FLOSS Weekly 732, 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.
Doc Searls (00:00:00):
This is Floss Weekly. I'm Doc Searls. This week, Simon Phipps and I talked again cuz we had him on just a few months ago with Greg Crow Hartman, who is the highest level you could get, being a Colonel Maintainer and is behind so much stuff that we're doing. We talk about Europe, we talk about open source, we talk about controversies, we talk about lots of stuff we didn't last time, and it's a really cool show. And that is coming up. Next.
Simon Phipps (00:00:34):
Podcasts you love from people you trust. This is TWiT.
Doc Searls (00:00:38):
This is Floss Weekly, episode 732, recorded Wednesday, May 17th, 2023. Update Your Colonels. This episode of Floss Weekly is brought to you by Fast Mail. Reclaim your privacy, boost productivity, and make email yours with fast mail. Try it now free for 30 days@fastmail.com slash twi. Hello again, everyone everywhere. I am Doc Surles. This is Floss Weekly and I am joined this week by Simon Phipps himself from his bunker in in the office of the week in somewhere in the uk. <Laugh> one of the Hamptons.
Simon Phipps (00:01:25):
I, I'm, I'm, I'm just watching as the video zoom's past me in front of me. Hi, doc. I'm, I'm in my upstairs office. I can see the sky out there in trees. And and II you can see I'm, I'm You went boy,
Doc Searls (00:01:38):
<Laugh>. I
Simon Phipps (00:01:38):
Was, I'm on the edge today. So you're on
Doc Searls (00:01:41):
The edge today. You were saying the
Simon Phipps (00:01:43):
Le I have the gray leopard print and I'm already, it's
Doc Searls (00:01:48):
Really, it makes you look like, makes you look like parts of four people as well as, you know, invisible your invisibility cloak for the, for the, for the spies of the world. We wanna get going in a hurry this morning. So our guest is Greg Crow Hartman. Did you listen to the last show by any chance? I
Greg Kroah-Hartman (00:02:06):
Did not.
Doc Searls (00:02:07):
<Laugh> Greg did not. I was asking Simon actually.
Greg Kroah-Hartman (00:02:10):
Oh, sorry. No, just,
Doc Searls (00:02:12):
We'll just go ahead and bring Greg in. <Laugh>, let's,
Simon Phipps (00:02:16):
Let's bring Greg in. Yep. <laugh>.
Doc Searls (00:02:18):
I'm sorry, <laugh>. That's alright. Not sure talking Those who don't know. I mean, Greg is an alpha linux curl developer. Actually a major latest curl developer because Alpha is to suggest to something else. He's been the colonel maintainer for the stable branch, the staging subsystem, driver cord, debugs many, many other things. And he was on here a few months ago and the world was very different. <Laugh>. So we're having him back. How are you doing, Greg?
Greg Kroah-Hartman (00:02:45):
I'm doing great. Thanks for having me again. And
Doc Searls (00:02:47):
You and Simon are only one time zone apart right now. I guess we are.
Greg Kroah-Hartman (00:02:51):
Yeah. On this side of the pond. It's nice over here. Oh, it's good. I'm gonna be in in his continent or island next week, so
Doc Searls (00:02:58):
It's all good. Right. It's a subc
Simon Phipps (00:03:00):
And I'm going over to Brussels next week, so you'll probably miss
Doc Searls (00:03:02):
Me.
Greg Kroah-Hartman (00:03:03):
Oh, so real switch. Yeah.
Doc Searls (00:03:05):
<Laugh>. Yeah. You're, you're both in, in, in Europe. <Laugh>. It's funny, we, we move around a lot and I'm in the thing that is technically our home, but when our son was younger, he would refer to the second home as alt home. And the next one after that is alt shift home <laugh> <laugh>. So, so I feel that way. It's like that a little this way here. So, so how you doing, Greg? You're I'm doing
Greg Kroah-Hartman (00:03:29):
Good. Yeah. Almost good little jet lagged. I was in Korea last week, so Yeah.
Doc Searls (00:03:33):
And you said you just released seven, whatever. That, that's
Greg Kroah-Hartman (00:03:38):
Six, seven colonels I did, yeah. Little over lunch time today. So <laugh>, how, like I said, it's, I joke's just a Wednesday for me, so it's all good.
Doc Searls (00:03:48):
<Laugh>. Well lemme so what I mentioned to some people who's having a back on the show actually had Dave tat here last night. Who, huh. You know, he's an instrumental in network, everything. He debugged this network and then it didn't work this morning and he's not here <laugh>, so we got it working again. But he wa he wanted to ask a lot of things about embedded, and of course he's not here to do that either. I'm wondering how, you know, this clock over here or the, that appliance or that thing on Mars, do they get updates or they just done, are they frozen in place? I, I don't actually know. Is, can you generalize about this at all?
Greg Kroah-Hartman (00:04:27):
They should get updates and, and in fact there's a new law in the EU that's coming out for all these mobile devices that run Linux or anything run or anything, have to get updates. They have to get security updates. They have to stay alive for, I think it's five years past when they were bought, not when they're offered to be sold when they're bought. So current companies have to update them. It's actually gonna be the law. So it'll be, it'll be interesting to see.
Simon Phipps (00:04:52):
So I I I, I was, I've be watching the amendments coming through this week, Greg, and oh,
Greg Kroah-Hartman (00:04:57):
This is a different one. No, that was, last one's already passed.
Simon Phipps (00:05:02):
All right. So you to go about N two,
Greg Kroah-Hartman (00:05:04):
This is the mobile one. And the mobile one passed last year and it's going into implementation. Yeah. We're c I'd love to talk about cra that's different probably why you're going to Brussels. Yes. But mobile devices have to be kept alive for a certain point in time. So Yep. They will have to update their kernel. They're gonna have to move to a new kernel revision because we're not supporting long-term kernels for as long as they wanted to. But the joke about these kernels being supported for six years or so is it turns out nobody updates them. So I've been keeping some of these kernels alive for six years and nobody uses them and I know nobody uses them. Cause when they break accidentally, nobody tells me. So it's a mean little thing. I was in Korea last week talking to some, i, a company there that uses Linux and lots of things and trying to help them figure out how to do more updates in a better way. But iot all, I mean, my washing runs Linux my, my power meter runs Linux. All those guys, they all should be getting updates. And they do need to get updates because the world changes. And that's the thing. If they're sitting a little box and they don't interact with the world, sure don't update 'em. But if the world changes, you need to keep your kernels update.
Doc Searls (00:06:15):
Do, do you is that a gating factor when you go hunting for appliances? <Laugh>? I mean, are you
Greg Kroah-Hartman (00:06:22):
Looking? No, it's just random. Like again, I bought my bought washing machine and happened to run Linux. I I didn't know that <laugh>. So it's just everywhere. I can't get away from it. It's just kind of funny. It's in every, it's in my car. I mean these are not things I go out and hunt for all the tv. I mean, some things you cannot get away from all the TVs in the world. They all run robotics.
Doc Searls (00:06:42):
<Laugh> back channel day stuck with it. <Laugh> Jonathan asked to do your jail break your appliances <laugh>, which is, no,
Greg Kroah-Hartman (00:06:49):
I don't do that. I don't want that. Some things that are fun, but I mean, I also just like using them for what they're supposed to be used for. So, no, it's fine. I have, I use Linux on my own kernels on my, some of my devices. That's about it. But yeah, putting my, I've run my own kernels on phones. It's no fun, but it can be fun. Cool. Some phones, I mean, I'll call out pixel six runs, lean's kernel tree plus a bunch of patches, which is really good. It's been kept up to date for the past two or three years, so it will run mainline kernels, which is cool. So you can, there are devices out there that can run the latest kernel, but there's some something, I was talking to one company that supports devices for 20 years and they talked about it in a way of you, you're keeping these alive for 20 years, but it's not like once you sell them, they're frozen in time. Every year they update the kernel, they update the thing because these are alive living things. And maybe the year 18 or 19, then they say, okay, we're gonna stop, stop updates and we'll just do only tiny maintenance stuff. But it's a lifespan of a device and that's part of the device and how they keep it alive and maintenance and understanding about that. And that's good. You want that <laugh>. So so yeah, all your wind turbines, they're all running Linux. I mean, so you want those kernels updated <laugh>.
Doc Searls (00:08:09):
Yeah. So
Simon Phipps (00:08:09):
How, how, how do you feel about that Greg? You know, the idea of legislate legislating to say that people have to update because you know, as I've been dealing with the CRA I've got quite a lot of impact from input from community members that they feel that they should make the decision about whether they update rather than it being a part of a regulation. Do you think it's smart to regulate updates like that?
Greg Kroah-Hartman (00:08:31):
Well, all they're saying is, if you look at the updates for CRA's, a little different. But like the mobile one is you have to provide security updates for the lifetime of the product and the lifetime of the product is gonna be five years. So you can't just run away. Right. And that's good. That's, I mean, I think that's a good thing. How those vendors provide the lifetime of updates is up to them. Now I will claim, and I've proven that you provide a secure kernel unless you're using the stable kernel releases, we have loads of documentation to back that up, et cetera, et cetera. So you have to, by virtue of keeping a secure device, you have to take these updates. Otherwise there's no other way. The CRA has a lot of other really weird things. My best one was you can't ship a device with any known vulnerabilities. Yeah. who is the known <laugh>, right. Known to me, known to you. Are you gonna just put your head in the and listen to anybody and then you're like, problem. So has lots and lots of issues. That is very interesting. And I've been involved in that little bit.
Simon Phipps (00:09:33):
Right. Yeah. So it's interesting you point that one out cuz I, I actually think that the open source community is having an impact on the cra cuz I, I've seen this week's amendments from the IT entry committee and in particular the, the, the point you just made about saying, you know, not shipping known vol anything with known vulnerabilities the, the, the text has been amended in the current proposal. And it's, well I saw,
Greg Kroah-Hartman (00:09:58):
I saw
Simon Phipps (00:09:59):
Known to the,
Doc Searls (00:10:01):
Tell us who the CRA is in the first place for those who don't know and the whatever it was, committee <laugh> as well, just so we have our acronyms. Right.
Simon Phipps (00:10:09):
So, so the CRA is the Cyber Resilience Act. It's an act of the European Parliament that's going through on a fast track. It will be law by the end of this year. And it basically in makes sure that manufacturers of network connected devices have a responsibility for at least five years to provide secure updates and to ownership secure versions of their products. And it achieves that. It's
Greg Kroah-Hartman (00:10:38):
A little bit more than that, but
Simon Phipps (00:10:40):
It, it achieves that by using a, by C marking the product. And there's a then a load of there's a load of conditions and variants that relate to things like whether your product has a function that's critical to the functioning of society, which gives you more responsibilities. Whether or not you are, you are a developer of open source software there is an exception in there that excludes, theoretically excludes open source software. And so, so the act is basically about making sure that manufacturers of network connected devices with, with software in them take responsibility for their products in the market. And and they will end up with liability for them as well. So the product liability directive in Europe is being updated to attach massive liabilities to people who failed to ship secure products. Did you fix that upgrade? What, what, what did I get wrong? <Laugh>?
Greg Kroah-Hartman (00:11:35):
No, that's good that that's, but I mean, think of it as this way, a nice way is the, there's rules about shipping a toy, whether it's harmful or not, right? You wanna make sure any device with electronics in it that's connected to the network is also not harmful. So it's, that is the goal. They try and carve out open source in some weird ways. I don't really want an exception for open source. I don't think we need an exception. I think open source should stand on its own, but there is some way that they're trying to check liability onto the individual developers of open source rather than the liabilities on the person who ships the product. And that's where some issues are involved. And they're like, so I'll call out Orange, the big phone company in France for saying specifically that we're willing to take the liability of the, all the software and the dice that we ship, don't push it onto these random other people that had nothing to do with the device we made, even though they wrote parts of the software. So that's, that's
Simon Phipps (00:12:29):
The goal. That was nice. You know, the, the guy who wrote the bill assures me that he didn't intend the developers to pick up the liability. I agree.
Greg Kroah-Hartman (00:12:38):
And it's written that they say they don't intend, but then if you read the words
Simon Phipps (00:12:42):
Yes, <laugh>, there's kind of friendly fire casualties that result in some of the other effects that are in there. I mean, so I don't think I,
Greg Kroah-Hartman (00:12:49):
The hard writing rules Yeah. <Laugh>,
Simon Phipps (00:12:51):
It, it, it, it isn't about having an open source exception. It's about saying that they should be regulating the software in the market, not the software in development. And Correct. Making that true, if
Greg Kroah-Hartman (00:13:02):
That comes
Simon Phipps (00:13:03):
The act is really hard.
Greg Kroah-Hartman (00:13:05):
Yes. I, I would like that to have, it doesn't, but there's also some other things that everybody keeps forgetting about, like the vulnerability issue stuff and like known vulnerabilities, well known by whom and known for how long. And then there's other rules about length of time and, and all this stuff that, and then, oh, all Europe is gonna spin up its own version of Mitra to collect vulnerabilities and hoard them and doll them out to people. And it's, there's some really, really scary things about that. I mean, for the colonel, I can't tell you if there's, if I'm fixing security bugs or not. Legally I can't because if I tell one person, I gotta tell everybody, right? So just say update your colonels, <laugh>. So that's gonna be very interesting as it goes on. And hopefully, I mean, you say the latest version, the latest, I saw a version yesterday that had 350 pages of proposed changes to the amendments. So <laugh>, there's a lot still in play here.
Simon Phipps (00:14:01):
Yep, yep. I mean, I get, I have in the lucky position, so I, I'm representing OSI and dealing with the legislators directly on this and, and I'm lucky enough to get like 177 page document every two days, which I have to go through and read and find out what changed. Well, cuz there's no, I think I'm
Greg Kroah-Hartman (00:14:18):
On that mailing list now too. I got that. Yeah. I skipped the meeting. So, yeah,
Simon Phipps (00:14:23):
So, you know, I, there's actually got a good news story here, which is that I think that for the first time that I remember we're actually having an impact. You know, I saw words that I'd written in response to one of the legislators in the last set of amendments. Good. You know, I've seen, I've seen people from the committee offices wanting to have their language checked by somebody who understands open source so they don't make mistakes. So I think we're having a real impact here. And I think this is the next stage in the maturity of open source and of Linux of you know, we've, I'd be very afraid we are finally important enough to be regulated by parliament and I'm, I'm not sure we quite ready for that yet.
Greg Kroah-Hartman (00:15:07):
Well, all software should be regulated by parliament. I mean it, I don't think Linuxes should be special here. There's nothing special about us. If you want to use Linux for your devices Yes. Follow these rules. If you wanna use any software for your devices, yes. Follow these rules. I don't wanna carve out, I just want us to play on the same level playing field. Don't hurt us more with some of these ways that it is written, it's gonna hurt us more. And that's, I know it's not their intent, but it's the way it's written. So if it works out, that's great. And hopefully the, the big, the big issue is this is gonna hurt companies a lot and companies aren't really aware. I mean, the interesting thing is like, look what the rules say for micro controllers AMD and Intel, anything that has microcode in it that's covered. And that's gonna be very, very interesting how those guys deal with that. Cause that's software software's in your chips <laugh>.
Simon Phipps (00:16:00):
So, so now, I mean there's another dimension to all this that, you know, when you, when you become essential to the economy you also have to begin to deal in geopolitics. And I saw the Linux kernel mailing list rejecting a patch from bike electronics recently, which is a, a Russian company that uses the Linux kernel. Do you wanna talk us through the, the logic behind rejecting that patch from bike electronics?
Greg Kroah-Hartman (00:16:27):
Well, first off, nobody can force any other maintainer to take a change. That's just by rule. So it's not against our code of conduct. I can reject a change anything for any reason. I want to, I don't have to reject it in a rude way that's against our code of conduct. But I can reject anything because once I accept a change from you, I am now responsible for it. If I don't feel comfortable taking a change, I won't take it. Right. so the person, the company in play here is a company that only works for a certain, for the Russian government defense industry. And they were adding support for a specific piece of hardware that they create and use. And the maintainer who of that subsystem was uncomfortable in taking that and didn't wanna take it. And so he didn't, and Lena backed him up and the company and then there was a private email thread. The company knows the issues involved. Lena explained them very well and discussed it over with them and asked the company for the specific details. But generally you can't force us to take a change if we don't want to. Right. that's as much as it comes down to as far as that goes.
Simon Phipps (00:17:31):
So, so, so to what extent
Greg Kroah-Hartman (00:17:33):
For we <laugh>. So,
Simon Phipps (00:17:35):
So, so to what extent is that a matter of the, of personal taste on the part of the developer own? To what extent is that a matter of the Linux Foundation being subject to export controls by the US government?
Greg Kroah-Hartman (00:17:48):
It has nothing to do with export controls. Nothing at all. There are no export controls with regards to the MUAs government with regards to open source. It, it's treated as a publication. Just like when Chinese companies were banned before they had their work enclosed specification committees where they had to sign contracts removed. But open source contributions and open source interactions are all covered under the open laws of publication and academia. So those are not covered and we're not covered by any type of that. This is purely a personal decision. Again, we make personal decisions not to take changes all the time. You have to kernel development and development in general is a personal interaction. I have to trust that you're going to be around to fix the problem when you get it wrong. So if I don't trust that you're necessarily gonna be around or I don't trust, your motives are good, I don't have to take your change and you can persuade me otherwise, which is great and I encourage you to do so, but it's at a per, everything we do is at a personal level.
(00:18:50):
This had nothing to do with governments per se.
Simon Phipps (00:18:53):
So, you know, personally, I completely support the, the the decision that that Yaakov made there. And I would encourage all of your maintainers to make the same decision every time personally. But do you think, but we also
Greg Kroah-Hartman (00:19:06):
Take patches from other, we take passions from Russian people all the time that we know are Russian. So it's not like we're banning people from the country by any means. I took some the other, so it's, I support Jacob. This
Simon Phipps (00:19:20):
Company is a, this company is a very specific case who, you know, it's, it's, it's entirely reasonable, I think, to have concerns. Well, the question I was gonna ask is, how long do you think you're gonna be able to get away with it remaining a personal decision if governments are gonna start regulating what you can do with regard to defect resolution.
Greg Kroah-Hartman (00:19:42):
So they can't regulate what we do. They can only regulate what people use our devices for, right? <Laugh>. So think about it that way. So defect resolution it'd be interesting that if you're going to discuss how we have to meet certain guidelines and whatnot, that if a company wants to use Linux, they're gonna have to support that. So take a specific example like software bill of materials. Under the rule, the CRA is devices have to ship a software bill of materials that contain a list of everything that they have in it. Wonderful thing. I've had companies come to me and say, there's no way we can possibly do this. We ship too much software, we don't know what we're doing. I'm like, that sounds like a management problem, <laugh>. You need to get ahold of that. You need to figure that out.
(00:20:25):
And they're like, well, we need the community to provide software bill of materials. I'm like, no, you are the community. Do this work and provide it. Give us a, give us a S P D X line. Give us a software bill of materials and stuff. So just contribute and in a way that meets your requirements. If your requirements now are these new engineering rules, these new directives by different governments, provide the work to do that. If you wanna use Linux, nobody's forcing you to use Linux. You use Linux because it's free and it meet your needs. <Laugh>. That's why you use
Simon Phipps (00:20:56):
Linnux. Yeah. Ok. Cause you know, I've got this, I've got this instinct in my bones that it's gonna get more tricky than that. Over, over time. You can always,
Greg Kroah-Hartman (00:21:06):
I'm not, I'm not disagreeing with you, but we only accept, so I'll say one thing. Like the Lennox Kernel security team cannot sign any legal agreements. We can't get into any con anything like that because we all contribute on our individual merit independent of anybody we work for. We're not covered under any employment agreements. We can't, no employer can tell us what to do or have access to what we do. So that is an individual thing. So no company can tell the, the external security team what to do per se. That being said, I have been subpoenaed by the US government and had to talk to the Senate about things, but they are agree with how we do things. So we're OK with that. <Laugh>
Doc Searls (00:21:45):
Well I I'm
Greg Kroah-Hartman (00:21:45):
Sure Dutch government will talk to me soon cause they're right across the street. <Laugh>
Doc Searls (00:21:49):
There, there's more to cover around around what's happening in Europe and how that's driving the world, at least politically and to some degree technically. But we'll get to that. But first they let it half, let everybody know that this episode of plus weekly is brought to you by Fast Mail. Make email work for you with Fast Mail. Customize your workflow with colors, custom swipes, night mode and more. Fast mail now has quick settings from the Quick Settings menu. So you can easily choose a new theme, switch between light mode and dark mode and change your text size without leaving the fast mail screen. You're looking at Quick settings will also offer options related to the fast mail screen you're viewing. You can generate a new masked email address, show or hide your reading pain. Switch between folders and labels and more. Choose to auto save contacts or choose to show public images of senders from external services like Gravita.
(00:22:48):
Set default reminders of events. Change how invitations are handled, or turn notifications for calendar alerts on and off. Now add or buy a domain through Fast Mail and they will set all the records up for you so it works immediately. Fast Mail gives you the ability to send and receive emails from your own domain and manage multiple email addresses in one space, which helps keep you organized and protects your personal data For over 20 years, fast Mail has been a leader in email privacy. The Fast Mail team believes in working for the customers as people to be cared for, not products to be exploited. Advertisers are left out putting you at the center. You pay for free email with your privacy at Fast Mail. Your data stays yours with better productivity features. For as little as $3 a month, fast Mail has better spam filters and absolutely no ads.
(00:23:45):
And privacy isn't all you get with Fast mail superior productivity tools with scheduled send snooze folders, labels, search bar, et cetera. Plus keep track of all the important details in your life easily with fast mails. Powerful sidebar works as password managers like bit worded and one password to make it easy for you to create unique passwords for every account and safely store them on your device. It is great on desktop and mobile, especially when you download the Fast Mail app to get the most out of your email. It's easy to download your old data and import it to your new Fast mail inbox. Fast Mail is moving email forward with new internet standards and open source innovations that power many email services other than their own. Don't get left behind by substandard email providers. Reclaim your privacy and boost productivity with fast mail. Try it now free for 30 days@fastmail.com slash wit. That's fast mail.com/twit.
(00:24:50):
So there's a we could talk about AI for hours, I suppose, because that's the big, I mean, you only hear like a few months ago Craig <laugh> and, and and it's like AI has, has, there's always a topic. I I dunno if you're familiar with Don Norman who wrote things that make us smart and design of everyday things and stuff like that. He, he, he once told me that that there's always a subject that is a black hole. Meaning if you bring it up, the event horizon of the topic exceeds the dimensions of the room and all conversation falls into it in no light escapes. And I feel like AI is sort of that right now. But as long as we're on Europe there's a story that the EU has this AI act already an amended act, an an amended act. And by the way, the US House is meeting today on, on ai and there is, it's actually live. If you want to leave here and go there, I don't recommend it, but I get a note about that earlier. But it, it said that this will have a significant impact on open source. And I'm wondering if you're paying attention to that or if that matters to you much.
Greg Kroah-Hartman (00:26:10):
Well, I mean, AI is just pattern matching. I I I just made a bunch of people real mad by saying that <laugh> statistics at a higher level. But I mean, why is ai, I mean again, pattern matching somehow special and resolved from data collection laws, which is why the EU is put their new thing or privacy issues or copyright laws or fair act of derivative works and fair use sector. Why are they somehow special and they're not. They should be <inaudible> all the same other rules and laws that we have. I mean, AI again, buzzy logic. My, my rice cooker 20 years ago did that, right? That's just the same idea, <laugh>. You can do it at much more things and you have semi sane ways of summarizing stuff, but, and it works great for translators, but I mean we, we've been using it for ker stable work for, I don't know, a decade now. Has anybody noticed? I don't know. So it solves a good, it solves a problem for some things. Yeah.
Doc Searls (00:27:06):
When are you using it? What are you using? I mean, you said pattern matching or statistics at a high level, which is when we, which great. When we
Greg Kroah-Hartman (00:27:12):
Have a stable, so leis has a development tree, right? We patches nine changes an hour to that and then we backport fixes to the up from that tree to our stable trees. A lot of those changes are marked by developers saying this needs to be back boarded, right? Some whole subsystems. And some people forget to mark those. Some don't even wanna participate, so they don't. So researcher Julia Lule and Paris came up with the idea of have we take the body of work of these are known fixes for the colonel, that developers have said these, these should be back recorded and apply that to everything and see how they match up. And she had a research student go off and do that. They wrote some good papers. Sasha who works with the stable me, we do the stable work together, took that on, refined it and refined it and refined it some more.
(00:27:58):
So now he runs these pattern matching logic. Basically it's AI stuff on the kernel and picks out all the patches that look like they need the other fixes and manually reviews them, sends 'em off for other people to review. And we have it. But we've been doing that for a long time. So again, I mean they saw it's fuzzy logic match, say fuzzy matching, right? They solve good problems that way, but we're not immune from an algorithm being why should it not follow the same rules of data collection, fair use, copyright, et cetera, et cetera. And I think when people somehow think that this, this idea of the day can ignore all the laws is very dangerous for all sorts of other reasons. Cause we live in a society why, which we all need to follow these rules, right?
Doc Searls (00:28:42):
So, so I'm, I'm wondering, I I I love this statistics at a higher level panel pattern matching. And so, but there are lots of people out there, really reputable people, important people. You all know Harari who is quoted by everybody. This is the end of the world there. Generative AI becomes general AI becomes our overlords and the matrix comes true. And, and there's always this logical gap between it's all gonna end and how it happens because how it happens is never described <laugh>. And does any of this concern you at all? Where I'm going with this though is I I love that you pointed out that the is run by people <laugh>, it's individuals, it's a collection of individuals is not beyond that an organization. It's the commons of a sort, but it's, it's remarkably human in how it works. And so I wanna tease out in general, I mean not just here what is it that can only be human, that can never be pattern matching by a higher statistically adept machine.
Greg Kroah-Hartman (00:29:52):
I, I know. I mean this is what philosophers has been discussing for decades. Are you centuries, right? I'm not, I'm not a p philosophy major. I know there's people I
Doc Searls (00:29:59):
Was, can't help it. Okay,
Greg Kroah-Hartman (00:30:00):
So then, I mean, you, you have the rights. I mean you have the, the discussion of what is the consciousness and what is in the consciousness and, and all those discussions and what makes human a man person, human and the soul and all those things. I'm not gonna answer that. I'll say from a technical point of view, if you look at technology, it's cool. Does statistical pattern matching, is it a hot dog or not? Is it a tank or not? I mean, those type of things. And then applying those patterns to other things that a human asks it to do, right? Like I would love for, I mean code autocomplete based on current patterns, but somebody has to write the code in the first place to be the auto completion. Do that. I mean, if you look at the chat stuff, they're, they just scrap the web.
(00:30:42):
They scrape my my domain lennox.com also and stuff. And if you feed questions based on articles I've written, it'll spit back those answers. Wonderful. That's a great summary. Wikipedia index, right? But I mean that's a, somebody had to create the summary in the first place. Who's gonna make the new stuff from that is gonna be the interesting point. Everything's a remix. Yes, that's what art is, but kinda make the next logical step to do things in a way. I mean, they've messed with certain things they can do procedural generation of antennas, right? Maybe you can do that for art. I dunno. Somebody has to say, is this good or bad art? I dunno. Again, that's a philosophy argument. I'm not gonna get into that <laugh>
Simon Phipps (00:31:25):
<Laugh>. So Greg, to to switch away that subject from you. <Laugh>, I, I know that you are sit you are sitting in a place where you are seeing a lot of inputs and changes and trends because of the LinuxCon is at the center of everything. So where should we be looking for the next big deal after the digital agenda wave in Europe and the the, the the wave of AI use of software? What's the next thing that has caught your attention for where we're going to have to be concerned about a societal impact
Greg Kroah-Hartman (00:32:05):
Af Oh, I mean, Lena's always gets asked and I get asked, what's gonna be, what's the next kernel? I mean, I don't know. It's when it, while our job is to make the hardware, all these crazy people come up with work, right? So you were giving you the hammer and you make the thing with that. That being said, we see hardware coming. I mean, we've seen hardware that's in the process for a long time. So more pro more chips, different processors, risk five everywhere. We see the low level stuff that's going faster memory buses, memory buses of span cabinets and span, things like that. Giant number of CPUs compressing down to tiny ones so you have more logic. But I mean, I will say ai, they're starting to put more of these micro controllers and micro, these AI engines, which are just extreme fast PA math engines into chips themselves.
(00:32:53):
I mean, the pixel has that with their tensor chip, which is cool. You can do a lot of really cool stuff on your device. So I see a lot more, I don't wanna call it edge computing, but I see a lot more local computing of these types of things. I mean, your phone can do all these crazy pattern matching and recognition than your camera app. You can do that and turn that into other types of stuff as well. And so I see a lot of lot more computing happening in devices, which is cool. Yeah. Cause we're running a really low power. Cause Lenux runs really low power. We give it all to the user to do things and they can do some really neat stuff that way.
Simon Phipps (00:33:28):
So, yeah, well I, I mean I've, I've loved the the tensile chip in this thing and or the, the capabilities it produces. And I, I've, I very much like the idea that when I'm using an AI capability, it's running locally rather in the cloud. I think that that makes me feel more confident using the device, if anything.
Greg Kroah-Hartman (00:33:46):
Yeah. Some interesting thing is they're, they're dividing up Linux. So Linux is now running multiple times on the same device, so they want to carve that out so that one program can't steal that AI model from another program because you're running in another virtual machine. So we're seeing more and more use of, instead of applications running on top of Linux, Linux being I don't wanna say Unikernel, but it's Linux and an application talking to Linux on an application, talking to Linux on application to get a little more security boundaries. So we're seeing increased security that way. They have to use these chips for something and they're providing much more, better and harder security boundaries that way, which is cool. Wow. So that way you can't have some random, you can't have TikTok stealing your banking application or your pictures. You know, that's a goal. That's a good thing to have.
Simon Phipps (00:34:32):
So you know, I've, I've in the user chat or the viewer chat here, I get people begging us not to talk about AI cuz it's so boring. <Laugh>. and, and the the person who said that would like to know what work you're doing to make Linux run faster on the new Apple chips how, how, how is that going?
Greg Kroah-Hartman (00:34:51):
I'm not doing any of that work. There's some other people that are doing really good. I mean Apple Chip is just yet another cpu, right? That they don't give us a pro, the specs for people love reverse engineering it than the engineers doing that are doing an amazing job. And it's cool and I'm more power to them. I'm using stuff that people actually document <laugh>, so
Greg Kroah-Hartman (00:35:13):
I'm not doing anything. I see their patches cool stuff is they're writing some of the drivers in Rust and they're hitting all the issues that we thought they'd be hitting. It comes to using Colonel Work and Rust. It was just fun to watch the community come to grips with a lot of this stuff. But it, it's interesting. It's, it's good seeing that.
Simon Phipps (00:35:32):
H how about gaming? You know, we, we've got a bunch of people here who are, are very by what steam is doing on top of Lennox. Does, does that, does that float <laugh> there it is.
Greg Kroah-Hartman (00:35:44):
I mean, Val, I, I've known these guys for a long time. I used to visit them. I live in Seattle. They're good. They do really good stuff. I mean, this little steam deck is the most amazing thing I've seen in a long time. I, you can run Windows games on this thing. The engineering and the work that these guys have put into this thing is absolutely amazing. And I'm blown away by it, especially the price. I highly recommend everybody getting this thing. I went to customs last week in, in France, and they dug through my bag like they always do and pull this thing out and the the guy said, oh, this is nice <laugh>. So that, that's unusual for somebody for the, for the security people, especially in France to show emotions <laugh>. But it's a great, I mean, I, I will call out Steam.
(00:36:30):
What Steam has done is amazing. And it's an ecosystem that they've done, but the layers and emulation, the wine developers, the proton developers, it's the graphic skies. They've done a really good job and Steam has backed all this stuff they're doing wonderful. I, I'm amazed it and I'm very happy to see them succeed and again, strongly recommend all with him. But it's cheap. I mean, the, the, the price of this thing is so, so cheap and worth it. And my son got one actually before me somehow. And he loves the thing too.
Simon Phipps (00:37:02):
Yeah, I'm very concerned that my daughter might be watching actually, because I know what she wants for her birthday in July. And you might just very expensive.
Greg Kroah-Hartman (00:37:11):
I mean, the thing runs everything. I mean, I don't know of any games that doesn't run. There are some odd ones that it's like, oh, this is really a weird, but then I just plug it into a doc and I throw it on my monitor and you can plug other things on that. Some people use it as a laptop and other things, which I think is a little ridiculous. But the processor in here is interesting. The, the GPU in there, I, I'll call out AMD for doing a really, really good job in the community and Val pushing them to do this work, to get this kernel work and this driver work upstream, and it's all open source and they're doing a really good job. And that's how companies can get this stuff working in a way that makes them succeed. I'm, I'm really, really pleased with that. It's a success story as far as a community based goes. Hopefully it's a commercial success for
Simon Phipps (00:37:54):
Them. Yeah, well I'm very, very impressed. I'm also very impressed they're running a business that is doing gaming without getting into the cutthroat waste of money, of developing games. You know, they're, they're, they're busy providing a surface for you to leverage other people's investment. And their business is all about, I mean, Val always technically clever.
Greg Kroah-Hartman (00:38:14):
Yeah, Val has been that way a long time. I mean, they understand. They're also a bunch of X Microsoft people there, and then they'll really know their stuff. They're really, really good engineers. They know how to also work on creating an ecosystem and a developer ecosystem and a community and supporting other people because you don't know what other people are gonna do. And that seems in their ethos, I mean, take a cut of it, like it from any other store that takes the cut, but right. They seem to be doing pretty good.
Simon Phipps (00:38:42):
Now, I heard you mention Rust in passing there. So, you know, is, is Rust a good thing for developing on the kernel? You seem to suggest that there may be some issues.
Greg Kroah-Hartman (00:38:52):
So Rust is in there is in the kernel, we have the basic framework. You can write a hello world driver in the kernel. That's the simple part. The hard part is going to be actually doing something that works. People said, oh, it's just a driver. It's not, it's gonna be easy. It's a standalone, it's a leaf. But I mean, if you think of a tree, a leaf, it's at the top of a tree. It depends on everything. The tree, the branches, the trunks, the roots, and it interacts with them in a way. So these drivers have to get bindings and ways to talk to the rest of the kernel. And the interaction between the Kernel C code has its own idea of memory management and ownership and lifetime rules and the rest idea of ownership and lifetime rules and whatnot. That intersection there is <laugh>.
(00:39:37):
It's, it's tricky. It's tricky to get right. They're doing a great job in getting it right. They're proposing a lot of good stuff, but doing that is hard. It's hard engineering work. They're getting there. It's not gonna solve the world tomorrow. But it's nice to see that they're trying and trying something interesting. And it's so far as seems to be working. There's some direct comparisons. Like somebody wrote a driver in buttocks, a driver C and driver and rust, and tried to compare them and it seemed pretty easy. It'll be interesting to see how it works over time, because the trick of the kernel is in Linux it evolves, right? We change its APIs, we change how it works, we change it, its interfaces. And if me as a seed developer have to change the interface and then it's gonna cross that rust, cross that rust boundary, I better know enough rust to make the changes and cross the rust boundary and fix all that up there. That's gonna be some just basic engineering and community building and interaction issues over time. We're seeing some pushback from some subsystems saying, I don't want rust bindings for my subsystem right now because I got all the work I can do right now. Just keeping the C code alive. Interesting to see if all the companies that are supporting this actually pony up and give the developers, allow them to do this work, which in the end, if they wanna see rust in there and Russ support it, they need to do the work.
Doc Searls (00:40:55):
We have a question here. Go go ahead Simon.
Simon Phipps (00:40:58):
No, I, I'm, I'm heading in a completely different direction. You do that question. I see it.
Doc Searls (00:41:02):
Okay. Yeah, it's it's some Jonathan who was on the show last time will the Colonel ever have a stable a p i and then adding, adding regression testing question mark.
Greg Kroah-Hartman (00:41:14):
So I wrote Stable p i nonsense about 20 years ago. It's in the colonel if you wanna read it. You don't want the colonel to have a stable api, it'll kill Lennox. Our stable API is up between user space and the kernel. We guarantee that we've said we will never break it, et cetera, et cetera. We, that's our one rule. You can't break that without anybody noticing it. If you break it and nobody notices it, it really wasn't being used. That's fine. That's our API internally. No. And you don't want us to, I lay out why you'll want us to do, if you wanted us to have a stable api, what, 20 years ago we've stuck with that, we would've failed. You know, even Windows changes their API all the time. If an operating system isn't it's debt I'll say that.
(00:41:57):
So what was the other one? Regressions actually. We have somebody sponsored and is being paid to track regressions and Thorston is doing a wonderful job of that beating on all the maintainers saying, Hey, here's a regression, here's a list of 'em when they're fixed. When they're not fixed pushing changes to Lena saying, Hey, this was a regression, this subsystem that was over here. The maintainer hadn't picked it up yet. Take it and let's go. He's doing a wonderful job with that. He's been doing that for a little bit over a year now. Total success story. He always needs more help by seeing some now interns from some of the Lins Foundation projects starting to help him out with just tracking things down and doing regressions are key. And we have a database of it. You can up, you can see all the regressions that are currently there, their status of them, if you have a regression reported and he'll track. That's the best thing that we ever have asked for. It's doing better.
Simon Phipps (00:42:50):
So you, you used to be at Susa and I did, you know, I'm, I've been watching red Hat busy becoming IBM over the last month. The, the, the landscape has really changed for you know, the, the ordinary user of a computer thinks of Red Hat when they think of Linux. I think what do you think is, is, do you think it's good? What's happening there? Do you think this is a, a natural part of growing up? That we are seeing Susa becoming basically a, a, a corporate entity and Red Hat laying off all of its community people? Do you think that's a natural evolution or do you think that that's a crisis?
Greg Kroah-Hartman (00:43:29):
Well, I also, I was at IBM before <laugh>, so I, I've seen both sides of I mean, red Hat would love for you to say that they think their marketing department says Red Hat Linux is Red Hat. You know, that's, that was their goal's. Goal is for Europe, no, Linux, isa. But I mean, one thing that, that the Dish shows have to contend with is people just want to solve a problem, right? And if you want to use a Lenux distribution to solve your problem, wonderful. If you wanna buy a box with a distribution on it and get it commercially supported and Red Hat or IBM is there to do that for you, IBM has been in the software support business for forever. That's where they make their money. So that makes sense. Suza has been in that business for 25, 30 years now.
(00:44:14):
That's where they make their money. That's a good business to be in. It's not a huge business. The majority of the world actually by far runs Android. Take all the Android stuff out of it. Everything else around the error majority runs Devin and the World Cloud systems run Devin over 70%. It's, it's insane what Devin, what the world runs in Devin and Devin's of great distribution and solid and there. But if you look at, if you want to solve a problem, you go rent yourself a cloud computer. So a free distribution on there or a distribution from that cloud provider. I mean Google, Amazon they all provide you a free version of Linux for their Oracle does as well for their cloud. Cause they don't want you to have to pay any extra money. Anyway, you go from there. So you solve your own problems. Is the enterprise distribution business has always been a small business. It's interesting thing to see if you can hold onto that, but it's still probably the same about size that it always has been. But just the other size is so much bigger. I mean, even Microsoft has said they're not making as much from the Enterprise cloud on Windows as they are on Lenox.
Simon Phipps (00:45:24):
So, right. But, but I mean I, you know, I was very disturbed to see the, the head of Fedora being laid off by Red Hat and also to see so many of the community people being laid off from the osbo and to see opensource.com being shut down. You know, did do, do you think we've, we've, we've, we've seen this is now the end of the era for community opensourcing in companies like Red Hat and suse and IBM M
Greg Kroah-Hartman (00:45:55):
No, I do not think that. I mean, companies open source com, open source program groups and com companies are getting bigger and bigger and there's more and more of informing. I think you're gonna see it spread out to more and more companies. I mean, look at the big giant industrial companies, Phillips, Siemens, Bosch, they have open source development group. They have developers on there working on open source projects in the community, right? So these are, it's just spreading out to be more and more the Links Foundation posted something saying job prospects of open source developers are even higher now than they ever have been. Yes, it's always sad to see companies lay people off. I do not understand those business decisions. I'm not in that type of work. I think I do know a number of people who were laid off and got a job instantly afterwards.
(00:46:39):
So if you have the skills, I worry for the new graduates coming out. My son's in university and he's, his friends and peers are worried about that. And I mean, as someone who grew up in, in outta college and one of the major recessions in the world, I've been there, seen the com booms and good skills of developers are highly needed and whatnot. And companies realize it's cheaper and faster to work in the community and use open source tools than ever before. So I was at the open source the cloud conference in AM an Amsterdam a couple weeks ago, and there was 10,000 people there. It was insane. It was like, wow, I guess this Linux thing is actually gonna succeed.
Simon Phipps (00:47:20):
<Laugh> <laugh>. So, so what, what should we be telling our, you know, what should we be telling our kids to focus on if they're gonna head into software now? My son Web, web
Greg Kroah-Hartman (00:47:34):
University for ok. I
Simon Phipps (00:47:36):
Mean, so what have you told, what have you told them to go do? Cause we should all do that.
Greg Kroah-Hartman (00:47:40):
<Laugh>. I mean, I would say systems level work, because that's what I, that's the level I work at, right? I see all these hardware companies trying to solve problems and they are all companies using software now, web development, everybody needs a web development and stuff like that. That's also a great skill. But I mean, the generic skill of programming, of breaking a problem down the smaller pieces and applying that to a language no matter what the language is, that's the skill to learn how to debug, how to handle these types of data structures, how to solve a problem. And then you can get a job in any language doing any types of things. But there's so much, so many jobs out there. I mean, just take music for instance. Like all these electronics companies are now, all their guitars and all these amplifiers, everything is all software now that's, there's tons and tons of speakers.
(00:48:25):
All the speakers are all software now. I mean, doc's favorite company, Sonos has been running Lenux for, what, 20 years now. That's all Lennox and software heavy stuff. And that's now in all speakers. I mean, there's a lot, a lot of software jobs out there for a lot of industries that never had it before. My little amplifier here that I get has software infirm. We're running a USB stack handling the signals and whatnot. It's not running Linux, but it's running other operating system. Zephyr Zeph is a wonderful little operating system that's tinier than Linux that I'd recommend people using. So that, yeah, but just the basic skills of programming. I mean, as a parent, your job is to hopefully impart skills onto your child that they can be successful and stand on their own, right? So give them a skill.
Doc Searls (00:49:11):
<Laugh>, <laugh> there are a couple questions. What one of them is do you have any que wait, there's a, a smile with it, but do you have opinions on Wheland versus Cord?
Simon Phipps (00:49:25):
Exo
Doc Searls (00:49:26):
Exor. Exor, sorry.
Greg Kroah-Hartman (00:49:27):
Oh, Exor is dead. Yeah, I mean, that's not an opinion all the developers have left. I I've actually deleted code out of Exor. Yeah, I mean it's not, it's not an opinion, it's just that ex Wheland does not work on Nvidia because Nvidia doesn't wanna do the work and they haven't wanted to do the work for 20 years. So just don't buy their hardware. I mean, it's as simple as that. So if you want to run Nvidia hardware, then you're kind of stuck with this weird mess and yeah, Wayland doesn't work so well, but I mean, Wayland's been working for me for decades a decade now or so, and there's no XOR developers. All the xor developers are working on Wayland. So why is this even an argument? I dunno, I shouldn't be.
Doc Searls (00:50:05):
Yeah, there, there's a <laugh>. There's more, more on she G B T all abandon all Hopi. You learn to coach. She G B T has become shi of the destroyer of Code monkeys there. I I, I, I thi am I wrong in assuming that this isn't gonna do anything? People who wanna code are gonna learn to code, it's just gonna happen and we're always gonna need that. We don't, we don't. You gotta know it in order to look at the, the code that, that the chat GBT comes up with and see where it sucks or doesn't work.
Greg Kroah-Hartman (00:50:39):
Sure. You have to see if it's right and how to even run it or debug it. Right? How's it even working? The debug, I mean, chat GBT is great for spitting out frameworks of languages and stuff that has to do a lot of frameworks like Java. It might be great for turning out things, common patterns in Java, but you still, I mean, programmer, there's been the goal of program list program programmer list programming for what, 50 years now, if not more than that. Marketing or managers want to be able to say, Hey, take my business logic and run this program. Right. Turn into a program. You're always gonna to somebody to turn that into something that'll work. MD Bug. Yeah. All you
Simon Phipps (00:51:14):
Have to do is code. All you have to do is specify it well enough that the computer can write the code for you. And of course, the specification that you're writing is in no way a program.
Greg Kroah-Hartman (00:51:26):
No, I've been there before. There's this specification rule. Yeah. Let's not go there. But that's been proven to not work. Let's just say that, you know what, what about u l all those years ago, right? Remember u l Yeah. Yeah. Follow the little diagrams and the boxes and they'll spit out the code on the other end and everything will be perfect. Turns out you need a feedback loop, right? Ai, I remember Loop,
Simon Phipps (00:51:48):
I remember a software product in the eighties called the last one that was released in 1981 by a company called DJ Systems that took input from a user and generated an executable program in basic and apparently that in its marketing was gonna be the last software tool anyone was ever gonna use.
Doc Searls (00:52:11):
It's
Greg Kroah-Hartman (00:52:11):
Funny. The cool thing is things like, I mean, visual Basic brought the idea of programming and control of solving your problem to a huge swath of other people, right? And that was great. I know a number of those people that turned into doing other work, but they were solving a problem that they had. Chap, t b t solves the problem people have with summarizing texts or looking things up or writing a flyer for your birthday party. I mean, that's great. That's a cool tool. Wonderful. Is it gonna become sentiment and take over the world? That's again, like Doc said, where's the middle step there?
Doc Searls (00:52:46):
I dunno. It's, it's funny, when I was when I first moved to Palo Alto in 1985, my landlord was a guy who worked for Lockheed and he was an antenna scientist. And he said to me it was getting boring because antenna science was completed. It was a done science <laugh>. We knew it all <laugh>. And, and I, I kind of believed them because they came out of radio and I I know antennas. I understood it pretty well. I, I, and I made them know,
Greg Kroah-Hartman (00:53:16):
Well that's an interesting industry that the whole idea of genetic algorithms has proven to be wrong. I mean, they have genetic algorithms. Once you have that feedback loop of, is this better than that? And you can model it in software, you create antennas that look like nothing the antennas you used to see, but work even better today. Right?
Doc Searls (00:53:31):
Right, right. Yeah. I mean, and the thing is too, the, I mean the, the frequencies that carry data best are, are so high that all the antennas are small. You don't see them <laugh>, you know? Yeah. For the most part, you know, there's a little nub on the top of your car. That's it. That's listening to satellite. I think. I think even the, with
Greg Kroah-Hartman (00:53:48):
That nub up on the top of the car is built in this weird converted way that you never would've been able to imagine. I mean, the antenna is, if you ever taken a cell phone apart the way that, you know, that winds around the circuit board and this certain crazy way that's designed that way for a reason and
Doc Searls (00:54:02):
Or as part of the case, that's another one. You know, it's and and it knows if your hand is on it and is using another one there, something like that.
Greg Kroah-Hartman (00:54:11):
Yeah. You have to do that. So, so yeah, so the idea of programming is a dead, I think it's okay. Don't, my son is going into that <laugh> business,
Doc Searls (00:54:19):
So we're getting toward the end here. We're not there yet. I mean, what, what do you anticipate, I mean, how much is anticipation involved in, in your work? It seems a lot of it is reactive. You're people come in with patches, people do bug reports, is you're pushing things into trees and, but, and are, what are you anticipating, you know, like looking forward the next few years?
Greg Kroah-Hartman (00:54:42):
I mean, all my work is reactive, right? It's always reactive because we don't, so, I mean, interesting thing is we don't deal with high-minded ideas. Like somebody can't come to say us and say, wouldn't it be great if Linnux did this? Our response would be that I would be like, yeah, it would be cool. Go out and try it and I'll review your attempt. That works, right? Yeah. because I don't have the time to do anything else, but, and the idea that you have to prove something that actually works in order for us to accept it is a good thing because we don't go off and on crazy directions and we're able to see and pick from the technologies that work. So I see the hardware lines cuz you see hardware roadmaps and CPU roadmaps that are out stretching out into the future.
(00:55:24):
So I see these goals of these hardware companies and what they wanna do, and again, it's more, more cores, less power, more interconnects. And then on the server side, crazy number of interconnects memory on, on a system over there that's migrating to a system over here, it's all through fast interconnects and there's all sorts of insane networking things happening. So I see that, you know, the constant evolution of 10 faster and accommodating our code to handle that. I mean, networking stack is famous for, you only have so many microsecond or milliseconds to handle this packet before you have to get the next one. And taking those ideas to other parts of the kernel like storage io, because now your storage is on a bus that, that fast your old network used to be. And transferred that into another idea of how does it handle other types of buses and other types of things in the system in order to be able to accommodate that.
(00:56:23):
That's the things I see in finding areas that need to be work. I mean, the cool thing about Linux is the core data structures and infrastructure, the kernel get rewritten all the time. Linus rewrote the locking subsystem a couple years ago or so. No other operating system would ever do that. You can never get a manager to sign off on that. So we're adapting to the world of what's coming, right? Or what's there. Cause we see the, we see the hardware that's coming out tomorrow per se. Yeah. We see the patches that are coming out tomorrow. We're just trying to make sure, I mean, my goal is have Linux work on all pieces of hardware and make sure that keeps working and happening. And then the CPU guys are making sure it runs on all the different types of CPUs. And we meet with the CPU companies and see the roadmaps and they give us heads up of what's coming and we say, no, that's gonna be crazy. They're like, yeah, we know it's crazy, but the did it so now we have to make it work. I'm like, oh no. We see that happen a lot. <Laugh>.
Doc Searls (00:57:15):
I I wanna ask a since you brought up Leis you know, I, I look, I was just looking at the, at the colonel mailing list, which I do from time to time. And I see Linis is all over it. I mean, he's still top dog obviously, I think, and h that doesn't, I mean obviously it scales. So he is, it scales as long as he's in charge. I mean, I I I gather you don't think about that this much and there are other people that can stu step into his shoes. I'm just wondering where, where this goes. If he gets hit by a bus tomorrow, what does it make make much difference? The bus,
Greg Kroah-Hartman (00:57:51):
The bus question now that's only been coming up for the past 25 years. I
Doc Searls (00:57:54):
Know coming up for less 30 years that I I I'm personally you know's answer before. So yeah,
Greg Kroah-Hartman (00:58:00):
You know's answer right to that. You asked him, you've asked him that before. Yeah,
Doc Searls (00:58:04):
I've, I've asked him before. He says somebody else has stepped in basically. Yeah. He's
Greg Kroah-Hartman (00:58:07):
Like, I don't care, I'm dead. Right. So <laugh>
Doc Searls (00:58:12):
Yeah,
Greg Kroah-Hartman (00:58:13):
I mean Colonel development is a process of trusting other people, right? I trust 10 people, leaders trust 10 people who scale down the thing that we trust that they're gonna be there to fix the problem when they, when they mess up. Cuz we all mess up. And that's the whole way Lenox development works. We have 44,000 people that how somehow we trust in different ways. Yes, we've talked about this on what happens if, I mean this gets hit by a bus or I get hit by a bus or anybody gets hit by a bus. Colonel Kernel developers, we have a plan, we know we'll handle it. We all have backups, we all have other people have access to our trees. I have somebody who help can get access to my trees. There's a num a short number of people that can have access, toes, trees. We can, we have ideas and how to handles if it could, if something were to happen. So don't worry about it.
Doc Searls (00:59:04):
I, I, I, rob, I'm going with this a little bit and just in my own head is how much of this is a model for civilization done well enough, if not right? This is never right, right. We're, we're always working things out. I mean, this is part of being human. We are, we're learning, we we're changing all the time. You know, we're, we're not, we're not finished. Like what you said earlier about you don't want a stable API because Yeah, it's not really,
Greg Kroah-Hartman (00:59:30):
I mean that's the only thing if we stop changing, we're debt. You know, if you stop evolving, you're debt. I don't know about a model for a society. It's a model to solve a specific problem very well. Cuz I mean, we have 4,000 people. We do not agree on everything. We don't get along even in some things, but we all agree that we want to solve the problem to make Lennox better. That's about all we can agree on. We barely agree on the license at times. <Laugh> specific nuances of the license specific technical version of things like that. That's, I mean, we are can all agree on this one little tiny thing and we all work to make that better, the best that we can. Because it, and it's nice that companies pay us to do that because it solves companies problems for them in a cheap way. You can use Linux because it's free and it actually saves you money than if you were to ever do it yourself if you do it right. So it's a good model to solving, I think technical problems. Maybe if you, I wouldn't say it's a good model to solve anything dealing with people problems. Cause people and interactions with people are much, much different and much more complex and much more just difficult. Like I
Doc Searls (01:00:35):
Said, yeah, I forget who said it, but it might have been sart. Who the hell is other people? <Laugh>. So there's, yeah,
Greg Kroah-Hartman (01:00:42):
Yeah. It's like the phrase people are people, it's like, yeah, people suck <laugh>. So I mean, we all suck. It's, it's just, that's the nature of human behavior and we're all big squishy bags of water and we can't control 'em. And we have feelings and emotions and hormones. I mean, that's what makes life interesting and good and I think that's a wonderful thing to do. But solving a cool technical problem is about all we can agree on, on this and it works well. So we're, we're pretty lucky that way and Lena does a great job. But also the cool thing about Lenas and hopefully all the other colonel developers and me, is you can convince them that they are wrong and they'll change their mind. And that happens numerous times and that's a rare thing. And somebody who's expressing good leadership skills that way.
(01:01:21):
I know people who have worked for other companies for other operating systems whose managers could never change their mind <laugh>. And so they left and now they work on Linux. So I think that's one reason Linux has seeded so well, and Lena does a really good job in that, is cause we are open to new ideas and change and grow and continue to allow to do this. And the cool thing is we can do this. I'm encumbered by a whole lot of politics and managerial issues because we're interacting on a personal basis, not on a company basis. And I don't know, it was us to solve this problem in a very unique way. It's, it's pretty cool that that's happened, but it's also spread. I mean, look at Kubernetes, Kubernetes is a good success story for taking the idea of how the kernel was developed. I helped work with those developers and setting up their model and not, and succeeded in a totally different technical space. But again, thousands of developers, huge numbers of companies using it, solving a cool problem in a very complex area. Very well. Those guys are doing great.
Doc Searls (01:02:19):
That's great. And, and we actually are outta time now. <Laugh> wrap when ready? I think we're ready. So we always close the question we asked you before what, what your favorite text editor and scripting language are. So you just repeat yourself on that. Or maybe it's change since we talked last.
Greg Kroah-Hartman (01:02:38):
Text editor, Silvi Vi or Vim and then scripting language. I, I still use Bash. I mean I, I'll drop down to Pearl as my other favorite, but I'm being forced to learn Python by cause the colonel.org webmaster and who does all our tools for infrastructure writes Python and sometimes you have to help them out or I have to try and add new features. So I'm learning Python finally.
Doc Searls (01:03:01):
That's great. Alright, well Greg, it's been great having you back on the show. We'll have to have you back again as always. And thanks for all you all you do. This has been fantastic.
Greg Kroah-Hartman (01:03:12):
Thanks for having me and you guys do a great job. It's up too. And good luck Simon in Europe. If you need any help in Brussels, I'm gonna start training.
Doc Searls (01:03:18):
Yeah, you guys may have to team up <laugh> versus the bureaucrats. Well, Simon, that was cool. That was great.
Simon Phipps (01:03:29):
Could we can The conversation can roam in many directions when you're dealing with folk who are at the center of things. It's the the, that old saying a captain we are surrounded and the answer is fantastic. We can attack in any direction. And <laugh> that, that's really how it is when you're talking with folk like, like Greg who can see attacks coming in from every direction.
Doc Searls (01:03:56):
Yeah, there's I, I'm actually in this conversation. I'm really, I'm almost for the first time really missing writing for Linux Journal because, because Linux is continuing to evolve, even though Linux Journal, you know, I guess it's still there, all hailed to them. They, whoever's running it now has kept it alive. So everything we've written that was published is still going back to 1994 is still there. So that's cool. I you know, one of my big awful discoveries for all of us is that we thought the internet was a, the web was a library and it turns out to out's been a whiteboard. So that's a, that's been an issue for me cuz I want, I want it to be a library. But there it is. Anyway. We'll, but this has been a great show.
Simon Phipps (01:04:42):
Mm-Hmm.
Doc Searls (01:04:43):
So what do you, what do you plug, man,
Simon Phipps (01:04:46):
I, you know, I'm, I'm pretty, pretty plug free this week. I just encourage people to, to keep on supporting your, your favorite organization that's intervening on your behalf with legislators. At the moment, I, I would suggest that people might wanna join osi to, I'm working for osi. They're, they're paying for my time in Brussels cuz nobody would do that stuff unless somebody was at least paying their transport fees. And if you join OSI as a member by going to opensource.org/join they will have more money that they can spend on making sure you are represented in Washington DC where my colleague Deb Bryant is doing the work or in Brussels where I'm doing the work. So do please become an RSI member today. Yeah.
Doc Searls (01:05:29):
Support these people.
Simon Phipps (01:05:30):
Put your weight behind us.
Doc Searls (01:05:33):
Excellent. Thanks a lot Simon. And I have a small plug I think of, there's a thing called Dwe Camp. Just look at a Dwe camp. I'll be going to that in June. I wasn't gonna be, and now I am. So, and I'll be saying stuff there. So that's, that'll be fun. So next week, let me look at my thing here. I have to, I, I always, Tim Bonk, I think is it, and I'll go back here and yes, it is Tim Bondman of Open Source Science is going to be our guest next week, so come back for that. Until then, have a good time. I'm Doc Surles. This is Les Weekly. See you there.
Jonathan Bennett (01:06:12):
Hey, we should talk Linux. It's the operating system that runs the internet, bunch of game consoles, cell phones, and maybe even the machine on your desk. Then you already knew all that. What you may not know is that Twit now is a show dedicated to it, the Untitled Linux Show. Whether you're a Lennox Pro, a burgeoning csit man, or just curious what the big deal is, you should join us on the Club Twit Discord every Saturday afternoon for news analysis and tips to sharpen your Linux skills. And then make sure you subscribe to the Club twit exclusive Untitled Linux Show. Wait, you're not a Club Twit member yet. We'll go to twit.tv/club twit and sign up. Hope to see you there.
