Open Source, AI Sovereignty and the Future of the Stack | Techstrong Gang

Sound check. Hey everyone. Welcome to a special Thursday edition of Techstrong Gang live from SUSECON in Prague.

Well, truth be told, it's not really live. We recorded this about two hours before you're watching it. Right.

But it's as close to live as you're going to get. Mm-hmm. We were live when we filmed it.

Let me introduce you to our gang today. To my far right, we've got our Chief Content Officer, Mike Vizard. Hi.

And the man in the middle, Steven Vaughan-Nichols. Frequent, well, regular contributor to the Techstrong sites, but the man needs no introduction if you follow tech media over the last... How long you doing this?

A year or two? Closing in on 40, God help me. Yeah.

Yeah. God bless you. Good.

Steven, welcome. Thank you. Glad to be here.

Thank you. So, Mike, as we wrap up today, tomorrow we're off on our- Merry way ... merry ways.

Mm-hmm. What do you want to harp on today? We are coming to the close of an open source software conference, essentially.

Yep. And I'd love to get Steven's opinion on, well, what is, in your assessment, the state of open source these days? Because there's a lot of chatter around AI and- Right ...

are people still going to need open source? Because I'll just write my own code and- Right ... I'll just download that and away we go.

Mm-hmm. So, as you were talking to folks around here, what was the mood? Still positive about open source.

This is a SUSE show. They've been at this since 1992. So, yeah, they're still big on open source.

I don't see a whole lot of worry here about how AI and open source are going to interact. I know outside the walls of this hotel conference center, though, there's a lot of concern about just what you talked about. Because you can sort of roll your own, do you really need open source?

Now mind you, everything that's, quote, "rolling your own" is actually based on earlier open-source work. Yep. Which, that's something for the lawyers to fight about.

I'm not going to worry too much about it. But it's interesting. I think open source is going to survive.

I think eventually there will be an alliance between them that will work well, but in the short term, it's going to be kind of confusing as people figure out exactly how these two can work together. Because you're right, there is a lot of, well, I can just roll your own out there. That's why all the software as a services stocks went crashing down recently was, well, you can roll your own.

To which my response has always been, yeah, you could always have rolled your own, and yeah, it's a lot easier now with AI, but still you have to manage it, you have to secure it, you have to do all this other work, which is a real pain in the rump. And AI, it's great, but it can't do all that for you. " And it all worked together.

Under the heading of, just because you can do something, doesn't mean you should do something. Exactly. Yeah.

Exactly. I think there's a couple of things to unravel on this one. First of all, you don't hear them say the word open source here.

You hear them say open. Open. Yep.

And open infrastructure. Yep. It's almost like they're trying to extend the open- Right ...

beyond the software. Yeah. And to them, the open...

My friend Ron Guler always used to tell me about open source. It's free as in beer and- Right ... free as in freedom.

Right. The freedom, it's free as in freedom is their version of open source because it's tied into the whole sovereignty issue. Right.

And so it's open to pick what platform you want to work on, what software you want to use, what AI you want to use. Right. These are all open.

Right. Some of them are software, some of them aren't. Right.

But they're open. Right. Secondly, I think you're right.

We are going through this sort of dance right now between AI and open source and roll your own or smoke your own or whatever you want to do. But here's what the usage patterns are showing. Less people, less real humans- Right ...

are logging into open source projects, downloading open source components- Right ... and so forth, code snippets. And that's causing some people to get a little concerned, right?

Right. Are we going to stop supporting these communities? On the other hand, the amount of code being downloaded from these open source communities is up.

Mm-hmm. Which- Insanely up. Right.

Which says to me, something, or it's not someone, but something- Mm-hmm ... is downloading this code and using it. Right.

Yep. And I think as we move more into agentics, the agents, and we make them more efficient- Mm-hmm ... an agent's going to say, "Well, look, it's a lot easier, cheaper for my end user- Right ...

" Right. And so you're going to see open source code usage shoot up. Right.

Mm-hmm. There may be a little dance, but when the music stops, that's what happens. So are humans writing code to train machines that are then using other machines that are downloading the code?

Is that kind of... Yeah. Well, if you look at the amount of code in open source repos and projects, there's millions of open source projects.

Mm-hmm. There's so much code there. To think you're just going to toss it aside and clean room everything.

Yeah. And the maintainers are being overwhelmed by people- It's not efficient ... who want to contribute now and using code they developed creating- Right ...

the AI, right? Right. And a lot of that is truly garbage code.

It's slop. Yeah. But that said, I was talking to Greg Court-Hartman at KubeCon about a month ago And he tells me, and he should know, he's the guy who maintains the Linux stable kernel, that the code from AI is actually within the last two months- Gotten a lot better ...

it's gotten a lot better. Yeah. And they're so surprised, they do not have an explanation for why it's happening.

" It's not that. They're not honestly sure what it is. They think it's probably a combination of people have finally figured out how to use these tools properly, and the tools are improving.

But in any case, it actually has become a lot more helpful. Yeah. That said, maintainers are still overwhelmed, and particularly when you look at things which, again, they didn't really talk really about at SUSECON at all, but now there's this whole stuff about how these learning models are going to be able to basically zero day any project.

You look at it- Right ... and you get 100 plus errors. Right.

Mozilla the other day announced that they used, I'm blanking on the name of the pro- Nikos? Yeah. That one.

And that they got something like 180 bugs found. Yeah, but how many of them were real, and- Yeah ... how many of them were exploitable?

Yeah, exactly. Which is always another question. Yeah.

Because yeah, you can always find lots of bugs- Yeah ... but how exploitable are they? But that's the issue.

So couple of things again, Steven, you got me thinking. 6? I don't remember.

And then the next release of Codex- Right ... by OpenAI. Clearly there was an uptick in the quality that these things can generate.

Yeah. I think that, as you said, there's a couple of people are getting better at prompting and how to using it. Right.

Things are getting better. So a switch was thrown. Right.

Mm-hmm. And what happened was, look, we tend to think of human-generated code as the gold standard. It was far from gold.

No. Yeah. Right?

No, there's a lot of garbage. Human code, for X lines of human-generated code, you're going to have Y vulnerabilities on average. Sure.

Mm-hmm. With AI, we had for X amount of lines of code, it used to be Y to the third. Right.

And sometime in December, January- Mm-hmm ... the human line and the code generated, or the AI generated line sort of met. Mm-hmm.

The thing is that human line is always going to stay steady- Right ... and the AI line continues to go down. Right.

So it is getting better, but it's not the amount of code that is choking the system, it's the vulnerability reports. It's killing bug bounty programs, not just in open source, but in commercial. Yeah.

Yeah. Who can keep up with all this? Mm-hmm.

And a lot of it is- You're going to need an AI to keep up with the AI eventually. Yeah, but the AI has to figure out what's worthwhile and what is just noise. Yeah.

And then you start running into the whole AI model collapse thing. When it takes in all this stuff that's AI, and some of it's slop, then it starts learning on slop- We're seeing it ... and around.

And slop begets slop. Right. And garbage- To get biblical.

Yeah. Garbage in and garbage out. Yeah.

Now mind you, there's always been human slop as well as AI slop. Let's be real about this. But I forget who it was who told me this yesterday, but the one thing, though, that AI does have a real advantage over people is that people don't scale, we go to sleep, we eat.

Mm-hmm. We try to have a work/life balance. I'm not sure really what that is, but I've heard about it.

I saw some people out last night, so maybe they must have some balance somewhere. I guess so. I was working last night actually.

But- How long will it be until AI demands some time off too? Yep. You never know.

Maybe the AI needs some time off, because that will reduce the slop it's taking in. I don't know. Yeah.

No, but Steven's point about the collapse is, I also read a few reports on this. Yeah. Mm-hmm.

After a while, I had it. I did a site with AI, I redid the Textstrong learning site, and all of a sudden yesterday, it had some sort of stroke or something. Mm-hmm.

It just lost its CSS. Right. Oh.

" And it was like, "Oh, I'm so sorry. " You know how AI is, they're nothing but polite. Mm-hmm.

Yeah. Yeah. Yeah.

And it did. It recreated it, thank God. Right.

I have issues with talking to machines, but that's- Yeah ... another thing altogether. Yeah.

Let's talk about something else that I heard barely a whisper about here. Mm-hmm. But they talk a lot about Broadcom and VMware, but- Right ...

not much about Red Hat. Mm-hmm. What is your assessment of when you compare SUSE and Red Hat, how do you stack these things up, and how do they look to you and...

I think they're both very strong. I think their product lines are both very strong. I think both corporations are really strong.

And I see them, they're very much competitors- Mm-hmm ... but at times they've been very much putting the other one down, and I don't see them doing that so much anymore. I think with the whole Broadcom, VMware thing, that's much more of a tactical thing.

So, but there's still a lot of VMware customers out there who they hope they can persuade to come over to some sort of a Kubernetes or, in SUSE's case, Rancher-based approach- Mm-hmm ... to do it, and that makes good sense to me. I'm a little surprised, though, that they're still emphasizing it, because again, that's, to me, is more of a tactical thing, and it's been a couple of years now.

I would've thought that all the people who were interested in coming over from VMware would've already have made that move. Of course- Yeah ... that's easier said than done because it's not easy to shift over.

Yeah. But they're hoping. Yeah.

They're trying to make it so, but- Yeah, they're trying to make it so, as is Red Hat, as is- They all are ... Mirantis. A lot of them.

Oh, yeah. Anyone who has a virtualization platform is. Yeah.

Right. Yeah. Anyone who has or anyone who's involved with any OpenStack people.

Yeah. All those companies- All of them ... are wanting to play the game.

To your point, walking around here, I think I encountered a handful of people who used to be VMware customers who are here, but it wasn't by droves by any means. Right. Yeah.

I interviewed a few. Here's the thing. When I look at Red Hat versus SUSECON, the thought that occurs to me, is it a case of parallel evolution or are they copying each other?

Right? Because they both have this big sovereignty push. They're both moving into cloud native.

They both moved into cloud native- Cloud native ... and have been in it for a long time. But if you look at their messaging, if you look at the pipe, product roadmaps as we used to call them- Mm-hmm ...

very similar. Right. Right?

It's Macy's and Gimbels at this point. Mm-hmm. And so I don't know how many people get the Macy's- ...

Gimbels thing- I was faking it ... but that's okay. Yeah.

But- What about EJ Korvetts, man? I worked at EJ Korvetts- ... cashier in the record department.

Mm-hmm. But, anyway, so I do think there is that. Mm-hmm.

And they're both... Look, it's healthy, I think, for the market to have a healthy SUSE and a healthy Red Hat- Mm-hmm ... pushing each other and moving the whole open source and Linux and cloud native space forward there.

I think when it comes to the VMware, though, thing, here's the thing, and I forgot who I was talking to yesterday in the wrap. I think it was in our Techstrong gang. Mm-hmm.

Maybe it was Stephen Dickens, I don't remember. But, you had a product that if you were doing virtualization in a private data center or your own closet or what have you, not in a public cloud- Right ... 95% of you were using this product.

Right. Mm-hmm. That's a big chunk of the market.

Right. A big chunk of the market. And so yes, it may be tactical- Mm ...

but it's such a big, fat- Mm-hmm ... money target- Right. Right ...

that you can't help. Right. But, if you're the CRO here at Red Hat, you've got to be saliv- at SUSE- SUSE, right ...

who came from Red Hat- Right ... you've got to be salivating over that- Mm-hmm ... big prize.

Right. And so as we spoke about yesterday, I'm not saying VMware's going, that Broadcom's going out of business. No.

Look, they bought it to even out their revenue stream because hardware goes up and down, and- Mm-hmm ... software is more stable. " Right.

"Yeah, it's going to cost me more. I got some bloatware in there. " Mm-hmm.

But there's enough people, and I haven't seen it just here at SUSECON, we did a whole series with AWS partners- Right ... who say, you're not going to change because of the money. You're not going to change because of the bloatware.

You're not going to change because of the attitude you get, take it or leave it. Right. But you are going to change because this is a good time to look at options- Right ...

to decide, do I want to do cloud native? And if I do cloud native on the other one, I'm kind of stuck with- Mm-hmm ... their path forward to it, Tanzu or whatever.

Do I want to do... And you know what, the guy from, oh, shoot. Maybe it was V...

Not V. Oh, you interviewed him, too. Traffick?

Traffick. Right. He said it well.

Your choices when it comes to this is you could migrate, right? You could migrate from VMware to another virtualization. Mm-hmm.

It's not much sanctions, that's just- Right ... doing a migration. Right.

You can modernize- Mm-hmm ... which I think is kind of code word for moving to the cloud- Right ... or you can transform- Mm-hmm ...

which I think is a cloud native of- Mm ... not microfocus, multithreaded, micro- Microservice? Microservices- Services ...

based architecture. That's a transformation. That's a big lift.

Yeah. Mm-hmm. But that, so where on that scale are you?

Mm-hmm. If it's just a migration, it's not as compelling. Right.

If it's a modernization, you're moving to the cloud. Right. If it's a transformation, you're in for the whole enchilada.

Right. The other thing that I did hear a fair amount about here is this whole edge computing push- Right ... from these guys.

Yeah. And they bought Losant, and they're going to open source that platform. Right.

But is edge computing kind of a new frontier for open source? Because when I look at all those platforms, there sure is a lot of proprietary software running out there on those things. So that kind of...

Well, edge computing, there's always been a lot of open source there. There's, what is it, Stirling X from the Open Infra Foundation. There's a lot of people involved with that.

Wasm. Wasm. Yep, exactly, which is sort of kind of in there, too.

What I see happening with the edge push is I really see that more as not so much an ordinary edge. I see that as SUSE wants to get into the industrial, the factory edge. Mm-hmm.

Yes. Which makes sense to me. That's a real market, because that market, that vertical, has never really completely gotten into it, but with AI coming on, which means the first AI robots that are really going to be big are not going to be cute little dog robots, well, that make real money.

It's going to be these factories using AI to automate. And for that, you're going to need a networking solution that will integrate that into the rest of your structure. And I think edge for SUSE will fit into that.

Yeah. I mean, that's where I see this going. BMW, I'm a BMW guy- Right ...

and so I kind of follow what they do. Right. They actually have been using robots.

Right. And some of them are now autonomous. They don't use the AI word more, they use autonomous, but really advanced robots in building- Right ...

the new class, they call it- Right ... the latest version of BMW. Right.

So I think you are going to see a lot of that. And they did mention, I did interview their person with the edge here from SUSE, and they clearly distinguished the industrial edge offering- Right ... from their- Right ...

normal edge offering. Right. So they do see that, Steven- Right ...

I think you're dead on. However, here's my take. I think AI is going to create a whole new edge use case.

People who want to run Claw, Open Claw, or some other agentic stuff, not on someone's cloud- Right ... where that token thing is spending like a one-armed bandit. Right.

If I could run it on a smaller module, I'd want to be able to run AI on the edge- Right ... on my own- Right ... self-contained thing, and my costs are fixed, and it's better.

And you know who the big winner there is? Apple. I think the edge device of tomorrow is the Mac Mini, the Mac Studio, and I think Tim Cook stepping down and this guy coming in from their hardware, you're going to see a lot of Apple AI on the edge devices.

I don't think I'm going to see Apple devices in factories. I think I'm going to see some sort of- Not in factories. That's industrial edge.

Yeah. Different product. I think I'm going to see what, to his point- Right ...

a need to normalize all that data, which is what Lucent's really about. Yeah. And then that's going to be used to train some local model- Right ...

that's going to be running on those devices, not in the cloud, and probably running something that feels like Open Claw, but maybe not Open Claw. Well, you're assuming we can get DRAM chips, memory chips. Yeah, and I'm assuming that the core edge device is going to be a lot more powerful in the next few years because we're getting better at taking those things down.

A lot of the edge computing devices today can barely run the protocols that they have. Right. But you're going to be able to process and analyze data at the edge, or near it.

But there's such an installed base already. What, you're ripping them out? You're going to start taking down SCADA?

Yeah, but this could be more of a transformation thing, though. Yeah? Because I could see, okay, you're right.

A lot of IoT devices, whether it's consumer or even industrial, are pretty darn stupid. But if we start getting, say, cheaper NPM chips rather than the cheapest possible X86 or ARM chip, which is where IoT is now, with a moderately powerful NPM chip, you can put a lot of smarts into a... Well, actually, the Mac Mini has already shown you can do a lot- Mm-hmm.

That's all, yeah ... on a small, relatively cheap platform. Yep.

And maybe it's not Apple Mac Mini, but a clone- Right ... a similar sort of architecture- Yeah ... similar sort of form factor.

Yeah. And we won't have an 8 or a 9 chip any time soon. I don't think you need a mini mainframe or a mini computer at the factory to run these things.

Right. Yeah. That's the whole point of the edge.

And the thing at the very end can still be stupid if I've got a mini server- Right ... close to it. Yeah.

So I don't necessarily have to put everything on there. So that's one thing I'm looking forward to. I think we are going to see a tsunami of robots in all different shapes, sizes, and form factors- Mm-hmm ...

as they get more affordable- Mm-hmm ... cheaper. Right.

You're going to see them everywhere. You're going to see them cleaning the streets. You're going to see them cleaning your house.

You're going to see them in the factories. Mm-hmm. You're going to see them on the UPS truck.

Mm-hmm. They're going to be everywhere. They're going to be chasing me around.

You think- Around the house. No, but you think knowledge workers losing their jobs or being threatened by AI is something? Wait till these robots hit.

Yeah. Yeah. Three to five years is my estimate.

Yep. Yeah. So, last topic, but I heard a little bit about this, but not a lot.

But platform engineering is definitely a major trend. And there was some discussion of that here, but it wasn't like they were kind of banging the drum. They kind of said GitOps was the future, and they left it at that.

But I know you have a piece talking about- Yeah. Well, I think they didn't send the memo to the platform engineers. But this whole sovereignty thing- Mm-hmm ...

comes down to a platform, right? Right. Someone's going to be responsible for this.

Right. And this is not the first time I've seen this pattern recently with platform engineers. They are the proverbial downstream.

And you know what happens when stuff runs downstream, right? Mm-hmm. It falls on you.

And-So I think these platform engineers and the platf-- and you know, plat-- the platform engineers, it's a dubious term. The SREs, the IT ops people that- Right ... we've known for most of our careers.

Mm-hmm. They're going to get stuck holding this bag, and they've got to... I'd like to see them get in sooner than later.

But I do think that ultimately, that's where this sovereignty thing lands. I don't think a developer- Mm-hmm ... gives a crap- Yeah ...

to tell you the truth. He's maybe working on his local machine or what have you, but someone's got to platform this sovereignty, and it's going to be them. Yeah.

And the, but the nature of the IT job is just changing so dramatically. A lot of the traditional IT and admin tasks and what, or what we call ClickOps, are clearly going to be done more and more by AI, which is going to be managed by someone who feels like a platform engineer. Right.

So there may be fewer IT people at the end of the day managing much more at higher levels of scale. Mm-hmm. But I don't know, what's your assessment of how this plays out?

Oh, it's going to be interesting. I started off managing on IBM mainframe 360s and Unix machines. Linux is still far in the future when I started up.

So I know all this stuff, which is still, I still have VI in my fingers, embedded in my fingers, and I know sed, awk, and grep and how, and all that junk. And all that sort of- There's a- ... antique knowledge now.

There's a therapy group here. I know, I know. Yeah.

Old techies, that'll be us. And now, and then of course came DevOps and Ansible, which is my personal favorite, and Puppet and all the rest, and Chef. The list goes on.

And now we're moving to yet another layer of abstraction. And one of the things that sort of worries me about this is, if I'm, say I'm someone new, though, and I'm working at this really high level of abstraction, I really don't understand what's going on underneath me. So if the AI feeds me garbage, can I tell that, "Wait, that doesn't make sense"?

I don't know that I'm going to be able to do that, whereas for all that, I do not suggest anyone start with IBM 360 assembler as their first language. No, don't. That was my mistake.

" And, uh- To me, this is a case of where are we coming into a near-term future where Grandma died, she took the recipe with her, and now we don't have chocolate chip cookies anymore. Mm-hmm. Right.

We covered an article last week on the Gang, something I wrote. I don't know if you saw NASA- Right ... open source the command module and lunar module- Yep ...

source code. Yep. Used to work for NASA, so I follow them closely.

Four kilobytes. Yep. 4KB.

Yep. Itty and bitty. Yeah.

And that was a lot of cards back then. Yeah, but four kilobytes. Every bit and byte- Mm-hmm ...

meant something. Right. If you asked coders today to try to do that in the same footprint, you're SOL.

Mm-hmm. Right? You're not even getting close.

Right. We've lost that level of- Mm-hmm ... you want to call it craftsmanship.

Right. Coding skill. As we continue to abstract away- Right ...

we lose more and more of that. Right. And so, does that make us worse?

Well, no, because I think it's all relative. Right. This is the lesson that I learned when I bought my first hard card, hard drive, a hard card, hard drive.

It was like this big. Right. And it was 40 megabytes, and I said, "My God, I'll never buy- ...

" Nope. And then, six months later, that was full. Right.

It's all relative. Yeah. Yeah.

So we write bloated code, we abstract more, but we have machines that are better able to handle that thing. Yeah. We don't, you know.

And the dirty little secret of all of this is that- Mm-hmm ... the utilization rates for the servers sucks. Right.

Yeah. They're basically idle for 80 to 90% of the time. Right.

And so, guess we're writing more sloppy code, but meanwhile, we have all this hardware that we're installing that's sitting there idle, probably not good for the carbon equation, but- Mm-hmm. It's like the physics law of fluids, right? Exactly.

It fills its container at the lowest level. The more power you have, the more- Mm-hmm ... trust, the more sloppy you write.

Yeah. Yeah. Thinking of, touching on chocolate chip cookies again- Uh-huh ...

there's the infamous Oak, what is it? The Uncle Amos chocolate chip cookies, which started out great, and then, of course, they were great cookies, and then they were bought by one factory- A big company ... and they modified it because they abstracted away from the homemade qualities to make it cheaper and more efficient.

So eventually, you end up with cookies that nobody wants, and there went, and that was the end. It reminds me of this AI coming full circle, the AI- Right. Exactly.

So are we in danger of something like that happening? I don't know, but it's, you know- I miss Famous Amos chocolate chip cookies ... some people, yeah, Famous Amos, that was it ...

they were the best. They were. And then they did, you're right.

They went, they... Same thing, Ben & Jerry's. Yeah.

Right? It's still not bad ice cream, but it's not the Ben & Jerry's- Right ... that it was.

And that, are we going to see that with this AI sort of collapse? Mm-hmm. Is it falls in on us?

I-The thing is, AI keeps getting better- Mm-hmm ... or more powerful or however you want to describe it. Right.

So it can hopefully correct itself. Keep those cookies. Yeah, keep the recipe.

Keep the good recipes. So is this going to be, let's using that baking metaphor, right? So if you go to a bakery, maybe we'll pick France, but there's artisan bakeries where it's made by humans and it's locally owned and created, and that's great food.

And then there's all the stuff made in the factories- Right ... and all the commercial stuff. Is that what's going to happen with code and IT, right?

There'll be human artisan code- ... and then there'll be machine code. Well, there's actually sort of that.

It's already there. See, we're talking about how people don't get really artisan code, very low level. But actually, there is at least one project, FFmpeg, which if you're watching this now on your video, an FFmpeg bit of code is translating that code.

Yeah. And that is still, to this very day, written in Assembler. They don't have compilers.

C, they sneer at as being much too sophisticated for them, and they are proud of the fact that they are still writing this at an extremely low level in all these different architectures. Bless their hearts. Meanwhile, though, I think the machines are going to start running or writing code in lower level languages, right?

Mm-hmm. We created C++ and Java- Right ... to create a higher level of abstraction for humans to talk to machines.

Right. Machines don't need that. They're just going to run in some lower level language.

So maybe everything comes- Which goes- ... back to Assembler. Maybe.

I think we will see machine-to-machine languages- Mm-hmm ... that cut out the human middleman. Right.

Yeah. Whether that'll be faster, better, I don't know. But I do think that's coming.

And then here's the thing about that that gets really interesting, right? So if the machines are better at writing the code in their lower level languages, maybe they'll be much better at the utilization of the servers, and we might actually need less infrastructure because the machines are going to use what we have more efficiently. There's the thought.

Mm-hmm. Now who's being naive, Mike? NVIDIA doesn't like that thought.

No. No, no. I'll tell you that.

No leather jacket for him. No. Hey, we got to wrap this up, though.

I think we're about out of time. Steven, thank you so much for stopping by and the gang today. My pleasure.

You can catch Steven on any of our Techstrong sites, but he's well, not red, well- Respected ... produced across the- Right ... entire tech media scene.

Yeah. That's going to wrap up our SUSE coverage. We are going to have a wrap-up newsletter tomorrow with final thoughts.

And we are here the rest of today, well, till about 2:00 Prague time, with even more coverage. But as far as the gang goes, that's it. Tomorrow, I believe John Swartz is running the bus.

John Swartz is running this thing. He's driving the bus. I think he'll probably be talking about Apple, but what do I know?

Well, John doesn't- Odds are. like to talk about it. So stay tuned for John and the gang tomorrow, and then Monday you'll be back in charge.

I'll be back. I'm on the road for the next couple of weeks, so I'll see you. I'll try to drive in when I can, but until then, this is Alan Schimmel for Techstrong Gang.

Thanks, everyone. Have a great day. Bye.