Windows Weekly 1007 Transcript

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

Mikah Sargent [00:00:00]:
Coming up on Windows Weekly, Paul Thurrott and Richard Campbell are here, and midway through the show we talk about Microsoft's earnings. But before we get there, we talk about the state of Windows with 4 new builds for the Windows Insider Program, plus an update on how Paul's doing with the RTX Spark. Well, things are improving at least. And we mourn the loss of the Arc browser, plus a little bit of Xbox news, but really not a lot, and our tips and picks of the week. Stay tuned. Stay tuned.

Mikah Sargent [00:00:46]:
This is Windows Weekly, episode 1007, recorded Wednesday, October 28th, 2026: A Sea of Blues and Greens. Hello and welcome to Windows Weekly, the show where we talk To 2 of the foremost Windows watchers in the world. I am Mikah Sargent, subbing in once again for Leo Laporte while he is on vacation. Uh, happy to be here with those Windows watchers. Let's start with Richard Campbell. Richard, where are you coming to us from this week?

Richard Campbell [00:01:21]:
It's, it's weird that that's a question, isn't it? But yeah, no, I mean, I'm in Utrecht, uh, in the southern part of the Netherlands. for the Techorama conference, which just wrapped up a couple hours ago.

Mikah Sargent [00:01:33]:
Nice. And Paul Thurrott of thurrott.com, from whence do you hail?

Paul Thurrott [00:01:40]:
I'm in the depths of my despair. No, I'm sorry. I'm in— I'm still in Mexico.

Mikah Sargent [00:01:44]:
The 4th layer of hell.

Paul Thurrott [00:01:45]:
Yeah, exactly.

Mikah Sargent [00:01:46]:
The 4th circle of hell.

Paul Thurrott [00:01:47]:
Yeah, yeah. No, I didn't pass go. I went right to the whatever the final— what is it? 9 layers? I don't know.

Mikah Sargent [00:01:52]:
Right to the center. Yeah. Well, I'm glad that you're here with us and that the— apparently that circle of hell, whichever it happens to be, gets— Dante scholars out there, tell us about— gets reception. It's good. In any case, yes, this is the show where we talk about Windows, Microsoft, and the surrounding technologies. So let's get into it. Apparently we're starting out with a conspiracy theory? No, just a theory.

Paul Thurrott [00:02:21]:
A conspiracy theory? Oh no, it's not. I've actually had a theory confirmed. Uh, before that, I just want to mention real quick, uh, Microsoft is going to announce their quarterly earnings today. Um, this will be toward the end of the show, maybe right after the show ends. We'll see based on timing. But

Richard Campbell [00:02:34]:
did we talk about the $60,000 then? Because that sure sounded like positioning for the quarterlies.

Paul Thurrott [00:02:40]:
Yeah, well, I do— wait, actually, do I— I don't— did I not put this in here? I must have.

Richard Campbell [00:02:45]:
I don't think you put it in there.

Paul Thurrott [00:02:47]:
I meant to— hold on, let me see. Uh, no, I do it. I have it in there. Yeah, no, it's in there. Yeah, we'll talk about that in a moment. Um, all right, so, uh, yeah, we don't have to talk about that further, but, uh, there have been some, uh, reporting changes. Remember, Microsoft changed their, you know, corporate structure, etc. So it's going to be kind of interesting.

Paul Thurrott [00:03:05]:
They're— from my sort of cynical viewpoint, they are finding unique— I'm gonna even call it innovative ways to hide the money. Um, and so, you know, it's going to be kind of interesting to see what that looks like, uh, for the first time. So whatever needs—

Richard Campbell [00:03:20]:
There's this real sense by the Street that they're having— they're slowing down, and now they seem to be reacting in like every way that they cannot show that they're slowing down.

Paul Thurrott [00:03:30]:
There, there's a lot of indications of kind of, uh, finally, by the way, uh, an industry-wide slowdown across the board, right? So meaning like data center rollouts and, uh, CapEx spending, etc., etc. But 2 things that just are contrary to each other, which I find to be very strange, is there's some consensus I've seen, meaning I've seen it in more than one place, that Microsoft is in some ways compared to their peers in this space, in the big tech, big AI space, the hyperscalers, right? Yeah. The adult in the room, that they're in fact handling this in some mature fashion, to which I can only laugh like a Bond villain. That's crazy. And then the other, but the other one is that There's this other stuff, like Microsoft has almost lost the thread on what the point of what they're doing is. To me, they seem so disconnected from reality. Like, I don't even— to call them the adult in the room makes— what are the other guys, like Batman villains? Like, what are we talking

Richard Campbell [00:04:28]:
about?

Paul Thurrott [00:04:29]:
It's bizarre. Yeah. But anyway, so we'll get to that, hopefully, as it occurs. I'll just say just upfront, because assuming these results occur before we end the show and we're able to discuss it briefly, that in no way will be an in-depth look at what's happening because you can't just look at the headline stuff they announce. You have to go—

Richard Campbell [00:04:51]:
No, they're so sculpted. They're so designed to send a message that doesn't necessarily reflect what the data shows. It takes time to get between the lines on the stuff they write these days. And half the time you can't even do that.

Paul Thurrott [00:05:05]:
Right, right. So, um, yeah, all right, let me get to— let me get to some good news or whatever. Maybe, maybe just at least not negative news. Um, so late last year we got the first indications of this, and then early this year we've got the formal Pawan Davaluri, uh, announcement: we're going to address the pain points of Windows 11. And

Richard Campbell [00:05:30]:
yeah, we like this. This is good news, right?

Paul Thurrott [00:05:32]:
If you are a Windows 11, uh, user or a distro, whatever This is all good news, right? The interesting thing is though, and literally from the first second I saw this, I was like, okay, 2 things. Why? I mean, literally, they've ignored—

Mikah Sargent [00:05:49]:
First of all.

Paul Thurrott [00:05:50]:
All these complaints for years and years and years.

Richard Campbell [00:05:52]:
Why now?

Paul Thurrott [00:05:54]:
And why now? Yeah. Why and why now? Okay, so I guess 3 things. And then what's really happening here? They're not addressing the uncertification of Windows 11 almost in any way, shape, or form, right? So if you go back a couple years, I wrote an article about this, like here's the 7 or 8 ways in which Microsoft has completely uncertified Windows 11. Now they're addressing pain points. Do any of those pain point fixing things match up to any of those certification points? And I believe one of the 7 or 8 did, and that was the whole thing.

Richard Campbell [00:06:26]:
Yeah.

Paul Thurrott [00:06:27]:
And it's a little more nuanced than that, but if, you know, you could weight these different ways, like the most important things that are terrible in Windows 11, you know, versus what they're doing, and it's, it's really not even close. That said, everything that they're doing in Windows 11 as part of this campaign is great, right? You can't complain about it. It's— no, it just is, right?

Richard Campbell [00:06:48]:
And I tend to agree, but I also feel like part of this is recruiting engineers back. I think Pavan, when he got the 2 teams back together and sort of looked at the situation, said, how many people had just checked out? So I'm gonna— When he saw that outcry, it's like, hey, this is an opportunity for me to pull my engineers in.

Paul Thurrott [00:07:07]:
Okay, so that's, that's an interesting insight, by the way, um, because that actually corresponds to something I just heard from a friend. So I'm not going to read this verbatim, but I— the— you and people listening or watching may remember that I've probably said this multiple times. At some point in going over what they're doing and, and just, you know, understanding Microsoft as a company and seeing what's going on in the world and everything, I was like, you know, I think they're doing this because of AI, right? I think they can't say that the local AI story is Windows 11 unless Windows 11 actually works, you know? Yeah, that's fair argument. There's a PR thing going on right now. Sorry.

Richard Campbell [00:07:43]:
Because when he tried to pitch the idea of adding AI to Windows, all hell broke loose.

Paul Thurrott [00:07:49]:
Yeah. Look, one of the constant refrains in my world is that I will write a— I could write a story, we'll talk about it at Windows Weekly, whatever it is, and I'll say, hey, Microsoft made this small change to some app like the new Outlook, which everyone loves. You know, everyone, you must have heard this, big fans of the new Outlook.

Richard Campbell [00:08:05]:
Absolutely, my favorite. And my favorite part was that button that switched it back to classic. That was great. I like that button a lot.

Paul Thurrott [00:08:13]:
You could vibe code an Outlook that's nicer than the new Outlook. But, you know, other than that, whatever. But I'll write an article like, hey, they fixed, or they more likely not fixed, but brought back some feature that has been missing in this product since it began. That was in the— is in the classical look. And then you will inevitably get the feedback from someone who'll say, okay, that's nice, but what about dot, dot, dot, right? Because everyone has their kind of pet peeve kind of a thing.

Richard Campbell [00:08:36]:
So many pet peeves.

Paul Thurrott [00:08:38]:
Yep. So anyway, trying to make sense of our world, I had at some point, I don't remember when this happened, but at some point over this year, I was like, I think this is because of AI. They're talking about putting agents into the OS. Mm-hmm. Um, I, I— who would trust this? Like, I mean, Windows is not trustworthy, right? Or, you know, I mean, it's getting, getting better, right? But, um, so that was, that was my theory. Now, the interesting thing is I got a text, rather, from a friend who used to work at Microsoft and still has many connections inside Microsoft, who shall go nameless, but his initials are— no. Uh, and, uh, and this reads Very clear. First of all, I can tell he doesn't watch Windows Weekly, which is fine.

Paul Thurrott [00:09:23]:
Why would he? But he basically verifies what I just said, right? But with far more detail. And Microsoft started caring about Windows again because Satya and Amy realized it was losing big time to Apple as the place to do AI tasks.

Richard Campbell [00:09:39]:
AI.

Paul Thurrott [00:09:40]:
And that this was just fundamental to the OS, right?

Richard Campbell [00:09:44]:
Yeah.

Paul Thurrott [00:09:44]:
So Judson and Pavan put together a plan, which has a name, which I'm not going to say just in case. to get the garbage out of Windows, meaning, you know, all the widget— the terrible stories that, you know, are in the widget thing, the Bing crap that's behind search highlights, you know, all the terrible stuff. Yeah. There's some bit of information about how unhappy the Bing team is about that because they just lost a lot of money, but I'm not going to get into that. But part of this effort is about moving a bunch of evangelists and engineers back into Windows to actually make this make sense and to make sense of it for the world, right? In other words, to communicate that Windows, well, first of all, it has to become the best place to build AI apps and whatever goes around that, but then market it as such, right? The Copilot thing has been a disaster for them from a PR perspective. I mentioned this last week, interestingly, but— Oh yeah. I think I did. But right at the beginning of that Copilot wave in 2023, they had an offering for consumers called Copilot Plus, which they got rid of pretty quick.

Paul Thurrott [00:10:53]:
And I paid for it immediately. I tried to do stuff with it. I think we talked about custom GPTs, which we could all agree is a terrible name for anything. Mm-hmm. But that was a complete failure. But by virtue of the fact that Apple in Apple Silicon has unified memory. And because this came up out of the phone world where this kind of thing is actually super important, it just does the memory allocation automatically.

Richard Campbell [00:11:20]:
Yeah.

Paul Thurrott [00:11:21]:
Meaning that if you need the GPU—

Richard Campbell [00:11:23]:
That's the way that Apple architected the M series. And then they didn't do it with this in mind. Not because— It's just the way they did things.

Paul Thurrott [00:11:29]:
Nope. Not— had nothing to do with AI, but that's the way they did it.

Richard Campbell [00:11:32]:
No grand plan. It's just what they did.

Paul Thurrott [00:11:34]:
And it really, it's an interesting just artifact, I don't know, byproduct, whatever.

Richard Campbell [00:11:41]:
Yeah.

Paul Thurrott [00:11:41]:
That this is in fact very, very good for Local AI. So, you know, whatever, people are buying up Mac Minis, et cetera, et cetera. Yeah, it goes on, but this stuff is not necessarily related to this. But yeah, there's a whole kind of organization that's still being spun up in some ways around It's like fireworks going off outside because it's Mexico. And around this stuff. And okay, then it goes on. There's other stuff I don't wanna get into, but that's interesting to me. And I don't remember, well, but this would've been before we started recording the show, but I think we were live for the Discord.

Paul Thurrott [00:12:18]:
But I wonder about this stuff with RTX Spark, right? Mm-hmm. Because if you think about Apple as a threat for local AI, That's Microsoft, right? Because they have Windows and that's the whole ecosystem. That's a big problem there. But think about Nvidia too, right? That's a problem for them too. And there are signs that Apple might be getting back into the server market. They might be putting Apple Silicon up in servers, right? Very interesting. There are also news stories about different big AI companies like OpenAI adapting their products to work on different hardware architectures, including the TPUs that Google uses, right?

Richard Campbell [00:13:01]:
Right.

Paul Thurrott [00:13:01]:
Because they want to— yeah, they have— it has to be out everywhere. So Nvidia is in this place in time where they are running the world, literally. They're— if they're not number one, you know, briefly, they are number one or two in the world from market cap perspective, richest company in the world. They are making real money on AI, unlike everyone else, and They want to protect that, right? You mentioned, I think you said, did you say this last week? I think it was you. Someone made the comment to me recently that Nvidia has essentially become a bank, meaning they're bankrolling the AI efforts at companies and partners that they believe to be strategic to make sure they can survive and thrive using their stuff as a way to kind of spread the wealth in the ecosystem, et cetera.

Richard Campbell [00:13:46]:
We saw the same thing in the dot-com boom, right? It becomes a circular effect. You invest $100 million into this company so that you have the investment, but then they have to turn around and spend it on Nvidia chips anyway. So you just made $100 million in sales. Your $100 million just became $200 million.

Paul Thurrott [00:14:01]:
I think it was last week whenever recently I referenced a joke from Scrubs, but one of my favorite jokes of all time applies to the AI financial situation we have today. So Steve Martin back in the 1970s said, I'm going to write a book called How to Make $2 Million in Real Estate. And the first step is make $1 million in real estate. Step 1, get $2 million. And that's kind of what this whole thing feels like.

Richard Campbell [00:14:28]:
Best way to make a small fortune in AI is to start with a large fortune.

Paul Thurrott [00:14:33]:
Yep. So yeah, so it's interesting, right? But I wanna be super clear about the Microsoft improving Windows 11 thing because it, on one level, it doesn't matter. As long as you use Windows 11, Windows 11 is getting better. And, uh, oh yeah, I skipped over that part of this text, but I mentioned last week, I said, you know, Microsoft's never said this, but they pretty much completed the list of stuff that they had said they were going to fix, right? I mean, as of today, pretty good chance you have all of it on your computer, right? Um, but they're doing further improvements. So I refer to this as a sort of phase 2, meaning They didn't really say it was a phase 2, but, you know, here's a couple of the Photos app and whatever else we were talking about last week. And he actually said that— he said he didn't say phase, it was a synonym for phase, but it was, uh, the same. And I was like, yeah, that's what— that's also what I said.

Mikah Sargent [00:15:25]:
That's fascinating.

Paul Thurrott [00:15:26]:
Yeah. Um, so that's, that's kind of cool. But, um, so there are more improvements coming. I— there's like people like Scott Hanselman have talked about some of the actual enshardification stuff, like, you know, being able to sign in normally as a local user. You know, if that's what you want, right? Right. And I'll get to that toward the end of the show as well. But there are kind of workarounds or even fixes for most or all, I would say, of the certification stuff. It's an effort.

Paul Thurrott [00:15:50]:
You have to maybe fine-tune a Windows ISO before you install the system, or you can later use the utility and try to blast stuff out of there. Both, you know, work. It's like you can get Windows to where you want it. It's not completely horrible. And then they've made all these changes to things I didn't really care about, like the Start bar, the Start, button in the Start menu and the taskbar and whatever else. And that's fine. I mean, that's— it's cute. It's, it's all good.

Paul Thurrott [00:16:14]:
It's, it's fine. It's cute. It doesn't, doesn't hurt me. It helps me in some ways, whatever. But it's nice. I will say coming back to Mexico, because I have— the one thing, the one feature they changed that I use all the time is you can have that smaller taskbar, which I very much prefer. I just spent like 2 years experimenting with like hiding the taskbar all the time because I hated how big it was, you know. So on all my computers, I have this like small taskbar, which I love.

Paul Thurrott [00:16:42]:
It's almost as small as like the menu in macOS. You know, it's nice. It's like, it just doesn't take up too much space. But then of course I come here, I have several computers, and you bring them back up, you get them back up to date, and they all have these gigantic taskbars on the bottom because they're, you know, pre-26H2 or whatever. Yeah. And it's like, ugh. And then you see it again, you're like, oh, that's so awful. Like, and we've been dealing with this since like 2021.

Paul Thurrott [00:17:05]:
Like, it's been like this the whole time, you know?

Richard Campbell [00:17:08]:
Yeah.

Paul Thurrott [00:17:08]:
Windows 10 had resizable taskbar.

Richard Campbell [00:17:11]:
You've had a few weeks of it gone away and then you miss it terribly.

Paul Thurrott [00:17:17]:
You get used to it. It's funny, but you know, as, look, as any complainer will tell you, or actually, you know what? Maybe most complainers aren't this self-aware, but I will tell you as a complainer, when something's wrong, you'll notice and you'll complain, right? And you'll complain. And you complain and you complain and you complain. And then if that thing actually gets fixed, which I gotta tell you, pretty rare, right? But when that wonderful thing happens, right, this should be a wonderful moment where you can be like, oh my God, they listened and they fixed something. It's really great. But my God, you get used to it so quick. Like you just, you go from like years of complaining to like, oh nice, it's fine. And then you stop talking about it, right? Right.

Paul Thurrott [00:17:53]:
But then you go back to an older computer, you open these things up and they're all on the old one. You're like, oh, that's awful. And so it's just so terrible. So anyway, like, I, I, the point here is, I logically, it made sense to me that Microsoft was trying to fix Windows because strategically for Microsoft, the biggest thing in the world right now is AI. And a big part of that for them is transferring the cost of running AI tasks in the cloud to the desktop as quickly as possible. And that can only make sense for them if that computer is their computer running their system. Because then they can do that integration thing and have the orchestrator or whatever it is that will make sense of local AI versus cloud AI and all that good stuff, whatever it is.

Richard Campbell [00:18:36]:
And that's the path we're headed on now is you're going to try and run everything locally and ship out to the frontiers where you can't do it otherwise. And whoever controls that router controls their revenue stream.

Paul Thurrott [00:18:47]:
Right. And so one thing I've been trying to get a handle on, this is supposedly available either now or will be by the end of the month. If you use GitHub Copilot, which is a developer tool, they don't call it an orchestrator. I forgot the name already. I'm so sorry. But it's something like an orchestrator. And it's supposed to do that kind of thing, right? And it may be, I haven't seen it yet, but it may be almost like an early peek at what should be a fundamental part of Windows 11 or Windows, whatever, you know, maybe it's Windows 12.

Richard Campbell [00:19:19]:
The next version of Windows.

Paul Thurrott [00:19:20]:
Whatever, yeah, whatever version. Where, you know, in the same way that, you know, we have an NT kernel that does scheduling of tasks and things and handles, you know, multitasking and back and forth between apps and which things get RAM and yada, yada, yada, you know, that it has to work that way for AI models as well, you know.

Richard Campbell [00:19:39]:
Right.

Paul Thurrott [00:19:41]:
So, and yeah, I mean, the default, you'll know it works when that is the default, like it's local first. You know,

Richard Campbell [00:19:49]:
right?

Paul Thurrott [00:19:50]:
Uh, we are so far from that though, and I just, uh, we got a ways to go. Well, for, um, what I would call the, um, uh, more, well, less simplistic task. In other words, if you want to connect someone, you could do this on any computer today, just LM Studio, whatever QWEN model that fits on there, it doesn't matter. Um, right. And you can ask your questions, you can, uh, You know, have it write you a little cute little report or whatever it is. Like, that stuff's fine. There's no big deal. But when you get into something more complex, like a coding project, which is maybe the ultimate complexity, or a creative task where you're like, here's an image, I want you to remove the people from the background, or blah, blah, blah, whatever it is, or you're making videos, or you're doing whatever you're doing, that's where it really falls apart right now.

Paul Thurrott [00:20:35]:
And yeah, I kind of ranted about that last week, but don't worry, I'm going to rant about it again this week. And more. But we'll get to that. We'll get to that later. Anywho, regarding Windows 11 still, we've had one more set of Insider Builds since we spoke last week. There are only 2 semi-interesting features in here based on looking through all this kind of stuff. But in both cases, I can't say I care too much or whatever. But if you're familiar with how phone integration works in Windows 11, you know that you get a really basic experience with the iPhone, which is lacking in many ways.

Paul Thurrott [00:21:13]:
And then depending on the Android phone, you get a better experience, including some really good experiences if you have like a high-end Samsung phone or one of the flagships it supports for this.

Richard Campbell [00:21:24]:
Pixels.

Paul Thurrott [00:21:25]:
Yeah, and over time they have, they, Microsoft, have really improved the interface and also added things that were missing in the beginning. Like if you go into the Settings app and then into like, Yeah, Bluetooth and devices. There's a mobile devices section in there now, and it used to pop up this window where you had all your phones listed and what the capabilities you were gonna pass through, et cetera, et cetera. Most people will have one phone. I always have like several, so I have this kind of bizarre, you know, search set.

Richard Campbell [00:21:51]:
I was thinking, ah, designed for multiple phones, just the way Paul likes, but nobody else.

Paul Thurrott [00:21:56]:
So it does, I mean, you know, yeah, I mean, it's, so yeah, so I have a few phones, but, In the beginning, one of the weird things when they first started doing this work was you couldn't remove a phone. Like you could add a phone, but you couldn't remove one. So in my case—

Richard Campbell [00:22:12]:
Phone never goes away.

Paul Thurrott [00:22:13]:
I had like 11 phones in there at one point. And I'm like, guys, like, could, you know, I could just disable it sort of, but it was always in the list. So they fixed that over time that you can in fact remove a phone now, which is nice. Like, which is to me feels like baseline functionality, but You know, anyone who's used the, it's the, you know, the Phone Link app and then this mobile devices thing is kind of the backend of that or whatever. And so you've got settings in both places, which is a little confusing, but, you know, you get whatever level of integration you get with the system and there's some good stuff in there, right? You can enable things if you have the right phone, like copy and paste across devices. So if I copy something to clipboard on my Pixel phone, it will, and the paste on Windows, it will paste the thing from the phone and vice versa, which is really nice, right? You know, wireless or wired file copying, et cetera, et cetera. There's a bunch of stuff, right? And I feel like, you know, they've added that sliver thing to the side of the Start menu. We learned recently they're going to improve that so you can see more stuff in there, which I'm not— actually, my iPhone's connected now, so it's pretty much empty.

Paul Thurrott [00:23:14]:
But eventually that will be full of more stuff. And one of the things that's in at least some of these builds, certainly these— I think it's experimental is pretty much what gets everything. Yeah. Is mobile device controls in Quick Settings, right? So this is the Windows key plus A, little thing that comes up in the corner. And they have Quick Settings buttons in there for like Wi-Fi, Bluetooth, et cetera, et cetera. They're going to add one, like a— it's actually going to be like a panel for your phone, like just for mobile devices. And then from there to your phone, and the phone will have some buttons and it's probably going to vary by phone, but you can do from there, say, put your phone into Do Not Disturb mode. Right.

Paul Thurrott [00:23:50]:
'Cause we need like 117 different ways to put a phone into Do Not Disturb mode. But okay, so that's something I don't really care about, but okay, that's fine. This is, that's fine. Like we need like another way to do that. It's okay. The other feature there, this is crazy to me, but some, for 26H2, so it's in all versions of Windows, they added an internet speed test. option to Quick Settings, right? So if you go into Wi-Fi, I think is where you can see it. Yeah, there's a little button in there.

Paul Thurrott [00:24:21]:
It says test internet speed. And so they announced this like, oh, we've integrated like, uh, internet speed testing in Windows 11. I'm like, okay, that sounds kind of cool. And then you click the button and it loads a webpage where you go to bing.com.

Mikah Sargent [00:24:33]:
Are you kidding me?

Paul Thurrott [00:24:34]:
So I'm like, that's, that's not— that's— so they're actually—

Richard Campbell [00:24:38]:
That's not integrated anything.

Paul Thurrott [00:24:39]:
Not at all. But now they're they are integrating it into, uh, that little quick settings window, I guess. So, you know, small, small things, nothing, nothing major, but they are, you know, they're whatever. I'm sure they're getting a lot of feedback and for a change are listening to it. So maybe these things were on the list, I don't know, but I can't say they impact me too much, but they're, they're fine. Like, I don't— it doesn't bother me that they're there. I just don't care.

Mikah Sargent [00:25:05]:
Paul, can I ask you how often Over the course of doing this podcast, have you gotten like messages from employees who go, come on, man, we're working real hard on this, I promise it's coming around? Or do you get angry people? Or do you get like, oh God, I get it, man, it is this way,

Paul Thurrott [00:25:29]:
right? So the tenor of those messages has changed over time, because if you go back to when Windows was kind of the center of what Microsoft was doing. Any criticism of anything— like one of the big falling outs I had in my life was with this guy Mark Croft, who worked in the Windows team at the time. And I described Windows XP, to give you, you know, 25 years ago, as a sea of blues and greens.

Mikah Sargent [00:25:54]:
Which is not inaccurate.

Paul Thurrott [00:25:56]:
I thought it was a pretty pithy way to say that.

Richard Campbell [00:26:00]:
It was Windows 2000 with the Fisher-Price interface.

Paul Thurrott [00:26:04]:
Yeah, right. So he was not happy with that, and he didn't tell me that. He just didn't talk to me for 10 years.

Mikah Sargent [00:26:12]:
Oh my goodness.

Paul Thurrott [00:26:14]:
So later I ran into him.

Richard Campbell [00:26:16]:
That's a lot

Paul Thurrott [00:26:17]:
of energy. Yeah, I ran into him at a show, probably Ignite. I don't remember. It was a Microsoft show. And I didn't realize that this was like a problem for him, but seeing him, I was reminded of this. And I actually said to him, I said, this is kind of random, but you probably don't even know that I wrote this, but I wrote this kind of thing and I kind of feel bad about it because I know you were heavily involved. And he was like, dude, I can't tell you how much that means to me. This has bothered me.

Paul Thurrott [00:26:42]:
I've— Wow. For a decade. Yeah. So, but since then, and look, I have a lot of these kind of interactions, not me being a jerk and then someone feeling bad, but I have a lot of those interactions too. But as far as people on Windows or people at Microsoft, So, you know, we're talking about 25 years later now, right? So over time, people I've known for a long time at Windows or elsewhere at Microsoft have kind of started to go. Like some people have, a lot of them are retiring, you know?

Richard Campbell [00:27:12]:
Yeah, yeah.

Paul Thurrott [00:27:12]:
Some of them have been laid off. They've had those layoffs, right? Some of them—

Mikah Sargent [00:27:15]:
Oh yeah, the voluntary. Yeah, the voluntary, yeah.

Paul Thurrott [00:27:18]:
Yep, that's all happened, right? And, you know, and there are a few left, right? There are some people who are contemporaries who are still around, who've been around a long time. Thank God, but it's not like it was in the old days, right? And then tied to this, Windows is not even close to being the center of Microsoft anymore. So the kind of B-teamer type that would end up in Windows up through about 10 seconds ago is not someone I cared to know or talk to. And things like when Windows 11 came out, Microsoft made a video, which I mocked on this show, where they had the designers who made the thing. And it's like, these people, they're sitting there, they all use Macs. They've never used Windows very clearly.

Mikah Sargent [00:27:55]:
Right.

Paul Thurrott [00:27:55]:
And they're like, the Start menu is an environment for creativity, or whatever they were saying. And I was like, you guys, use Windows and then talk about it, please. Like, it was disgusting, right? So I don't have any relationship with those people, obviously. But today, you know, it's possible because they are bringing more engineering back and they're working on it. Maybe this will shift again, I don't know. I don't know. I mean, I used to listen— I used to have late night chat, like text messages or phone calls with people like, um, Terry Myerson and Steve Sinofsky, who's an insane person. Um, but then you go through like a period of time where there's no leadership there at all, and then it's got people like Panos Panay, who was a keyboard maker and shouldn't have been there.

Paul Thurrott [00:28:42]:
And I don't know what the heck that was, and I don't have or want a relationship with those people, right? Yeah. So Pavan Davalori, I love. I don't know him personally. I don't ever chat with him or anything like that. yet, but he seems like an engineering-focused, well-intentioned person, smart. I really like him, the current guy, right? I don't know him, but what I see, I like. One thing that was a refrain for several, many, maybe more than several years, there were a lot of people at Microsoft frustrated by what was happening to Windows but couldn't do anything about it.

Richard Campbell [00:29:17]:
Right.

Paul Thurrott [00:29:18]:
And Scott Hanselman's one of those guys. He's been very public about this. He's kind of happy to see that this change is occurring. And he's thrown himself into it. Like, he wouldn't have normally been part of this work, but he has felt for so long that this is important that he is actively—

Richard Campbell [00:29:34]:
Yeah, I wonder if Pepijn didn't say, like, come help us, because he's like 50% Windows now.

Paul Thurrott [00:29:39]:
And we need engineering folks like him. And this is a guy who understands the programming end of stuff. And, you know, we need— it's not enough to have Mark.

Richard Campbell [00:29:45]:
And how people use it.

Paul Thurrott [00:29:47]:
Yeah, he's a great— he's a really nice kind of front end for some of this stuff. You know, they brought— Marcus came back and he's doing a lot of the UX stuff now in Windows. I love him. I've always, you know, he's always been great and that's great too. So I see a lot of positive signs. But yeah, I mean, most of the people I've known and still know and am friends with today, the complaints are parallel what I say, because they've all been frustrated with it as well. Even if you were in some other part of Microsoft, you have to—

Richard Campbell [00:30:18]:
Everybody uses Windows.

Paul Thurrott [00:30:20]:
And they see it too, and they hate it. And why is this thing so crappy? And it has to do with whatever is the strategic focus of Microsoft at the time. Cloud computing was an awesome thing for Microsoft as a whole, but Windows was the one core product at Microsoft that didn't fit into this world at all. Mm-hmm. And what that amounted to was like, how do we do what Microsoft would call software as a service, what I would call subscription service, in the context of Windows without charging you for Windows? Because we know that no one's gonna pay for Windows. And so it's a lot of upsell stuff around Microsoft 365 and OneDrive and whatever else. Clipchamp has a paid tier. Ads in Windows are putting— like forcing you to use Edge so it goes to Microsoft sites and then you get the ads on the backend and they make money that way.

Paul Thurrott [00:31:07]:
Bing integrations, a lot of KPI nonsense in there. And, you know, at some point, you know, it's like anything else. Like, you use a product or service, it's really great, and it's a no-brainer. And that thing could be Windows, it could be Netflix, it could be Spotify, and it's reasonable. And you're like, yep, I like it for the price I'm paying for this thing or whatever it is. Like, it's great.

Richard Campbell [00:31:28]:
Work in pieces.

Paul Thurrott [00:31:29]:
And it's not so great, and it's not so great. And then one day you wake up and you're like, when did this thing become a steaming pile of crap? Like, what is this thing? You know, and so like I said a little earlier, what the good news is, other than all the work Microsoft's doing, which is wonderful, is that if you still have problems with whatever is in Windows, like, and I do, those things, like I said, you can fix or work around, um, pretty much 100%, right? I don't think there's any major area where that's not the case. Um, so whatever bothers you, whether it's privacy and telemetry tracking, or, uh, I just want to sign in with a local account, or whatever it is, I don't want ads everywhere, like whatever your pet peeve thing is, whatever you run into every day and you hate, um, you know, it's all fixable. So

Richard Campbell [00:32:09]:
yeah,

Paul Thurrott [00:32:10]:
that's good.

Richard Campbell [00:32:10]:
And,

Paul Thurrott [00:32:11]:
um,

Richard Campbell [00:32:11]:
which is why I wrote like a little—

Paul Thurrott [00:32:13]:
yeah, yeah. Okay.

Richard Campbell [00:32:16]:
I mean, also layer on the security apocalypse is the other layer of this,

Paul Thurrott [00:32:22]:
right?

Richard Campbell [00:32:22]:
Like the, uh, the update channel has never been more frantic than it is right now.

Paul Thurrott [00:32:27]:
Yeah. So I, I'm not going to be able to give you these kind of dates off the top of my head because that's not how my brain works, but We have discussed this and this will be familiar to you at least, but Microsoft has been actively working to improve the security of Windows for several years now, meaning Windows 11, in very overt ways. This is, you know, David, they pulled David Weston out of the woodworks and he slept on a stage somewhere. It could be a Rust conference, it could be a security conference, it could be Ignite, it doesn't matter. So they have the Secure, well, the Secure Future Initiative, which Microsoft broadly—

Richard Campbell [00:32:57]:
Yeah.

Paul Thurrott [00:32:58]:
And then

Richard Campbell [00:32:58]:
there's a Future Secure Initiative, FSI.

Paul Thurrott [00:33:01]:
Yeah. And then there's a Windows version of that. There's like a Windows Resiliency Initiative, I believe, is the name of one thing. But anyway, but they're doing a lot of deep work in Windows 11. And so even before the pain point stuff in Windows 11, I'm just going to bring up this so I can just see what is there exactly. They've added like a fairly dramatic number of recovery tools and security things to Windows that were not there before. You can fix problems in Windows using troubleshooters. Obviously, that's been there for a while, but you can fix problems through Windows Update, which basically involves downloading the Windows version, the version of Windows you already have, and reinstalling it without blowing away all your apps and settings and data and all this stuff.

Paul Thurrott [00:33:41]:
Obviously, the, the ability to do like reset the computer, etc., but point-in-time restore, which is a modern replacement for restore points or system restore, which we've had for for 25 years. Quick Machine Recovery, which is a brand new tool, which runs in the Windows Recovery Environment. I mean, this is all just bread and butter security resiliency stuff. It's good. Yeah, it's clever. Yeah, and it predates any of this pain point stuff. So yeah, these things have kind of been happening in parallel. You see little signs of light in the darkness, you know.

Paul Thurrott [00:34:13]:
There's security stuff, like you said, because you brought it up, but that's come, you know, not quarterly, but every, you know, 2, 5 times a year, whatever. Um, the Pavan stuff from last year.

Richard Campbell [00:34:22]:
Now every month.

Paul Thurrott [00:34:24]:
Right. But, you know, if you want to, you can make the argument that Pavan kind of played his hand a little bit because we keep referencing this point about a little over a year ago where he tweeted, I'm going to be at Ignite and we're going to talk about agents in the taskbar in Windows 11. And the world lost their mind on this guy. Yeah. Uh, to a degree that was Completely unreasonable and just horrible to this human being. And it's like, guys, like, obviously they have to do this, right?

Mikah Sargent [00:34:53]:
Right.

Paul Thurrott [00:34:53]:
And if you go and, I mean, Apple just, oh, actually it's gonna be in the show later, but they have to, guys, you know? So there's this whole negativity around this. Ignite comes, it's the one session where he appears at and they discuss. This is what the UI is going to look like for agents in Windows. And I disagree with a couple of things, but there is fundamentally, if you're trying to teach someone anything, one of the easiest ways to do it is to compare it to something they understand. Sure. An AI agent is like a background process. You're like, oh, okay, I get that. Okay, that makes sense.

Paul Thurrott [00:35:31]:
I don't agree with this part of it. Agents are like apps, so they'll have a button in the taskbar. And if you need to talk to it, you can right-click the thing, a menu comes up, you get your stuff, you can chat with it, whatever. If it needs to reach you, it will put up a banner notification, which is a familiar user interface in Windows that's existed for, you know, 30-ish years, whatever. Um, you know, it's changed over the years, but it's the same type of notification any other app would throw you in Windows 11. And you're like, okay, like, I get it. You're, you're taking something that's complex that people are maybe a little nervous about or don't quite understand, and you're making it look and work like the thing they get already and use every day, and it makes it familiar. And I'm not saying it's the right UI, I'm just saying it's not a— it's a reasonable approach.

Paul Thurrott [00:36:15]:
Um, is there a better way? Yeah, maybe. I'm not a UX expert, but I— there was all this anger over this, and it's like, guys, A, they are going to do this, wake up, and B, it seems it's not that controversial, you know.

Richard Campbell [00:36:30]:
Yeah, no, but a lot of that anger was just AI lashback too. It really doesn't have anything to do with Windows per se. It was a senior person at Microsoft talking about AI and people are angry.

Paul Thurrott [00:36:42]:
Yeah, yeah, exactly. And you get, yeah, it's that South Park thing where people are like, like, you know, he said AI, you know, it's always like this. So whatever. I look, I, the thing, you know, this is Mozilla did this and This is the case in Windows too, really. So Mozilla, that audience, open source, freedom, privacy, right?

Mikah Sargent [00:37:05]:
Freedom!

Richard Campbell [00:37:06]:
Freedom!

Paul Thurrott [00:37:07]:
And Mozilla's like, guys, we gotta do, we're gonna have to do something.

Mikah Sargent [00:37:10]:
We gotta do the thing.

Paul Thurrott [00:37:11]:
And people are losing their minds, right? So they have an AI kill switch. You can just turn it all off at once. Great. It's your choice, which is correct. But they talk to people that use the product and they talk to the people who hate AI. And it's almost like a Jordan Klepper skit on The Daily Show, if you watch that, where they're like, we hear you, we hear you. We're not going to— don't worry, you don't have to have AI. But when you turn off AI, you're not going to get spell checking, you're not going to get language translation.

Paul Thurrott [00:37:41]:
And they're like, whoa, whoa, whoa, whoa. No, no, I want that. Yeah, but that's AI. No, but that's okay. And you're like, dude, you're so against this and then you didn't even know what it was. Yeah. You know, and so the way they organize— Yeah, they organize the UI so that the things that most people do want and use are right at the top of that list. So you can turn all the AI off and then turn those 2 or 3 things on and it works.

Paul Thurrott [00:38:04]:
It's fine. Like, so that to me, that's a smart, you know, interface. Um, I think the term— we've talked about this before— I think the term AI kind of poisons the pool a little bit.

Richard Campbell [00:38:13]:
Yep.

Paul Thurrott [00:38:13]:
You know,

Mikah Sargent [00:38:14]:
yes,

Paul Thurrott [00:38:15]:
if you go into, uh, the Photos app or, um, The Paint app in Windows, especially if you have a Copilot+ PC, there's a bunch of AI-based features in there like remove background, remove an object from the background, extend the picture out in whatever direction, the AI upscaling stuff, blah, blah, blah. I don't think there's a person on this earth that would complain about that functionality being in those apps. It's wonderful. Even if you don't use it, anyone objectively would look at that and say, yeah, no, this makes sense. It's all AI. Oh no, I hate AI. Yeah, I hear you, I hear you. But I mean, you're stupid about it, but you understand these features.

Paul Thurrott [00:38:53]:
Like, I hate AI, but what about the background removal? Oh, that's awesome. It is awesome. It's awesome. It's AI.

Richard Campbell [00:38:59]:
Well, the thing is, my experience has been anything that's called AI is stuff that doesn't work, because as soon as it does work, it gets a new name like background removal.

Paul Thurrott [00:39:07]:
Thank you. Yeah,

Mikah Sargent [00:39:07]:
exactly.

Paul Thurrott [00:39:10]:
Easily 2 years ago, I was like, if we just talk about this like it's technology, you know, most AI is basically spell checking. Yeah. I don't remember anyone picketing the Microsoft campus because there's a spell checker in Word. Like, what are we talking about here? You know? So yeah, look, there's always going to be good implementations of whatever technology, bad implementations, just frivolous implementations, you know, whatever. We're burning this planet to the ground so we can make pictures of ourselves where we look like squirrels or something. And it's like, okay, maybe that's not the best use of AI, but whatever. It's like when Apple and Adobe—

Richard Campbell [00:39:49]:
I may have turned one of my guest bios into a Shakespearean sonnet yesterday.

Paul Thurrott [00:39:54]:
Yes. Right. Right.

Richard Campbell [00:39:55]:
Just saying.

Paul Thurrott [00:39:56]:
One of the things I wrote in college was, let's see if I could pull this out of my butt, it was the Canterbury Tales but written as if it was a Shakespeare sonnet. Um, today AI could spew that out in 1 second. You know, I spent an entire weekend writing that thing. Um, but you know, when the laser printer came around and you could all of a sudden— and the Mac obviously had proportional fonts and bitmap display— but you could get this on the paper now with the LaserWriter and, um, uh, what became PDF— what's the file— uh, PostScript. And you know, people just started making every Printed page looked like a ransom note. It was like, we're not just going to have different fonts everywhere, we're going to have different fonts in every letter. It's like, what is this thing? Did you kidnap my child? What is this? Oh no, it's just an invitation to a party. Stop doing that.

Mikah Sargent [00:40:42]:
Did you kidnap my child?

Paul Thurrott [00:40:44]:
Yeah. What is this? But I mean, I think that's—

Richard Campbell [00:40:48]:
There was a lot of web pages built like that in 1998 too. Like,

Paul Thurrott [00:40:52]:
geez. Yeah. Or you could—

Richard Campbell [00:40:53]:
He's called him once, he's Marki tag back.

Paul Thurrott [00:40:56]:
Yeah, the blink tag was like this version for web publishing. It's like, you, you— one thing you can't do on a piece of paper is make it blink. Yeah, but that's not a bad thing. Like, I, you know, I don't—

Richard Campbell [00:41:07]:
why are you giving me—

Paul Thurrott [00:41:09]:
I have epilepsy. What are you doing? You know, like, so there's a— anyway, there's a lot of, um, it's maturity, right? You have to get used to it. You have to become comfortable with it. Like, you, you've mentioned this, it has to work. Right. And I think one of the early problems with AI, and by early I mean like 10 seconds ago because it moves so fast, but, you know, people would try, you know, Copilot when it first came out or whatever AI, and they're like, this is garbage. It can't do math. It's not doing what I asked.

Paul Thurrott [00:41:38]:
It's inventing things. You know, we all love these like hallucination terms, whatever.

Richard Campbell [00:41:42]:
Oh yeah.

Paul Thurrott [00:41:43]:
But you know what, man? You know, you know, take a bathroom break, come back. It's got better. It's gotten monumentally better. I'm not saying it's great across the board, but of course someone has posted a sonnet about Windows, which is beautiful. That's funny. So I've gotten off track here because that's what I do. But we, look, we live in a time where Windows is getting better. And it's important to me because of what I do that I understand why.

Paul Thurrott [00:42:19]:
But I guess, as I said before, for most people, the important thing is it's getting better. And it is getting better. Like, objectively, it's better. It's already better. It's a lot better. It's better than it was a year ago, you know?

Richard Campbell [00:42:33]:
Truly.

Paul Thurrott [00:42:33]:
So, and then whatever. So we have these stupid 2 new features in these builds, and they're not important, and they're not whatever, but they're not important to me. I mean, depending on who you are and how you do things, one of the— one of both those, you're like, yeah, no, actually I've always wanted this thing and that's great, that makes my life better. And there's nothing wrong with that, it's fine. I just— it doesn't really change the, uh, the equation for me. But, um—

Mikah Sargent [00:42:58]:
Do you see that as the difference between when you are writing for your site and your books versus this podcast? here is where you get to talk about for you, and there is where you try to be more generalized?

Paul Thurrott [00:43:15]:
So if I review a product, right, like a laptop, I try to be obvious when it's something I personally prefer, and then I try to be more objective, and I try— you try to put yourself in people's, you know, place. I mean, I've been reviewing laptops for probably 25 years now. I mean, I There's a place for commentary. Here would be one, I guess, in my site. I could do that, of course, and do— but when I write a book like the Windows 11 Field Guide, for the most part, I'm not here to entertain you, or I'm not here to editorialize.

Mikah Sargent [00:43:47]:
Right.

Paul Thurrott [00:43:48]:
It's like, here's how it works, or doesn't work.

Richard Campbell [00:43:51]:
Here's how to do it.

Paul Thurrott [00:43:52]:
Here's how you fix it. And it's really just a reference and a how-to guide. And it's not that— this is not the right place for me to Editorialize too much, I guess, if that makes sense. I don't know.

Mikah Sargent [00:44:04]:
Yeah, no, that makes sense.

Richard Campbell [00:44:04]:
You can make this work, but it sucks.

Paul Thurrott [00:44:07]:
Yeah, well, yeah, right. Um, yeah, I, I, I, um, this is tied to the Local AI stuff a little bit, but, uh, because of the way Microsoft develops Windows today, it's not like the old days where we have released a new version or some update and here are the features and everyone has them. Mm-hmm. It's like, we have some new features, you will get them eventually. I have— I've had several laptops arrayed side by side, all on the same build of Windows, all with a different matrix of features, because that's how this stuff rolls out. And that presents a challenge for anyone like me who, A, likes to find out how things work and then document that or communicate that and be able to say explicitly, If this is the problem, do this to fix this. Or if you go here, you will see this, click this to do this, whatever it is. Like, it's very, um, it's, you know, it's very logical.

Paul Thurrott [00:45:03]:
I don't know. Um, but now I have to write things like, well, you know, open the Settings app, you may see this thing, may, but you may not. And if you don't, you know, it gets into a weird area.

Richard Campbell [00:45:15]:
Yeah.

Paul Thurrott [00:45:15]:
And then you get to

Richard Campbell [00:45:15]:
AI and it's non-deterministic software.

Paul Thurrott [00:45:18]:
Right. And AI is the ultimate for that. It's like

Richard Campbell [00:45:21]:
Yeah.

Paul Thurrott [00:45:21]:
You can't— there is no version of do this prompt and it will do this. You know, it just doesn't work that way. Unless it's a simple question with a simple answer, like a one-word thing. Yeah, I just— it's hard for people like me, who I believe is a lot like people in our industry, a lot of people in our industry, meaning some form of on the spectrum ADHD, where it's super important for you mentally, for you, you know, to be able to say, if I do this, this will happen. If that's your life, AI is going to be a hardship for you. Um, it's just— it just is. Like, it's— I'm sorry. Like, um, there's also like a little bit of a Christmas present kind of aspect to AI, like especially if you have it make something for you.

Paul Thurrott [00:46:07]:
And that could be something as complex as an app or just a simple image. You're like, I would like a picture of myself maybe, and— but I want to look like a farmer and I want it to be a watercolor painting And I want there to be an elephant or whatever stupid thing it is you're wasting tokens on. And then it's like, and it comes up and you're like, you know, you're like the kitten, you know, that like in that video, right? Because you don't know what it's going to— it will usually match what you asked for. It may not be what you imagined, right? And then you can screw around with it. Sure. But you could do that same prompt, same AI, different AI, whatever. And you'll get a different thing every time. I mean, they may have similarities, of course, but you're not getting a deterministic outcome.

Paul Thurrott [00:46:49]:
Like, it's just— it's either depending on— some people like surprise parties, some people do not, right? You know, it's that kind of thing. Like, it's— so I don't know. You know, it's— we live in interesting times, I guess.

Richard Campbell [00:47:07]:
We live in interesting times.

Paul Thurrott [00:47:08]:
Yeah, this is maybe the only way to say it. Okay, do I need to— I don't want to—

Mikah Sargent [00:47:14]:
I was going to say, do you want to take an ad break for a moment?

Paul Thurrott [00:47:17]:
Yeah, we can, of course.

Mikah Sargent [00:47:18]:
All right, let's take an ad break and then we'll be back. Thank you, Leo. All right, let's pick up where we left off in the world of Windows. Oh, looks like you were muted.

Paul Thurrott [00:47:33]:
I am muted. I am muted.

Richard Campbell [00:47:37]:
You're not.

Paul Thurrott [00:47:39]:
You are unmuted. There we go. Sorry, someone muted me. Um, so yes, um, so this is interesting. Um, back in July, Microsoft announced a buyout for, I think it was about 7,000 to 7,500 US-based employees. Um, 30% of them took Microsoft up on this offer, which apparently is a good percentage for that kind of thing. And I guess that worked out pretty well. And I should say, and to put that in perspective, of course, 2025 especially, but also part of 2024, a lot of kind of chaotic, random-seeming layoffs across Microsoft.

Paul Thurrott [00:48:20]:
It was really bad. So they clearly finally— took them a while, but they clearly learned from this and They're doing this kind of thing. And the chief, what's her name? Chief person off of, whatever her name is, Amy Coleman.

Richard Campbell [00:48:36]:
The HR person?

Paul Thurrott [00:48:37]:
The HR person. Yeah, they don't call it that, but yes, HR person. Has explicitly talked about this. Like, look, we're trying to be better about this kind of thing. Obviously, this is a sensitive area when you're trying to get rid of employees. But Microsoft announced this week and they announced to employees, I think they basically published the newsletter, the email to employees publicly, that they're going to have a new voluntary exit offer for eligible employees that they're offering to 60,000 employees. 60,000 globally is roughly one quarter of all Microsoft employees. I think the—

Richard Campbell [00:49:12]:
I thought it was 60,000 could qualify.

Paul Thurrott [00:49:15]:
Would qualify, right. No, it's not that they're happening because First of all, there are certain organizations where they might say, we're not doing this for some reason, or they can explain why to upper management or whatever. It's not the entire world. It's eligible employees in participating teams in the US, UK, India, and Canada. So, you know, whatever. It goes all the way up to the vice president level, which is kind of interesting, by the way. Mm-hmm. And so we'll see.

Paul Thurrott [00:49:45]:
So if the number is 60,000, and that's actually over a quarter of all Microsoft employees worldwide, and it's only those 4 countries, I don't actually have the numbers for each of those places. I know in the United States, like over half of the Microsoft employees worldwide are in the United States. But if you add in those other 3 countries, sorry. You know, what is the actual percentage of people where you could get this offer that are getting the offer, it's much higher than 25%. It's probably between 33% and 50%, I would imagine right now. Yeah, if they— if the same 30% figure holds true, um, then we're looking at about 8,700, you know, employees that might leave the company. And 8,700 employees out of 223,000 is not a huge percentage, but—

Richard Campbell [00:50:35]:
No,

Paul Thurrott [00:50:35]:
um,

Richard Campbell [00:50:36]:
they've done 10,000-person layoffs. The only difference here is that it's very self-selecting, so you're losing the people who could probably get another job.

Paul Thurrott [00:50:45]:
Right. So yeah, one of the— first of all, the people who are— I don't mean to say left at Microsoft, there's 223,000 people, but there are people like both of us know who are at Microsoft who have withstood these multiple rounds of layoffs, have seen to what it has done to all the people around them. There was a period of great fear inside Microsoft. because it just was so illogical and random and didn't make a lot of sense. It was handled really poorly. There's a ranking system they still have. You have performance reviews, et cetera. If you're kind of on the line, so to speak, this might be an opportunity for you to be like, I'm going to benefit as much as I can and get out of here this way, do it voluntarily.

Paul Thurrott [00:51:26]:
Yeah. Because the alternative is you turn this down and then the earnings come out today and we're 2 weeks into the next quarter. And you get laid off, you know. So these people have choices to make, um, which is kind of interesting. So we'll see what happens.

Richard Campbell [00:51:40]:
And I, and I, I've talked to a bunch of folks already. First thing we all found out was that it hit the press before it hit the company. So most company— most employees found out about it from the press, which to me speaks to this has got everything to do with the quarterlies. Yeah.

Paul Thurrott [00:51:58]:
Well, yeah. So the timing of the July thing I just referenced was right before they announced the earnings from that quarter and also for that fiscal year. This was timed earlier this week. So it's just before they're about— like I keep saying, they're going to announce earnings today. Look, no one is ever going to say this explicitly at Microsoft, meaning in an announcement for the rest of the world to see, but this is very clearly AI-based, right? And I mean that kind of broadly in that there's this infrastructure spend and they want to save money somewhere. They're doing a crazy job of trying to hide the money they're spending, et cetera, et cetera.

Richard Campbell [00:52:35]:
Mm-hmm.

Paul Thurrott [00:52:36]:
But there's also, as much as they may want to almost deny this, and I kind of resist this myself a little bit, but that in the notion that AI can take jobs away, I mean, Microsoft is kind of a top-heavy company. There's a lot of people there. I mean, it's possible that—

Richard Campbell [00:52:55]:
It's also a great CEO. Microsoft used to be a people company, right? When you're a software company, your people is everything, but they're not a software company anymore. They're a cloud company and their infrastructure is more important than their people.

Paul Thurrott [00:53:08]:
Yeah, that's horrible, but yeah, that sounds about right. So, you know, I Look, no one likes to see people get laid off, right? Obviously. This kind of thing is better than that. If you're going to downsize, I guess this is a better way to do it.

Richard Campbell [00:53:26]:
No, that's a warning. We're going to do layoffs. We don't know if you— but we might not need to if enough people self-select. I think this is also a broader appeal because the other one was very much about voluntary retirement. That's not what this is. So I think there was a group of folks who saw the voluntary retirement, says, I don't qualify for that. I wish I did.

Paul Thurrott [00:53:48]:
That's right.

Richard Campbell [00:53:49]:
I think I'll pick up a few of those.

Paul Thurrott [00:53:51]:
So I don't have—

Richard Campbell [00:53:52]:
everybody I've talked to who's talking about taking this has another job to talk about. They're already talking about.

Paul Thurrott [00:53:58]:
Oh, okay. That's good.

Richard Campbell [00:54:00]:
Yeah, they're just going to take the package so they can go on to the next thing with a little extra money in their pocket.

Paul Thurrott [00:54:06]:
So I can't find this and I don't have it in my notes, but A couple of months ago, Xbox came out. They had a bunch of layoffs. There were layoffs across the company. And at the time, the way it was described was, we have plans over the course of this fiscal year to lay off whatever the number was. And the numbers they announced on that day were like half. So there are layoffs coming, right? I mean, this is not a prediction. I mean, just based on what they communicated, uh, previously. So whenever, whenever this kind of concludes, this, um, buyout program essentially, right, um, for a much broader swath of the company and some people outside of the United States, um, like Rich has been saying, you know, there will be some round of layoffs and there will be the stuff we kind of knew about, so to speak, because Xbox talked about it a lot.

Paul Thurrott [00:54:59]:
And I think there might be some I think some of it will be the kind of end result of this. The people that maybe they had earmarked but didn't take the buyout offer. I don't mean— I'm sorry, why am I laughing? That's terrible. But might be part of that as well. Yeah.

Richard Campbell [00:55:17]:
Yeah. And I don't know that they earmarked per se. I do know that their layoff mechanism does not involve the direct report. Like, I have plenty of friends who are in that role where they are told, Here's the person you're laying off and here's what you're allowed to say to them. Go do it.

Paul Thurrott [00:55:31]:
Lovely. Okay.

Richard Campbell [00:55:32]:
Yeah.

Paul Thurrott [00:55:32]:
It could also— I mean, if I was in some kind of management position and was forced to deal with this, I would— I mean, you'd want to put a human touch at the end of this, but I mean, you kind of look at numbers, right? I mean, we have whatever metrics we have for measuring performance of employees, et cetera. You kind of look at the bottom, whatever percentage, because maybe that's what you were directed to do. Then you actually look at the people and say, well, hold on a second. This person is doing things that are kind of outside of our formal reporting structure, measurement structure, whatever it is.

Richard Campbell [00:56:04]:
Yep. Or they contain unique— they're a unique resource of some kind. It's like, yeah, or they've already had a turnaround. There was definitely a wave of layoffs for people who had poor results the previous year, got laid off even though they were now working well.

Paul Thurrott [00:56:20]:
Yes, that's right. Because in that case, they actually changed the measurement period, which is something they had not done before. Usually those people were rewarded because they had turned it around. That's right. In the past year, a lot of those guys were among those that got laid off, which is not great. Yep. I mean, I worked at a big company for 15 years-ish and witnessed a lot of this corporate nonsense. Had a conversation with like the senior vice president who was 3 or 4 levels above me.

Paul Thurrott [00:56:50]:
I don't remember, but I called him directly to complain that they'd laid off a guy. And, um, who was our—

Richard Campbell [00:56:57]:
That was irreplaceable.

Paul Thurrott [00:56:58]:
Yeah, he was our connection to Microsoft for— we used to do the Best of TechEd Awards and they laid him off. And I was like, dude, this is the only— this is one of 2 people I actually deal with on a regular basis and he's wonderful. And all I did was sort by probably by pay and age, do that little matrix and got rid of them, right? They can't say that, that's illegal. But the guy responded to me, oh, he's like, Paul, I had no idea. And I'm like, I know, I know. So maybe do it a little smarter next time, idiot. You know, like, like, I like—

Richard Campbell [00:57:29]:
Well, that's part of the mechanism. Normally you go to their direct report and say, hey, this is what we're looking at doing.

Paul Thurrott [00:57:35]:
Yep.

Richard Campbell [00:57:36]:
Tell us how you might do it differently, still get to the same goal.

Paul Thurrott [00:57:40]:
This is the problem with any big company is 'cause you have all these layers of management and you have people who have goals like you do, but their goals are very different from your goals. You know what I mean? And at some point as you go up the chain, they're completely disconnected from the actual human beings on the other end. So from my perspective in the editorial part of that business, and I knew all the team and I loved everybody and it was great, it was always a shock when they came down from on high. Whereas this guy's looking at a spreadsheet and from him, From his perspective, this made total sense. And you have to be a little smarter than that. So we'll see. You know, this is, look, this is a little bit like Windows 11. Kind of doesn't matter why, but it is better than it was like a year ago, right? When they were trying to reduce employee count a year ago, they would just like put people against a wall basically.

Paul Thurrott [00:58:28]:
Like, you know, it was terrible.

Richard Campbell [00:58:29]:
But the joke was in the 2025 fiscal, they laid off 30,000 people in what, 3 rounds? and hired just as many. Like, the employment numbers for 2025 did not change.

Paul Thurrott [00:58:40]:
That's right. Yeah, that's a curious thing.

Richard Campbell [00:58:42]:
That's when I said, I think this is employee terrorism. I think they're just trying to keep the workers afraid.

Paul Thurrott [00:58:48]:
So if you— so yeah, there's that aspect of it. And then you have to— this would be impossible because you can't really have all the numbers here, but I mentioned, you know, it was very clear at the company I worked at the time, Penton, which is basically the tool of the Antichrist, that they were just looking at age and salary. Literally, that's how they got that. And you could see it in the order they got rid of people. It was horrible. And that's unfair and it's terrible or whatever. But imagine— this didn't happen to my knowledge— but imagine if the overall Penton employee count didn't change over that year. And you're like, well, what happened? Well, they hired other people.

Paul Thurrott [00:59:20]:
A lot of those people they would have hired would have been fresh out of college. They would have made next to nothing. They didn't have all the, you know, the healthcare stuff had changed completely or whatever, the retirement stuff, whatever it was. And, um, you could save money that way too, right? You know, um, hire a bunch of, uh, college interns and get rid of the, uh, you know, the 30-year veterans. Um, not a good idea, you know, but you could do it.

Richard Campbell [00:59:44]:
Well, and as you recall, that entire division you used to participate in does not exist anymore.

Paul Thurrott [00:59:50]:
Yeah, right. I know.

Richard Campbell [00:59:52]:
So it went— it, it went well.

Paul Thurrott [00:59:55]:
Oh, it went great. I mean, sometimes you need a push. So in my case, it went fine. You know, I would have stayed there for the rest of my life if they didn't completely screw up the whole thing. But, you know, whatever. Yeah. All right. So I don't remember.

Paul Thurrott [01:00:11]:
I think I said a couple of things last week about RTX Spark and Local AI.

Richard Campbell [01:00:16]:
You were having a great time. They were the best machines you'd ever used. I remember.

Mikah Sargent [01:00:19]:
Yeah, I remember that too. I was like, I don't know why everyone hasn't bought something from this hardware line. It's wonderful.

Richard Campbell [01:00:26]:
$5,000, you've— it's just all love and kisses, right?

Paul Thurrott [01:00:29]:
I don't remember it, but maybe I blacked out. Um, so look, I, I spend time on this every single day, and I, I've had multiple days since then where I, I— it was the morning or the afternoon depending on how things were going. I'm like, all right, I'm gonna solve this. I'm gonna get this, you know. And I think I mentioned this, but I have this list of several to 10 goals, you know, I'm gonna— because it's not just programming. Programming is just one of the many things, you know. I want to look at the creative stuff. I want to look at productivity and agents and all the other stuff that goes into this.

Paul Thurrott [01:00:58]:
And, and, um, and it's hard with the way my brain works to run into this, you know, brick wall in step 2 of 10, and you completely lose focus of everything else, not just in the other 8 steps, but in everything else in your life. Like, if it's I'm working on books. I'm working on— I have daily— I still work every day. I have to write. I have to do whatever I'm doing. And this whole thing has just spiraled like out of control. And I'm still struggling with it. I've talked to people that understand this stuff better than I do.

Paul Thurrott [01:01:30]:
I've talked to some friends who are heavily involved in this. Like, Brad, I have failed more at this than I failed at anything I've ever done. in my entire life. Just this morning, I finally decided to do the thing I wasn't going to do in the beginning, which was just kind of go back to the command line and just kind of muscle through this.

Richard Campbell [01:01:52]:
Work your way through it.

Paul Thurrott [01:01:53]:
Yeah, and I mentioned this to Richard, I think before the show started recording, but I spent the morning with 2 laptops. One of them is just a Copilot+ PC. The other one is the highest-end RTX Spark, doing the exact same thing on both. Which is you install the Ollama CLI, you install the cloud, the Therapeutic Cloud CLI, you download a model. Now this is where they diverge because the 128-gig RTX Spark thing can download this freaking giant enormous thing, whatever it is. And then the other one, just something smaller basically, you know. And this is something, you know, again, this is a little bit of a peek toward like kind of something I was thinking 2 and a half weeks ago. Like I wanna get like, how does this actually differ from a normal computer? computer.

Paul Thurrott [01:02:37]:
One of the ways it doesn't differ is they both failed exactly the same way, which was awesome. So that was great. And I spent— most of the time was spent just downloading the models because they're humongous. And I don't know why they— honestly, despite the size, I don't know why they take so long. But it's kind of a hurry up and wait situation where—

Richard Campbell [01:02:57]:
So nothing happens quickly.

Paul Thurrott [01:02:58]:
Yeah. And I make the mistake of, well, I'm not sitting there watching the download progress. I do other things, right? So This morning what I was doing is writing the article that was going to become the description of what I was doing, of course, but then it fails at the end and I'm like, okay, that kind of changes the tenor of this article. Good thing you wrote an article. Yeah. So we'll see where that goes. I'm going to work on that, but I'm not done with that. But that was what happened this morning.

Paul Thurrott [01:03:23]:
And last night we were watching TV. It doesn't really matter, but I decided to go back to the one thing that did work. And the one thing that did work— and I used work with an asterisk, but it worked in the sense that it just didn't throw up in my face— which was with Hermes Agent and whatever that gigantic— I think it's the Quen Coding Next or whatever it is, the gigantic model.

Richard Campbell [01:03:47]:
Yeah.

Paul Thurrott [01:03:47]:
Um, I left this thing running. It didn't— I, I set it up in such a way that it didn't bother me ever. I was like, you have complete control of this computer, do whatever you got to do. I don't— do not want you to Ask me any questions. I don't want you to tell me to—

Richard Campbell [01:04:02]:
Dangerously ignore permissions.

Paul Thurrott [01:04:04]:
Yep.

Richard Campbell [01:04:04]:
Do it all.

Paul Thurrott [01:04:05]:
Well, because at some point you just like, I just want it to work and then maybe I'll get it going the right way. I don't know. But I asked it, this is what the thing I'd run into yesterday, the 2 days, whatever the past couple of days was, tool use. And the thing I've learned, because I'll go to a model and say, can you just write this for me and actually run the code on the backend against whatever Visual Studio runtimes, .NET runtimes, whatever it is, and not have me copy and paste code because then it has 29 errors and I have to describe each one of these. Stupid. I want you to do the thing. They all did that thing I was joking about last week. They were like, oh, Paul asked us if we wanted to— well, I think I— let me take a look.

Paul Thurrott [01:04:46]:
They talk and talk and talk to themselves and then they're like, no, I can't do that. What I realized after some period of time, because I think it was the third one, that said something that suggested it might be the app. And that was like, I have to say, I'm a little embarrassed by this. I didn't really consider that part of it. I assumed it was the model that was incapable. But as it turns out, if you think about these apps, so these are things like LM Studio, Hermes Agent, Llama Windows or Windows Llama, whatever it's called, OpenCLAW, you know, whatever these things are, there's a lot of them. I sort of assumed like tied capabilities solely to the model in some ways. And I realize now that was a big mistake because one of them, it didn't say it explicitly, but it was the commentary it had was something to the tune of, this could work if you use the right, they don't call it, we don't really have a good term for this, AI app, an AI front end or whatever.

Paul Thurrott [01:05:38]:
And I was like, interesting. So then I went back to my notes and I was like, all right, so I got this, I got it to sort of work. once. What was that thing? And it was Hermes Agent, and I'd broken down all the barriers, you know. And so I did this with it last night, and I asked it to make Notepad, and it churned literally for 1 hour and 55 minutes. It heated up the room 20 degrees. It sounded like a jet engine was going to town in there.

Richard Campbell [01:06:02]:
Fans were whizzing out the side.

Paul Thurrott [01:06:04]:
Yep.

Richard Campbell [01:06:05]:
Plastic was melting into the thing.

Paul Thurrott [01:06:07]:
Yeah, the thing at the table it was sitting on melted right down to the carpet. I said to my wife, actually, I set it up before we went to dinner. That's what it was. And I jokingly said to her, I said, the good thing this building is made out of concrete, otherwise we'd come back to the thing burned down to the ground. And then it'd be like, the good news is I made—

Richard Campbell [01:06:22]:
Your apartment is going to be burned out.

Paul Thurrott [01:06:24]:
Yeah. Yeah. That's like, but I made the app. So anyway, after it was done, it was almost 2 hours and it did it. It made an .exe and I ran it. And it looks like— it looked like Notepad looked in like 1997. You know, there's no tabs, no anything. It's not a modern UI, nothing.

Paul Thurrott [01:06:43]:
A lot of the features didn't work, things like keyboard shortcuts, Ctrl+O for open, etc. Um, it was written in WPF, which is great. Um, it didn't get the, the line column count thing correct, which you can do in WPF, is pretty much impossible in the Windows App SDK. Um, anyway, but I looked at this, I was like, okay, So I've been working on this. Yeah, exactly. But this is like 2 and a half weeks into this. This is not the first time I got something to work, but it's one of the few times. It's the same thing I had already gotten to work, one of the things I had gotten to work.

Paul Thurrott [01:07:19]:
And it wrote a version of the app that I wrote myself by hand 8 or 10 years ago, uh, and have since well surpassed, you know. And it's like, okay. Interesting. And then like you said, yes, iterate. Absolutely. That's going to be a story. And I'm really curious about that. I also, it occurs to me with Local AI, it might be a good idea to start smaller.

Paul Thurrott [01:07:45]:
Meaning instead of like, don't duplicate Notepad, but make a very simple app that does a couple of simple things with a text editor. And then I'll just start adding features from there and maybe that will go better, right? That's an approach. But Okay, whatever. But it did it. And I was like, all right, well, it's not perfect. It's not even close, but it runs and it sort of, okay. So there's some—

Richard Campbell [01:08:05]:
Well, it's what you were after was the end-to-end, can this thing make code according to my instructions? Anything.

Paul Thurrott [01:08:14]:
Right, right, right.

Richard Campbell [01:08:15]:
That even partially operates. You got that.

Paul Thurrott [01:08:17]:
Yeah. There is some version of this where, depending on the tool you're using, where you could very specifically specify every function. I could do something like, here's a picture of every one of the menus. You know, all of these things have to be here. Here's what each one does. You could do that, right? I mean, maybe that's a bit— that's almost certainly a better approach, especially right now.

Richard Campbell [01:08:39]:
Detailed specifications. Well, and one of the trades here we're finding more and more is like the more detailed the specification, the smaller model I can use.

Paul Thurrott [01:08:47]:
Yes, right. That's true too, right? And part of the reason I wanted to do the 2 computers side by side Which again, I intended to do probably 10 days ago or more, but, um, was this— the Snapdragon one is running a smaller model, and I wanted to kind of compare the quality, if that makes sense, the speed, you know, whatever the things are. Yeah, you know, what does this look like? And, um, because, you know, how little can you get away with? I mean, um, I—

Richard Campbell [01:09:11]:
yeah,

Paul Thurrott [01:09:12]:
it's possible.

Richard Campbell [01:09:12]:
You have to have these monster machines with 128 gigs, or—

Paul Thurrott [01:09:16]:
Right.

Richard Campbell [01:09:16]:
Something smaller work?

Paul Thurrott [01:09:17]:
There's something to be said for you probably don't. And I'm not saying 32 is right, but possibly 32 to 64 might be the right place, so to speak. But that is predicated on models improving. And, you know, we'll see how it goes. 'Cause one of the problems with having too much RAM is you could load some stupid big model in there with a gigantic context window, and then it just happens at glacial pace, right?

Richard Campbell [01:09:39]:
Well, it also just overthinks. Like, I've had this, I've been using the OpenZen add-in for Home Assistant. So I don't have to set up a separate agent and so forth. It's all self-contained within HA. And then you can switch out what models you want to use. And it's like, dude, do not take a high-end model and put it into this problem space.

Paul Thurrott [01:09:59]:
Do the smallest model you need that will actually accomplish what it is. Yeah. Sure. That's right. So that speaks to the experience I'm having, which is that when it comes to this kind of thing, it's, first of all, it's brand new. Like, we just, not a lot of, It's still improving and there just isn't— it's just not well established enough to be like, do this, if you do this, this will happen. It's that thing I was talking about up top. And I feel like it does get there.

Paul Thurrott [01:10:26]:
Mm-hmm. Whether I live long enough to see that happen is unclear, but that ties into my next thing, which is that if you do the math, and this is a fun little project for everyone to do, 128 gig machine, it's about $7,000. If you gave Anthropic or any of the companies $20 a month, how much— how many months would you have to do that before it added up to $7,000? And it's 350 months, which, if you do the math on that, is over 29 years.

Richard Campbell [01:10:59]:
Yeah.

Paul Thurrott [01:10:59]:
So I'm gonna turn 60 soon. 89? I don't think so. It would be way smarter for me to just pay them $20 a month if I could somehow lock in that rate.

Richard Campbell [01:11:10]:
Yeah, you really want the $200 all-you-can-eat max account.

Paul Thurrott [01:11:14]:
Okay. Yeah. I mean, so that's still 10 years. Still years. It's almost 3 years. But I bet I could get away with the $20. And this is the interesting thing in the sense that big tech is mostly abusers and terrible. And we know that this $20 a month or whatever the cost is, is not covering the cost to them.

Paul Thurrott [01:11:38]:
This is a golden moment where you can actually take advantage of something before it becomes insertified, um, with the knowledge that 100% it will become that thing. There's no doubt about it.

Richard Campbell [01:11:47]:
Sure.

Paul Thurrott [01:11:48]:
Um,

Richard Campbell [01:11:48]:
well, how many of us in the self— these various cell phone eras got an all-you-could-eat data package and held on to it for absolutely as long as you could because all the packages that came after it were worse?

Paul Thurrott [01:12:00]:
So I, um, when the iPhone came out in 2007, I switched from Verizon to AT&T because you had to, right? That was the only, uh, carrier at the time. They, they had a, an actual unlimited, not with an asterisk. It was actually unlimited. And that was fine for— I don't remember the number of years anymore, but 3 to 5 years somewhere. But at some point, AT&T was like, we're getting destroyed by this. Um, yeah, turns out people with iPhones are on the internet all day long. It's not the model. Remember, you used to pay for minutes, used to pay for Texts, that was their model and this was not working out.

Richard Campbell [01:12:33]:
It's not a thing anymore.

Paul Thurrott [01:12:34]:
Yeah. They switched to a different plan. And if you wanted to get whatever the iPhone was, or any phone that year, you had to switch off of that plan. And I held out as long as I could, but I'll never forget this. I went into the AT&T store in Dedham, Massachusetts and explained, I'm going to buy this new phone and I know I have to switch plans. And he goes, wait. He goes, you still have one of those Unlimited iPhone plans from 2007? I'm like, yeah. He goes, I feel like I'm killing a unicorn.

Paul Thurrott [01:13:04]:
I was like, yep, I understand. I'm not happy about it either, but what are you going

Richard Campbell [01:13:10]:
to do? I've got some sparkles here to sprinkle as we walk.

Paul Thurrott [01:13:14]:
Yeah, exactly. I guess I was one of the only ones left at the time. I don't know.

Richard Campbell [01:13:17]:
Yeah.

Paul Thurrott [01:13:18]:
How do I keep this for you?

Richard Campbell [01:13:20]:
How can I keep this going somehow?

Paul Thurrott [01:13:23]:
Can

Richard Campbell [01:13:23]:
we do it as a warranty replacement? Like, what can we do?

Paul Thurrott [01:13:26]:
No, I have to upgrade the phone. So, you know, whatever. But

Richard Campbell [01:13:28]:
yeah,

Paul Thurrott [01:13:29]:
but yeah, when it comes to AI, I mean, the one thing that really hangs over me is that, and because I use AI a lot less than a lot of people, I think, but I'll use it for things like generating images and stuff. I used it to create, take a photo I had taken and turn it into like a kind of a, like a vintage travel poster style thing. It was, it's beautiful. And I will use some version of that for the book my wife and I have, some version that's coming later. But it's wonderful. Like, that's just great. But I don't use it day to day for work or anything. I mean, if I was a professional coder, I would absolutely use it.

Paul Thurrott [01:14:08]:
Yeah,

Richard Campbell [01:14:10]:
but then you'd be setting up Ralph loops and a bunch of other things too, right? I understand. What's interesting about these RTX Spark laptops is that most of the efficient methodologies, that processor should be headless. It should just be fighting with itself in that adversarial model. And your machine is loose to do other things.

Paul Thurrott [01:14:33]:
I just said this to my wife. I said, when this is mature, we'll have a box in a closet, you won't even see it. And what I did at the time, we have a little set-top box from the TV service that we never use, but it's just sitting there under the TV. And I'm like, it's going to lay that box. You don't even know it's there. You don't think about it. But the— and one of the many reasons you want that is because it's going to be on all the time. It has to be.

Paul Thurrott [01:14:55]:
We talked about this last week, I think, you know, for AI agents to do their thing, their background processes, they have to run. They— if the laptop lid closes, they don't run. Um, I don't know.

Richard Campbell [01:15:06]:
And they're running their fans. Like, I feel bad about closing it when it's working that hard. now, I know it's got to cool itself, right?

Paul Thurrott [01:15:13]:
Yeah, the only— it— like, if anyone here— well, you do, right? So if you play games on a PC and it— that, uh, especially if it's a laptop, if it has a laptop GPU like an NVIDIA RTX now, whatever they have today, I don't remember.

Richard Campbell [01:15:25]:
4060 in this thing.

Paul Thurrott [01:15:26]:
Um, yeah, you know that that thing cranks up. It's like a, you know, it's a turbine.

Richard Campbell [01:15:32]:
But when you're done playing— Plus degrees Celsius.

Paul Thurrott [01:15:35]:
Yeah, you can— you— I have to raise the volume just to hear the damn thing. Like, it's unbelievable, like, how loud that fan can get. But Whatever, it's cranking out this game, looks awesome. Um, but when you're— when I'm done playing, I don't just close the lid. Like, I actually let it sit there for a few minutes. Like, I go down to the desktop, obviously, and then I leave it alone. I, I don't want to do anything that will harm the thermals. I want it to cool down, you know?

Richard Campbell [01:16:00]:
Yeah.

Paul Thurrott [01:16:01]:
Um, and that's what these computers are like. Um, and I, I talked about this last week, you know, my My wife came in the room one time, we were going to watch TV, and she's like, is there a jet engine landing in this room? Like, what is going on in here? And I'm like, oh, it's failing at AI. What do you mean? It's like nothing's going on. It's fine. Yeah. Anyway. Okay.

Richard Campbell [01:16:20]:
Well, and then the next part is because it is going to crank the fans nonstop, like, you don't even want it in the room. I would've put it under the stairs close to the concrete to try and keep it cool.

Paul Thurrott [01:16:30]:
I came to Mexico. One of the laptops was here is that awesome ASUS Pro A16, A16 Pro, I think it's called. It's a 16-inch screen. It's lighter than— it feels like an engineering sample. You almost like go flying with it. It's so light. And then I have these RTX Spark things, which are like bricks made of lead and same size screen, you know? So this thing's sitting there completely silent. This one's over there the whole time, you know?

Richard Campbell [01:17:00]:
Yeah.

Paul Thurrott [01:17:01]:
And the battery life on this thing lasts like 12 hours. The battery life on that thing lasts like 5 hours if you're lucky, unless you're doing anything on it. And it's a different experience, you know? And look—

Richard Campbell [01:17:12]:
But I also get a sense that Microsoft was desperate to not be left out of the desktop AI experience.

Paul Thurrott [01:17:19]:
This is, yeah, but this is a cart horse problem because Windows is not ready in my opinion. And—

Richard Campbell [01:17:25]:
No.

Paul Thurrott [01:17:27]:
I actually don't know that local AI is ready for yet. It will be. I mean, we all understand this, but I just feel like the cloud is where it goes.

Richard Campbell [01:17:34]:
And then that is the right form factor. And it should be something else that your machines speak to that do certain workloads.

Paul Thurrott [01:17:41]:
One thing that is present and becoming more common in this world is this notion of like, I think Mikah was talking about this last week from his phone. You have the computer use thing. And so you can do things that are occurring on your, computer, but you're doing it from the front. You think of something, you're like, I wanna have it do this, whatever. And that thing's running all the time back home or wherever. In the NVIDIA world with these computers, there's a, I can't remember the name, it was a software/service thing that it will look on your local network and see what other NVIDIA GPUs there are out there. And it will actually chain the work across those GPUs. So you could have multiple computers, you could even do this at home, right? Yeah.

Paul Thurrott [01:18:19]:
And okay, that's kind of cool, but this is the thing you were talking about, Richard, which is like, I could be on a computer that has no valid local AI capabilities. It could be a Chromebook, it could be anything. But as long as I can connect to the thing that's occurring on the computer, I can give it the instructions from there. And so I have this thing that's thin and light and I can carry around, which is nice. But back home or wherever I am, there's this thing that's on a shelf and churning through tokens and electricity or whatever it's doing. But, um, that's kind of like where it should be, right? This is the, the old home server thing or the NAS thing or—

Richard Campbell [01:18:54]:
Sure.

Paul Thurrott [01:18:55]:
Uh, whatever. And, you know, unfortunately the NASes we have today are objectively not designed for this. They all have like the crappiest lowest-end Intel and AMD, or maybe some of them are ARM now too. But, you know, we almost need like a bank of what is essentially headless GPU stations, you know, or whatever. I guess maybe that's— maybe that is what some of this stuff is. I don't know. Um, it's a weird— it's a weird infrastructure to think about as a home user. But, um, but carrying around this thing, my God.

Paul Thurrott [01:19:26]:
I did travel to Mexico with these things, and, um, we're lucky the plane didn't go flying out of the sky, I guess. But, um, none of them would have lasted for the duration of the flight, I can tell you that.

Richard Campbell [01:19:38]:
Yeah, well, these are not built for battery life.

Paul Thurrott [01:19:42]:
No, I mean, the battery life's okay. It's not, you know, a joke, but it's, you know, like you wouldn't play a game— like you wouldn't get in Call of Duty multiplayer on a laptop and not be plugged into the wall, you know. Ditto for this stuff. Um, yeah, just wouldn't do it. You would do it. And I

Richard Campbell [01:19:58]:
know

Paul Thurrott [01:19:59]:
half-heartedly looked at—

Richard Campbell [01:20:02]:
sorry. I could see— I have friends I'm not necessarily one of them. I might, but I definitely have friends that would carry a laptop and the NUC.

Paul Thurrott [01:20:12]:
And a NUC. Interesting. Okay.

Richard Campbell [01:20:14]:
Yes. The NUC with the great— with the GB20 in it, right? With the RTX Spark is. They want an RTX Spark NUC.

Paul Thurrott [01:20:23]:
When you say carry, what do you mean? Carry it where? Like when they travel?

Richard Campbell [01:20:26]:
In their laptop bag. Yeah.

Paul Thurrott [01:20:28]:
Okay.

Richard Campbell [01:20:28]:
Because they want local compute, right?

Paul Thurrott [01:20:31]:
Yeah.

Richard Campbell [01:20:32]:
And if they, you know, or more importantly, they're conference speakers, they want to demo that on stage.

Paul Thurrott [01:20:38]:
Yeah, but I mean, other than latency and that and internet problems, I mean, if it was just at home at an office, wouldn't that work fine? I mean,

Richard Campbell [01:20:45]:
no, just fine. And you'd VPN into it.

Paul Thurrott [01:20:47]:
Yeah, I mean, the nice thing about it being a laptop is you can carry it. It's not necessarily the most, uh, portable thing in the world, but you and I and a lot of other people have traveled with enormous laptops before. I mean, this is It's not an uncommon thing.

Richard Campbell [01:20:59]:
I can't even tell you how many times I've abraded my shoulders because I'm carrying so much laptop in my pack.

Paul Thurrott [01:21:04]:
I still have a, I have a, like a crick or whatever you call it in the back of my, it's like my below my shoulder that could be related to a lifetime of laptop carrying. I don't know.

Richard Campbell [01:21:12]:
Yeah, it could, could be. Yeah, it could be that P5 era we were once.

Paul Thurrott [01:21:16]:
Yeah, exactly.

Richard Campbell [01:21:17]:
With laptops with power supplies you could kill a small animal with.

Paul Thurrott [01:21:21]:
A million, million years ago, I was on a plane. I would've been living in Phoenix. It was a Dell laptop. um, 32 megabytes of RAM. I remember I had specified this so I could run NT on it. NT 4.

Richard Campbell [01:21:35]:
Laptop.

Paul Thurrott [01:21:36]:
Yep. And, uh, Quake had come out, so it would have been '96, '97, somewhere in there. And, um, so 20, 30, 30 years ago. Jesus. Uh, and anyway, but Quake had come out and Quake ran on Windows. Like, there was like a WinQuake, and then, you know, they did very various permutations. But because it ran on Windows, like, it actually ran in NT. So I was— Quake at that time was basically just an offline— I mean, it was an online thing, but it was the, you know, there's a game with whatever the levels are and stuff, and I was playing it on a laptop.

Paul Thurrott [01:22:04]:
And that laptop, that crappy little laptop, and this, this woman behind me like stuck her head between the seats and she goes, is that VR? I was like, no.

Richard Campbell [01:22:16]:
In the '90s.

Paul Thurrott [01:22:17]:
Yeah, but yeah, I mean, like, no, not technically, but She had never seen anything like it. Yeah, it looked like, like terrible, but whatever, it worked, I guess. I don't know. Okay, I guess we can move on. I won't turn this into a 2-hour debacle.

Richard Campbell [01:22:32]:
Are you sure this is not a PepCak problem?

Paul Thurrott [01:22:36]:
I'm not sure.

Richard Campbell [01:22:36]:
I think you're mostly— I think, yeah, I think you're mostly upset because you, you have this sense that maybe it's me. It's not that these—

Paul Thurrott [01:22:43]:
Oh, definitely. I, I, um, I don't remember if I said this, but there have been 2 major moments in my professional life where I just came up empty. Like, I just didn't understand it, right? And I've definitely at least mentioned this here and there, but one was when they announced what was then called Windows Azure. And I'll never forget this because I sat in that audience, tens of thousands of people, and I'm like, I don't know what's happening. I don't understand this.

Richard Campbell [01:23:09]:
I don't get it.

Paul Thurrott [01:23:10]:
When we left the room, there were PR people out there and they were, They were pulling people— they pulled me over and they said, hey, if you have any questions or need to know more, we're going to have the guys come down and do a PR session for the press, whatever. We're in— it was a room of maybe 20 to 30 people, and they talked to us for about an hour, maybe longer, I don't remember. And I walked out of there and the same woman who had gotten me to go in there saw me coming out and she goes, oh, was that any good? Was, you know, did— yeah. And I said, yeah, I have no idea what's happening. I, I— Hmm. It didn't matter how much they explained it to me. I just didn't understand it. The other one, which I definitely told the story of because it was hilarious, was the quantum computing Ignite.

Paul Thurrott [01:23:48]:
Remember this one? Yeah. They booked this. They booked a meeting with me, and I met all those guys who were on stage. We sat in a room for 30 minutes at least, maybe 45 minutes. I don't remember. They tried to dumb it down as much as you can. It's like Stephen Hawking writing a book. It's like, it doesn't help that you're dumbing it down.

Paul Thurrott [01:24:04]:
It's still a million years ahead of me. I don't understand this thing. Quantum computing was like nothing. So I sat down in the press room and I opened up Word probably at the time and wrote the headline or title was Everything You Need to Know About Quantum Computing. And then the first sentence said, is the article I wish I could write, but I don't understand this topic. And it was just so far. And I still to this day, like I get Azure now, no problem, but I have never caught up on quantum computing. Like I just don't understand it.

Paul Thurrott [01:24:33]:
It's beyond me. I'm just not that smart.

Richard Campbell [01:24:36]:
Azure in 2009 with Web Roll, App Roll, like that was, pretty obtuse stuff. He was essentially describing serverless computing a decade before anybody wanted it.

Paul Thurrott [01:24:46]:
Right. It just didn't— the world that had evolved to that point had made sense to me. You have individual computers, now you have workgroups. Now we need a central server and we need to have something that became Active Directory and we have identity management and we have permissions and we have whatever it is, right? This all made sense to me. And then, you know, by the time the cloud made sense, I got it. But when they announced Azure, it was several years ahead of that. And I was like, so you have a different Windows Server for some reason? I'm like, what? I don't understand. Like, I just didn't understand it, you know? Today I do get it.

Paul Thurrott [01:25:24]:
But so this stuff, the weird thing is I feel like I do get AI, but the local AI a bit, you're like, well, will translate, this will be easy, you know. And it's like, I get that I have to do a few more things. Like, I, I'm in charge of figuring out the client, I'm in charge of figuring the runtime, I'm in charge of even picking the models, I guess.

Richard Campbell [01:25:42]:
Yeah.

Paul Thurrott [01:25:42]:
Um, I'm unqualified for that, but I— okay, fine, I can pick out a car and have no idea how to drive a car, it's no problem. But man, it's not worked out. It would be like, you know, you just kill the metaphor, like you drive out of the lot and then you just drive into a wall and they're like, what are you doing? Like, I turned the wheel, it didn't turn. I don't know, I thought it worked the way I thought it was going to work. I— it just, you know, that's been my experience. And luckily, it's just like punching my face— self in the face. Um, okay, sorry, I, I promised I wasn't going to beat that to death, but I can't help it. So, um, okay, do you want me to continue? We have a couple of software things, or do you want to take a break?

Mikah Sargent [01:26:18]:
Let's continue, and right after these software things, we'll take another break.

Paul Thurrott [01:26:22]:
Okay. These are mostly quick anyway.

Richard Campbell [01:26:25]:
Yep.

Paul Thurrott [01:26:25]:
So I assume this audience is well-versed in Microsoft Authenticator. I actually prefer other authenticator apps. Single-factor, yeah. Yeah. But the one thing that the Microsoft Authenticator app is uniquely good at is when you have a Microsoft account, meaning not just an MSA for consumers, but a Microsoft work or school account. It's just that experience is just, it's perfect. So even if what I do, and I would tell other people to do the same thing, is like, If you use a different authenticator app or prefer a different authenticator app, that's fine, use it, but use this for your Microsoft accounts. It's like, it's just awesome.

Paul Thurrott [01:27:00]:
Unfortunately, the Android version of that app— I guess this didn't impact the iPhone— had a major security flaw. And some security researchers from a company called Enclave discovered this and did the right thing. They contacted Microsoft, showed them the flaw. Microsoft confirmed it. Issued a fix. It actually went out in September, by the way. This thing's been available for over a month and a half, so you have it. If you have—

Richard Campbell [01:27:23]:
I have it.

Paul Thurrott [01:27:24]:
Authenticator on Android, you're good. It's already happened. And then they described what it was, but it was basically just an attacker could exploit the vulnerability. You have to get an app on the phone for this to work. So there's always that little asterisk with any— well, most things security-related, like you have to do something stupid, or unknowingly, I guess, is maybe the fair way to say it. But this thing was communicating authentication in plain text, basically. And this app could take control of not just the phone, well, not really the phone, I'm sorry, but your work-related email, calendar, Teams messages, files, your OneDrive.

Richard Campbell [01:28:01]:
Anything that depended on the authenticator.

Paul Thurrott [01:28:03]:
Yep. It was just seamless pass-through. So honestly, in that case, it worked really well for an authenticator app. But—

Richard Campbell [01:28:09]:
Well, but also, like you bring up the point, You use a Microsoft Authenticator, especially against Microsoft tenants, because it mixes up its authentication strategy all the time. It's not just enter the 6-letter code. It's like, I'm going to send a number to the authenticator. You pick from 3 numbers, or—

Paul Thurrott [01:28:24]:
Right. I love the way they do it. I actually like it because it's— You never know how it's going to go. Exactly. It's random. So sometimes you type numbers into a thing. Sometimes you— it gives you numbers.

Richard Campbell [01:28:33]:
It shows you numbers and you select one.

Paul Thurrott [01:28:35]:
I love that. I love how random it is. It makes it— it feels more secure.

Richard Campbell [01:28:40]:
But that chunk of code for all of those strategies had a bypass.

Paul Thurrott [01:28:45]:
Right. Now, I don't know that this security company or this researcher used AI to find this, but this is a little window into what I think is and will become an even bigger problem, broadly speaking, with security flaws in software, which is that you can see why this is serious. In this case, we did the right thing. And we fixed the problem before anyone found out about it. Great. It was never exploited. That's wonderful. But this is one thing in one Microsoft app/service, and my God, is the Microsoft ecosystem enormous, the software ecosystem.

Richard Campbell [01:29:19]:
Are we gonna talk about PHP to shell? Because you forget just how much PHP there is out in the world until suddenly there is a major exploit to PHP.

Paul Thurrott [01:29:30]:
Yeah. Now, you might remember this better than me because I can't think of the date, but I wanna say Roughly 3 years ago, it was roughly this time of year, Microsoft had that big security problem where some hackers had gotten into their infrastructure. And the way they had done it was through some random VM that some Microsoft employee had hoisted up in Azure. Yep, demos count. Yep.

Richard Campbell [01:29:51]:
Maybe that's the wizard.

Paul Thurrott [01:29:52]:
Yep, you make it super easy to hoist a VM with whatever properties, you use it for whatever demo at some show or some meeting or whatever, some work you did.

Richard Campbell [01:30:01]:
You never shut it off and then you don't pay for it.

Paul Thurrott [01:30:02]:
Yep, you forget all about it. Yeah, there's no sense of anything. And so there's a lesson to be learned from that. It's fairly obvious.

Richard Campbell [01:30:08]:
That is the root of FSI right there.

Paul Thurrott [01:30:11]:
You're right, right, right, exactly. Thank you. Exactly right. And this was not related in any way to CrowdStrike, and I don't remember if it was plus or minus before or since, but

Richard Campbell [01:30:20]:
it was— It was before CrowdStrike.

Paul Thurrott [01:30:22]:
Okay, yeah. And it was like— and I think it was— they revealed it in January that year, and then CrowdStrike was probably July that year or June that year, whatever timeframe. But it was all part of this kind of growing re-awareness, because this was trustworthy computing 25 years ago too, where Microsoft was like, we've been doing things a certain way, we got lazy, we need to be better about this stuff. And I think the situation there is much better, but now we have AI finding these problems, and I feel like it's just become exponentially worse, right? Or more, maybe.

Richard Campbell [01:30:53]:
You could also argue that it's better. I would also say this, think about the security stories you've just run down about authenticator and so forth. People aren't messing around, they're just fixing them. You know, for the PHP to shell, Cloudflare whipped up a bidirectional firewall patch. So, if you were running PHP, you could just turn that on and it wouldn't allow the handshake for the exploit.

Paul Thurrott [01:31:19]:
So, is this the PHP thing? Is that what you're referring

Richard Campbell [01:31:22]:
to? Yeah, this is the PHP one. And people just took their sites down and fix them.

Paul Thurrott [01:31:26]:
Right. So if you read the internet, if, um, well, it'd be a miracle if you could because the story was you couldn't get on the internet.

Richard Campbell [01:31:33]:
But yeah, yeah, because every WordPress site was affected.

Paul Thurrott [01:31:35]:
Yeah. And, you know, someone I work with was asking about this and I was like, look, we, we were hosted, we do this stuff, and this is someone else's thing and they're already working on it. We don't have to worry about it. But, um, but yeah, it's, it's, we This is an era of transparency for security, which I think is super important because a lot, depending on the company, the organization, whatever the decision makers, I think in a lot of places there would be this impetus to be like, just, you know, it's not being exploited, let's not worry about it, you know. Or maybe we don't have the urgency because it's not being exploited, we can take our time on this. Whereas now there's this understanding like, no, if—

Richard Campbell [01:32:15]:
Everything is urgent.

Paul Thurrott [01:32:16]:
Yeah, this thing happened. If this thing found it, we have to fix this right now.

Richard Campbell [01:32:19]:
Yeah.

Paul Thurrott [01:32:20]:
And so you're right. I mean, in that sense, it is a good time, you know, or good.

Richard Campbell [01:32:22]:
Yeah. And consistently then, by the way, that exploit's been there for decades.

Paul Thurrott [01:32:26]:
Yep.

Richard Campbell [01:32:28]:
And consistently these exploits turn into and you get the keys to the castle. Like, it's not I get to pass a little data. It's like it's called PHP to shell for a reason. You have shell access to that machine now. You can write anything you want.

Paul Thurrott [01:32:42]:
The AI version of one of those choose-your-own-adventure books would be the worst story of all time. It's like, all right, which way do you want to go? And you're like, I'm going to go A or B. You're like, B, you own the keys to the kingdom. The end. Wait, what?

Richard Campbell [01:32:52]:
Okay.

Paul Thurrott [01:32:52]:
You know, like, it's like literally like we're just going to bypass the whole story. You just did.

Richard Campbell [01:32:56]:
Yeah.

Paul Thurrott [01:32:56]:
You know.

Richard Campbell [01:32:57]:
But I also get a real sense in the community. It's like, hey, we all live in glass houses right now. Put your rocks down. Got no time for that. Just fix it. And that's because it's just happening so often. So in some ways, security culture is getting better because we're simply all in a foxhole, right? We're all at war.

Paul Thurrott [01:33:19]:
Yep. Yeah, this is— look, Microsoft, I referenced this earlier, but Microsoft did this with Trustworthy Computing. This all started out from Jim Allchin going on vacation right after they shipped Windows XP. He's in France on a boat. Someone calls him and says, we got a big problem. It was a UNPP. UPnP vulnerability, which is tied to the thing I didn't understand the security implications per se, but sitting in the audience during the Windows XP launch, which was almost exactly 25 years ago, by the way, in New York, I had my first experience where you bring up what was probably still called Network Neighborhood at the time, I think, or My Network Places or whatever it was called. Mm-hmm.

Paul Thurrott [01:34:02]:
And could see all of the wireless networks and devices. So you could sign in, there was no security. When they implemented Wi-Fi at that time, it was just, there was no security on these networks, they were all open. So you could get onto any network and then browse all the machines. And I downloaded some files from other computers, Microsoft computers, just to make the point without really kind of understanding the ramifications of how serious that was. Yeah. So he came back home early from his vacation. Trustworthy computing was the net result.

Paul Thurrott [01:34:34]:
But then, like I said, they got soft, they got lazy, whatever. They moved on to other things. And so SFI, Windows Resiliency, whatever these initiatives are named, is the modern version of that. And I think AI has put a nice—

Richard Campbell [01:34:47]:
Well, and then remind you, why do we call it Midnight Blizzard? Because those are the code names we give Russian state actors. The bad guys now are not kids in a basement.

Paul Thurrott [01:34:58]:
Right. Or if they are, they're working for the— yeah, they could be working for the Koreans or the Chinese

Richard Campbell [01:35:03]:
or the

Paul Thurrott [01:35:03]:
Russians. Yeah, exactly.

Richard Campbell [01:35:05]:
Yeah.

Paul Thurrott [01:35:05]:
Yep. Yeah, this is the nuclear détente theory. It's like, well, he has 13 missiles, we have to have 13 too. And it's like, okay. I mean, it's like even if we're going to do things the right way ourselves, we have to We still have to pay attention to this stuff. Like, they're going to get in. So we need to look at it. It's good.

Paul Thurrott [01:35:24]:
The net result, I think, will be mostly positive. But Richard, some months ago, you put an idea in my brain, which I now cannot escape from, which is how will this change some of the software that we use? Like, imagine if there were some gigantic endemic problem in the Windows code base that could not be fixed.

Richard Campbell [01:35:44]:
Yeah.

Paul Thurrott [01:35:44]:
How do you—

Richard Campbell [01:35:45]:
Without a major change of work.

Paul Thurrott [01:35:47]:
How are we gonna get past that?

Richard Campbell [01:35:48]:
I have every evidence that those are stacking up right now, that they're waiting to get through all of the test suites, all of the fable work or mythos work to get to the bottom of this so that they can structure, here's how we're gonna change. I'm gonna guess authentication, right? That fundamentally, there's been a flaw in these authentication strategies. It's never been detected. And until they get the comprehensive view of what that whole thing is, they don't want to go out forward with, here's how you're going to have to change authentication.

Paul Thurrott [01:36:19]:
A lot of it's going to be based on security boundaries. Like, if you think about things like UAC and what that entails, and then you think about what Microsoft has said about agents just this month, about agents running in Windows are going to run under some user account that's unique, not just to the agent, but to agents in general. It says specific. properties, it's a different security boundary.

Richard Campbell [01:36:40]:
Yep. And many of them are going to be short-lived. They're going to apply for a credential, do an effort, the credential is going to be deployed, they're going to do their work, and then it's going

Paul Thurrott [01:36:51]:
to die. I mean, when XP came out or whatever, most people, normal individuals running Windows 98 or something, weren't signing in with an account for the most part. If they were, they didn't use a password. They were running as an admin. Everyone's still running as an admin. You know, UAC was a step to sort of fixing that. Asking people to set up a standard account is not— is a non-starter for 99.99% of the world.

Richard Campbell [01:37:15]:
For civilians, for sure.

Paul Thurrott [01:37:17]:
Right. And now we have in Windows 11 a system that we're still working on it. You know, the administrative protection is like this. It's the, the idea is that you're going to deal with it with, in this case, Windows Hello or Windows Hello ESS, which if you thought UAC was annoying, let me Let me make that an exponentially more annoying experience.

Richard Campbell [01:37:36]:
You're going to love ESS.

Paul Thurrott [01:37:40]:
Yeah. If you like things to move quick, oh, you're not going to like this. It's going to be a problem.

Richard Campbell [01:37:46]:
If you're feeling like you just didn't have enough nag dialogs in your world, I've got a solution for you.

Paul Thurrott [01:37:51]:
Welcome to administrative protection. Yeah. Okay. So there's that. And then these 2 are just super quick. PowerToys, I think everyone understands. Awesome. latest version came out.

Paul Thurrott [01:38:01]:
They're still on like 0 point, like the version is 0.102. Like, guys, whatever. Yeah. But tons of improvements across existing utilities and then 2 new utilities, one of which is like kind of duh, like cursor hide. The idea is that the mouse cursor actually disappears. I thought Windows just did this, right? But then I was looking at the program I'm writing and I realized it doesn't do this. If I'm typing, make it disappear. I don't want to see it.

Richard Campbell [01:38:31]:
Don't put it on screen. Get out of the way.

Paul Thurrott [01:38:33]:
You ever do that thing where you're like, something on your screen, you kind of reach it, you're like, oh, that's the cursor. Like it's right over a letter partially and it makes it look like the wrong letter maybe, or, you know, anyway, that's one of them. And then mouse button lock, which is an accessibility thing. So I don't really necessarily need this, but it's a way to more seamlessly handle like handle drag and drop of things, whether they're files or things within an app or whatever it is, where you can kind of lock the mouse button on so you don't have to, you know, like hold it down as you move the thing around. Like, okay, you know, it's fine. Like, I don't, I don't need it. I'm not there yet, but I'll be 80 in about 2 weeks, so I'll need it eventually. Um, and then the inevitable, sadly, uh, you know, with the browser company had come up with their Arc browser, which is, uh, unique.

Paul Thurrott [01:39:20]:
And one of those things that was kind of hard, like, how do I explain this? It was like if you polled any 10 people, 9 of them would be like, no, this thing's not for me. The 10th one was like, oh my God, this changes my life. Like it was—

Richard Campbell [01:39:31]:
Right.

Paul Thurrott [01:39:32]:
Kind of a hard little hurdle because they changed a lot of kind of common browser interactions. Things like Ctrl+T in Windows that would normally open a new tab would bring up that kind of command palette type thing instead. You could type and it's like, oh, that's different. But a lot of people, and I think all people if they could get used to it, just found it to be incredibly productive and wonderful and whatever. But then they realized that the architecture they had created for this wasn't gonna scale for this AI stuff that was then starting to happen. And so they created a second browser called Dia, which I have to say, I'm not even sure if it's— I don't think it's come to Windows yet, has it? I don't believe it has. I think it might be Mac only still, but— No. They were also bought.

Paul Thurrott [01:40:17]:
I don't remember the company. Who was the company? It was kind of a— I don't remember which company. So they were purchased. So far it hasn't seemed to impact much, but for a while the story has been like, look, we're going to— Dia is the focus. We're going to keep going with Arc, but this is where all the new work's going to be, et cetera, et cetera. But they announced this past week that they are inevitably retiring Arc. And so certain features will be disabled next month. and then it stops working on December 22nd, just in time for Christmas.

Richard Campbell [01:40:48]:
So, sorry. It's only because Arc was so beautiful, right? That's the only reason we care is because Arc was so very beautiful. Yeah, it was beautiful. I would argue it's almost so beautiful, we were not worthy of it. Like, it was a challenge to use it because it just, it took away all of the conventions of browsing.

Paul Thurrott [01:41:09]:
It's

Richard Campbell [01:41:10]:
too soon, but also we did not deserve you.

Paul Thurrott [01:41:13]:
If you think back, so when Google made Chrome, which was probably what, 2012, somewhere around there, 2011, something like that, you know, the reason they did it was because they couldn't trust the browser makers of the day to optimize their browsers for web apps, which was all they had. This was their world, that we make web apps. We want this to be sophisticated. We want this to be very performant. They couldn't trust Microsoft with IE or maybe Mozilla with Firefox, whatever it was. So they made Chrome and it was super fast. It was awesome. And then it's changed a little bit since then.

Paul Thurrott [01:41:46]:
I don't know if you've ever looked at Task Manager when you're running Chrome, but it's a slightly different kind of browser now, but whatever. But I feel like Arc was to that day, a couple of years ago, what Chrome was back then, which was, this is a very small team, small company, but looking at the world and saying, look, we spend most of our time in the browser. Let's make this thing— make that make sense. Beautiful. Look, they're not in a decision-making place for the world, but I think they made something really special. So there are alternatives that are sort of like it, like Zen Browser, I think is the name of the Firefox-based Arc-like browser, if I'm not mistaken, and there's probably others. It's, yeah, it's the only browser I ever wanted to use vertical tabs.

Mikah Sargent [01:42:33]:
Yeah, I've tried using them in others and never works.

Paul Thurrott [01:42:36]:
Never works.

Mikah Sargent [01:42:37]:
Arc's got it.

Paul Thurrott [01:42:37]:
I always go back. I often go back in one second. Like, I'll toggle it on and be like, nope. Nope. Like, see you later. Yeah. There's something. Dipping your toe in the water.

Paul Thurrott [01:42:45]:
Nope. No, they did such a great job with it. Like, it was so well done. But anyway, it's going away. I haven't looked at Dia recently. Initially, I was very interested in it, um, but part of the problem is they always go Mac first. And I, to my knowledge, I'm not— maybe someone else knows. I don't know if someone has it in the chat, but I don't know that they ever came out with a Windows version of it.

Paul Thurrott [01:43:08]:
Yeah. I don't think they did. I mean, they probably will, but who cares? Oh no, it says— let's see, it says Windows. Let's see. Download. I'm sorry, I'm wrong.

Richard Campbell [01:43:19]:
But it was never feature parity with the Apple version.

Paul Thurrott [01:43:22]:
Yeah, this is a big— yeah, this is, this is something that does happen. If you're familiar with iA Writer, which is a Markdown editor for the Mac and the iPad and now for Windows, um, the version on the Apple platforms is dramatically better. Um, it performs better, it has more features, um, and it's just, you know, they started there. That's their, that's their world.

Richard Campbell [01:43:40]:
That's their home.

Paul Thurrott [01:43:41]:
And I love that they brought it to Windows.

Richard Campbell [01:43:43]:
It's the one the devs use, so of course it's gonna work.

Paul Thurrott [01:43:45]:
So it's just kind of optimized in a way. That we don't see on Windows, unfortunately. Yeah, I'll look, I'll look at the— again, I guess. But, or yeah, right, because I had run it on the Mac originally. Um, anyway, it is on Windows. I apologize.

Richard Campbell [01:44:01]:
Cool.

Paul Thurrott [01:44:02]:
Oh yeah, Mashed Potato mentions Ladybird was another one. Yeah, like that. There's a bunch. I mean, the, the world of, um, alternative browsers is bizarre right now. There's so many.

Mikah Sargent [01:44:12]:
Yeah.

Paul Thurrott [01:44:12]:
Um, yeah, so many. Like the browser is very, very long

Richard Campbell [01:44:15]:
tail.

Paul Thurrott [01:44:16]:
Like I have a— I use Helium and there's a 90% chance most of you people have never even heard of this browser. I think it's, uh, it's fantastic. But

Mikah Sargent [01:44:26]:
all right, well, I think that means it is time for a little appearance from Leo. Uh, let's get back into it with AI.

Paul Thurrott [01:44:39]:
Uh, real quick, the version first because this— I just got the alert while that ad was playing. Um, yeah, I, I don't— this is early onset dementia for me. You would think I would be attuned to the cycle of Windows updates, right? But more often than not, Weekday or Patch Tuesday, that day I will be surprised to discover there are updates for Windows. And it's like, yeah, this happens twice a month every month for the past Many years, whatever, but—

Richard Campbell [01:45:06]:
and still comes out of nowhere.

Paul Thurrott [01:45:08]:
Yep. You're like, where did this come from? It's like, okay. Um, so yesterday technically is the day they would normally release the weekly preview updates, but it actually just occurred, so they're out now. So if you are downloading preview updates, you might want to go take a look. So there's 26H1, which is a standalone island by itself, and then there's 25 and 26H2, which are getting the same updates, right? And so I'll Maybe we'll discuss this next week, but real quick, I'll just say the big thing for most people, meaning those on 25 or 26H2 now, is the new context menu in File Explorer, which you can also edit to make it to be what you want, which is really cool. So that's actually pretty good.

Richard Campbell [01:45:48]:
What about the whole icon thing? Are they getting rid of the icon thing? Can we get rid of the icon thing? I'd love to not have the icons. That'd be awesome.

Paul Thurrott [01:45:57]:
Richard, I think 2027. No, you can't. I don't know how that works, actually. Yeah, let me just look at the— yeah, like every freaking menu item has a stupid icon. So stupid.

Richard Campbell [01:46:08]:
Yeah, they're labeled now, so it's even stupider.

Paul Thurrott [01:46:12]:
This is beautiful because Microsoft, in its infinite ability to be inconsistent, is you right-click on the desktop and every item in that menu has an icon next to it.

Richard Campbell [01:46:26]:
Nice.

Paul Thurrott [01:46:26]:
And you go to the View menu and every item in that submenu has an icon next to it. And then you go to the New menu and every icon in that sub— every entry in that submenu has an icon next to it. And then you go to Sort By and there are no icons. Look, guys, there were literally icons for those things. They exist.

Richard Campbell [01:46:44]:
You just didn't display them.

Paul Thurrott [01:46:45]:
You just didn't put them in there because you're Microsoft and you can't do anything right. But okay, that's hilarious. And then on the 26H1, I don't think Yeah, this one's kind of minor, so that's good. Anyway, I'll write that up later, I guess. I'm going to have a hard time today because Microsoft earnings, Google earnings, and then this. And it's like, that's the rest of my day right there. Okay. AI.

Paul Thurrott [01:47:10]:
Mm-hmm. So Anthropic filed for their IPO this week. Strongly recommend reading this document, which they filed with the S-1 filing with the SEC. It's hilarious. Look, we may end humanity. You should invest in us.

Richard Campbell [01:47:33]:
Don't you want to be on the ship that goes down?

Paul Thurrott [01:47:38]:
Yep. Their revenues are improving year over year, not exponentially, but pretty dramatically, right? As you would, because they were nothing last year and now they're next to nothing this year, but that's a big jump. Their losses are also going up fairly dramatically, which is pretty consistent with all these AI companies. But

Richard Campbell [01:47:53]:
that $80 billion, wasn't that a big write-down? Like $65 billion of it was a write-down of some kind? Yeah.

Paul Thurrott [01:47:58]:
The problem for Anthropic, which is the same problem for anyone that is not the infrastructure company, like the Microsoft, the Google, the Amazon, the hoster, if you will, the hyperscaler, but you rely on them right, is that they have, um, I think it's— I think the number is— it sounds impossible, but I want to say it's $800 million in, um, uh, like promises basically for hosting. And, and the, the problem for them is that if their customers don't need that, and I think we all kind of understand most of them don't, um, they're still on the hook for paying for it. And so one of the— Sure. problems that we see for these companies, and OpenAI falls in this category as well, is that they may be on the hook for some hundreds of millions of dollars that they can't actually afford to pay. So we'll see what— which is, they've done everything they can to raise money privately through private funding, and OpenAI thought they had as well, but then they realized that's not the case. And so they actually pushed back their IPO. So we'll see what happens. Yeah.

Paul Thurrott [01:49:02]:
But Anthropic's kind of an interesting Google-type company, meaning Google from 20, 25 years ago. Um,

Richard Campbell [01:49:08]:
yeah, we'll see. You know what, Netscape IPO'd too.

Paul Thurrott [01:49:13]:
Yep, and they're growing gangbusters, um, or something. I don't know. Anyway,

Richard Campbell [01:49:19]:
okay, I expect all of these companies to end up acquired for pennies on the dollar as they run out of money.

Paul Thurrott [01:49:26]:
Yeah, yep. Um, hey, look, as long as I can make that, uh, squirrel picture, I don't care. Uh, Use it while you can, that's all I'm saying.

Richard Campbell [01:49:38]:
Dude, my Shakespearean sonnet as a bio is my life.

Paul Thurrott [01:49:42]:
Hey, you have an awesome whiskey theme song at the end of the show, which is just a thing of beauty.

Richard Campbell [01:49:46]:
Wouldn't exist otherwise.

Paul Thurrott [01:49:48]:
Would not exist otherwise, it's beautiful. Canva is an interesting company, obviously trying to take on Adobe. I wanna say 2-ish years ago, they purchased Affinity, which at that time made a suite of low-cost alternatives to 3 Adobe apps. It wasn't the whole suite, but they had a designer app, a photos app, which is the one I used, and then I guess a, like a vector illustration type app, like a vector graphics type app. Um, after Canva bought them, they combined all 3 of those apps into a single app they just call Affinity, which is actually Affinity 3. And, uh, they made it free, which is crazy.

Mikah Sargent [01:50:24]:
Wow.

Paul Thurrott [01:50:24]:
It used to be really inexpensive, um, and I paid for it. And when you bought this app, you could use it across multiple platforms. So you bought it once and you could use it like Windows, Mac, and iPad. When they went to Affinity 3, they made it free, but the iPad app disappeared, and they were always going to do that. And it's actually coming out next week. So if you have an iPad and want to use Affinity, it's coming soon. It's a full experience, it looks like. That's going to be awesome.

Paul Thurrott [01:50:52]:
But concurrent to this, Canva, which is the parent company, announced Craft, which is its first-ever AI model. So if you think about, this is a company that's trying to compete with Adobe. You can see where that's going. But it's an image model—

Richard Campbell [01:51:07]:
While Adobe's happily being copied, like it's such a good time to be a software monolith.

Paul Thurrott [01:51:12]:
Well, yeah. So I was talking to my wife about that the other day because I like to bore her as much as possible. And I had said that 30 years ago plus, I was working in a computer lab at the college I'd gone back to. go back to school to become a programmer, and that worked out great. But Linux had become a thing. A friend and I were in the computer lab downloading Linux floppy disk images for Slackware, and a stack of floppies on the desk, right? And then inserting them one at a time, writing them, whatever. I had said to him semi-rhetorically, but I was really wondering this, what happens to Microsoft when some combination of Linux, which is free, and some free office productivity suite arrives, and how do you compete with that if it does 90% of what your stuff does? Like, and it's just free, right? I saw that as kind of an existential threat, and history shows nothing happened. Like, it was— that stuff exists.

Paul Thurrott [01:52:11]:
It's existed for a long time. You can— you don't have to. I mean, and there are reasons for why Microsoft and Office and Windows are still things, but whatever, uh, some combination of business users and inertia and whatever else, who cares, whatever. Obviously it didn't work. So yeah, I mean, my knee-jerk reaction to, I think Artcraft, I think is the name of the company that's making these AI vibe-coded alternatives to Adobe and Microsoft apps and other apps, is yeah, I mean, I'm surprised they haven't sued them yet. But then again, like, if this doesn't amount to a financial threat to their business model, maybe they don't have to do anything about it. You know, maybe the people using this were never gonna use those apps anyway.

Richard Campbell [01:52:50]:
Yeah, and suing them is not a good look. Right? They're sensitive to this thing. They've been here before.

Paul Thurrott [01:52:57]:
That's a good point. Yeah. I mean—

Richard Campbell [01:52:58]:
I'll bet you I know what they are doing. They're trying to hire him with a gigantic signing bonus to make him go away.

Paul Thurrott [01:53:05]:
That's fascinating. Yeah, maybe—

Richard Campbell [01:53:07]:
It's a great solution to the problem.

Paul Thurrott [01:53:09]:
Right. I mean, I will say I don't like it from maybe a moral standpoint, but suing a company with infinite resources, suing an individual with zero resources, It's disastrous. Well, it's bad from a PR perspective because you're admitting something, but it's also smart in the sense that you could shut this guy down pretty quick because he's never going to be able to spend you.

Richard Campbell [01:53:30]:
Not even close. But the reaction of that universe will be everybody will make one. It'll be everywhere. Well, by the way— The anti-authoritarian software industry is all over that. Don't attack them. It's a stupid thing to do. It will get you nothing.

Paul Thurrott [01:53:47]:
Okay. So I, yeah, I mean, I'm really curious about this. I, because when I saw this initially, I was like, A, this is amazing. B, he's going to be sued. There's no doubt about it. But you know, it's been—

Richard Campbell [01:53:57]:
The same personalities that want that software are the ones that would have pirated it anyway.

Paul Thurrott [01:54:02]:
That's what I, yeah, that's what I'm sort of talking myself into. So we'll see. So, but yeah, Canva is not that. I mean, Canva is a more traditional approach, like Affinity was a more traditional approach to competing with the market leader in that market, right? Um, uh, and Canva as well. And then you combine those 2 things, you're like, okay, this is like a, a free to low, you know, depending on what you're doing, free to low-cost alternative to some of the Adobe stuff. Um, but doing their own models, and it's starting with this one model, there's going to be more versions, and, and they're designed for what Canva is mostly used for, which is design. You know, kind of like that book thing I was talking about, made it like a book cover. You could make, you know, you make posters, you make marketing materials, whatever it might be.

Paul Thurrott [01:54:46]:
I guess it's pretty good at that. I mean, the output looks pretty good. It's the first time they've ever done it. There's a public API. They're gonna have an MCP server. I mean, they're doing the whole thing. So we'll see. We'll see what happens.

Paul Thurrott [01:54:57]:
I don't know. And then this one's kind of interesting to me because I saw the kind of headline and I was like, wait, didn't they already do this? But Opera announced that they're bringing Local AI to their desktop browsers, which is Opera One, GX, and Neon. Neon, is that the right name? Neon? Mm-hmm. I guess it is. Why does that look weird? Neon. I think I have that wrong. What about Opera Air? Anyway, it doesn't matter. So, okay.

Paul Thurrott [01:55:22]:
And I'm like, wait, didn't they already do this? I looked it up and yeah, sure enough, a couple of years ago, there's an experimental feature still in Opera One anyway that lets you use what they now call Opera AI, but using it against a local model so it runs on your machine. So private and useless, whatever, but that's fine. But apparently in the interim, some proposed web standard called WebAI has appeared. Microsoft and Google and other companies back it. And so they're using WebAI now to bring local AI to their browser. So it's kind of an extension of what they were doing before. And of course, in those 2 years, I mean, local AI has actually gotten a lot better. Sure.

Paul Thurrott [01:56:02]:
My experiences notwithstanding. I mean, for the types of things you might do in a browser. Right? Yeah. So you can try this now. It's worth looking at. I mean, it's kind of an interesting idea. They also have some cool kind of, well, they're extensions, but they're extensions that do specific things or experiments, but it's just designed to kind of give you an idea of the types of things you can do with the local AI, which I have to say is needed because this is the problem with like Copilot+ PCs. It's like, sell me on this thing.

Paul Thurrott [01:56:34]:
You're like, oh, it does this and this and this. Yeah, I don't— what are you talking about? Uh, it's even worse with RTX Spark, but, um, it's, it's an— it's, it— anyway, this is a good idea. So, um, to me this makes sense. And then, so Apple, uh, just before we had the show, announced, uh, their 3rd day straight of new hardware announcements. And today's, uh, products were 14 and 16-inch MacBook Pros with OLED screens and multi-touch, that thing that Steve Jobs and Tim Cook pledged they would never do because it's impossible to hold your arm up here like this and it doesn't make any sense. This is the, you could combine a toaster and a rabbit, I don't remember the comparison, but it doesn't mean anyone wants to buy it thing, whatever that was. They made fun of Microsoft Surface from the moment it was first announced. Yeah.

Paul Thurrott [01:57:28]:
I didn't, it happened close enough to the show, I didn't get to see. I will, I'll look at it later, but you know, what's the justification for this in this world? The only rationale I can think of is that to develop apps for Apple devices, you have to have a Mac. And it actually would be a lot easier if you could use that emulator with a touch, with a touchscreen, right? 'Cause it would emulate the device. Yeah. I don't believe that's gonna be how they market it, but okay, whatever. Yesterday, I think it was, they also announced like an update to Creator Studio, which is this combination of what used to be called iWork and some like creator apps like Final Cut Pro, which is famous, obviously Logic Pro and some other things. But they, as part of this, they announced they're bringing AI into these apps. And I'm like, all right, let me get this straight.

Paul Thurrott [01:58:16]:
I mean, I just wanna make sure I understand what's happening.

Richard Campbell [01:58:18]:
Copilot, you get it. Yeah, exactly.

Paul Thurrott [01:58:20]:
So we live in a world in which Microsoft is, well understood to copy Apple all the time. And yet you just released a touch laptop and you're putting AI assistance into your productivity apps. You know what? Screw you, Apple. Like, come on, man. Are you kidding me? It's not even like, what's, what's, how do you get like, you have hypocrisy, you have irony. What's the next, what's the next step? Because that's what—

Mikah Sargent [01:58:46]:
It's a German word, surely. Yeah, like a schadenfreude.

Paul Thurrott [01:58:49]:
It's not schadenfreude, but there's a word. Yeah, I think you're right, actually. There's some word like— like, it's actually very much like the Elizabeth Holm thing we were talking about, uh, last week. And by the way, my wife and I have watched this movie since then, and, um, that— I, I mean, first of all, her husband's a psychopath, um, full stop. Yeah, but she is also a psychopath. Yes. And I have to be honest about this, they kind

Richard Campbell [01:59:13]:
of deserve each other.

Paul Thurrott [01:59:15]:
It's, it's— but it's weird because the thing— look, I don't mean to give him any credit for anything, this guy, because he's a useless sack of garbage. But, but you know he is a useless sack of garbage, right? She, she is tricky.

Mikah Sargent [01:59:28]:
She's very tricky.

Paul Thurrott [01:59:29]:
She fools everybody. And you could find yourself just like— That's what I—

Mikah Sargent [01:59:34]:
and that's what happened to me during the film a few times was—

Paul Thurrott [01:59:37]:
Yep. Look, she cries, she's losing her mind.

Richard Campbell [01:59:40]:
Just last week. You're empathizing with it.

Paul Thurrott [01:59:42]:
But now I see it and I'm like, you find yourself falling into this trap. Another thing we need a word for is this, and this happens a lot in technology. I wish I could think of a really good example, but they tried to invent this thing that could do multiple tests against blood, a single drop of blood from one individual.

Mikah Sargent [01:59:58]:
It can do tests. It can do any test.

Paul Thurrott [02:00:00]:
Yeah. Define the word the. Are we really doing this again? You don't understand that for future tense, Are you kidding me? Anyway, whatever. Okay, it's fine. She didn't— she doesn't understand. She doesn't understand who's on first. It's okay, supposedly. Um, my, my— the, the, the— well, I was gonna say the worst thing.

Paul Thurrott [02:00:21]:
It's not the worst thing. Her being a tricky psychopath, I think, is the worst thing that came out of this for me. But, um, they lied. They never solved this problem. There's a lot of chemistry that has to occur. It may be impossible. There are a lot of physics physicists at the time, or biologists, whatever they were, who were like, this is not— including her student advisors who were professors at whatever, Stanford, I think, who were like, yeah, just to be clear, this can never happen. But one interesting thing has happened since Elizabeth Holmes and Theranos, and that's AI, that thing we've talked so much about.

Paul Thurrott [02:00:51]:
And there's a moment in this documentary, or whatever it is, where there's a guy who starts criticizing their work before, and he gets shut down by the husband, which is one of the most screwed up things in the whole show. But the premise here is like, AI is gonna help us solve these problems, right? And so I want you all to imagine a future world where Theranos 2.0, whatever they call this thing, actually does solve this single drop of blood problem because of AI. And then it then can stand there and say, see, told you. This platform didn't exist when you were doing this. This is not the same thing. And, and you know that's coming if it ever happens. I mean, I, I actually feel like it's never gonna happen, but

Richard Campbell [02:01:31]:
no, I

Paul Thurrott [02:01:31]:
don't. Goddamn, that was so irritating. It's just— okay, sorry. Um, all right, I— there's no Xbox news this week. I—

Richard Campbell [02:01:41]:
Wow.

Paul Thurrott [02:01:42]:
I know. I'm not 100%. I know. I don't know that this has ever happened. I, I—

Richard Campbell [02:01:47]:
no,

Paul Thurrott [02:01:48]:
we've had light weeks for sure. Um, But I will just say, so the new Call of Duty is now broadly available. It might have happened last week as we were starting this. I didn't have a chance really to dive into it until afterwards, but it's good. It's honestly, it's one of the highest rated Call of Duty games ever made, which makes sense to me. It obviously was influenced by Battlefield, the latest 6, I think, or whatever the latest one was. Yeah. Um, it doesn't solve the fundamental problem with, um, Call of Duty though, which is that, uh, it's just the same thing over and over again, you know?

Richard Campbell [02:02:27]:
Yeah.

Paul Thurrott [02:02:27]:
And, uh, as a— but I'll— I don't know if you're any kind of an addictive personality or whatever it is, like, I'm not sure what to say about this thing. Like, um, it's interesting to me that it's different and definitely an improvement over the last couple of games especially, but It's also interesting to know that it's not being played as much as the game from 2 years ago, Black Ops 6. It's actually not that popular, which is kind of fascinating. It's possible that it's—

Richard Campbell [02:02:59]:
Yeah, but all those numbers in and around the pandemic are distorted.

Paul Thurrott [02:03:02]:
Yeah. I mean, 2 years. Yeah. Okay. Yeah. I mean, I don't know. So it's— I mean, I'm going to play it because I hate myself, but I don't— I don't know. I don't think it solves the fundamental problem with Call of Duty, I guess, is what I'm saying.

Paul Thurrott [02:03:17]:
It's not as bad as local AI, I will say that. It's not that bad. But yeah, I don't know what happened. I went back, I went back to Xbox, to the site. I'm like, they must have— no, nothing has happened. Nothing.

Richard Campbell [02:03:31]:
Hmm, that's really interesting.

Mikah Sargent [02:03:33]:
Yeah, where I, I didn't get a chance to— and if you've done this whole thing already, then we don't have Um, but actually, you know what, let's, let's take a break, an ad break. But I want to ask you, Paul, when we come back, or not, depending on if you've already talked about this, okay, how you actually see the future of Xbox, because they're supposed to be celebrating their 25th anniversary and instead they're laying off staff and trying to have somebody come in.

Paul Thurrott [02:04:01]:
Mikah, you couldn't be more wrong. They just released a bunch of new wallpapers. They're very clearly celebrating, right?

Richard Campbell [02:04:07]:
Oh, and the nano controller. Everything is different.

Mikah Sargent [02:04:10]:
Like, I'm curious to hear if it's— if you think it's going to be cloud or handheld, or if there will be an Xbox console.

Paul Thurrott [02:04:16]:
Um, sorry, before I get to what you asked, um, so Microsoft earnings are out.

Mikah Sargent [02:04:21]:
Oh,

Paul Thurrott [02:04:22]:
um, top line is, you know, it's going great, right? So revenues $92.3 billion, up 19% year over year. Wow. Uh, net I got to look these up because I'm not sure if these are Capadanga, but we're going to say $36.6 billion in net income, which is— they say it's up 32%. That seems—

Richard Campbell [02:04:40]:
It's a lot of money.

Paul Thurrott [02:04:42]:
So remember, they combined their—

Richard Campbell [02:04:43]:
60,000 layoffs because a lot of money.

Paul Thurrott [02:04:46]:
I know that's always been the tough thing around layoffs. It's like, but you just made, you know, another record

Richard Campbell [02:04:51]:
profit. Lay more people off.

Paul Thurrott [02:04:53]:
Yep. They combined their 2 biggest business units into one, and now they have 2 business units instead of 3. Right. So The third of those, the one that Windows is part of, is now called Devices and Consumer, and that actually shrunk year over year by 3%, but $15.5 billion in revenues. The rest of the company is $77 billion in revenues.

Richard Campbell [02:05:14]:
Right.

Paul Thurrott [02:05:14]:
And that's up 25%. And that's everything, right? Like Azure, Microsoft 365, the big stuff, right? I'm looking— I can't do this now, obviously. I can't do this now, but we're going to have to— and then I'll have to pay attention to the post-earnings conference call, see what they say about CapEx spending. How they lie there.

Richard Campbell [02:05:31]:
Yeah, next week's conversation will be great.

Paul Thurrott [02:05:33]:
It will be long. I think it's— I don't know, great. I can't promise great, but, um, that's that. All right, and then to answer, uh, Mikah's question, so Xbox would be a viable, lucrative, profitable business if they didn't make hardware. Um, there are a lot of Xbox fans that demand that they make hardware and don't understand why they would do anything other than make hardware, the thing that just loses money. Because if you think about about it. Xbox without hardware is Activision plus Bethesda plus whatever other gaming studios. And Activision was wildly profitable, you know, forever, at least for the past, like, I'll call it 10 to 15 years.

Paul Thurrott [02:06:12]:
So, but they can't seem to escape this. So last year, Microsoft at that time, Xbox at the time, which had been run by Phil Spencer, wanted to release an Xbox-branded handheld gaming, uh, device, and was— that was nixed by the senior leadership team. So they farmed it out to a third party, uh, ASUS, that makes the ROG Xbox Ally and X and, you know, whatever, those, those 3 of them now. But, um, it's Windows, right? And so they had to adapt Windows, just as they have to adapt Windows for AI. They had to adapt Windows to be like a pure-play gaming, uh, device. And that means getting rid of background processes and notifications and all the other nonsense that might happen. Project Helix is the code name for the next console. And one thing that just came out of a— I don't remember if it was a rumor or a meeting that Asher Shama, who runs Xbox, now had with the team, but next year we're gonna see an Xbox-branded gaming handheld.

Paul Thurrott [02:07:06]:
And then the year after that, we're gonna see the living room console. And there are rumors that that thing could cost $1,200 to $1,800, which is a—

Richard Campbell [02:07:13]:
Well, it's the same thing that happened with the GabeCube. Yeah, I mean, it's expensive.

Paul Thurrott [02:07:18]:
Every device— Amazon has announced a bunch of new devices over the past whatever months. Apple is coming up with new stuff. It's all way, way more expensive. It's not 10% more expensive. This is— these are dramatically more expensive.

Richard Campbell [02:07:29]:
Much more expensive.

Paul Thurrott [02:07:31]:
Yeah, it's bad. So obviously that's gonna impact them. So my take on Xbox is they could be a wildly successful business if they just let go of the hardware. And one— and I know that rubs people the wrong way, but one of the many ways you can do that, and there are definitely other ways, would be to farm it out to companies like they did with ASUS for the handheld that know how to make hardware and know how to be profitable making hardware. And that's not something that's a skill for Microsoft, which Xbox and Surface have proved since they've existed. So you could have, you know, whatever set of specifications, requirements, whatever it is, and, and allow third parties to do this. HP, Lenovo, could be consumer electronics companies, whatever. And I think that would be better for the ecosystem.

Paul Thurrott [02:08:10]:
I think it'd be better for everybody. But, um, and I almost feel like Project Helix, which is based on Windows, so they're doing the Windows architecture on the gaming console and on the gaming handhelds because Windows is the biggest gaming platform other than, you know, mobile, uh, like little silly games. Mm-hmm. And it opens up worlds in both directions. So when we see things like Xbox backward compatibility, which right now is just the original Xbox come to the PC. That's part of this, right? And we— and on consoles we have 360 and Xbox One as well, and that will come to the PC. And then on the reverse end, we have a future Xbox console that will be able to play Windows games regardless of what store they come from. So you could get a game from Steam, you could get a game from Epic Games, and it will work.

Paul Thurrott [02:08:58]:
You know, we'll see how it goes, but it will work on your Xbox console because it's Windows. And so that part of it, that makes sense, um, to me. I— but I also think if they just got rid of the hardware, that would make even more sense. But I think they'll do this and then we will see. And if this fails badly, this will be why— how they get out of it. And they can explain— they were like, guys, we did what you wanted, it's not working out.

Richard Campbell [02:09:20]:
Yeah, but you know, and Sony's struggling too. They don't know when their PS6 is going to come along. They put out the Pro not that long ago.

Mikah Sargent [02:09:27]:
Are they firing as many people?

Richard Campbell [02:09:28]:
Same problem.

Paul Thurrott [02:09:29]:
No. And, and by the way, they've sold dramatically more consoles. You know, it's not close. It's not, it's not even 2 to 1 anymore. It's like, it's not close. Yeah, it's not even close.

Richard Campbell [02:09:37]:
Yeah.

Paul Thurrott [02:09:38]:
Okay, that— I think that answers your question.

Mikah Sargent [02:09:40]:
You know, that does. Yeah. I, I, I— it's hard seeing Microsoft as a, as too much of a hardware company overall.

Paul Thurrott [02:09:49]:
Yeah.

Richard Campbell [02:09:50]:
Xbox,

Paul Thurrott [02:09:50]:
even big experiments, even aside from that, because there's always been like Speaking of experiments, like little hardware products, there was like an Outlook phone that looked like a phone, you know, back in the day. Yeah. Like a desktop phone. There's been hardware here and there over the years, but Xbox increasingly looks a little anachronistic within the broader Microsoft portfolio. Like it doesn't, you know, I get that people play games on Windows, okay, but Windows primarily is that kind of productivity platform, et cetera, et cetera. it feeds into the Microsoft 365 stuff and, you know, services that you need like OneDrive, uh, that you might pay for or whatever. But Xbox, it's like, like, I don't know, it just, it's always seemed like a strange fit. But, um, but, you know, they've done the cloud thing, you know, so they have cloud gaming.

Paul Thurrott [02:10:43]:
Um, the, the Windows thing makes sense to me. I, you know, we'll see, but I, I really feel like this is going to be used as the the justification to, um, finally get out of hardware, which I think would be beneficial for the business. So, but I know, I know that's not popular. Don't, don't shoot the messenger, even though I invented it. Um, okay, so— It

Richard Campbell [02:11:03]:
was never about the console, it was always about the game.

Paul Thurrott [02:11:06]:
Yeah, yeah. This is one of those tricky things in the video game space. People always complain about like exclusives and stuff, and I just knee-jerk hate that kind of thing because it just restricts the number of people you can play with, and it Creates these little isolated communities. And it's like, one of the most wonderful things, you know, playing Call of Duty for so many years is at one point they allow people on Xbox to play with people on PlayStation. And you're like, awesome. It dramatically expanded the audience and made it better because when you go into any lobbies, whatever game types, there's many more people there. And then since then, they expanded that to include PC as well. And PC is bigger than both of those things almost combined.

Richard Campbell [02:11:43]:
Sure.

Paul Thurrott [02:11:43]:
And so the audience expands yet again. And you're like, I want exclusives, I want exclusives. And I just point to this, I'm like, or we are inclusive instead of being exclusive, and everyone gets to play with everyone. And I think that's just better, healthier, whatever. I don't

Richard Campbell [02:11:57]:
know.

Paul Thurrott [02:11:58]:
So I get— I guess I get it, but yeah, whatever. Okay, um, do we need to do a break?

Mikah Sargent [02:12:05]:
Just do the pause that refreshes before we head into the back of the book. With the tips and picks of the week up next. All right, it's time for our tips and picks of— well, not my tips and picks of the week, your tips and picks of the week. Uh, tell us what you have for us this week, Paul.

Paul Thurrott [02:12:25]:
One of the things tied to like the uncertification of Windows is that not only is the product getting— or was getting, it's gotten better now, but it was getting worse for a long time. And this would expose itself in weird ways. And one of them was like they would restrict the ways in which you could install Windows, and they really were directing you to use an online account, meaning a MSA, right? For most people, a Microsoft account. And there have always been workarounds, and it seems like every year new versions of Windows comes out, and it's like, oh, they got rid of this workaround, whatever. And I feel like I live in an alternative reality because I install Windows more than anybody. pretty sure. And I write a book about Windows that I update every year. And I always check these workarounds.

Paul Thurrott [02:13:08]:
And I'm not saying I've never run into a workaround I used that went away, but I think I haven't. But the workarounds I document in the book, you know, like the UBI/BypassNRO stuff, that still works. It works fine in 26H2. So I've been updating— I'm doing a new edition of book for the latest stuff. And, um, of course you start with the installation stuff, so I've been going over that multiple times, multiple device, multiple PCs, uh, ARM PCs, x64 PCs, different architectures. But 26H2, it still works fine, so I just don't stress over that is the, uh, is the point. Um, originally my tip was—

Richard Campbell [02:13:50]:
Consolidation point, like it's ARM and x86, everybody lives happily together in the same package?

Paul Thurrott [02:13:58]:
What do you mean? I'm sorry.

Richard Campbell [02:14:00]:
Like, because we've been talking about how ARM was kind of its own thread for some time when it came to the updates.

Paul Thurrott [02:14:04]:
Oh yeah, no, I mean, for this stuff, it's the same. Actually, that's— you cannot download— am I missing something? Yeah, no, you cannot download a 26H1 ISO to my knowledge. So it is true that the 2026 ARM releases are right now an outlier, and Microsoft has not explained what happens there. Um, actually, that's something I need to look into. Um, hmm. No, thanks for throwing that little wrench into my world, jerk. Um, but, but yeah, but other than that, um, uh, if you, uh, have a first-gen X Snapdragon X computer or any x64 computer, there are different ISOs, you know, for both. Um, the big difference there is just that the drivers are not on the ARM ISO.

Paul Thurrott [02:14:51]:
So there's a couple of ways around this. One is you could go to your PC maker and find some driver bundle download. If you can do that, that's great because you could actually integrate that into the installation media. But the thing I found is that when you use like a utility like Rufus to kind of bypass some of the Microsoft nonsense, you often find yourself booting into like a near-empty desktop with no icons, There's nothing there, but you also can't connect to the internet. So what you have to do is connect an Ethernet cable, which will connect immediately, go to Windows Update and update, and then you'll get all the drivers from there. And so at some point you'll see Wi-Fi appears in Quick Settings and then you connect to the Wi-Fi network and you're fine. But when you first start out, there's nothing there, and that's what Windows on ARM is like, right? And so I've done this on 2 different— yeah, yeah, 2 different ARM computers here, and in both cases it was that experience. There was nothing there, uh, Ethernet connection, Windows Update, and then one I think was pretty quick and the other one I think I had to reboot and try again.

Paul Thurrott [02:15:50]:
But, um, the drivers all eventually come down and you can look at, um, what do you call it, Device Manager, and, and ensure that that's what happened. But, you know, you see Wi-Fi appear, you know, the mouse cursor sometimes isn't there, you have to use the keyboard, uh, which is always entertaining in Windows. But it works. And then, yeah, right, I don't— it's funny, it's bizarre I didn't think of this. 26H1 is different, right? So I just mentioned that there's a bunch of other recovery tools in Windows that have come over the past few years. So if you do have a 26H1 computer, meaning you have an X2 or actually now an RTX Spark computer, yeah, you can't download an ISO. Yeah, you can't. So that's, that's an interesting problem.

Paul Thurrott [02:16:35]:
Um, this is, you know, remember when Windows on ARM first started, that was a problem for years where there was no ISO.

Richard Campbell [02:16:43]:
Yeah,

Paul Thurrott [02:16:43]:
they

Richard Campbell [02:16:43]:
kept it installed.

Paul Thurrott [02:16:44]:
Yeah, you have more options now than we did then. So I, I don't think anyone's going to be— well, I should say anything could happen, but, um, you know, we have offline cloud restore from the Windows Recovery Environment. Chances are very, very good that you'll be be okay. But, uh, yeah, no, thank you. You screwed up my whole day. Thanks. Um, I'm gonna look into that. Um,

Richard Campbell [02:17:04]:
okay,

Paul Thurrott [02:17:05]:
that's beautiful. Thank you. Uh,

Richard Campbell [02:17:07]:
I'm here to help.

Paul Thurrott [02:17:08]:
And then, yep. Um, I feel like every other week I, I recommend some web browser and it's some web browser I've already recommended many, many times. But, um, I, I really want to—

Richard Campbell [02:17:19]:
No exception.

Paul Thurrott [02:17:20]:
Exactly. Yeah. Uh, yeah, there's a reason I stay to script. Um, So this past weekend during NFL football, you might have seen, if you watch football, you will have seen some Firefox ads, which is the first time Mozilla's done a major ad campaign for Firefox. Wow. But in many years, I'm sorry, they used to.

Richard Campbell [02:17:39]:
So much for starving open source.

Paul Thurrott [02:17:41]:
Yeah, well, they've spent the past couple of years really. And so if you think about Windows and pain points and stuff, they've been doing this with Firefox, right? So obviously over time, they've seen their usage share declined dramatically. It's kind of sad. I've always appreciated what Mozilla stands for as an organization, just the kind of ethics and just the moral stance of the company or whatever you— or the organization, however you want to say it. But they've done the hard work to fix— they haven't totally fixed it, but to fix Firefox. And they just rolled out that new—

Richard Campbell [02:18:12]:
300-something patches, right? Like it was—

Paul Thurrott [02:18:15]:
Yeah, they're doing that. Yeah. But just making the browser Modernizing it, modern look and feel, modern features. They did the right with Smart Window. They have the AI stuff in there that works like an incognito window, but for AI. So it separates it from everything else. They've gotten a lot right. There's still, to me, performance and I'll call it resource usage, like RAM usage problems with Firefox.

Paul Thurrott [02:18:39]:
But the ad campaign they have is fun. They brought back these guys who, I guess we would've called, not memes, but they were, um, like the first viral internet personalities. You know, before, you know, today they would probably be on social media. We didn't really have, you know, our Facebook or anything like that back then. But, you know, the, uh, it's like the Grumpy Cat type stuff, although he's not— that's not part of it. But, um, they brought back these people, you know, 20, 30 years later, I guess 30 years later, um, who were, who were, uh, viral back in the '90s. And it's, it's kind of— it's a fun, it's a fun game. Like, I wish them well.

Paul Thurrott [02:19:17]:
I really do hope they can get Firefox over the hump. I think for a lot of people, it's just fine, you know? But I— and I love the organization. I love what they're doing. So I like— I'm, you know, we'll see. We'll see if they can make a comeback.

Mikah Sargent [02:19:31]:
Definitely rooting for them. I knew that their campaign was a success because I was standing in the kitchen and my partner comes in and goes, Hey, so what about Mozilla Firefox? And I said, wait, how does he know about Mozilla? And he says, uh,

Paul Thurrott [02:19:46]:
like, um,

Mikah Sargent [02:19:47]:
should I switch to Mozilla? Like, I'm using, I'm using Chrome for work. And so I kind of went into it and explained how it's like a different browser engine and everything like that. And he's like, okay, so it sounds fine that I'm using Safari. I don't need to switch to Mozilla.

Paul Thurrott [02:20:02]:
But I guess Safari's probably fine on a Mac or whatever.

Mikah Sargent [02:20:04]:
Yeah, exactly. And he uses an iPhone and everything. But it was just so cool to see because I think of Mozilla as for us nerds. And so for him to be like, hey, so tell me about Mozilla. I was like, where have you heard about Mozilla?

Paul Thurrott [02:20:15]:
A lot of web developers still use Mozilla, which is very interesting to me. But

Richard Campbell [02:20:19]:
yeah,

Mikah Sargent [02:20:19]:
that's where I got my taste of it, was in the, uh, back when I was doing web design in middle school and I started using Firefox regularly. It's a thing.

Paul Thurrott [02:20:28]:
It's still a thing. I— the guy who does our web stuff is a Mozilla or Firefox I'm not a Firefox guy, but there's some really good technology in there. They do containers and things like that. You can have multiple accounts for the same services and contain them. And it's worth looking at. And I do wish them well. Your partner coming back and saying, what do you think of Mozilla Firefox? Is like someone walking in the room and saying, hey, crazy idea. I'm thinking about buying a Saturn, like a car.

Paul Thurrott [02:20:55]:
Exactly. And you're like, what? Or like, what? Windows? I'm gonna buy a Geo. And you're like, what are you talking about? It's like, remember this thing we had in the '90s that looked exactly like a Toyota, but it was like $10,000 less? You're like, yeah, they haven't made those in 20 years.

Mikah Sargent [02:21:08]:
Oh my God, a Saturn.

Paul Thurrott [02:21:11]:
You know what I mean? It's like the Saturn of web browsers. Good stuff. It's just worth looking at, you know, support the companies that support them.

Mikah Sargent [02:21:23]:
I always keep it around too. If something is going weird in Chrome and something is going weird in It's always good.

Richard Campbell [02:21:30]:
Safari first. You

Paul Thurrott [02:21:30]:
should always have at least 2 browsers. And I,

Richard Campbell [02:21:33]:
yeah,

Paul Thurrott [02:21:33]:
I really like the new look and feel. I— the Nova, I think is the name of it, is— it's a nice looking browser.

Richard Campbell [02:21:39]:
Yeah.

Mikah Sargent [02:21:41]:
All right, I think it's time to run as though we were radio. Oh wait, that's not it.

Richard Campbell [02:21:47]:
It's a pun on the run as administrator, which is hilarious. It usually means run wearing a Lego t-shirt.

Paul Thurrott [02:21:54]:
Nice.

Richard Campbell [02:21:55]:
Uh, This week's Run As Radio, episode 1060, I recorded a while back at Kansas City Developers Conference, and I got a chance to sit down with Elizabeth Christensen, a fascinating woman, very, very intelligent, but she's a PostgreSQL person through and through. And so it was fun to have that conversation with her about, you know, a lot of shops that have moved to the cloud, but they're still running SQL Server, maybe some Azure data, you know, Table Store and stuff like this. Why would you involve PostgreSQL in anything? And so, we talked through the sort of scenarios where it comes about. One place is from the dev, that devs like some of the features in PostgreSQL, makes things a little bit easier in some respects. And so, if you get pressure for this, like, what's your right response as an administrator dealing with backups and high availability, security integration with things like Entra? Like, there are solutions to all those problems. And so, Microsoft hosts a version of Postgres for a reason. It's a good player in the Azure landscape. And so, it should be on your radars, at the minimum, to have the knowledge on how to use it and how to secure it, but more importantly, that it actually does deal with a bunch of problems.

Richard Campbell [02:23:08]:
And so, not just, you know, the basic things, because it'll run in any cloud, but also, its analytic tools are extraordinarily good. So there's a bunch of different open-source products like Metabase and Streamlit and so forth that can really use Postgres for a ton. So it was great to sort of catch up and talk through all those things with her. She's very bright and part of the PostgreSQL Association as well. So it is an organized group, not just a simple open-source project, something much more than that.

Mikah Sargent [02:23:38]:
Cool. And works at Snowflake? I remember that company for some reason.

Richard Campbell [02:23:43]:
It's, uh, it's, you know, data analytics in the cloud, you know, born in the cloud.

Mikah Sargent [02:23:49]:
Born and raised. Oh, right. Let us now get into the drinking. I mean, the liquor, brown liquor of the week.

Richard Campbell [02:24:00]:
I've had a pretty nutty whiskey couple of weeks, you know.

Mikah Sargent [02:24:04]:
Oh, there it is. I

Speaker B [02:24:12]:
am. Come gather 'round the cupboard, there's a bottle in the way. Do I have to play the whole thing? No. No. That's gone astray, Richard's got the tasting glass.

Mikah Sargent [02:24:25]:
You get the idea. Nice.

Richard Campbell [02:24:27]:
You got it. You know what you're up to.

Speaker F [02:24:28]:
I had to make an effort.

Richard Campbell [02:24:30]:
Yeah, yeah. I appreciate that. So I've had a fairly stupid conference week. Last week, I flew to Trondheim, Norway, to do the keynote, one-day show, very fun, lovely group of people, very regionalized show. It's mostly locals that go to that show. They make a one-day show because many of the dev offices in Trondheim close for that day so that the entire teams can go to the event. Beautiful day. Then, you know, went to the party in the evening.

Richard Campbell [02:25:00]:
Next morning, got on an airplane, flew to Warsaw. That conference, the Developer Days Warsaw, was already underway. I did a closing keynote for the end of the first day. I went to their party, had a great time. Next morning, got on an airplane, flew back to Norway. I went to Stavanger to the Hello Stavanger conference, where I was able to do a talk on their final day closing session, and then went to their party. And the next morning, had day off for the first time. 3 conferences in 4 days is a stupid thing to do.

Richard Campbell [02:25:33]:
Do not recommend. I knew when I put it together it was a dumb thing to do, and now that I've done it, I feel bad about doing it, uh, because you don't really get to participate in the conferences all that much. You're just racing around thing to thing. But at the end of all that, I finally had a decompression day, and my buddy Nile, who lives in Stavanger, uh, who last year we had a little party at his place, we're trying some very unusual whiskies. extremely rare stuff, actually organized an excursion for us where we went to a distillery in Eggersund. So this meant getting on the train out of Stavanger, riding a good hour into an even smaller town. Eggersund's way in the southwest corner of Norway, pretty much looking straight across the North Sea at Scotland. They're basically at the same latitudes at that point.

Richard Campbell [02:26:20]:
And there's the whiskey they make, they call Vardberg. And Vardberg is a hill, a very old hill. This is ancient geology in this part of the world. This is one of the few places that has a north site, which is part of the original rock that formed the planet 4 billion years ago. And then the ice sheets came through and scoured it down and changed the shape of it as well. But humans have been there for a really long time. In the Vardberg Hill, down at the bottom of it, there are carvings from the Bronze Age of ships, because this is proper Viking country we're talking about, right? This is the southwest of Norway. Everybody got around by ship there, and literally for thousands of years.

Richard Campbell [02:27:07]:
So we were in that land of the Vikings, even the pre-Vikings. You know, the Viking era is really the Iron Age, 1000 AD, but There were people in ships traveling around there 2,000, 2,500 years before any of that. And the ship story is an important part of the equation because the, the Berestands who make Vardberg whiskey actually start with a Captain Wilhelm Berestand that in 1895 retires from sailing and sells his last ship off and goes into the beverage business. in Eggersund. So he founds this company called Eggersund Mineralvannfabrikk, aka the Fabricators of Mineral Water. And so they make soft drinks and coffee products. In 1895, because he had this experience sailing, he knew about the importation process, he knew where the products came from. And so he sort of brought coffee and soft drinks to Eggersund.

Richard Campbell [02:28:06]:
They made all kinds of unusual sodas, even then with the mineral water, vanilla, pomegranate. One of the most famous ones is called Beersje Duk, except they don't know how to make it anymore. They've lost the recipe, but it's famous in their documentation. It's a very, very popular product. So Wilhelm has, has a son and he got a guy named Berndt. And in 1924, Berndt takes over the day-to-day management of the company. At this point, it's still relatively small, this Eggersmann's mural of Van Pabrik. And so, uh, Berndt is also a fisherman.

Richard Campbell [02:28:41]:
He catches herring and shrimp and lobster. He actually does lobster exportation to France alongside some of his drink products as well. This is back when they're delivering these soft drinks by horse and cart. Fast forward another generation to Wiggo Berntsen taking over in the 1950s. He is very post-war Norway. Uh, big boom time and actually expands operations to the point where they build the big factory down by the waterfront in Eggesund where it is today. So Wiggo does the massive expansion, and then his son, 4th generation, 1979, Harald Berntsen takes charge when he was only 22. And Harald to this day is still CEO of the company.

Richard Campbell [02:29:26]:
Now, they only made soft drinks through this whole time because it was essentially possible to make alcoholic beverages. Norway had very strict rules around alcohol, and they, they modernization the rules in 1990, what they call the Modern Alcohol Act. They were having problems with alcohol consumption and overconsumption. I mean, it's Norway, especially up there where it's dark a big chunk of the year, and it's easy to get into a situation where you drink too much. But as you get in the 1990s, you also get the beginning of the ECC, of the European cooperation network, and Norway becomes a participant in that. And so ultimately by 1996, the EU is putting pressure on the fact that Norway is operating all alcohol distribution in a monopoly called the Vinmonopolet, which is literally the government monopoly over alcohol comprehensively. And so in '96, Norway splits up that monopoly into separate units of importation, production, and retail selling to give them a bit more flexibility with that. And our man Harold, who's still alive in 1998, starts talking about making beer.

Richard Campbell [02:30:39]:
They're clearly doing beer experiments, but under the situation with the laws at that time, they're not allowed to make anything. Vinmapolet, even in the broken up form, has total control over any alcohol of 4.7% ABV or more. So you could only make weak beer. And there are some companies that are making wheat beer and now they're talking about getting into that as well. And the timing on that is excellent because as much as Harald started to explore getting into making beer, one of the big beer makers in the area back up in Stavanger is called Tau Beer. Now Tau is actually an island across from a way of the fjord. It's on another chunk of the peninsula that had been supplying beer to Stavanger for nearly 150 years. So it's the middle 1800s, right? But it was all weak beer, but they were a longtime producer.

Richard Campbell [02:31:29]:
And in 1990, they got acquired by a much larger beer maker, Ringnes. And Ringnes, right out, not long after acquiring Tau, also became part of the Carlsberg collective. And this is the west coast of Norway. It's kind of nowhere, you know, it's a very quiet place other than it's the oil country. So they have a lot of money, but it's small. But in 2003, the Ringnes Group announces they're going to close the brewery in Stavanger. And it's not like that brewery wasn't doing well. It was, it was busy.

Richard Campbell [02:31:59]:
It employed more than 100 people in the area, which is a relatively small town, but it wasn't quite as profitable as some of the other facilities. And so Ringnes is going to shut this thing down and the workers protest. In fact, the whole city gets behind not closing down Tau. In fact, they start smashing Ringnes polls at the pubs and things. But there's no stopping ringness. They do it, and by June they shut the whole thing down. And there's— in the midst of this outcry, they talk about, like, what's the Stavanger beers? And one of them is Berendsen. So Berendsen gets a huge break making weak beer, and so they start producing beer at scale.

Richard Campbell [02:32:39]:
And in the meantime, Norway starts responding to the modernization of alcohol in general with their new version of the Alcohol Act, which actually allows for craft distilling and the making of stronger beers. Today, Bear& Sons makes a huge array of beers, including sort of classic Belgian-style Christian beers, uh, or Christmas beers. They— some of them are up to 19% alcohol. I— that's not beer. That's— maybe you can't even call that malt liquor. That's like serious stuff. It'll poke your eye out. The company gets renamed as Berensten Bryghaus, or Brew House, because of the beer they're making in 2007.

Richard Campbell [02:33:19]:
And this is when Harald's children start to get involved in the company. So there's a Fred and a Frederik, male and female. And Fred actually goes to study beer manufacturing in Copenhagen. And when he comes back in 2015, he takes over as head brewer and his sister is also working in the company dealing with other aspects of the business. So Harold's still very much in charge, but their beer production is very serious and good business alongside still making soft drinks. That's still a thing. But they all— Harold always had an eye to wanting to distill. Now, this doesn't actually happen until 2018 when they purchase Holstein Stills and, and start making gin and, and aquavit and, and start laying up whiskey.

Richard Campbell [02:34:09]:
So the first cask for whiskey, it laid down in November of 2018. Now, much of the story I got directly from Fred, who was our tour guide when we went to the distillery. They have a beautiful visitor center. They have a nice barrel room that you walk through. We got to go back and see their stills. But of course, since they're beer making, They have all the beer-making equipment for doing the processing of barley in the first place. So their mash tun and washbacks are all stainless steel, which is very typical of beer-making because you're not going to distill it, because you're not going to heat it up. You use stainless steel tanks rather than wooden ones because they're safe.

Richard Campbell [02:34:43]:
You have to sterilize them in between making batches of beers. You don't want bad bacteria to creep in that'll actually incubate in the bottle and ruin your beer. So adding stills to the equation was not that much of a stretch. You've And if you add hops to that and do the additional processing, you really are making beer. Or you take that brewer's beer and you pop it into the stills. And Fred introduced us to the stills. They're Holstein stills, they're out of Germany, small ones, but they're combination pot column stills, but they don't use the column still. They literally do a wash distillation and a sparge distillation in the pot.

Richard Campbell [02:35:21]:
spirits distillation in the pot stills they have there. They have another stack just for gin. So why did he have the column stills? Because he bought them already, uh, in the standard format, which was a combination pot and column still. We ended up having a conversation long enough that the others started getting tired and just wanted to get to their drinks about using the combination of column and pot stills to make spirits. But he's just suggested doing pot still. distillation. These are small stills, about 1,500 liters, because he only lays up a few thousand bottles at a time. And there's only a couple of Vardbergs actually available today, specifically Vardberg Born, which is their inaugural whiskey.

Richard Campbell [02:36:00]:
First editions were made in 2023, and that initial edition is actually hard to come by. There were only 500ml bottles and they only made 2,000 of those bottles total at 48%, aged in 3 ex-bourbon casks, which they were getting from Buffalo Trace. And a couple of Oloroso sherry hogsheads. They now have more barrels to start to do wine finishes and so forth. So we sampled a bunch of their experimentals, but this Vardberg Born, this is the one you could actually buy, including in America. And it won the Silver Award at the World Whiskey Awards in 2024. You're looking at about $100 a bottle if you can find it only in select locations, but it's a legit Norwegian whiskey with a sherry finish. All right, so it's hard to go wrong.

Mikah Sargent [02:36:49]:
Wow, what a story.

Richard Campbell [02:36:51]:
Yeah, fun, huh?

Mikah Sargent [02:36:52]:
Yeah, absolutely. And with that, I do believe we are here at the end of this episode of Windows Weekly. Uh, if you would like to get this show every week, well, head to twitch.tv/ww. That is where you can go to subscribe to the show in audio and video formats. All right, first and foremost, Paul Thurrott of thurrott.com, anything you would like to plug this week?

Paul Thurrott [02:37:20]:
Um, I'm a local AI expert and, uh, no, I guess not. I, I'm doubting the meaning of my life. I don't know, not really.

Mikah Sargent [02:37:32]:
No? Okay. And Richard Campbell, Run As Radio. Anything else you'd like to plug this week?

Richard Campbell [02:37:39]:
Well, I am hanging around at Tekorama, so they gave me one of the fancy hotel rooms. So let me spin the camera around for you and say, like, I could have done the show from this hot tub in my room.

Paul Thurrott [02:37:50]:
Wow. Are those Christmas lights in the window or whatever?

Mikah Sargent [02:37:51]:
I was gonna say, those lights.

Richard Campbell [02:37:54]:
Uh, no, I think there's— yeah, there's just lights outside.

Mikah Sargent [02:37:56]:
It's nighttime. Oh, it did look like Christmas lights to me too. I thought, well, that's fancy.

Richard Campbell [02:38:00]:
Well, I appreciate your restraint. Yeah, I just figured out having a laptop sitting over top of a big pool of water like that seemed like a dumb thing to do.

Mikah Sargent [02:38:09]:
Uh, well, gentlemen, thank you so much. And, uh, I think I'll see you again next week before the inevitable return of Mr. Laporte.

Richard Campbell [02:38:19]:
Yes, I will be. And I will be in Sweden.

Mikah Sargent [02:38:22]:
Sweden next week.

Richard Campbell [02:38:24]:
Yes, Malmö.

Mikah Sargent [02:38:26]:
Goodbye, everyone.