into it to be able to do more math. Mmx yeah, keith says MMX, thank you. Thank you, keith. Sse is one of the other ones. Yeah, there's been a couple of these different extensions that get added, and so this is sort of the same thing in RISC-V, where they have added one of these extensions, one of these sets of instructions to the processor, to these new designs, instructions to the uh, to the processor, to these new designs, and so you know, you, you can, you can buy the old 386 equivalent, but they're making the piniums now.

01:15:20 - Jeff Massie (Host)
So, uh, that was interesting to call out if you use the deep computing risk 5 main board and you run deep seek on it, have you gone off the deep end?

01:15:33 - Jonathan Bennett (Host)
Swimming in the deep end of the pool at least. Wow.

01:15:37 - Jeff Massie (Host)
Oh goodness.

01:15:37 - Jonathan Bennett (Host)
Anyway.

01:15:39 - Jeff Massie (Host)
I'm a dad, I could have a dad joke in there. I know, I know.

01:15:43 - Jonathan Bennett (Host)
So I was very interested in this framework, but risk is moving a little bit faster than framework is at this point. But risk is moving a little bit faster than framework is at this point, and so maybe this is not the risk board that you're looking for, I guess. At the same time, I hope it does do well, because I would like to see more experimentation like this, particularly in the framework space.

01:16:05 - Rob Campbell (Co-host)
You're saying we shouldn't take a risk on this one.

01:16:07 - Jonathan Bennett (Host)
Yeah, yeah, skip this one, it's too risky. It's too risky by not being risky enough. It's just not enough risk. Risk isn't risky enough. Wow, all right, let's move to some command line tips. And, rob, every time I go to write a cron command, I have to Google the cron format and you're about to tell me that there's another way to go about this.

01:16:33 - Rob Campbell (Co-host)
Yes, there is another way. So a little bit of background for those who don't know what cron is. Cron is just the Linux scheduler or not the Linux. It's a way to schedule jobs in Linux. It's the other Linux scheduler. Yeah, it's not other Linux scheduler. Yeah, it's not the process scheduler sometimes we talk about. It is the scheduler to schedule tasks User scheduler.

01:16:59 - Ken McDonald (Co-host)
It's a task.

01:17:00 - Rob Campbell (Co-host)
Yeah, task scheduler is not the right anyway, but you can schedule things like this command to run every night, every Monday, every, every day, every minute, if you want. And so you know. A quick, quick little example here, and I'll just remind you that, uh, jeff did talk about cron and cron tab in episode 37 way back then, if you want to go back and check that out. But if you look at this, this is the default crontab. I just uncommented what was there and you see like a 0, 5, asterisk, asterisk, 1. And then this is the command here is tar, dash, zcf, and then two files, so it's tarring up the home home directory. But you take a look at this and you're like what's that? Zero, five star? If you want to build your own, it's like how do you do it? And, like jonathan said, like I always do, always google it and and find some formatter.

01:18:04
But there there is a new tool on GitHub, written in Go, that you can use to. You don't even have to leave the kernel, you can use a tool called EasyCron. And it's still early, it's basic. It will give you the format and tell you a few things. I'm kind of hoping that EasyCron will progress further and actually write it into the Cron tab for you, but right now what you could do, so for those watching, I'm going to run EasyCron and it comes up. It has your five stars or asterisks there and it's pointing me on the first one that says minute. So you know, I want that to be every minute of the day. I I space over to hour, I want to be at three, and that day a month, if I do star, that's going to be every day, if I do month, that's going to be every month, and if I do day a week, that's going to be every day. So once I have that filled out, it's going to say it tells you at the bottom there that what I've written in there is going to run at 3 am and then it tells you when your next ones are going to run.

01:19:25
Now you know, if I back up and say one, it's gonna do 3am only on monday, so 3am every monday. And you know you can. You can go through there and say let's say that also it's 3 30 day of the month. That's gonna be weird. So at 3 30 am on day four of the month and monday. So and that's a weird one, I am assuming what that's. Well, it tells you 210, 217, 224.

01:20:01
I don't know how that one actually works because to me, my logic tells me that they should only be when the fourth day of the month lands on a Monday. But maybe that's the fourth day of the month and Mondays. That's what it must be, I think. But anyway, this little tool here, that's all it's going to do. It's going to show you how to format it properly. You can kind of go through and see what it's going to do. So right there, like I didn't know what that did, I guess that's the fourth day of the month and Monday. And then you can take that and put that into your cron tab. And I'm hoping, like I said, this first posted in January on GitHub, the first release. There's been 25 commits since then and I'm really hoping they add more features like the ability to input your commands and write it right to the cron tab. I think that'd be really slick but it's a good start.

01:20:59 - Jonathan Bennett (Host)
It's neat, little neat, little tip like maybe it only go.

01:21:02 - Jeff Massie (Host)
It calculates that out only where your cursor is. So if you went over and filled out, maybe it would it doesn't have anything to do with that.

01:21:09 - Rob Campbell (Co-host)
But let's see what if I did the ninth, so there it's going to do it on the 9th, which is tomorrow, because it's that, and then the 10th, because it's a Monday, so it's an, and is how that works. So I've learned something myself right here just by playing around with EasyCron how that actually works. So it's an and not an, or no. It is an or Sorry. So it's a. It's a and not a, or or no. It isn't or. All right, it's an or not an and logic is hard.

01:21:42 - Jonathan Bennett (Host)
Sorry to interrupt your tip, rob, with my maniacal laughing that I quickly muted. But I read keith's 512s comment that everyone playing the risk drinking game is now under the desk drunk. And that got to me me. He broke me All right, jeff. What if we need to dig?

01:22:01 - Jeff Massie (Host)
You know it's funny that Rob mentioned me and his command tip because Rob actually did this one in episode 48, long time ago, but it was part of kind of a toolbox. So dig, which stands for domain information groper, it's a powerful network tool used for querying dns servers and it's essential for diagnosing and resolving dns related issues. So if you're a network maintainer, this is a must-have in your toolbox and, like I said, we covered this in the past as part of a group of tools, but I thought it deserves some individual attention and for our newer listeners, since it was back in episode 48, so quite a while ago, it's good to be aware of this tool. So you can find the article in the show notes that explains how to install and set up Digg across several major distributions. They give you cut and paste if you're a little unsure of the command line. But given the depth of DIG, I'm just going to touch on the high-level workings and leave the detailed research to you. So, for example, the basic syntax for the command is DIG space, server, space name, space type Now server's optional.

01:23:14
It can be the IP address or name of the DNS server you want to query, but if it's omitted it just defaults to the servers listed in the slash etc. Slash resolve dot config file. The name this is the domain name to query and then type. This is also optional. Some of the supported record types are A or MX and NS. But if you leave it blank it defaults to the A record.

01:23:45
Now, that's not all it supports. It's just giving examples of some that it does. There are several options available to customize the output and the article lists several of the common ones. It's not exhaustive Now I'm not going to dive further into the details here, but the article continues with examples of, you know tracing paths, querying specific servers, formatting output, performing reverse DNS lookups and verifying your DNS sec. It also discusses practical applications of the program. So there's a lot of good information in there and it's a lot of. Here's what you're pasting in. Here's the expected output and it talks about the output of what it is. So if you're a network administrator and have never heard of the dig command before, definitely check out the article in the show notes yep, very good, I dig it all right.

01:24:41 - Ken McDonald (Co-host)
We've got ken up next with another pipe wire command, I believe yes, this time we're going to be I'm going to be covering how you can dump that information into a JSON file, json format anyways, in fact. Let me go ahead and bring up my command line.

01:25:13 - Jonathan Bennett (Host)
That doesn't look like a command line. No, it doesn't look like a command line.

01:25:17 - Ken McDonald (Co-host)
No, it doesn't Not the command line I'm trying to show anyways.

01:25:28 - Jonathan Bennett (Host)
We could just talk through it. You've got a link up to the pictures.

01:25:32 - Ken McDonald (Co-host)
Yes, I do. In the show notes you'll see. I have a link to the screenshots I covered. It starts off by showing what you see when you type in pw-dump space dash h, which, as with the other commands, it gives you the pw-dump. Then where you'd put the options and, if you needed to, where you'd put an ID dash h or dash dash help will show you that help information. You can also do dash dash version to show what version you're using. In my case it was showing that it's compiled with libpipewire1.2.4.

01:26:18
And, as with the other commands, if you are wanting to do a remote pipewire daemon, you do a dash r or dash dash remote, followed by the daemon name or the daemon ID. It can also be used to monitor changes on the command line, though there's so much changing that it's kind of hard to keep track of it. I actually found that I prefer doing a PW-dump than using the greater than sign and then putting the name you want to put it on In the show notes. The screenshot I took is showing me putting dumping it to 2025-02-08-01-13.json. That way I've got a date time stamp for when I did that and then you can open it up in whatever text editor you want and let it display that, and it's got quite a lot of information in the JSON format which can actually be useful when you're using some of the other commands where you may need to use the JSON formatted metadata as part of the command, one of which I'll be demonstrating next week.

01:27:59 - Jonathan Bennett (Host)
Very cool.

01:28:01 - Ken McDonald (Co-host)
But it's a great way to get a snapshot of how your pipewire setup's looking instead of trying to do a graphical one. But you can do that from the command line as well, and that's going to be one of my future commands.

01:28:18 - Jonathan Bennett (Host)
Yeah, I don't know if we've talked about it. I've got a command that I'm now inspired to talk about and remember to do this next week, because I've got a different one for this week, continuing on with my rock m theme. Um, let's see here that button. There we go. So I have here a command running and that is rockm-smi, and that will give you this output, which is just super useful for monitoring your Rockm GPU. And you can see, right now, mine is running along at 51 degrees and 33 Watts. Oh, excuse me, I've got, uh, is that my? Yeah, yeah, yeah, yeah, 33 Watts of power. It's using right now 50 degrees. Uh, it is capped at 211 Watts. It's not what it's using now, but it can. Uh, the fan is only running at 14% and it is tooling around somewhere between two and six, maybe up to 10%, on the GPU usage. It's because I've got a video that is playing back behind me. You can't see it because the TV does not show up on the camera, but that is what that is doing, and so this is super useful for keeping track of what your AMD card is doing and how hard you're pushing it. So, if you think that someone is mining Bitcoins on your rig RockMSMI might be one of the ways to check it out and find out.

01:29:52
But I'm going to have to talk about JQ. I don't know that we've talked about JQ on the show and it would be a super useful for dealing with the output from uh, from the pipe wire command that we just talked about. So maybe next week I can get to that. Uh, all right. Well, that is our command line tips and I think we're going to dive into our closing and we're going to let Rob first. He's he's itching ready to go. I saw he jumped to it.

01:30:19 - Rob Campbell (Co-host)
Uh, rob, what you got the plug so first I want to follow up with about my command line tip. It is not the easy croncom ornet that somebody pointed out. You will have to go to GitHub and it's by Elliot40404 and look there for EasyCron. Or the easiest way is to find our show notes and there is a link to that. Now on to my closing.

01:30:50 - Jonathan Bennett (Host)
On to your regularly scheduled closing notes.

01:30:52 - Rob Campbell (Co-host)
On to the regularly scheduled closing notes. My name is Rob Campbell and my website is Robert P Campbell. If you go there, you can find some great links, such as the one to my LinkedIn, the one to my Twitter, the one to my blue sky, the one to my mastodon, and you can come here and you can connect with me and be my friend, because I need friends and I left the best one for last. I always do right here. If you love what I do and you think it's just amazingly awesome and you want me to keep doing it, or you just want to show me some love for you know, whatever reason, you could donate a coffee right here in this little coffee cup icon, um, and I'd much appreciate it.

01:31:46 - Jonathan Bennett (Host)
Yeah, Very cool.

01:31:48 - Jeff Massie (Host)
And on to Jeff uh, you can find me at my go fund me page for the society to help Rob over his coffee addiction, Taking pledges there. No, but actually this week it's Poetry Corner and I have a haiku. It's not DNS. There is no way it's DNS. It was DNS. Have a great week everybody.

01:32:17 - Jonathan Bennett (Host)
Yeah, I think they make that one in T-shirt form. All right, Ken, anything you want to plug?

01:32:23 - Ken McDonald (Co-host)
Yes, I want to let everybody know that issue 217 of PC Linux OS Magazine is out and it's got some good articles and two recipes you may find helpful.

01:32:38 - Jonathan Bennett (Host)
All right, very cool. I do want to mention, first off, that you should know about Club Twit and of course, that is the way that you can support the show and you can support Twit it's Club Twit. There is a QR code right there. You can scan that to join the club. It's about the cost of a cup of coffee per month. Definitely worth looking into. If you want to find my work otherwhere, I am on Hackaday and we've got Floss Weekly there, a really cool episode coming up this week and a really cool one coming up the week after that and every week after that. Of course, I've also got my security column goes live there every every friday morning. That's a lot of fun too. Keep track of the the newest happenings in the security world, at least the things that I find interesting. Uh, other than that, we just we appreciate everybody being here watching the show, coming along with us and uh, those will get you next week on the Untitled Linux Show.