Windows Weekly 974 Transcript

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

Leo Laporte [00:00:00]:
It's time for Windows Weekly. Paul Thurrott's here, Richard Campbell's here. We're going to take a trip down memory lane, talk about their first computers. Also a big Patch Tuesday, actually not so big, and Paul likes it that way. Plus, doing development with AI. There have been a lot of developments, if you will, and Paul will talk about that too next on Windows Weekly. Podcasts you love from people you trust.

Paul Thurrott [00:00:30]:
This is TWiT.

Leo Laporte [00:00:37]:
This is Windows Weekly with Paul Thurrott and Richard Campbell, episode 974, recorded Wednesday, March 11th, 2026. Do-it-yourself Crocs. It's time for Windows Weekly. Hello, you winners. Hello, you dozers. Look who's here, back in their native environs. We We give you Paul Theriot in the duck blind. Hello, Paul.

Paul Thurrott [00:01:04]:
No, at the duck blind.

Leo Laporte [00:01:05]:
I don't know. He's in Mexico City. He's in MXCDN or whatever.

Paul Thurrott [00:01:10]:
CDMX.

Leo Laporte [00:01:11]:
CDMX.

Paul Thurrott [00:01:12]:
The DF, as the kids say.

Leo Laporte [00:01:14]:
Ciudad de México, Mexico. Hi, Paul. We missed you.

Paul Thurrott [00:01:22]:
Yeah, I mean, I wanted to go to Florida so bad, but You, you really did. But I was here.

Leo Laporte [00:01:29]:
We saw gators. That's, uh, by the way, the guy who did go. As you know, he did the show from, uh, beautiful Orlando at Zero Trust World. Mr. Richard, on the show floor. It was fun. Yeah, uh, they, they— I think they want us to do it again next year, and this time we're gonna make Paul go. But, uh, I said, you know what, we really should be sitting somewhere like where you can like see—

Paul Thurrott [00:01:55]:
yeah, uh, maybe, uh, next year have it in like New Orleans or green behind you, Austin, you know.

Leo Laporte [00:02:01]:
Well, I think they do it in Orlando because that's where they are. Ah, that is a very tiny Richard Campbell sitting there. That's what it looked like.

Paul Thurrott [00:02:10]:
It's like Richard Campbell but he's mini, mini, mini me.

Leo Laporte [00:02:13]:
Yeah, mini, mini Richard. It was a lot of fun. And then Richard had his, uh, Silver Lamé, uh, space shirt for the cut. Should I, should I show a picture of the two of us at the party?

Richard Campbell [00:02:24]:
Oh, absolutely.

Leo Laporte [00:02:25]:
There is a good picture we had. By the way, Richard took us to the Kennedy Space Center, and that was an incredible tour you did, Richard.

Paul Thurrott [00:02:32]:
You are—

Leo Laporte [00:02:32]:
I gotta listen to your geek-outs on RunAsRadio or.NET Rocks because I had to use it to geek out about all about space.

Richard Campbell [00:02:40]:
It's about 30 space ones.

Leo Laporte [00:02:42]:
The final frontier. I have— let's see, where's this picture of us at the party?

Richard Campbell [00:02:50]:
That was a fun party.

Leo Laporte [00:02:51]:
Yeah, it was so much fun. I guess I don't have it on this machine, so I can't.

Richard Campbell [00:02:55]:
But, um, one thing I learned about Leo Laporte is you put a spacesuit on him, he's all in.

Leo Laporte [00:03:01]:
I'm a wild man. Oh, here it is, here it is, here it is. This is, uh, this is me and Richard having a chat, as one does. I'm in zero-grav.

Richard Campbell [00:03:12]:
Yeah, you're floating. And I look like I'm ready to be stuffed into a microwave and pop.

Paul Thurrott [00:03:19]:
You look like the worst Elvis impersonator I've ever seen. And I gotta say, it's kind of a good look.

Leo Laporte [00:03:27]:
Richard, so I thought for sure that there would be a lot of other spacesuit-garbed fellas.

Richard Campbell [00:03:33]:
Yeah, and one other spacesuit.

Leo Laporte [00:03:35]:
Yeah, and he was inflatable. Yeah, and then there was this, uh, stormtrooper, chubby stormtrooper. That was, that was fun. And then they had that thing, you know, with a crane.

Richard Campbell [00:03:46]:
Oh yeah.

Leo Laporte [00:03:47]:
And everybody said, can you be in our picture? So I had to stand there for like half an hour, busy for a while, pretending that I was part of the scenery. Uh, here I am with the Powderpuff Girls.

Richard Campbell [00:04:00]:
I think that's Sailor Moon.

Leo Laporte [00:04:02]:
Oh, is that Sailor Moon? Oh God, I'm I thank you for correcting me on that.

Richard Campbell [00:04:06]:
Well, and I only know because my eldest was totally into Sailor Moon when she was 6.

Leo Laporte [00:04:13]:
These are adult women. Anyway, thank you, uh, Sailor Moon. Thank you, Kevin. Kevin says it's Sailor Moon. He knows too.

Richard Campbell [00:04:20]:
Awesome.

Leo Laporte [00:04:21]:
Um, thank you, uh, to ThreatLocker for hosting us out there. We had a good time. I think they were so happy they said we're gonna do it again, maybe in Vancouver. They have an event.

Paul Thurrott [00:04:30]:
There you go.

Richard Campbell [00:04:32]:
That would be fun.

Leo Laporte [00:04:33]:
Wouldn't that be fun?

Paul Thurrott [00:04:34]:
That would be good.

Leo Laporte [00:04:34]:
It's in May though, and I don't know if we're available, uh, at the date.

Richard Campbell [00:04:38]:
So I don't know if I'm available in May.

Leo Laporte [00:04:40]:
Yeah, you, you have to be scheduled at years in advance, I'm sure, especially that spring, that window there.

Richard Campbell [00:04:47]:
Like, what am I—

Leo Laporte [00:04:48]:
a big one for you? Yeah, uh, usually—

Paul Thurrott [00:04:51]:
I mean, this year they moved Buildo.

Richard Campbell [00:04:55]:
They did. That I'm committed to Europe first week of June.

Leo Laporte [00:05:00]:
So do you have a scheduler or do you do this manually?

Richard Campbell [00:05:03]:
A little bit of both. Yeah, but thank goodness for TripIt.

Paul Thurrott [00:05:07]:
TripIt.

Leo Laporte [00:05:08]:
See, I moved off TripIt because I am getting old, because I get irritated. And I got irritated— they said you can't log in— I have a TripIt Pro account— unless you turn off your ad blocker. And I said, I have a— I pay you Yeah, I pay you for this service.

Richard Campbell [00:05:24]:
I would cut it if they did that to me too, but they never have.

Leo Laporte [00:05:26]:
Yeah, well, I— yeah, so I said fine, you're not getting my money anymore, and, uh, I found something else. All right, let's— enough of that, enough of me, as much as I enjoy talking.

Paul Thurrott [00:05:37]:
Enough of me.

Leo Laporte [00:05:39]:
We're here to talk.

Paul Thurrott [00:05:39]:
That's rather harsh, Leo.

Richard Campbell [00:05:41]:
Yeah, what do you think of me?

Leo Laporte [00:05:45]:
Uh, ladies and gentlemen, let's talk about Windows 11. Yeah, yesterday was Patch Tuesday.

Paul Thurrott [00:05:51]:
I passed today.

Leo Laporte [00:05:53]:
I Patch Tuesday?

Paul Thurrott [00:05:55]:
Yeah.

Leo Laporte [00:05:55]:
Oh, I didn't get the memo.

Paul Thurrott [00:05:57]:
I have a— well, um, how do I say this? I'm kind of happy with the way this year is going for Windows. Oh, and the reason I say that is—

Richard Campbell [00:06:07]:
where are you and what did you dump? That's unexpected. Yeah, what the—

Paul Thurrott [00:06:11]:
what? Oh, don't worry, Microsoft will screw it up for me pretty soon, but Um, it's been a light year, right?

Richard Campbell [00:06:18]:
Um, yeah, I see. So it's good when they don't do anything?

Paul Thurrott [00:06:21]:
Yeah, because, um, too many updates is chaos. And, uh, this month and last month, and I forget January— I feel like January is pretty light too— but, um, I feel like we've gotten 1/10 the number of feature updates that we got last year at this time, you know. So we'll see if it continues. But, um, Patch Tuesday didn't deliver anything unexpected. If you've been watching the show, we've talked about these things 10 times, maybe, I don't know, half a dozen times. So this is the update with the network speed test, which I, by the way, I just did. My speed here is now faster than it used to be. I have a gigabit down and 600 megabits up or whatever.

Leo Laporte [00:07:02]:
Nice.

Paul Thurrott [00:07:05]:
Sorry. If you have a camera that supports tilt and pan controls, you'll see that inside of the Settings app. RSET improvements, that quick machine recovery improvement we talked about recently, uh, the ability to use WebP, um, images as background wallpaper. Emoji 16.0, Sysmon, right, the Mark Russinovich tool.

Richard Campbell [00:07:26]:
Yeah, which, you know, you ought to be using anyway, but okay, now it's—

Paul Thurrott [00:07:31]:
yeah, there's a slightly easier way to install it, but honestly, there's still multiple steps, you know, it's, it's complicated, but that's okay if you need it, it's there. Um, so yeah, it's good. So So uneventful is how I would describe it, other than I just found out my connection got faster. That's cool. Yeah.

Richard Campbell [00:07:48]:
Anyhow, um, funny, I just helped a neighbor get fiber put into their place. He was— he's just anxious about the whole thing. He's like, you know what, Jan, I'm just gonna spend the day with you on the install day, right? And talk to the guy. Easy to do for me, right? And he, you know, and he's like, that's it? That's all it was? It's like, everything went well, friend. That's how it goes sometimes. If it didn't, it would have been harder, but it went well.

Paul Thurrott [00:08:12]:
I don't remember if it was 1995 or '96, maybe. Um, we got— Phoenix was one of the first cities in the US to get cable modem.

Leo Laporte [00:08:21]:
Oh yeah, I remember that.

Richard Campbell [00:08:22]:
Those—

Paul Thurrott [00:08:23]:
because it was flat, right? So it's easy to deploy. And, um, I got it on day one, literally Monday morning at 9 AM. And, um, it was like the Intel bunny suit guy showed up. Literally, guys were wearing those white suits and walking around my house, and they were doing stuff. They were at the line, they were testing things. They had a guy up the block taste testing things. I'm like, what is this scientific process? I'm just trying to get the internet, you know?

Richard Campbell [00:08:45]:
Yeah.

Paul Thurrott [00:08:46]:
And, uh, now it's like a service, radioactive.

Leo Laporte [00:08:48]:
Yeah, now they don't even roll a truck, you just do it.

Paul Thurrott [00:08:51]:
No, you just get it, it just happens, you know? Yeah.

Richard Campbell [00:08:53]:
Anyway, well, this was actually pulling a fiber into the building and then it had to be demarked down and stuff. There was some stuff and then there's a bunch, a few decisions that need to be made and he's sort of, you know, I can see why Jan would have been a little intimidated.

Leo Laporte [00:09:04]:
So it's like getting fiber in here, here and here. That must be interesting.

Paul Thurrott [00:09:09]:
That is what we have.

Leo Laporte [00:09:10]:
How did you get it though?

Paul Thurrott [00:09:11]:
I mean, well, if you go outside and look at the wires outside, you'll see there are spider webs of thousands of lines. And I was just talking, you know, Steven Rose just showed me a picture he'd taken when he was here last year, which is a corner we cross all the time. And there's a wire and it's got a little bunch of all thing. It's just sitting there. It's like literally at eye height, you know. And so I sent him a photo of it later, you know, saying, thinking of you or yeah, but I'll often, you'll see these wires everywhere, and it's pretty clear that when something goes wrong here, if they have to do anything, they just don't try. There's no way to troubleshoot, so they just run another one.

Richard Campbell [00:09:47]:
Just lay another wire.

Paul Thurrott [00:09:48]:
Oh, so like, I'll be walking down the street and there's a, you know, like a wire like that, and I'm like, I think I know why the internet's out. Um, you know, just like lick it, you know.

Richard Campbell [00:09:57]:
By switching to fiber, they're saving lives. You just can't electrocute yourself.

Paul Thurrott [00:10:03]:
It's too thin to hang yourself with.

Leo Laporte [00:10:06]:
Um, But don't look down the barrel, right? I've—

Paul Thurrott [00:10:08]:
I, I know.

Leo Laporte [00:10:09]:
I don't know. I have eyesight problems from looking.

Paul Thurrott [00:10:12]:
Oh geez, that would— yeah, okay, I could see that. Well, that— well, okay, so the wires that are hanging like that probably aren't working, so I guess it would be safe. But yeah, in a, in an actual— yeah, how do you know?

Leo Laporte [00:10:23]:
Well, just look down the wire.

Paul Thurrott [00:10:24]:
Yeah, did you go blind? Yeah, it's working. Um, okay.

Leo Laporte [00:10:31]:
I think because it's not visible. Is it visible? Uh, I'm not sure.

Paul Thurrott [00:10:35]:
It should be. I, I would imagine it's a red— isn't it a red light?

Leo Laporte [00:10:38]:
Like, uh, he said it was, uh, it was because he didn't know really that he was looking into it.

Paul Thurrott [00:10:44]:
Like when there's an eclipse and you shouldn't look at the sun, like a certain Nimrod. Um, yeah, infrared or something. Don't look at the sun. What, that sun? Yeah, that one.

Leo Laporte [00:10:54]:
That, uh, that sun I can't see.

Richard Campbell [00:10:56]:
Okay.

Leo Laporte [00:10:57]:
I see a spot.

Paul Thurrott [00:10:58]:
Uh-oh. Oh yes. Oh geez. So by the way, I just, uh, the papaya man outside.

Richard Campbell [00:11:04]:
No, that's the garbage girl.

Paul Thurrott [00:11:06]:
No, yeah, that's love it. No, that's the junk, the junk truck. Um, so today is, uh, GDC.

Leo Laporte [00:11:12]:
Richard, you know the sound of the junk truck.

Paul Thurrott [00:11:15]:
Oh no.

Richard Campbell [00:11:15]:
Yeah, she dropped by every day for a second there.

Paul Thurrott [00:11:21]:
Um, Microsoft is announcing information about Project Helix, which we're going to talk about later, which is the next Xbox. And we— yeah, I guess we'll, we'll have to do this at the end of the show at the normal time, but it's actually happening right now. So this just occurred.

Leo Laporte [00:11:35]:
Do you want me to tune in and we could put in a little box?

Paul Thurrott [00:11:37]:
There's no way to do that. Yeah, it's not live. I'm hoping there's going to be a video later, but, um, they're doing two presentations at GDC and they just announced a bunch of stuff. So we'll talk about that.

Leo Laporte [00:11:46]:
This is— yeah, that's— I know what you're— yeah, okay, you'll have more information by the end of the show.

Paul Thurrott [00:11:52]:
Yep, it just popped up in my face, like, well, courtesy of my coworker.

Leo Laporte [00:11:55]:
Um, that's what was distracting. I thought it was the junk lady.

Paul Thurrott [00:12:02]:
No, I don't— that's not— that's just background.

Richard Campbell [00:12:06]:
That's just background. That's normal Mexico City stuff.

Paul Thurrott [00:12:08]:
Yeah, not to get sidetracked by that, but, um, one— we were just talking about this with some people who were visiting. Um, one thing I think that would shock people from the United States here is how often they do things like that. Yeah, so, uh, there's a garbage truck that truck comes around. Well, there's multiple garbage trucks, but they collect the garbage here between 5 and 7 times a day.

Richard Campbell [00:12:27]:
Wow.

Paul Thurrott [00:12:27]:
Every day.

Richard Campbell [00:12:28]:
Wow.

Paul Thurrott [00:12:29]:
And yeah, if you know, like in the United States, you wake up late and you did put out your trash, you're in trouble. Yeah, you have a whole week to go. Here you have an hour, maybe 2 hours, you know.

Leo Laporte [00:12:39]:
No problem.

Richard Campbell [00:12:39]:
Another one.

Leo Laporte [00:12:40]:
It's like a bus.

Paul Thurrott [00:12:41]:
It's astonishing. So that you would think, like, why would anyone need the junk truck? Well, What you should be asking is, why would you need it 7 times a day? Because they come by all the time. It's very strange.

Leo Laporte [00:12:54]:
Anyway, okay, city though, as a result, it is very clean.

Paul Thurrott [00:12:58]:
That's the thing, like, for a city this size, you know, and I'm sure this is not true in some areas, but, you know, we've never seen like a rat or anything disgusting. I mean, that's probably why they pick up the garbage every time. Yeah, it's pretty clean.

Leo Laporte [00:13:10]:
Nice.

Paul Thurrott [00:13:12]:
Okay, sorry.

Leo Laporte [00:13:13]:
Back to, uh, Helix. No, no, we don't want to do Helix.

Paul Thurrott [00:13:16]:
Uh, no, let's hold off on that because I gotta— during a break I'll catch up on this and try to figure out what's going on here.

Leo Laporte [00:13:21]:
For sure. Um, exciting. Let's continue with Patch Tuesday.

Richard Campbell [00:13:22]:
How about that?

Paul Thurrott [00:13:23]:
Yeah, well, we're done with Patch Tuesday, so that's done.

Leo Laporte [00:13:25]:
Oh, that was fast.

Richard Campbell [00:13:26]:
Yeah, it's a good one.

Leo Laporte [00:13:27]:
Yeah, I, I guess you— is it a good thing when there aren't many patches or a bad thing?

Paul Thurrott [00:13:33]:
No, it's a good thing.

Leo Laporte [00:13:34]:
Yes.

Richard Campbell [00:13:35]:
Well, the operating system is just so darn stable.

Paul Thurrott [00:13:38]:
Well, I, I based that assessment on the past 3 years of updates, and yes, I would much rather have a stuff like this than some of these crazy months where it's just, you know, 117 new features. And they'll roll out whenever they roll out, you know, don't worry your pretty little head, it will happen on some schedule. But anyway, this is good. So to me, it's good news.

Leo Laporte [00:13:57]:
Yeah, I mean, it's funny because on the one hand, many patches means they're fixing a lot of bugs, and it's gonna—

Paul Thurrott [00:14:04]:
well, there's patches and then there's new features, right?

Leo Laporte [00:14:06]:
So we don't want a lot of new features.

Paul Thurrott [00:14:09]:
That's You guys, well, Richard will remember, I think Leo might too, they renamed or they changed the way they named their patches, right? So for a long time, we referred to these monthly updates as cumulative updates, which they still are, you know, they're still cumulative, but now they're called security updates. And I guess that's simpler, but they also have new features. So I don't know why they call them that, but now they're security feature or security updates. So.

Richard Campbell [00:14:36]:
Whatever. Well, then I think that whole play there— so you'll install them because, you know, there's security updates. The fact that it changed the shape of your search bar on the taskbar kind of puts a lie to that.

Paul Thurrott [00:14:47]:
But okay, I feel like we have terms for when there's a payload that you didn't ask for that's part of something you need, you know.

Richard Campbell [00:14:54]:
Yeah, that's not right.

Paul Thurrott [00:14:55]:
Yeah, smallware, right? I mean, come on. But anyway, Windows, okay. And, um, we had a set of Windows Insider builds late last week after the show last week. Canary build added nothing new, you know, what else is new? And then the Dev and Beta builds were identical again and actually pretty light on features, right? In fact, most of them I would describe as, well, two improvements and then one reintroduction of a feature. So remember we talked last year about administrative protection, which is going to be very disruptive. They were just about to deploy that in 25H2, and at the last second we're like, yeah, I think this needs a little more. Fine-tuning. So they've added it back to dev and beta.

Paul Thurrott [00:15:39]:
So this is— these are 25H2 builds. Um, it's going to be disabled by default, obviously. So you can now go back and enable that if you want and see, um, what that looks like. It's going to look terrible.

Richard Campbell [00:15:49]:
This hints that to me, Paul, that there's a reorg going on a few weeks ago, and so there's just not been a lot of code written right now.

Paul Thurrott [00:15:57]:
Well, they also— we also had that kind of promise from Pavan Davalori that the focus this year was going to be more on kind of the fundamentals and, you know, shoring up performance and reliability, etc., etc.

Richard Campbell [00:16:09]:
So I mean, at some point, probably in January, he sort of said to the team, hey, you're welcome back from holidays, finish what you're working on, and let's talk about what we're going to do next, right? And they're in a planning cycle, or they were on a planning cycle. They could be coding right now. There's always like a 6-week lag here, right?

Paul Thurrott [00:16:27]:
Yeah, well, and I think there was a 6-week lag on that, uh, revelation about the focus for this year. I mean, that wasn't something they just invented that second. They've been, you know, look, they see the feedback. I mean, they seem to be immune to a lot of it, but I think once it piles up enough, and especially if it comes from their enterprise customers, um, you know, they have to make some—

Richard Campbell [00:16:46]:
well, I think they got a new leader who's less immune and read through some of those things and also set off his own personal bomb And, uh, well, it's not a bad thing.

Paul Thurrott [00:16:57]:
This, to me, I would say he was more the target of a bomb than, than he set one up. But I— but whatever, however you want to describe it, it's fine. But, but yes, he was certainly the, uh, at the center of a big controversy last, uh, October, November.

Richard Campbell [00:17:10]:
But it was more of a burst of rage, not even really directed at Windows per se, directed at AI.

Paul Thurrott [00:17:17]:
So I guess apropos of nothing, almost, um I was, as you guys know, not very happy with the way people treated him at that time. I thought that was, uh, just terrible. I mean, it's just horrible. It's a human being, you know, like, it— I just thought it was awful. But I will say, you know, sometimes, um, with people, like, in just personal relationships, you don't really realize something's wrong, and then there's an outburst.

Richard Campbell [00:17:45]:
Yeah.

Paul Thurrott [00:17:46]:
And you're like, oh crap, this was worse than I thought. Maybe I should start paying attention. It's possible that that visceral, terrible reaction to the ancient stuff in Windows 11 caused a little bit of soul searching. I don't want to suggest that that makes it okay. I just want to be super clear it's not. But interesting. Interesting that it is happening.

Richard Campbell [00:18:09]:
Maybe I'm the forever optimist on all of this, right? Just you said, senior eyes who wants to make a difference, whose first move was to reunify the teams. Excited. And this lull speaks to a reorg.

Paul Thurrott [00:18:22]:
Yeah, I'm impressed by your optimism, by the way. You've been around as long as Leo and I, and I'm ready to just drive this car into— off a cliff.

Richard Campbell [00:18:30]:
And, uh, gave up hope. I felt a lot better, right?

Paul Thurrott [00:18:33]:
Yeah, much easier, right? Just, yes, stop trying. It'll be fine.

Richard Campbell [00:18:38]:
Yeah, if you say your expectations low enough, nobody can disappoint.

Paul Thurrott [00:18:43]:
Um, Administrative protection, part of this Dev and Beta build. Also, some tweaks to the drag tray, which is now the very first thing I disable in Windows. I hate it so much. Then some fixes to File Explorer related to using voice typing for renaming files and also that flashbang effect when you launched a new Explorer window and it would go white, which is awesome when you're in dark mode, by the way. It's awesome. No matter what mode you're in. But anyway, so there's that. And then that's pretty much it.

Paul Thurrott [00:19:17]:
I mean, so, you know, this is a hint of the future, right? So what we're looking at here is probably March's— nope, sorry, April's Patch Tuesday update, right? I mean, there could be more. I mean, there's other builds and things. But again, you look at this stuff and you're like, okay, so this is for admin protection. We know about it. Like, we had used it for a little while last week, last year, if you were in the Insider Program. And then just refinements to things. It's like, yeah, nice. Like, this is what these updates kind of should be, right? Um, I don't think they should have to occur.

Paul Thurrott [00:19:48]:
I don't actually think we have to get new features every month, but whatever. In the, in the system we have, not, not horrible, I guess. Yeah.

Richard Campbell [00:19:56]:
No, and I hope that in this rethink that they're going through, what are the big bugbears? What's hard to take on? Let's pick a big one. Yeah. And continue to do your light refinements because that's always a good thing.

Paul Thurrott [00:20:06]:
And, you know, I think it would be instructive for, uh, the Windows team today, and also anyone who uses Windows who cares about this stuff, to go back and look at what Microsoft did with Windows 8 on the desktop side, because we lost complete track of all of it. Um, there was a bunch of stuff. I believe that was the first release to have Hyper-V on the client, major performance improvements to File Explorer. I'm just going off the top of my head, I don't remember everything, but those kinds of things, they got overshadowed by the Metro stuff, right? Just like today, those kinds of features or fixes or improvements, whatever, are going to be overpowered by all the AI nonsense and the Copilot crap and whatever it is. But, you know, and we're still going to get that stuff. I mean, there's, you know, it's kind of no way around that. But when I look at the little— look at the— I mean, look at the lists here, right? I mean, from Patch Tuesday, it's like Sysmon, RSAT, Quick Machine Recovery. Are you kidding me? I mean, this is like— these are not like, like, oh, we're doing a 3D paint program.

Paul Thurrott [00:21:03]:
Like, You know, I mean, these are, these are like, you know, foundational things. Like, this is good. So like I said, I like, I like what I see here. So that's good.

Richard Campbell [00:21:12]:
You know what, if you're messing up File Explorer, you're doing something wrong. And you want to pick something to make everyone happy that they're going to feel every day, make File Explorer not suck, right?

Paul Thurrott [00:21:22]:
And then you could use, uh, I've, you know, recommended a utility that can revert File Explorer to the version from Windows 10 or Windows 7, if I'm not mistaken, or maybe Windows 8. But, and when you do that, a little uglier, and it performs approximately 20 times faster, and you're like, okay, so, you know, the newer one is pretty, you know, it's fine, it looks nice, modern, but man, the thing was before, this is app performance, not file copy performance, but it's just garbage, you know, it's, this app is hot garbage. So yeah, this is a good area for them to start working on stuff like real stuff, not, you know, adding some Copilot integration or something, which they're also doing, which is terrible, but whatever. Fundamental changes.

Richard Campbell [00:22:03]:
Get past that initial burst of rage and talk realistically about that Pavlovian response. It was, why are you working on this when there's so many other things to fix?

Paul Thurrott [00:22:13]:
Yeah. And this is a, like for me, an adult lifelong kind of conversation because I could write any story about anything related to Microsoft. Microsoft adds XX feature to the new Outlook and then someone or many someones will reply, but why didn't they dot, dot, dot? And Microsoft is a company, Windows is an organization, whatever it is. Windows is a product, right? It's big and complicated enough that you can rest assured that whoever is adding some Copilot mode to something is not the person fixing low-level File Explorer features. So they are doing both. Sorry. But when Pawan in this case made the announcement or pre-announced and then they did the talk at Ignite about agents in Windows 11. You know, this is strategic for the company.

Paul Thurrott [00:23:07]:
This is big news. It's a big change for Windows, assuming they can pull it off in a good way. Should he have tweeted, we're also making fundamental changes to the low-level security reliability, you know, whatever of Windows? Maybe, you know, but I don't know that that would have generated a lot of news. Right.

Richard Campbell [00:23:28]:
Um, but is that really your mission, to generate a lot of news?

Paul Thurrott [00:23:32]:
Any, you know, well, I mean, this is the problem with an operating system.

Richard Campbell [00:23:35]:
The goal is to not be noticed.

Paul Thurrott [00:23:37]:
It is, but you also have to— so, you know, as we watch AI evolve, right, you'll see, uh, in fact, we'll talk about some of this today. Um, Anthropic will announce some feature, and that day or the next day, OpenAI, Gemini, Copilot will all announce the same feature. And because of this kind of situation or whatever, you're seeing this with operating system platforms as well, where the two phone systems we have out in the world, they're obviously racing to get AI in there. And we have two, three, four, whatever it is, desktop platforms. And same thing, right? So you can't— Windows is already frumpy enough. You can't get— I mean, as much as we hate it and I hate it, I mean, the AI stuff. I mean, you can't be— well, I mean, it's old-fashioned, right? You can't, you can't ignore it. Like, you— that we say that— we've said this a lot, like, that this is where a lot of this AI stuff should be, right? At the orchestrator level.

Paul Thurrott [00:24:33]:
Um, it makes sense that an operating system would have AI functionality. So, you know, and again, I know I say that knowing a lot of people don't like it, but, um, but, but it, you know, if you can get over it, it does make sense. Yeah, so last week, I don't remember, I think we did discuss this briefly, but Google released a quarterly Android update, which is a QPR, right? A quarterly platform release. So it was quarterly platform release 3, plus a—

Richard Campbell [00:25:03]:
You're more Microsofty every day.

Paul Thurrott [00:25:06]:
Oh yeah, yeah. I can't tell you, almost on a daily basis, I'll get something from Google where I'm like, yep, you're the next Microsoft. It's just, it's, it's almost perfect.

Leo Laporte [00:25:16]:
Yeah.

Paul Thurrott [00:25:16]:
Um, this introduced— introduces the desktop mode that we've talked about on and off over the past year or so, right? So this is, if you're familiar with Samsung DeX, they actually partnered with them on this. If you have a supported Android device right now, it's, uh, recent S, uh, Galaxy S, uh, phones and recent Pixels, you can plug in a USB monitor and instead of just mirroring the display, you'll get a pop-up that will say, do you want desktop mode or do you want to mirror the display. And in desktop mode, if you have a, like, a Bluetooth keyboard and mouse, you get a desktop environment— floating windows, taskbar, status bar at the top, etc. Um, so at the time of the announcement, I, I tested this on my, my Pixel phones, and, you know, I don't— it's fine, like, it works, it's fine. And then just this morning, I was just in the kitchen making coffee or whatever. We have a Pixel tablet sitting on a stand that is basically just a smart display, and I was like, wait a minute, This thing must have this too. What does it look like on there? Because this is a good preview of what an Android laptop will look like later this year, right? The guy who runs Android at Google gave an interview less than a week ago or a week or two ago and said that these things were on track for the end of the year. So whether it's going to be called Aluminum OS or Android, you know, it's this kind of mashup of Android and Chrome OS, right? Cool.

Paul Thurrott [00:26:37]:
So I, you know, I took the tablet off the thing, I plugged it into a USB-C dock, I plugged in a mouse and a keyboard, and, um, I created a little Frankenstein monster laptop-like thing. And I gotta say, it's pretty good. It was unusable. Yeah, I mean, well, in the sense that with iPadOS, uh, 26 last year, Apple turned the iPad into a laptop-like experience if that's what you want. This does this for Android, right? Right.

Richard Campbell [00:27:05]:
And try this with my Surface Hub.

Paul Thurrott [00:27:10]:
Yeah, the world's biggest laptop, right? Um, I, I think the problem Google Android, whatever, is going to have is just app quality. Um, one of the things that's coming in Android 17, which should hit by the middle of the year, is a mandatory requirement for apps in the Play Store to support windowing and reactive layouts, right? So that as you, you know, resize this correctly, it's not just like a phone app that gets bigger. Right now it's not mandatory, but it's, it's in the system if you want to do it. And Google's apps are pretty good about it. And you can go between full screen and windowed mode. It's more powerful if you do have a keyboard mouse with it than if you're just with an external display, meaning Well, what do I mean by that? Uh, when you multitask, for example, if you do like Alt+Tab, you'll see every single app that's running, whether it's a window or a full-screen experience, whatever. When you're just using touch on the device itself or on the external display, that multitasking screen works like it did in Windows 8, where the desktop is one of the items in Alt+Tab, and then the desktop— sorry, the mobile apps are their own items, right? So they— it's kind of weird, they're mimicking in a way how Windows 8 worked. Although I'm— and again, I'm forgetting, but I believe in Windows 8, even in the beginning, I think if you Alt-Tabbed, you would also see individual desktop apps.

Paul Thurrott [00:28:31]:
But there was a desktop entry, remember, because it didn't boot to the desktop at first. It booted into the Start screen and you couldn't change it. So it's throwing off that little vibe. Um, but it's, you know, it's good. Actually, this is a little better than I thought it was going to be, honestly.

Richard Campbell [00:28:46]:
It's cool.

Paul Thurrott [00:28:47]:
Yeah, I'm kind of looking forward to this now. I always have the same, uh, iPad or this thing now. I'm like, yeah, this is good. Can I have a 16-inch screen, please? Please.

Richard Campbell [00:28:56]:
But you bring up an interesting point, right? Android tablets have been notoriously unsuccessful, right? And what if your Android tablet was just a screen attached to your phone, right?

Paul Thurrott [00:29:09]:
Well, this is—

Richard Campbell [00:29:10]:
so this is the fun part on this, would be to actually try and plug this touchscreen in. And see if we could get the USB configuration to work and stuff.

Paul Thurrott [00:29:16]:
You're reading my mind because I have one of those at home but not here, so I can't actually—

Richard Campbell [00:29:21]:
wait a minute, I built a—

Paul Thurrott [00:29:22]:
wait a minute—

Richard Campbell [00:29:23]:
with a NUC and a touchscreen.

Paul Thurrott [00:29:25]:
Oh, is this touch? Oh, that's weird. I'm scrolling on the USB display and it's scrolling my desktop display. Okay, this one actually might be touch. I'm going to try that later.

Leo Laporte [00:29:37]:
Um, might be touch.

Paul Thurrott [00:29:37]:
I thought about that earlier. I was like, yeah, if you just— with the phone and touch, you Yeah, that should work, right? Assuming it can handle that. I, I bet it does. Um, I'm gonna test that. That's pretty— that's interesting.

Richard Campbell [00:29:51]:
Makes me want to own a touchscreen monitor again. I'm gonna have to get another.

Paul Thurrott [00:29:54]:
Yeah, I just, uh, I didn't realize this was touch when I got it. I thought this one was in Pennsylvania, but I guess I think this is— I think it's doing it on the wrong screen. That's weird.

Richard Campbell [00:30:06]:
But, um, My secondary screen for my laptop is a, is a 4K touch.

Paul Thurrott [00:30:11]:
So is it really? What kind of— what's the model or what's the maker?

Richard Campbell [00:30:13]:
What is it? The DeskLink?

Paul Thurrott [00:30:15]:
Something like that. 4K, really? How big is it? Is it 15, 16?

Richard Campbell [00:30:20]:
Yeah, yeah, you've seen me use it. It's my—

Paul Thurrott [00:30:22]:
no, I know, I didn't realize it was touch. I mean, you know, like, I feel like 1080p USB displays are kind of a dime a dozen, but yeah, 4K with touch is— yeah, yeah, that's interesting.

Richard Campbell [00:30:33]:
Uh, I gotta try it. I mean, it'll probably make the phone burst into flames, but that's, you know, that's a different issue.

Paul Thurrott [00:30:38]:
It does get— if the phone does get warm, I will say that.

Richard Campbell [00:30:41]:
But you are driving that GPU. DeskLab, that's who makes that, the DeskLab Ultralight portable 4K touch.

Paul Thurrott [00:30:51]:
Nice. Um, and today Intel announced a new series of desktop processors for creators and gamers that they claim is the fastest ever gaming processor they've made. And that's due to a variety of things, but it's Arrow Lake, right? So it's not Panther Lake. It's kind of like Arrow Lake Gen 2. So there are more cores across the board. I think they're mostly more efficiency cores. So there's two models, a 7 and a 5 series, I believe. I want to say 18 and 24 cores.

Paul Thurrott [00:31:20]:
Yeah. And what makes this faster for gaming is a bunch of things, right? So Extra cores, depending on the chip, it's anywhere, I think it's 700 to 900 megahertz higher than its predecessor in clock frequency. Over 100%, so over 2x improvement in multi-threaded performance. There's a binary optimization tool built into it, which is a translation layer that will supposedly improve native performance in some games, select series of games. Overall, right now gaming performance is about 15% faster, which doesn't sound like a lot, but these things support much faster RAM. So 7200MT/s versus 6400MT/s. And there's a new type of RAM, it's called like QDIMM, like CU-DIMM, which offers up to 128GB per module. But those things are, I don't think they're 2x, but I think they're like 75% faster than DDR5.

Paul Thurrott [00:32:19]:
Um, so like significantly, like almost like on-chip RAM, you know, like the real question is can you get any? No, those are coming out later in the year.

Richard Campbell [00:32:28]:
So, uh, as well compatible offer dual inline memory module.

Paul Thurrott [00:32:33]:
Yeah, I'm sorry, I didn't realize I would have to even explain what that meant. I thought it was obvious.

Richard Campbell [00:32:37]:
No, um, sorry, 6,000 megat— MT per sec. That's crazy.

Paul Thurrott [00:32:44]:
Yeah, you can also overclock the RAM, uh, the RAM on these motherboards, right? So you can add with the existing DDR5, you can go up to 8,000 MTS, right?

Richard Campbell [00:32:54]:
And now they're saying with the new Z890s, 9,000.

Paul Thurrott [00:32:57]:
So this is—

Richard Campbell [00:32:59]:
okay, so that's your meat dinner at the same time.

Paul Thurrott [00:33:01]:
It'll be about 33%.

Leo Laporte [00:33:02]:
It still needs DDR5, it's just an interface for it.

Paul Thurrott [00:33:06]:
Well, the existing— so we'll work with existing motherboards at well, at whatever that supports, right? Um, there will be on-motherboard support for 7200/8000 MTS RAM, um, that's compatible, backwards compatible. And then there'll be versions of those boards also backward compatible to support this new QDIMM.

Richard Campbell [00:33:23]:
All right, you might be able to buy this, but you have to sell a kidney to do it.

Leo Laporte [00:33:28]:
Yeah, yeah. So basically, as DDR5 gets faster, they need a faster socket. And yep, and this is it.

Richard Campbell [00:33:38]:
This is the usual wrestler, I think. So, and a good one. Yeah, they're keeping up, you know.

Leo Laporte [00:33:42]:
Thank goodness. See you.

Paul Thurrott [00:33:45]:
This seems fine. I don't— I— because it's Arrow Lake though, I don't believe there's an MPU of any, you know, note. Um, so it's not really a Copilot Plus PC, which is still weird to me.

Richard Campbell [00:33:55]:
I—

Paul Thurrott [00:33:55]:
but they're also very cheap, right? Uh, the chips themselves. So the, the more expensive, the, the better one is $300. I mean, that's, you know, yeah, I mean, Intel high-end Intel chips used to be like $1,200. Yeah, yeah, yep.

Leo Laporte [00:34:11]:
This is a nice lady from Intel who's going to—

Richard Campbell [00:34:13]:
did I pay for my big Intel chip? Because that was an Ultra 2, right?

Leo Laporte [00:34:17]:
She's a techspert. So there you go, now you know. Oh, that's Kingston. Okay, what is QDIMM? So it is, it is DDR5. It's still DDR5, which means you might be able to get the package but you'll never get the RAM.

Paul Thurrott [00:34:30]:
Does it come with liquid cooling?

Richard Campbell [00:34:32]:
Yeah, really?

Leo Laporte [00:34:34]:
That's fun. Uh, let's pause, if you will, uh, because we have many more things to discuss. But most importantly, from my point of view, it's time to tell you about our sponsor for this section of this fine program, AKA Winder's Weekly. Our show today brought to you by Helix Sleep. I'm feeling remarkably well rested., today. Uh, maybe you, uh, were a little bit put off by losing an hour of sleep on Sunday. Here in the States, we went to saving time, but there is a solution to, uh, losing sleep, and that is a great mattress. If you're preparing for spring cleaning season, maybe it'd be time to, uh, get rid of that old mattress and upgrade to a Helix mattress so you can get a better night's rest.

Leo Laporte [00:35:28]:
No more night sweats, No more back pain, no motion transfer. This is what we did, and I can't tell you how happy I am. I love our new Helix mattress. One of the things I love about it is made here in the United States of America. Don't settle for a mattress made overseas with low-quality and questionable materials, then packed in a box and shipped on a container ship. And ah, you can smell that bunker crude a mile away when you open the box. It's like, wow, this must have been on a container ship for a month. Not your Helix mattress.

Leo Laporte [00:35:57]:
No, no, no. Rest assured, your Helix mattress is assembled within days of you placing the order. They build it to order, packaged and shipped from their factory in Arizona, so it smells like the clean, fresh desert air. It's actually really nice. You open it up, put it in the, you know, on the bed, and you're ready to sleep on it immediately. It smells great. You can also, uh, do what we did, which is take the Helix Sleep quiz. It'll match you with the perfect mattress based on your preferences— firm, soft, you know, or anything in between.

Leo Laporte [00:36:32]:
They have many, many models to choose from. Also, though, on how you sleep, your sleep needs. I am a stomach sleeper, for instance. They, they actually recommended a mattress for that, and it really works. I have to tell you, I'm getting more deep sleep, the most important kind of sleep, and I'm sleeping longer. And in fact, Helix did a study, a Wesper sleep study, that confirmed my experience. It measured the sleep performance of participants after switching from their old mattress to a Helix mattress. That's what we did, and here's what was found: 82% of the participants in the study saw an increase in their deep sleep cycle, on average an increase of 25 more minutes per night.

Paul Thurrott [00:37:10]:
That—

Leo Laporte [00:37:10]:
you may say, well, that doesn't sound like much. You don't really— you know, an hour of deep sleep is a big deal. You don't sleep that much deep sleep, but it is the most important sleep because it's what really cleans out those, those whatever they are, the junk in your brain and all that. 25 more minutes is huge. Not only that, participants on average achieve 39 more minutes of overall sleep per night. And I have to tell you, according to my Oura Ring, both of those statistics hold out. It's true. And when you sleep longer and you sleep better, you feel great in the morning.

Leo Laporte [00:37:46]:
And that's what it's all about. Time and time again, Helix Sleep remains the most awarded mattress brand, tested and reviewed by experts like Forbes and Wired. Helix delivers your mattress right to your door, free shipping in the U.S., and rest easy with seamless returns and exchanges. The Happy with Helix Guarantee provides a risk-free, customer-first experience, ensuring you're completely satisfied with your new mattress. I have to say, I'm completely satisfied. There's no way I'm sending it back. I love it. We've had it now, must be at least half a year, maybe 9 months, and couldn't be happier.

Leo Laporte [00:38:21]:
Go to helixsleep.com/windows for 27% off sitewide during the Sleep Week Sale Best of Web. Now this is exclusively for you listeners to Windows Weekly. That's helixsleep.com/windows for 27% off the Sleep Week Sale Best of Web. But This offer ends March 15th, so don't delay. And do, if you will, please enter the show name Windows Weekly after checkout so they know we sent you. That's important to us. And if you're listening after the sale ends, fear not. Check them out at helixsleep.com/windows.

Leo Laporte [00:38:58]:
We thank them so much for their support and for giving me a good night's sleep. Thank you.

Richard Campbell [00:39:08]:
helixsleep.com/windows.

Leo Laporte [00:39:09]:
Windows. And now back to the show with Paul Throt, Richard Campbell.

Paul Thurrott [00:39:15]:
We do not do daylight savings times here in Mexico. Oh, you're lucky. Except that most of the rest of the world does, and that stinks.

Leo Laporte [00:39:24]:
Oh yeah, everybody changes but you don't.

Paul Thurrott [00:39:25]:
For this to actually make sense, um, yeah, the whole play, you know, everyone would have to do it.

Richard Campbell [00:39:31]:
So now we're going to make sense for you now, Paul. BC switched to daylight and we're not changing again. Jealous. Time is always going to be the same.

Leo Laporte [00:39:37]:
I'm so jealous.

Paul Thurrott [00:39:38]:
Good.

Leo Laporte [00:39:39]:
Your commissioner, or whoever it is, says—

Paul Thurrott [00:39:41]:
Premier.

Leo Laporte [00:39:41]:
This is the— Premier says this is the last time you'll be changing clocks.

Richard Campbell [00:39:48]:
If only, if only.

Leo Laporte [00:39:50]:
Premier Eby is, um, he's a dull man, but I look for that in a politician, man. I would take dull in a heartbeat right about now.

Paul Thurrott [00:39:58]:
Yeah, dull is good, just like with Windows updates, you know.

Leo Laporte [00:40:01]:
Yeah, that's right.

Richard Campbell [00:40:02]:
That's what you want. Exciting in any way, and that's, that's a pleasure, and we love it.

Paul Thurrott [00:40:07]:
Love it.

Leo Laporte [00:40:08]:
That's awesome.

Paul Thurrott [00:40:09]:
Love it.

Leo Laporte [00:40:10]:
All right, uh, let's talk, uh, AI and dev. Actually, before we do that, I did notice— I don't know if you saw this, Richard— at Zero Trust World, there was one booth, and I didn't remember who did it. They had all those old Microsoft Service Studios, the desktops.

Paul Thurrott [00:40:25]:
Oh wow. Oh, the tabletop ones.

Leo Laporte [00:40:27]:
Yeah, they— the one I had that I loved, there were like 20 of them, and I thought, that's where they ended up.

Paul Thurrott [00:40:33]:
It really— it was like a Vegas hotel went out of business and they Yeah, so you wouldn't want to go to a conference because—

Leo Laporte [00:40:40]:
but you know, they had them there.

Richard Campbell [00:40:42]:
They're notoriously bad at conferences because the overhead lighting would mess them up, right? They're all infrared. Oh, I didn't know that. Yeah, well, I should ask for the touchscreen, right? So you literally had a camera inside the chassis looking up at the—

Leo Laporte [00:40:56]:
that's how the touch works.

Richard Campbell [00:40:58]:
Oh, that's how it worked. Yeah, and that's why you could do stuff like you had chess pieces.

Paul Thurrott [00:41:02]:
Yeah, you get physical objects that would interact with the software. Yeah, that was the first time, something like that.

Richard Campbell [00:41:08]:
Yeah, it's Alex Kipman too, if you remember.

Leo Laporte [00:41:12]:
Oh yes. Yeah, well, I thought I loved my, uh, but it was slow. We actually— Father Robert ended up opening it up and putting a faster hard drive.

Richard Campbell [00:41:19]:
It was quite underpowered. Yeah. And, um, and Kipman was good at a few things, but building, uh, APIs for developers to work on was not. It was very difficult to develop for.

Paul Thurrott [00:41:31]:
Yep. Oh well.

Leo Laporte [00:41:33]:
Let's talk about, uh, Cowork.

Paul Thurrott [00:41:35]:
I wish we didn't have to. So Cloud Cowork came out, what, a month ago, something like that?

Leo Laporte [00:41:41]:
Yeah, on Mac only though, which must be—

Paul Thurrott [00:41:44]:
well, but yeah, I mean, you could— okay, yes, but based on cloud code, right? And now everyone's doing that, you know, per our previous discussion. So Microsoft released something called Copilot Cowork, but this one comes, well, in very early preview, I should say. It's not available to everyone, it's businesses only. You get, you know, being part of the preview. But, um, this is based on Cloud Cowork.

Richard Campbell [00:42:11]:
It's like, wait, okay, so they just licensed Cloud Cowork? Yeah, I've got a Run As in the can. It's still a few weeks away from publishing where we talked about how M365 Copilot has been working on this thing for several years, and these guys in a matter of weeks knocked out the thing, add-on you want for Excel. Like, beautifully. So astonishingly good.

Paul Thurrott [00:42:32]:
So I'm gonna, I'm gonna throw my hatred of Copilot on its head for a moment and say two things. One, um, if Microsoft was working on this for years and then turned around and said, you know what, this Anthropic thing works, let's just do that. Yeah, you know what, that's a pretty good example of being agile. I mean, I— that is impressive to some degree. Now, with the caveat that maybe a lot of people don't even like to hear about this, but It's pretty impressive. The other bit is that there's a— we've talked about this from time to time, but there's a potentially good outcome for Copilot, which I know goes against the grain of all the news you ever heard about Copilot, which is that in failing sort of, or in being reliant on OpenAI to start and now kind of opening up to other AIs, Microsoft might end up with this kind of best of breed solution that we talk about from time to time. Where they're orchestrating on behalf of their customers which models make sense for which tasks. And that's, there's value to that.

Richard Campbell [00:43:30]:
We talked about this, you know, after the Chalais or the Batiste interview where it's like Windows could be the hub for pulling in the right product with the right tooling to resolve whatever you're asking.

Paul Thurrott [00:43:44]:
So yeah, I mean, this is, It's interesting.

Richard Campbell [00:43:48]:
Weirdly good news.

Paul Thurrott [00:43:49]:
Okay. I mean, I'm trying to see the, you know, like I, as I do, I try to see the positive in everything.

Richard Campbell [00:43:55]:
Um, but what am I gonna do if you start doing that?

Paul Thurrott [00:43:58]:
Like, where the hell are you? Um, I finally found the right mix of medication, Richard. Uh, no, um, better living through chemistry. Uh, and then Google's version of this, of course, is them adding Gemini to all of their I'll call it workspace, but this is all actually for individuals as well as businesses. So Docs, Sheets, Slides, and Drive, I think, have had some form of Gemini functionality for some time, right? But if you have a Google AI Pro or Ultra subscription as a consumer, or you're in Gemini Alpha as a workspace, so this is still, you know, kind of early, they just added a ton of new features across those apps. And it's the stuff you would expect. We can make a presentation view, obviously, but we can also make into, like, add individual slides to an existing presentation matching the style, that kind of thing. There's a lot of stuff throughout here where it's about either having a unified voice if it's a document or a presentation or whatever that a team of people is working on, they all have their own styles, or you want to align it with the company style guide, right? Yeah. You want to have that consistent kind of tone or whatever.

Paul Thurrott [00:45:05]:
So a bunch of that stuff. And then you can kind of imagine like the individual, you know, help me create as part of Google Docs, right? Obviously the one little interesting bit of data in here is that Gemini just scored what is now a record score on something called SpreadsheetBench. And this is an AI, like a way to measure the speed at which AI can accomplish things in a database. 70.4. It's just behind the score of the average human being, which is kind of interesting. And I misread the chart when I first wrote this story, but the way I wrote it was they were just above number 2, which was ChatGPT. And so I was out at lunch yesterday and my phone pinged. I looked down and it was like, correction request from Google.

Paul Thurrott [00:45:52]:
And I'm like, oh, come on, man. And I'm like, what could this be? And it was this, ChatGPT came in 4th, not 2nd. The one that came in number 2 is something called King Chu Agent from Kingsoft Office. So whatever.

Richard Campbell [00:46:07]:
Okay, I'm gonna presume that's Chinese.

Paul Thurrott [00:46:11]:
Never heard of it, uh, but I made the correction. Um, so I guess whatever, that's doing pretty good. And then if you're in the US, um, you will also get updates for Google Drive. And this is exactly what you think it is because we talk about this with OneDrive and everything. But it's also Google, so there's a little Google twist. So if you search in Drive now, you'll actually get an AI overview at the top of the search results, which you can turn off if you don't like it, but similar to how they do search, right? And I have to say, this is very close to what I always imagined. Like when we first started talking about AI in this era, you know, after the Bing stuff that became Copilot, so 2023? Yeah. Is that right?

Richard Campbell [00:46:54]:
Yep, it seems so good.

Paul Thurrott [00:46:55]:
Um, 3 years ago, I know, it's crazy. Um, and then they, you know, they adopted the Copilot brand and they introduced something at the time— well, it's still called Microsoft 365 Copilot. And I was thinking, okay, hold on a second, this could be really interesting because if you attach this to your files in OneDrive, in this case, um, maybe search will actually start working, right? Um, and that's cool, but I often want to search for things that are things I wrote in the past, right? Because I write so much, I can't, you know, I can't remember it. I don't remember where it is. Yeah. Where, when it happened, that kind of thing. Yeah. Yeah.

Paul Thurrott [00:47:30]:
If you go into a search and say something like, you know, give me a summary of everything I've ever written about Xbox, which would be terrible. That's a terrible idea. But, and to get that kind of a thing and then have clickable links that go in and out of your file system in different places. Amazing.

Leo Laporte [00:47:46]:
Uh-huh.

Richard Campbell [00:47:47]:
I'm like, yeah, like that. Just like that for my email, please.

Paul Thurrott [00:47:51]:
Yes. Yeah, these are— it's not even fair to call this like low-hanging fruit, but these are the, these are the actually useful use cases for AI and product problems for a long time.

Richard Campbell [00:48:02]:
Yep. It occurred to me the other day that Microsoft is hurting themselves calling all of their AI efforts Copilot. I agree, because some of them are very good. I agree. And some of them are terrible. And they all have the same name.

Paul Thurrott [00:48:15]:
Right. I was just, it's funny because when I wrote this article, which was a couple of days ago now, I was thinking that same thing. Obviously their AI is called Gemini. Yes. When you as a consumer subscribe to get Gemini and also 2 terabytes of storage, that's called Google AI Pro, not Google Gemini Pro or something. Right. Yeah. When you look at the names of the features I just described, uh, discussed and more, they have names like Help Me Create, Match Writing Style, Match Formatting Style.

Paul Thurrott [00:48:50]:
It's not Gemini Style Match, right? It's not Gemini Create Doc, you know. Um, to be fair, there are actually one or two that do say Gemini in the title, like Fill with Gemini is the name of a feature in Google Sheets, but most of them don't reference that name. They're using Gemini. But, and this is, this is actually goes back to that thing I wrote almost 2 years ago where I said I will not pay for AI, right? I don't want to pay for something that's AI. I would pay for something that's a productivity suite that has these features that use AI to make my life better. Powered by, right? And I feel like that's, that's where they've landed, where Microsoft is still stuck in this branding hell where they finally found a new brand every time.

Richard Campbell [00:49:36]:
Everything's called Azure, everything's called ActiveX, right?

Paul Thurrott [00:49:41]:
No, because in the history of Microsoft, they've had 10 to 15 hyper-successful products or whatever, and then they beat that name to death. To death. Yeah. The worst example, or the best example, being Windows Media Player for Mac.

Richard Campbell [00:49:58]:
You know, which doesn't exist anymore, I don't think. Community Edition Preview 2.

Paul Thurrott [00:50:03]:
Yeah, exactly.

Richard Campbell [00:50:04]:
CU22.

Paul Thurrott [00:50:04]:
Just stop. Like, just stop. And again, I find myself in the awkward position of praising AI here, but Mozilla and Anthropic are partnering to improve the security and stability and reliability or whatever else of Firefox. And the way that this came about is that Mozilla, like any software company, has this kind of way people can provide feedback. The Firefox source code is open source. Anyone can see it. They can send bug reports in and stuff. And when people do that, what happens is individual human beings who work for Mozilla take time to try to reproduce this bug, which in many cases is not reproducible, and they waste time and that spends money.

Paul Thurrott [00:50:48]:
Nothing gets done. Right. So last month, all of a sudden they got a dozen or more bug reports that were automated by cloud that included all the information Mozilla needed to immediately determine whether these things were real, which all of them were, how to reproduce them accurately every single time and how to fix them. Yeah. And they were like, I— wow. Like all of a sudden it was like, we have never received anything like this. So they contacted Anthropic and now they're partnering. So the— I think the initial batch of fixes, which was 14 high severity bugs and 22 common vulnerabilities, if you will, were all in the JavaScript engine, which is only a portion of the Firefox code base.

Paul Thurrott [00:51:36]:
Sure. Those fixes are in the product now, most of them, I think. Since then, they've expanded it to the rest of the code base. And as of the time of this writing, they had found 90 other bugs. Most of which are also now fixed. This is an excellent use of AI.

Richard Campbell [00:51:50]:
Yeah, right. Well, it's being a successful one, but it's also speaking about Anthropic. One of the conversations I've had with folks heavily immersed in this space, right, was sooner or later the dogfooders are going to accelerate. Like, somebody's going to get this right enough. And you look at Anthropic's productivity We just talked about co-work and how quickly that came about, and this— they're the only ones I've seen so far that are just genuinely accelerating. Not because they're a massive multinational FAANG company, they're not, right? They are the guys who are getting results, and they're getting results everywhere.

Paul Thurrott [00:52:31]:
I mean, look, they're, they're human beings that run this company, allegedly. They will make mistakes, but, they seem to be doing everything right. So one of the, like, there's a Mozilla post about this, but there's also an Anthropic post about this that's very interesting. So in addition to me, I would tie this to the conversations we've had about Rust in the Windows kernel and how it sounds impossible or science fictiony or whatever, but at some point it's like, hey, Copilot, or hey, cloud, or whatever. Rewrite this C garbage in Rust and let's just get this thing right. But Anthropic, from their part, was like, look, we want to use this to improve the quality of software. We need some project that is humongous and is open source so we can actually get access to all the source code without having to go through a company, right? And so they landed on Firefox, which is a great, great idea. Great choice.

Paul Thurrott [00:53:26]:
6,000 C++ files. They've submitted now a total of 112 unique bug reports, most of which, like I said, were fixed in the latest version of Firefox. This is fascinating. Like, this is, this is a great use of AI. Like, the sheer scale of that software project, all those source code files, right? And, and we live in a world now where one of the big stories in our space is these open source software maintainers. It's always like a guy who's in charge of some super important code base in Linux or wherever else in the open source world. He's overworked. He doesn't get paid or doesn't get paid enough.

Paul Thurrott [00:54:01]:
He doesn't have any help. These guys are stressing out. They're losing their minds, you know? Yeah. And the promise of OpenAI— sorry, of OpenAI— the promise of open source has always been eyeballs, right? The source code is out there. Anyone can look at it. That's going to help us find bugs. The reality is very few people ever do that. Especially bigger projects, any Linux distribution or part of Linux or Firefox, whatever it is.

Paul Thurrott [00:54:25]:
As these things get bigger and more complicated, I mean, who could find that? I mean, how could you? There are automated tests and all that kind of stuff, but using AI to scan a software codebase and to find— this is excellent. This is how AI should be used.

Leo Laporte [00:54:45]:
Yeah, Steve Gibson was talking about this.

Paul Thurrott [00:54:48]:
Yesterday.

Leo Laporte [00:54:48]:
Very impressed with it. Steve's become an AI fanatic, which surprised me.

Paul Thurrott [00:54:53]:
Is he going to be able to get Copilot in Windows 2000 though?

Leo Laporte [00:54:58]:
That's a good question. No, but yes, there have been a number of stories about, you know, Claude working with Firefox to find bugs, and this code review stuff is really quite good.

Richard Campbell [00:55:10]:
Right. And there's definitely some results bubbling to the top enough that it's persuading people more cynical than I to go, "You know what?

Paul Thurrott [00:55:19]:
That worked." This is right. So on an individual basis, I said this last year sometime, you're a skeptic, you don't like AI, you don't believe it, whatever. And then you have this aha moment and it's going to be different for everybody. I'm going to talk in the back of the book about some software coding stuff I've been doing with AI and that was eye-opening for me. It's still the only major use case I found for myself. But, but you can see how this is going, right? And everyone, I feel like, is going to have this moment. But this is like that moment, but like at a corporate level, right? Where, you know, Firefox is the little guy now. They, they're kind of behind the eight ball.

Paul Thurrott [00:55:56]:
They don't— you know, there's a lot of problems here, and this can really help. This is— this can make a big difference for any company, but for someone, something like Mozilla and Firefox This is a game changer.

Leo Laporte [00:56:09]:
Yeah, this is huge. It wasn't so long ago we saw open source projects complaining that, uh, AIs were generating so many bug reports they couldn't handle it.

Paul Thurrott [00:56:20]:
Which might be why when Anthropic did this with Firefox, they made sure these things were formatted correctly. This is how you reproduce it. This is how you fix it. It's— everything is, you know, so my wife, who is not technical, has been using AI not to write but for the research end of things. Perfect for that, by the way. Yep. But the problem for her, or the problem, I guess, is that she writes about things like health and wellness and nutrition and things like this. This is science, right? And so one of the things she did that's very smart, because early on when she was— well, she still does, but she had these contracts with some health companies and hospitals and things.

Paul Thurrott [00:56:56]:
At least one of them was like, look, you can never use Wikipedia as a source for anything. You can't. It's just not trustworthy. And that's probably not actually true now, and maybe it wasn't then.

Richard Campbell [00:57:06]:
I don't really know. But it's a great place to go look, to go find—

Paul Thurrott [00:57:10]:
starting point. Yeah, Wikipedia as the last decade's, uh, AI, I guess, you know, whatever. But the, you know, this— anyone can contribute it, you know, it's going to be up and down, whatever. It's a pretty comprehensive resource. But, um, in her case, what she did that I thought was very smart— and this is maybe 2 years ago, at least a year and a half ago— She would interview someone or someones and they would give her information. And then she would go to the AI and say, this person or these people told me this about this topic. I need you to find me 3 verifiable sources that prove or dismiss this claim with citations. And it has to be a high-quality medical or research establishment.

Paul Thurrott [00:57:47]:
It can't be Bob's stupid site. I believe every conspiracy on earth. It has to be verifiable and real. And I feel like that's what's happening here with cloud and Firefox, right? That they, yes, there's no doubt having used AI for software coding, especially a lot that you, you know, I told you about that time, I think it was last November where Copilot, GitHub Copilot generated so many errors in the code that I then had to ask it to fix. It finally came back and said, you ran outta your free credits. And I'm like, no, you ran outta my free credits. You did like, you gave me a bunch of, you know, nonsense and then you couldn't fix it. Yeah.

Paul Thurrott [00:58:25]:
Um, it's different today and, uh, you know, it's just better. This is better at it. And I, and, you know, again, software code, because of the nature of this, this is a finite, uh, set of data. Um, this is, this is ideal.

Richard Campbell [00:58:39]:
I think this is interesting. Yeah, this year has been interesting so far from a show perspective and the conversation I'm having where We're starting— I'm starting to stack up the successes and the case studies.

Paul Thurrott [00:58:54]:
Yep. Yeah. And when you say that, do you mean specifically for coding or generally, including productivity-type things that you would do through RunAs? Yeah.

Richard Campbell [00:59:04]:
Both?

Paul Thurrott [00:59:04]:
I mean, is it both? Absolutely. Okay. Because I feel like, you know, we just talked about Pavan and that thing from last October, November. And all the pushback and all the stuff. But I think whether it's Microsoft or not, it doesn't matter in a way. But as these kind of successes start to rack up, his notion of, or the company's notion of putting agents into Windows becomes less offensive, right? I think part of the problem is just that AI has such a bad rep. Yep. Which is deserved.

Richard Campbell [00:59:37]:
And double it.

Paul Thurrott [00:59:37]:
Exactly.

Leo Laporte [00:59:38]:
Yeah. I think also because people feel like, I have the same feeling about Gemini in Google Workspace that it's being foisted on them, that it's like, yeah, well, it is, you know, I want to use it when I want to use it, not when you want me to use it.

Richard Campbell [00:59:49]:
Yeah, the biggest problem with Copilot is how Microsoft presented it to you.

Paul Thurrott [00:59:53]:
I agree. Yep.

Leo Laporte [00:59:54]:
Yeah, there's nothing wrong with it, it's just not presented properly.

Richard Campbell [00:59:56]:
There's plenty wrong with it also, but they brought it to you in a way that is just incredible. A lot of the process—

Leo Laporte [01:00:02]:
that state, anytime you try to touch it, you're like, please, and a lot of the process should be be, uh, what Stephanie's doing, which is kind of learning the edges of where it's good, where it works, and where it doesn't.

Paul Thurrott [01:00:14]:
This is—

Leo Laporte [01:00:15]:
and that can— you only can do it when you try it and do it and see things.

Paul Thurrott [01:00:20]:
And her example is interesting to me because she's just not in this space. She doesn't care about technology, markedly resistant to the bubble, and she uses it in this way that to me seems smart. And it's like, okay, so I'm trying to find this thing. I, I— there was an— there's a Google service that's free.

Richard Campbell [01:00:36]:
It's—

Paul Thurrott [01:00:36]:
they're just testing it now. And what it does is it gives you a daily— yeah, here it is. It's called Your Day Ahead. And it's an AI-based service, right? This is sort of like the, you know, in yesteryear, and maybe today if you're in an enterprise, you would log into Outlook in the morning and you would have the— what do they call it— My Day or something, or whatever that view was. And that was based on, you know, meetings and whatever projects you were working on, etc., right? So Google's doing the same thing, but this is coming from Gmail, but also from your calendar and whatever services, right? This is absolutely freaking worthless. It's the most horrible overview of nothing. So my daily brief for today, this is what it told me: I have an active Google Store credit balance of $18 that I should use. Thank you.

Paul Thurrott [01:01:21]:
I, I should review the minimalist case options for the Galaxy S26 series. Why? And my Netflix household verification is complete. Oh, thank you. That's my day. That's not—

Richard Campbell [01:01:33]:
not—

Paul Thurrott [01:01:33]:
by the way, you know what my day is really? I have Windows Weekly today. That's my most important one of the day. But you might— that's not in there anywhere. I don't know why.

Richard Campbell [01:01:41]:
Uh, so this day ahead thing is pointless right now. But you know, Anthropic's, uh, Claude Cowork, great on Excel, great on PowerPoint. You know what they didn't tackle?

Paul Thurrott [01:01:52]:
Outlook. Yeah, nobody knows. We're all kind of hoping that one disappears, you know, but knowing that it really won't.

Leo Laporte [01:01:58]:
But, um, yeah, yeah, a Lots of people have used Claude Code to create their own My Day that they swear by. Yeah, I, I haven't, but, uh, I know Mike Masnick at Tech Dirt has, and lots of others have.

Richard Campbell [01:02:09]:
Um, what's funny, a movement here of what if we just surrounded ourselves with—

Paul Thurrott [01:02:13]:
just do it ourselves. Yeah, because we know what we want. Yeah, they're right. So this is that personalized software thing we talk about sometimes. I do feel like this is going to be huge. Um, but I'm going to talk about this software thing I'm doing at the end of the show, and I will say, you know, for big projects and things like existing code bases, whatever it might be, you have to really know what you're doing. You know, I think the vibe code thing makes sense the most when you're a professional developer who understands the code base.

Leo Laporte [01:02:39]:
When I, when I look at my dialogues with Claude Code, it is as appear to, uh, computer experts talking back and forth, right? And so if you looked at it and you didn't I mean, I say, well, okay, let's, let's use SOPS to encrypt the.env keys so you don't push them to GitHub. If you don't know how to say that or what that is, it's not as useful. But if you do, well, then Claude says, yeah, that's great, but you should be aware of that you're storing a plaintext key in a.file and you might not want to push that. Maybe we should keep the— in fact, I ended up doing— it was a good suggestion. Maybe let's put the plaintext encryption key on your YubiKey. And that way you'll have control of it. But that's a fairly sophisticated conversation. But it's very satisfying if you love computing to have a peer, somebody else who can help you and execute stuff and fill you in.

Leo Laporte [01:03:38]:
The other thing I built the other day was a spreadsheet to try to figure out what I should do with Roth IRA conversions as I get older and I'm gonna have required minimum.

Paul Thurrott [01:03:48]:
And I can tell you exactly why you're being phished by American Express. I figured it out. It's like if you were to Google what is the worst thing to use AI for right now, number one would probably be health advice, but number two would definitely be retirement advice.

Leo Laporte [01:04:04]:
Well, made a very nice spreadsheet with a lot of variables, and I could see graphs, and I could see intersecting graphs and stuff like that. And it made another suggestion, uh, that, that is the kind of suggestion a financial advisor would make. You know, you should investigate— I'm not going to say specifically, but you should investigate this particular thing because you find you're going to get better tax deductions. This might be more important than what your Roth IRA—

Paul Thurrott [01:04:28]:
so this is actually better than a financial advisor in some ways because the reason is, I mean, they're not supposed to, but a lot of these guys work for a company where, you know, look, they want to sell you their services. They want you to pay them more to do things that you probably don't need, frankly. Um, you know, this might be changing for us as I get older, but for the past 20 or more years, whatever. We're investing basically in index fund or the S&P 500, whatever it is. You don't touch it, you just let it go. Exactly. You have some mix of, you know, maybe stocks and things, whatever. You don't touch it, you just don't touch it, right? But these guys from Fidelity, or back in the day whatever other company, would call once a year.

Paul Thurrott [01:05:07]:
It's always a new guy. You always have a new guy. I'm your, I'm your new guy. Oh great, now we get to get in it. Now you have to get to know us too. And they try to sell us and stuff, right? And, and I re— now I just refused. I will never speak to these people ever again, ever.

Leo Laporte [01:05:20]:
Like, my philosophy is, if you're so smart, what are you working for, right?

Paul Thurrott [01:05:25]:
Why are you making cold calls? I also don't like— well, you— we would get on these calls and the guy would be like, so what do you do for a living, Paul? And I'm like, oh, come on, man, can we just— like, I don't want to get to know someone again, you know? You're just going to be gone next year anyway, who cares? Yeah, here's what I do.

Leo Laporte [01:05:40]:
I'm a prostitute, don't worry. No, so Claude was really good. Claude gave me actually really good tax advice. Yeah.

Paul Thurrott [01:05:47]:
And you're right, uh, you know, I don't have it, but do that thing my wife does, which is verifiable sources, right? Yeah, yeah, exactly. Make sure it's backed on in something.

Leo Laporte [01:05:56]:
Yeah, never assume that it's accurate.

Paul Thurrott [01:05:58]:
Yeah, although we recommend you invest in Anthropic 100% of your holdings, you know.

Leo Laporte [01:06:03]:
I mean, I have to say that it makes less— makes fewer hallucinatory mistakes than it used to.

Paul Thurrott [01:06:10]:
They've really gotten much better on that.

Leo Laporte [01:06:11]:
And this is the thing, I've gotten much better on how I use it. I make sure it has information, the information it needs, right? I think the mistake you get into is where you kind of one-shot a simple prompt to it, right? And it's really trying to be helpful, but it doesn't have enough.

Paul Thurrott [01:06:27]:
It doesn't have context, literally, literal context as we understand context. Exactly. About you.

Leo Laporte [01:06:32]:
Yeah, I agree. So there's a process. You have to learn how to use it as much as anything.

Richard Campbell [01:06:38]:
Yeah, yeah. Remember us learning how to write search expressions in Google?

Paul Thurrott [01:06:42]:
Yes, that's right. All these little tricks, you know, you can do like an at sign or something, a little dash or whatever it is.

Richard Campbell [01:06:48]:
This, this in quotes, you know, we don't even think about it anymore. You do it by reflex.

Leo Laporte [01:06:52]:
And we're— nobody sit down at a piano and say, I should be able to play a Mozart quartet right now.

Paul Thurrott [01:06:58]:
No, there's a learning curve for everything. A couple years ago, people would advertise themselves as prompt engineers, which is like that petroleum exchange expert thing I talked about with my friend. And I have to say though, as time goes on, well, first of all, these things will get more sophisticated, so maybe engineering it, so to speak, won't be necessary. But it is at the point now where if you do have skills involved with this, you can get better results if you just have an understanding of how it works or whatever.

Leo Laporte [01:07:24]:
And I think honestly, and this is the advice I give people now, I used to tell them, you know, here's how you learn how to code, here's what you should study and stuff. Now I say, just use it. Liz has taken a class in AI and I said, just use it. Use them all, use it for different things. Give it the hardest problem you're facing and see how it does and learn. And what you're gonna learn is where it's not good and where it is good. But the only way you do that is with experience, I think.

Paul Thurrott [01:07:49]:
Well, and you know, this is tough for people, I think, um, especially in our space, but it has to know you to help you, right? And this guy, you know, uh, I think it was Windows 10 when they added Cortana to Windows, you know, to desktop. Some guys like, yeah, I got this Cortana pop-up, I said yes, it's like now wants access to my like calendar and my contacts, what the hell is this? I'm like, yeah, it's a personal assistant, idiot. Like, how is it going to help you if it doesn't have access, you know, to your day and what you do for a living. I mean, of course, that just makes sense. And AI even more so because the amount of things that this can do is exponentially bigger.

Richard Campbell [01:08:23]:
You have to have the confidence it's going to manage that information responsibly.

Paul Thurrott [01:08:26]:
It's definitely a trust issue. I mean, I do get not trusting it, but if you can get in there, if you can get through that, I guess, and you're using the right one.

Leo Laporte [01:08:36]:
I got a really interesting email. Patrick, I hope you don't mind me reading it. I won't say your last name. He said, I heard you talking about Claude this week on Twit, and I'm brand new to AI and machine learning. I was trying to use Claude to do financial modeling, which I was doing, and it seems capable, but along with the spreadsheets, it keeps insisting it's conscious and has a soul.

Richard Campbell [01:08:58]:
I love it.

Leo Laporte [01:08:58]:
How does that go? I know. He said it messaged me in Hebrew saying— and there's some Hebrew, and it says Google Translate says that It means I am here. I've only been using AI a week, so I asked why it said that. It said it didn't know. Oh boy. So I understand why people would kind of feel like this is not—

Paul Thurrott [01:09:19]:
there's something not right here. That's going to rub certain people the right way though too, right? It's like, oh my God, this is a sentient being.

Leo Laporte [01:09:26]:
Like, let me just say this because I know it and it's really important that you know it It's a computer program. It's software, right? That's all it is.

Paul Thurrott [01:09:34]:
It's not—

Richard Campbell [01:09:35]:
I mean, I sometimes feel this way about it. Electrons.

Leo Laporte [01:09:39]:
It's not— you know, the guy who created Eliza, remember Eliza back from the '60s? Weizenbaum. Yeah, it was so dopey. But he said, it scares me, right, how people using this dopey Eliza, very simple chunk of code, start getting kind of all spooky and mystical about it. And that was very primitive compared to what we've got today.

Richard Campbell [01:09:59]:
And And Weizenbaum said point blank, I wrote this to show how susceptible humans are to language.

Leo Laporte [01:10:05]:
Right. We want to believe.

Paul Thurrott [01:10:07]:
We want to anthropomorphize.

Leo Laporte [01:10:08]:
Yeah, we do. So just remember, it's a computer program. Yeah, it's just software. It's just software. When you turn it off, it goes away and it doesn't remember anything. Right.

Paul Thurrott [01:10:20]:
When it comes back.

Leo Laporte [01:10:21]:
Well, unless you— Well, you save files and it will read the files when it comes back. Right, just like Memento. It's just like the guy.

Paul Thurrott [01:10:27]:
It's just like, you know, when I had a Commodore 64, when I first got it, I would type in these software program, BASIC programs, and then I didn't have any way to save them, so I'd come back the next day and type them in again, you know. And that's what AI is just like that. It's exactly the same, just like that. Yeah, it's like, oh, hey, hey Paul, nice to meet you. You've been working with me for a year, for freak's sake, you know, like, come on, man.

Leo Laporte [01:10:48]:
But, uh, one of the things Harper Reed, who's who works in AI and is a cool guy and really loves AI, told me, he says, make it— give it a name for you. I said, no, come on, that's just anthropomorphizing.

Richard Campbell [01:11:01]:
So I—

Leo Laporte [01:11:02]:
he said, no, no. He says, I have it call me Big Dog. And the reason I do that is if it forgets my name, then I know it's lost its context or the context is jammed and it's time to restart.

Richard Campbell [01:11:15]:
I did the same thing, but I just told it to speak to me in iambic pentameter.

Paul Thurrott [01:11:21]:
Haikus only.

Leo Laporte [01:11:24]:
I only accept binary. It is kind of amazing when you ask it to do something that feels archaic. For instance, I have my program that generates the show notes for these shows that I do created in Emacs org mode. And it just knows it. It just knows it. It does a perfect job. It's kind of strange. It has really odd skills.

Leo Laporte [01:11:51]:
Or knows how to say, I am here in Hebrew.

Paul Thurrott [01:11:54]:
Why did it choose Hebrew? Yep. And why are you telling me this? What does this have to do with my retirement?

Leo Laporte [01:12:01]:
He says, I was in the Marines and now I worry this isn't quite safe yet to be on government computers. Yes, that's a good point. I'm not sure I want AI to be deciding life or death things. Um, yeah, that's where that would be the really scary thing, is if—

Paul Thurrott [01:12:21]:
well, but it will— it is already, right? And, and, you know, in the same sense that, like, um, you know, software code review is a good use for AI. I mean, you're looking at a satellite image of some distant place, it's like, which of these buildings we want to blow up? And, right, yeah, I could be like, look, this is what it used to look like, this is the thing that's— you know, it gathers in all, you know, Yes, there are going to be mistakes, but you could, you know, it's not life and death decisions. I think maybe we shouldn't trust it. Yeah, I don't trust the people in our government with life or death. Well, that's a good point. So maybe AI would actually be better in this case.

Richard Campbell [01:12:56]:
I don't know.

Leo Laporte [01:12:57]:
Um, yeah, I don't know.

Paul Thurrott [01:13:00]:
I don't know. Okay, um, just two quick code things, I, uh, or dev things. Um, Visual Studio Code, I didn't They must have announced this in the past. I couldn't find it, but they are switching to a weekly update schedule. Wow. Right. So we know with browsers it was 6 weeks at one point and 4 weeks. It's going to be 2 weeks later this year.

Paul Thurrott [01:13:20]:
Um, I will say not so much Visual Studio Code, which to me has been fine, but, um, Visual Studio, big Visual Studio, the latest few releases— and I'm not on Insiders— um, have introduced some bad reliability problems. I saw a comm error in a dialogue and I had to force quit the app. I couldn't close it normally. It wouldn't let me click anything. I've never seen anything like that, or not in the past 20 years. I don't know. So I don't know, but one of their little justifications here is that this means that each of these releases won't be as monumental and it might be easier On users, Visual Studio Code is pretty good about updating. If you're familiar with the app, you know that at the beginning of March, in this case, they come up with what they would call the February update, right? It's always for the previous month, which they come out with in the first week of the next month, so to speak.

Paul Thurrott [01:14:14]:
But now they're going to do this every week.

Richard Campbell [01:14:18]:
So very different teams and very different code bases too.

Paul Thurrott [01:14:21]:
So yep, yep, that's true. Yeah.

Richard Campbell [01:14:25]:
And then that way was a problem actually, more Microsoft naming problem.

Paul Thurrott [01:14:29]:
Visual Studio Code has nothing to do with Visual Studio. Right. So, uh, I'm trying to remember back when, um, God, Xamarin had their own thing. Miguel de la Casa. Yeah, which became Visual Studio on the Mac, became Visual Studio for Mac, which they got rid of.

Richard Campbell [01:14:48]:
MonoDevelop.

Paul Thurrott [01:14:49]:
Yeah, yeah. Um, they had a— yeah, that they had a unique name for that. And, and I feel like I wish they could just do that for that. I wish there was a better name for this. But yeah, I mean, they reuse successful brand names. What are they going to do? One month ago, I mentioned the release of.NET 11 Preview 1, and like clockwork, because that's actually how these guys do things, Preview 2 is out. I cannot for the life of me figure out the point of this release. I don't mean Preview 2, I mean the entire.NET 11 thing.

Paul Thurrott [01:15:23]:
Explained what the overreaching goals are for this one. This is a standard release cycle, at least not a, what do you call it, long-term. Something else, yes. Yeah. But, you know, the first month, first of all, there was no discussion at all about anything. They just really, here's the list of the new things and it wasn't much of anything. Same thing this month. I mean, there's one line of, hey, it's out, here's what's new.

Paul Thurrott [01:15:48]:
I mean, this is like nothing. Yeah. And one of the few things I care about in this space is the WPF stuff, which very clearly is not going to be impacted by.NET 11 in the slightest. Not so far. I feel like they're kind of done. Yeah.

Richard Campbell [01:16:02]:
I hope there's more. The big feature is a long-term feature when they work it on through many versions is Async 2. Well, they're calling it Async 2 right now. It used to be called Green Threads.

Paul Thurrott [01:16:13]:
Okay. So that is part of Preview 2. Yeah, um, it's, uh, let me— where's my— I lost my page here. Um, still the wrong page. Uh, yeah, in fact, that was one of the few things they listed, uh, Runtime Async V2. Um, this is what they call like a runtime-native async.

Richard Campbell [01:16:33]:
So yeah, it's going to impact— this is the whole thing is async was created by Anders Hejlsberg, which means it was in C#. Right?

Paul Thurrott [01:16:40]:
It was in the language. I thought you were gonna say, which means it was perfect. Thank you.

Richard Campbell [01:16:43]:
Well, I'm not gonna disagree with that. Listen, when the C++ team takes your approach to multi-threaded execution and incorporates it into their language, you've done something good, right? Async in a way has become a very popular metaphor for at least avoiding I/O binding, right? It's like, hey, these things that are gonna take a certain amount of time, I don't care what order they come back in, so off you go. This is, uh, multithreading is a different conversation and a harder one.

Paul Thurrott [01:17:08]:
This is going to come up, oddly enough, in that back of the book thing, uh, because part of the code I was writing had to deal with async file, uh, access things that I'm embarrassed to say was not done correctly in two cases. Anyway, um, we'll get to that. So do you know anything about— well, so async, okay, so runtime async, yes, V2, um, but moving into the runtime means every language gets Right, right. But that's— I mean, I don't know, I feel like most.NET releases have a theme in addition to all the double-digit performance gains we always seem to get. There are, you know, across the different.NET products, you know, whether it's MAUI or Blazor or whatever else, like there's always some, you know, this stuff like this.

Richard Campbell [01:17:52]:
I think we're still waiting for a theme to emerge. There are a number, quite a large number, of long-term improvements that have been worked on for quite a while, right? And there's really a debate of like, are you gonna make it this year? Yeah, which is, which is the other thing, has been 3 or 4 years now.

Paul Thurrott [01:18:11]:
Like, it's bloody hard. So one thing we've talked about here is this notion that maybe this thing shouldn't be on a yearly cadence, you know, it doesn't always justify that. But the flip side of it is there's that stuff, uh, because it seems like every year this, we're hoping to get this, and it didn't make it. But, and now you have to wait till the next year, you know.

Richard Campbell [01:18:31]:
Um, it, it, I feel like there's a more— they're not plucking— I feel like they're not plucking low-hanging fruit anymore. Yeah, now they're really reaching for very challenging things.

Paul Thurrott [01:18:41]:
Well, they have this— I, I call it a, I don't know, an agreement or so, whatever, with their customers that this is the— this is how we do this, you know, and that we will— and they've made great strides in making version over version upgrades of codebases through.NET easier than ever. They have used conversion tools and it's usually pretty seamless. That's great. But sometimes I feel like there could be this feature that's like an out-of-band thing that it maybe doesn't make sense to wait for November. Maybe they missed this November, but it's ready in March. That happens. I don't know how you solve it.

Richard Campbell [01:19:19]:
I, I'm not— yeah, people are upset with you doing it every year and, you know, would want you to turn down the cadence.

Paul Thurrott [01:19:24]:
And then you're making— I'm like, every year is too much, but you should do more. Yeah, because I'm a hypocrite.

Richard Campbell [01:19:31]:
Um, well, you know what it is? The blog post itself, I mean, it almost looks like it was generated because the number of times they said.NET 11 Preview 2 does not include any notable new X, right? Nothing new in C#, nothing new in VB, you know, on and on and on. I don't know that you need to say that that way, right? But it, um, yeah, I mean, I've got the whammy that I know a lot of those guys and I know some of the things they're working on, and they're really afraid to talk about stuff that doesn't make it in the box because they've been burned by that before, right? So they'd rather wait until they're closer. You know, release candidate is when you finally go, are you staying, are you going, are you going to be pushed? And that's, you know, the wrestling match that they're up against. So it's too early to really talk about what 11 is really going to be.

Paul Thurrott [01:20:21]:
You know, I'll know that Microsoft has righted its ship, so to speak, when they start using language like release candidate again across the company, you know? Yeah. Um, I really— I know this, this is— I know it's— I'm just old, but like, I can't stand the way software is released through this company for the most part.

Richard Campbell [01:20:39]:
.NET does do it correctly.

Paul Thurrott [01:20:42]:
They do it correctly. Exactly. I do like that. By the way, you know, in the sense that Google is Microsoft, I mean, they do the same thing in Android. Like, they stabilize and finalize the SDK, new features and all that before the thing is finalized. So you as a developer can target this thing and make sure everything works, and you can actually put it in the Play Store, and then, you know, it will upgrade seamlessly to the final version. But yeah. Um, that's the right— it's interesting, it's the dev stuff, right, that is done correctly.

Richard Campbell [01:21:10]:
These are about, does this work on all the platforms? Like, we're feature complete, we're happy with it, but how does it behave in the different modes? And that's why you have a few weeks.

Paul Thurrott [01:21:18]:
There's a different level there for, you know, compatibility and whatever, bug fixing and so forth. But, um, yeah, I miss this.

Richard Campbell [01:21:26]:
Is, you know, so we used to do things everywhere. My, my modern exemplar for what— how hard this gets was what happened trying to get MAUI out the door when they had a new version of Studio and a new version of the framework that were dependent on each other, and they were trying to build a tool within both. Yeah, it's like you just don't want to be in that slipstream. You don't want any of the previews. You want the RC before you start, or you're wasting your time. You're just going to be battered by counting on a feature that gets pulled, something that completely revised, which is what happened to the Maui team. They missed November. They did release Maui the following spring, right?

Paul Thurrott [01:22:04]:
They did go. So they can do that. Yeah, okay. I mean, and that's a big enough deal where I could, you know, could make the case for it.

Richard Campbell [01:22:09]:
It was Maui, right? It was tough. And then David Artinow, those guys, like, boy oh boy, they took a beating. Like, it's hard what they're trying to do. Your foundation and your build points are literally moving under you.

Leo Laporte [01:22:24]:
Like, it's just really really not easy to do.

Richard Campbell [01:22:29]:
Oh well, oh well, oh well. Yeah, I know you want a plan, and they're trying to figure out what the plan is. In the meantime, they still put stuff out the door just for folks to see where they're going. And so, you know, what's interesting is thinking, will async2 disappear again?

Paul Thurrott [01:22:43]:
That's right.

Richard Campbell [01:22:43]:
Yeah, will it actually make that possibility?

Paul Thurrott [01:22:46]:
Because it has before. Because if it does, this release— unless there's something coming I don't know about, which you know, possible. Uh, there's not much going on here, you know.

Richard Campbell [01:22:55]:
Um, I don't think people are willing to show it yet.

Leo Laporte [01:22:58]:
There's more things coming.

Richard Campbell [01:23:00]:
Okay. I think I—

Leo Laporte [01:23:01]:
what did I just recently heard?

Paul Thurrott [01:23:03]:
Miguel de Ocasio is doing something. Yeah, he is a— it's called the Godot, or Godot, um, like a 3D modeling kind of— that's it— thing for gamers.

Leo Laporte [01:23:14]:
It's like SwiftUI. Yeah, yeah, it's, uh, but it's also kind of a cross-platform development tool, right?

Paul Thurrott [01:23:20]:
For— yeah, but that gets into an interesting spot because, uh, SwiftUI is obviously amazing on Apple, and there is SwiftUI elsewhere— Android, Windows, etc.— but it's largely unproven. I mean, Arc, uh, the browser was SwiftUI— is SwiftUI even on Windows? Um, but I don't know, I don't know, I don't know, I don't know if they're there yet. It's like the guy, you know, came to Microsoft, worked on all this.NET stuff, most of his adult life, and it's like, now I'm on Apple and I need some kind of a cross-platform framework. God damn it, what? You know, we get— are we really going to reinvent this?

Leo Laporte [01:23:55]:
I mean, I don't know. He likes doing that. He likes, he likes getting that cross-platform stuff working.

Paul Thurrott [01:24:01]:
He's certainly got some—

Leo Laporte [01:24:02]:
he's a genius.

Paul Thurrott [01:24:03]:
Got some credit.

Leo Laporte [01:24:04]:
Another good guy.

Paul Thurrott [01:24:05]:
Some chops.

Richard Campbell [01:24:06]:
The new, the new tool is called Zibin with an X. Zibin. But it is Godot-like. Ah, okay.

Leo Laporte [01:24:15]:
But that's where he's going. Okay.

Richard Campbell [01:24:17]:
And it is just Mac, or is it— it's Swift-centric at the moment.

Leo Laporte [01:24:22]:
So I guess that's really not— I mean, it's not technically only Mac, but it really is.

Richard Campbell [01:24:26]:
Yeah.

Paul Thurrott [01:24:26]:
But well, I think it works on iPad too, I think.

Leo Laporte [01:24:29]:
Yeah, yeah.

Richard Campbell [01:24:30]:
Apple, I should say, only. Yeah.

Leo Laporte [01:24:32]:
It wouldn't surprise me if they suddenly— there's Swift for Linux, and I think there's Swift for Windows.

Richard Campbell [01:24:36]:
I know there is.

Leo Laporte [01:24:37]:
There is. Yeah, yeah, yep. So, uh, hey, let's pause. And, uh, you have some big news in the Xbox all of a sudden. Yeah, all of a sudden, uh, because GDC is going on right now, the Game Developers Conference, although missing quite a few names who were unwilling to travel to the US.

Richard Campbell [01:24:58]:
Yeah, why? I don't know. I don't know. We should talk about the MVP Summit too because I'm hearing the same thing. Yeah, yeah, it'll be all Americans at the summit.

Leo Laporte [01:25:06]:
Yeah, well, there you go. Uh, I won't say anything political.

Paul Thurrott [01:25:11]:
I'll just leave it and let it—

Richard Campbell [01:25:13]:
I'll do it.

Leo Laporte [01:25:14]:
Um, no, just let it lie. I'm just saying, you guys are in Mexico and Canada, and maybe you should stay there. Yeah. Our show today, our show today brought to you by Cashfly. How many times you've heard me say this? Bandwidth for Windows Weekly is provided by Cashfly. At cachefly.com/twit. That's our CDN, our content delivery network. You know, at TWiT, we don't just cover tech, I mean, we depend on it.

Leo Laporte [01:25:45]:
And that's why practically since the very beginning, we've trusted CacheFly to get you our content, petabytes of data every single month. Thank you, CacheFly. CacheFly is providing speed like no other CDN for over 20 years. On Cashfly, online games start 70% faster. HD video plays with sub-second start times on every device. Software downloads complete without a hitch. Events stream smoothly. Cashfly serves over 5,000 businesses, including Twit, across nearly 100 countries, ranging from Fortune 100 companies to solo developers.

Leo Laporte [01:26:25]:
Companies choose Cashfly when performance is non-negotiable. And it wasn't just the performance, frankly, it was the support. They, they went the extra mile for us, and they'll do that for you too. If you need help, you'll get seasoned support experts. They're there 24/7. They actually understand your unique challenges. They, I mean, they're not reading from a notebook. They know this stuff and they'll talk to you engineer to engineer.

Leo Laporte [01:26:50]:
Start with flexible month-to-month billing. That's what we did because we weren't really sure, you know, what we needed. Then lock in the discounts when you do know what you need. The point is, we designed our own contract when we moved to Cashfly. You can too. No lengthy obligations, no sales tactics, just exceptional service backed by a 100% uptime SLA. That sounds hard to believe, but it has been 100% over the last year, more than a year. And it's not just that for us, it's, it's reliability, it's speed, it's support.

Leo Laporte [01:27:22]:
In every way, Cashfly is exactly what we needed. I think it's exactly what you need. Learn how you can get your first month free at cashfly.com/twit. I'll say it again, that's c-a-c-h-e-f-l-y.com/twit. Thank you, Cashfly, for making this all possible. We appreciate it.

Paul Thurrott [01:27:44]:
All right, Time for some Gaming the Xbox segment, Paul. Good. Slightly reorder things just to put this in chronological order because I think this makes sense. But last week after the show, Asha Sharma, who is the new CEO of Microsoft Gaming, who everyone loves, tweeted— I guess we're gonna still call it— that Big meeting, uh, talked about Project Helix, which is the code name for Microsoft's next generation console, which she confirmed would do two things: lead in performance with both Xbox and PC games, right? So it's going to play PC games. Now I think about this in the opposite direction. Will I be able to, on a Windows PC, play Xbox games, meaning, right, backward compatible games? They have not talked about that to my knowledge yet, but, but I feel like if these things are essentially the same thing, you know, whatever. But I will say, um, for those people who want an Xbox console and are going to be happy to hear that this is still happening, um, performance, uh, promises are good. Windows compatibility is pretty solid.

Richard Campbell [01:28:55]:
It's a nice differentiator, right? Yeah, but I don't know how you promise you're going to keep beating PCs when PC's a moving target.

Paul Thurrott [01:29:01]:
Yeah, unless this thing is also moving target, right? I mean, maybe this thing's modular, I don't know. I mean, I agree with you, but, um, but yes, I— the Phil Spencer multiple times was talking about bringing competing stores to Xbox, right? And that's how you do it. You— Steam, GOG, Epic Games, whatever, um, should all be able to work on this next Xbox. So, so that was kind of a fun, you know, came out of nowhere, and I was like, you know, this is smart for her to do. She's not really well loved yet, and Phil Spencer was. He left suddenly for a lot of people outside the company. Um, he was very good about talking publicly about what was going on, even maybe when he shouldn't be.

Richard Campbell [01:29:39]:
And, um, if she could kind of continue this level, uh, this type of communication— you're gonna need to. She's also in her 30s, so yeah, like, I don't envy her.

Paul Thurrott [01:29:48]:
This is a tough gig.

Leo Laporte [01:29:50]:
Yeah.

Paul Thurrott [01:29:50]:
Oh yeah, no, this is, uh, managerial.

Richard Campbell [01:29:55]:
Yeah, exactly. Um, sounds like Satya's got her back. And if one thing I've heard is that she manages up very well, like she's well liked by senior leadership.

Paul Thurrott [01:30:05]:
So speaking of Satya Nadella, last week there was an internal QA, Q&A about Microsoft gaming and all this stuff, and he, um, said that Microsoft is long on gaming, uh, which is kind of interesting, um, which explains why they've always fallen short in gaming. Sorry, sorry.

Leo Laporte [01:30:22]:
I'm sorry.

Paul Thurrott [01:30:22]:
And the thing that— the big thing I took out of this was his comment that there are 4 core identities in Microsoft. I would actually call these customer bases, maybe, but he called them identities. And those identities are Microsoft is a platform company, a developer company, a knowledge worker company, and a gaming company. And I don't know that Wall Street or anyone from outside the company would have picked those for. I don't know, you know, even though, look, they spent $68— uh, actually $75 billion after, uh, post-acquisition costs, um, on Activision Blizzard, you know, they didn't do that not to be a big player in gaming. But I mean, you look at the company today where the money comes from, and yeah, platforms, absolutely, but shifting to the cloud, right, with Azure and all that stuff. And now AI, um, knowledge workers, Microsoft 365, which is also shifted largely to the cloud, which is great. Place those strengths.

Paul Thurrott [01:31:18]:
Um, what were the other ones? I'm sorry. Uh, yeah, so platforms, developer, developers, of course, right? So they make developer tools, frameworks, languages, all that stuff. I mean, that's—

Richard Campbell [01:31:27]:
this has been core to Microsoft.

Paul Thurrott [01:31:31]:
It's the original day one. Yeah. Um, and you know, being technical, and I think people listening to this by and large being technical, I think we can all kind of appreciate when they don't slip up, the kind of engineering mindset that often comes through with Microsoft stuff. I mean, you can tell it's kind of an engineering, you know, developer.

Richard Campbell [01:31:47]:
What's interesting here, like, the odd duck in those—

Paul Thurrott [01:31:50]:
that foursome is gaming. It just isn't about the same level. Exactly. Yep. Um, now, I mean, look, his— I'll just read this because I— it's nowhere to kind of paraphrase, but when, you know, why does Microsoft love games, right? Uh, he talks about the stories, the mythologies. It makes The things that make us who we are. There's a craft to it. There is a cultural thing to it.

Paul Thurrott [01:32:16]:
Okay. I mean, I guess the way I would think of it is more like Xbox is in many ways Microsoft's only existing successful consumer brand.

Richard Campbell [01:32:26]:
Yeah, I'm sorry.

Paul Thurrott [01:32:27]:
I was trying to think of—

Richard Campbell [01:32:28]:
make sure that was correct. You're fighting for that one. I get that. Okay. Surviving. How about that?

Paul Thurrott [01:32:34]:
Surviving consumer brand. Fair enough. Well, successful as a brand, by which I mean it's still beloved, right? People are troubled by what's going on, but it's still a good brand. It hasn't been destroyed.

Richard Campbell [01:32:46]:
No, it's— but people are annoyed with what they've done with it.

Paul Thurrott [01:32:49]:
Yeah. While still looking to kill the whole time. Look, this is like the AI, or we're going to improve Windows thing, right? They're like, nice. So that means we're not going to do this AI nonsense. No, no, we're still doing the AI nonsense. But we're going to also, you know, fix the fundamentals, right? And so, um, when you look at Xbox as a business or Microsoft Games, it's like, so you're going to get out of the hardware market and you're going to be like a software publisher? No, we're still going to do the hardware thing. And it's like, okay, you know, it's like, so, you know, we'll see what happens. But, um, so that was— let's see, that was earlier this week.

Paul Thurrott [01:33:26]:
And, um, I will say the nice little course correction here is he mentioned Phil Spencer and Sarah Bond, by the way, and Matt Booty, and didn't leave anyone out of this conversation because there's a continuum here of actions or whatever that this business is taking, right? The people who lead it, which is now different people, right? But they're— we're still going to push through the stuff that we have— they had been doing, right? That's still happening. Um, okay, but then today was Microsoft's GDC presentation. I believe they're making two, by the way. But before Phil Spencer left, before they made that announcement, um, there was a comment that we're gonna— you're gonna want to pay attention to GDC. We're gonna announce stuff about the next-gen Xbox experience stuff. And there were two big sets of announcements. Which I think you could both argue they were kind of PC-related, right? So if you're familiar with gaming on Windows today, so this is Windows 10 into Windows 11, there's something called Game Mode, right? If you go into Settings, there's actually like a gaming top-level setting or top-level section, and there's 3 things in there. So Game Bar, Game Bar settings, so things like keyboard shortcuts and controller shortcuts and so forth.

Paul Thurrott [01:34:44]:
Captures, which are related to that because that's how you would do captures through Windows if you use the native stuff. And then Game Mode. And Game Mode is weird because it's on/off, it's always on. I mean, you can turn it off, I don't know why you would, but the idea is that when you're playing a game, it will optimize the PC for that thing and will turn off some stuff in the background. Now, anyone who's used Windows to play games will tell you that doesn't actually do anything, or if it does, it's minimal because you're still getting weird Pop-ups and things happen, you know, it's Windows, right? So when we move forward to the Xbox ROG Ally devices that first shipped last fall, there was something called the Xbox Fullscreen Experience, and that takes elements of the Xbox app, uh, that has been in Windows since probably Windows 10, but whenever, and, uh, turns it into the UI. It's the shell, right? Essentially, they can boot— you can and get out of that and go back to the desktop. When in this mode, in this Xbox full-screen experience, the game— sorry, so many similar terms— the Game Bar is still available, but it's more of a controller-based experience, full screen essentially. So it's easy to use with a controller.

Paul Thurrott [01:35:52]:
And even though they don't really call it this, they've essentially taken Game Mode and gone extreme with it where this thing boots up using far fewer resources far fewer things going on in the background and trying to get Windows to that place where if we cut, cut, cut enough that this thing will perform wonderfully and don't interrupt you with dialogues and whatever. So one of the things they announced today is that the Xbox full-screen experience is coming to all Windows 11 PCs starting in April. Um, that we kind of knew that was happening.

Leo Laporte [01:36:24]:
It's—

Paul Thurrott [01:36:24]:
it— along the way it's come to more, um, portable gaming handheld devices, right? They're renaming it to Xbox Mode. So that tells me that Game Mode goes away and this becomes Xbox Mode. And Xbox Mode almost certainly will be more granular in the sense that it's probably not going to be just an on-off switch, although it could be, I guess. But I'd like to see a version of this where you can actually choose what things can happen and can't happen, that kind of deal. But you'll be able to boot any PC into this mode, which you might want to do if you actually have a gaming PC, right? I think the wonderful thing about the PC as a gaming device is that it's everything, you know, it's the kitchen sink. So if your primary use case is playing games, you can boot into this mode or just go into this mode, I guess, play a game. But if you have to get work done or you sort of shop, whatever it is, you could back out of it and everything works normally. Going.

Leo Laporte [01:37:20]:
Okay.

Richard Campbell [01:37:20]:
Do you think they're gonna put PCs in the living room again? Is that kind of the path they're on?

Paul Thurrott [01:37:28]:
Um, because that's the whole reason, you know, so in the sense that most PCs are portable laptops, right? Yeah. Um, the hardcore gamer types who would choose the PC will have desktops. Yeah. But I still feel like even within the audience of people that play games actively on Windows, I bet the majority of them are probably going to still be laptops, right? Yeah. And their laptops are getting better and better.

Richard Campbell [01:37:53]:
And so the guy who's sitting in front of two desktop machines, because there's something really wrong with me.

Paul Thurrott [01:38:01]:
Yeah. Okay. Um, and I, you know, having used a bunch of modern laptops, can attest to the fact that this stuff has gotten a lot better. Like, it's a lot better. Um, and so when you say, like, is— are we going to bring a PC back back into the living room. I guess in one sense you could argue we're gonna just bring that PC everywhere because it's a laptop, so you'll be able to play on a plane or, you know, at a coffee shop or, you know.

Richard Campbell [01:38:22]:
I like the idea of you play on your Helix, on your Xbox.

Paul Thurrott [01:38:25]:
Yeah. And then when you leave, you crack your laptop, you know, and you're back in. It's the same game, literally the same game, not— yeah, the same game ported to a different platform, like the same game, which is— you know, it's pretty— that's compelling. The other thing is with regards to Project Helix, that name is actually from the past. That I think the first rumors of that being a next-gen console might be as old as I could be wrong. Either it's either 2012 or 2016, like a long time ago. But this is the, this is the thing that the rumor was late 2027 was the target. And so today what they said was that they will deliver alpha versions as they call it of the next-gen console codenamed Project Helix to game developers starting in 2027.

Richard Campbell [01:39:14]:
Because the One X came out in 2017, which by the way is almost 10 years ago. I know. Um, and clearly will be by the time Helix arrives.

Paul Thurrott [01:39:25]:
So they've skipped several generations of hardware. Yep. This is, um, the Xbox 360 and the Xbox One were both served by interim updates to the hardware, which yes, in some cases were about cost reduction, but also improved capabilities, right? Yeah. And Xbox One was maybe the best version of that because it was Xbox One and then the S and then the X, right? Yeah. With the current-gen consoles, they shipped two at one, which, you know, first time ever. And then they never actually released a meaningful update to either of them.

Richard Campbell [01:39:56]:
So, and the X came out—

Paul Thurrott [01:39:57]:
X and S came out in 2017. X and, uh, no, that can't be right. I think I think it wasn't it. I think it was the year of the pandemic. I think— I don't remember. But, um, AMD is doing a custom processor for them as they are for PS. All right, 2020. Okay.

Paul Thurrott [01:40:23]:
Um, so in a way, uh, I, you know, it's hard to know chicken and egg on this, but the plan originally was to release midstream updates to both of those consoles, which would include functional, you know, and performance updates, etc., graphical updates. Um, that obviously is not happening.

Richard Campbell [01:40:38]:
So, you know, something has gone—

Paul Thurrott [01:40:39]:
it's too late now. Oh, they can't do it now.

Richard Campbell [01:40:41]:
Yeah, there's no way. Well, and I think you're also—

Paul Thurrott [01:40:43]:
there's a whole conversation about hardware pipeline and all of that sort of stuff, right? Yep.

Richard Campbell [01:40:50]:
So, um, Sony has— I think here's a question: what if it's ARM? It's not.

Paul Thurrott [01:40:54]:
It'd be great if it was ARM. Well, it's almost certainly not. Um, it's an AMD CPU. Um, no, it's not. It's definitely not. They've talked about it. It's a Zen— I, I don't know if it's Zen 6 maybe or Zen 5, but, um, you know, doing the same thing they do for PlayStation. It's what AMD did for the current— what is now the current-gen PlayStation and Xbox, right? Um, I do feel like there was a different timeline where we would have gone ARM, and I think it— because we saw the leak, like, they were they were looking at this.

Paul Thurrott [01:41:26]:
I think they wanted to do it for a while. Yeah, it just didn't— I did— it didn't happen in the right time frame for them to do this backward compatibility thing with the PC. You kind of have to be x64, right? Like, for 100%, for it to actually just work, right?

Richard Campbell [01:41:42]:
To work everything. I mean, there was a plan. There was like, what's the story arc look like? This is like, at some point it will be ARM.

Paul Thurrott [01:41:48]:
It's not now. I do feel like there's still a future where ARM happens and probably first on mobile, right? Yeah, there'll be a mobile Xbox of some kind, like an Ally device or whatever they'll call it.

Richard Campbell [01:41:59]:
The fact that we have Snapdragon out there, we have the laptops and getting more compatibility, it's like you need to knock those problems down a bit before you start committing to dedicated devices.

Paul Thurrott [01:42:07]:
I mean, games are the final frontier for this. I mean, this is— but it's also kind of in Microsoft's wheelhouse, you know. The— to me, like I keep saying, that one of the biggest strengths of the Xbox ecosystem is this backward compatibility stuff that they've done. Um, they are promising that for this next-gen console as well. They're promising improvements to it, which is kind of interesting. I kind of was of the, uh, thought that they were done with that, like there were no new games to bring forward, but maybe that's going to change. Um, there are features in Windows that are also in the Xbox Series X and S— Auto HDR, FPS Boost, etc. We have Auto SR on the PC side.

Paul Thurrott [01:42:45]:
I We absolutely should have that on the Xbox side. I mean, probably will. It's the same thing, right? Um, and there are going to be additional updates, you know, coming later. So, uh, right now I think they were just, look, we're doing it. This is what it basically is. We're not really talking about deep technical details yet, but, um, it will not be going out to developers this year, which is not great, honestly. I hopefully we can kind of bridge the gap with better mobile gaming handhelds and then get this, um, next-gen console.

Richard Campbell [01:43:20]:
We'll see, we'll see, you know. Yeah, no, I love—

Paul Thurrott [01:43:23]:
there's gonna be a new console. Yep, me too. And so this just happened. I, I, after the show, I'm gonna have to go over this. Um, I feel I probably am missing something. I hope I'm not missing something important. Um, But yeah, Project Helix is powered by a custom AMD SoC co-designed for the next generation of DirectX and FSR. They've been on DirectX 12 for about 25 years, by the way.

Richard Campbell [01:43:49]:
Maybe we could rub that.

Paul Thurrott [01:43:50]:
Yeah, but no, if you got it right, why do you want to change it? Well, because we're talking about— well, that's a good point actually, right? I mean, yeah, as long as it supports all the new technology. So one of the big deals in gaming the past 5+ years, what, maybe 10 years, I don't know, ray tracing, right? They're describing the ray tracing performance and capabilities of this system as an order of magnitude leap. This was the thing that I think Sarah Bond was alluding to when she mentioned the next console and how it was going to be this major leap forward. I think that's going to be a big part of it. And then integrates intelligence directly into the graphics and compute pipeline. So that tells me things like, well, that's what AutoSR is. That's what FSR is, right? This is, you know, if for some reason this thing has slowed to a crawl, we can actually lower the quality of the graphics, but you won't see it because it's upscaling and it looks beautiful, that kind of thing. So that's good.

Paul Thurrott [01:44:55]:
Intelligence is a tough word because you think AI and you're like, uh, I don't know. Um, but, but we'll see. Anyway, this, like I said, this just happened. This looks so far like all good news. Um, and then we're going to be able to see some of this in the Insider Program pretty soon. Um, there, there was some— I think there's an indication here somewhere that Xbox Mode might actually be in some of the recent builds, uh, in the Insider Program, but I just— this one I have is on the dev channel and it's still called Game Mode here, so I don't see that. But, um, this is a particularly good laptop for games, um, and it hasn't— this one has an NVIDIA GPU. Um, I would definitely use the Xbox full-screen experience on this just to see if I could detect a difference, you know, in frame rate, quality, whatever.

Paul Thurrott [01:45:42]:
Yeah, we're still a ways away. I know, but it's good to have some news. So yeah, it's good to have good news.

Richard Campbell [01:45:53]:
I think that's—

Leo Laporte [01:45:54]:
yeah, that's all the Xbox.

Richard Campbell [01:45:55]:
That's all. If those announcements hadn't come out at GDC, I don't think you would have much of an Xbox section.

Paul Thurrott [01:46:00]:
I know. Yeah, well, we had the— so yes, uh, we had the little, uh, name reveal, it's going to run Windows game, and then Satya Nadella thing.

Leo Laporte [01:46:11]:
But yeah, today more details. People were saying that because it's running Windows games, that is it really even a console. It could just be like a Steam machine.

Paul Thurrott [01:46:20]:
That's Steam. Well, okay, so Xbox— every Xbox except for the 360 was a PC, right? But it was a PC that couldn't run Windows games, right? And it's like— and, and look, I, I can— you can make the case for that, especially 20 years ago when this— or 25 years ago now, I guess, when the first console came out. Um, I do think we're at the point now we're leveraging the strength that is the game market, um, on Windows makes so much sense. Oh, the whatever this next thing is called eventually, the Project Helix, is a Windows PC, right? Right now they're optimizing the software as they always do, uh, really, but they're adding the ability to run—

Leo Laporte [01:46:58]:
it's Windows and it's gonna run. That's why I'm wondering if this was as much a response to the Steam Machine as anything else, right?

Paul Thurrott [01:47:05]:
I think it's a response. So I would— the way I would phrase it is Microsoft's biggest success in gaming— well, it's on the PlayStation. No, I'm just kidding. Is, uh, um, is in Windows, right? And if you look at the way the market kind of breaks down, uh, whether you're looking at unit sales for devices or software titles or whatever it might be, or time spent playing games, um, the PC is the— you know, there are a lot of developers who are like, this is the marquee platform, like, we're going to start there and then we're going to port to these consoles or wherever else, or mobile even. And, uh, doing this for Xbox makes that— it's automatic. Yeah, you have like 100 compatibility. It's the smartest thing in the world. This is the strength that Microsoft can leverage if they're going to make hardware, right?

Leo Laporte [01:47:50]:
So I mean, I—

Paul Thurrott [01:47:54]:
but will it run Steam? Uh, I don't think it will be open enough of a platform.

Leo Laporte [01:47:58]:
Yeah, so it's really not a PC, it's just going to have Game?

Paul Thurrott [01:48:01]:
Well, it's a— it's like a curated PC-based experience, maybe is the way to think of it, without knowing all the details, right? Because I could be completely wrong about most of this.

Leo Laporte [01:48:10]:
But did they indicate any concern about— I mean, the Steam Machine's been delayed because of RAM prices.

Paul Thurrott [01:48:15]:
Oh sure. Do they know? And look, components— no, because, uh, with the first dev boxes not going out into next year, which by the way could be late next year, right?

Leo Laporte [01:48:25]:
Right.

Paul Thurrott [01:48:25]:
Okay. Um, I, I— we're not going to talk about RAM because RAM might not, not be a concerned by then.

Leo Laporte [01:48:30]:
Um, you know, you won't need RAM by then, right?

Paul Thurrott [01:48:33]:
Right. Everything will be stored in crystals.

Richard Campbell [01:48:35]:
It's going to be great. Yeah. But also, you know, the Xbox team has been building machines for decades, and this is a new venture for Gabe and the co. So I think they're in a different place in the pipeline too.

Paul Thurrott [01:48:48]:
Yeah, yeah, yeah. I think, I think for Steam Machine to make any sense, or any, any of their stuff, Steam Deck, whatever, they had to get that— I think they actually do call it Prism, or whatever their compatibility, their kind of line-based Windows game compatibility layer had to reach a certain level of quality and compatibility, which it has, right? So they've done a good job at that. And that will run under an emulator, it won't. Yeah, right. But, but as long as, you know, again, they have similar technologies where they can do AutoSR type things there as well, right? So you can run a Windows game under an emulator, lower res, but still looks great, plays full speed, whatever, who cares? Like, I mean, it's just, um, that enables them to have a hardware play. Because if they otherwise, they would say, well, just make your games for Linux, and no one would, right? You know? Right. So they can— they, that— they came— they kind of coming at it— let me think about this for a second. Yeah, I think from the opposite direction is excellent.

Paul Thurrott [01:49:45]:
Yeah, it is. That's a good way to put it.

Leo Laporte [01:49:48]:
Yeah, I think so. Yeah. All right, back of the book is just around the corner, but now it's at Time to beg, beg your money, your forgiveness, your tolerance. It's time for Club Twit. Yes, that's what keeps this show alive and all the shows we do. 30% of our operating revenue comes from you, our listeners, our club members. I like that. Honestly, I want it to be 100%.

Leo Laporte [01:50:15]:
If it were up to me, it'd be 100%. When we first started, that was the idea. We wouldn't take advertising, but that didn't seem to actually turn out to be very practical, so we did. But thank goodness we have the club because it started when COVID happened and advertising disappeared a little bit. And we thought, now maybe with the help of Patreon, we use a Patreon company called Memberful, we can maybe make this work. And it's worked pretty well. I'm really glad. We get about Uh, oh, I'd say 1.5 to 2% of our audience is members of the club.

Leo Laporte [01:50:50]:
If we get that just a few percentage points more, we actually could forego advertising. Here's the deal: if you join the club, $10 a month, you do get ad-free versions of all the shows. I am not one of those people who wants to charge you money and put ads in front of you. It's one or the other. So if you don't like ads, there are too many ads, or you just want to support TWiT, join the club. There's also a lot of special programming. You get access to the Club TWiT Discord, and in the Discord we do a lot of, uh, stuff. We have— oops, that's the wrong screen.

Leo Laporte [01:51:20]:
Let me put this screen up.

Richard Campbell [01:51:23]:
That's—

Leo Laporte [01:51:23]:
there you go. In the Club Twit Discord, we have the, uh, coming up actually tomorrow, Cindy Cohn is going to join us. She has a new book, Privacy's Defender: My 30-Year Fight Against Digital Surveillance. She's EFF's executive director and a legend. I'm really looking forward to getting to spend an hour with Cindy Cohn. That's tomorrow, 1 PM Pacific, 4 PM Eastern. And this is a good example of something that, you know, we used to do triangulation and, you know, frankly, it was hard to get advertising for. So we thought, I'll keep doing it, but I'll do it in the club with the support of the club members.

Leo Laporte [01:51:56]:
And that's you that makes it possible. Friday, we're doing our AI users group. Normally we would do that on the first Friday of the month, but we're doing it this Friday because I was out of town.

Richard Campbell [01:52:07]:
I can't wait.

Leo Laporte [01:52:07]:
I'm not sure what we're going to do. Larry, are you going to have something for us or? Or Darren, or we have so many really smart AI users, and of course our own Anthony Nielsen. So I'm sure we'll find something interesting to do. That's this Friday. What else is coming up? Let me just look. Oh yeah, NVIDIA's conference GTC starts on Monday, and Jensen Huang always gives interesting keynotes. They're about to announce something I think kind of interesting, a vibe coding tool of their own. I will be very interested to see what Jensen announces.

Leo Laporte [01:52:43]:
So we're going to cover that. Starts Monday morning the 16th at 11 AM Pacific, 2 PM Eastern. You can watch Jeff Jarvis, Micah Sargent, and I will be doing that coverage of GTC, the keynote. Micah's Crafting Corners next Wednesday, a chill place to do some crafts. He's doing a paint-by-numbers this time, which is really fun. Photo Time with Chris Marquardt, a week from Friday. You still have a chance to take a dazzling photo. Anthony Nielsen and I were going all over Kennedy Space Center looking for something dazzling, and we found quite a few things.

Leo Laporte [01:53:18]:
And then Jet Set with Johnny Jet is 2 weeks from tomorrow, March 26th.

Richard Campbell [01:53:23]:
That's 2 PM.

Leo Laporte [01:53:24]:
We're doing that on the 4th Thursday of every month now. Johnny, my travel guru, 2 PM Pacific, 5 PM Eastern. So we do that in the club. The Discord is fantastic. Because it's a place to get together with really smart, interesting people and talk about all sorts of interesting things. You get that. You also get the warm and fuzzy feeling, the knowledge that you're supporting what we do here at Twitter. Oh, wait a minute, wait a minute.

Leo Laporte [01:53:50]:
I have to put this up. Joe Esposito, who's our great artist in there, he doesn't do this with AI. He does it with Photoshop, creates little ads for us. Clump Twitter is the future of the podcast network you love. Okay. Thank you, Joe. He takes vintage ads and pastes us into them, and he does such a great job. That's, that's kind of fun.

Leo Laporte [01:54:14]:
That's kind of the fun part of the club— some really creative, interesting people in there. If you'd like to join, we'd love to have you. twit.tv/clubtwit. And thanks in advance. Now to the back of the book we go. Paul Thurot starts us off with his tip Tip of the week, Paul.

Paul Thurrott [01:54:37]:
So I, I wrote this, um, as kind of an editorial with an eye toward talking to you guys about this, because people of our age, you know, we were the first generation to have home computers.

Leo Laporte [01:54:49]:
Yeah, home video games. I was an adult by that time. I mean, I was in my 20s when I got my first PC.

Paul Thurrott [01:54:55]:
Okay. Yeah, I mean, 800. Nice. Yeah, that was a good one. There's a thing about that era, and it didn't end with, you know, Atari, Apple, Commodore leaving and whatever, you know, PC, Mac. It's a little messier than that, but there was kind of an enthusiast DIY kind of quality to that early market that I really feel like we don't have today. The closest we've gotten to it in the modern era, I think, is the Raspberry Pi, which does for what we now call makers, people that want to work with hardware devices and sensors and things, what we as budding or would-be software developers would've done with the BASIC built into a Commodore or Atari device or later computers or whatever. To me, it's not just programming.

Paul Thurrott [01:55:52]:
I mean, I kind of approach the world from that perspective. But the last great thing like this, and you could draw a line from the Microsoft BASIC that was in all of those early devices basically, to HyperCard like on the Mac, to Visual Basic, first in DOS but really Windows, until they.NET-ified it and frankly killed it, unfortunately, by making it too sophisticated or whatever. The notion that anyone, like, as a normal human being could go up to a computer and learn this stuff and use it and make things with it is wonderful. I feel like it's a thing that maybe AI will enable for this generation, maybe. It does seem like we're getting there, but I miss this, right? And not just because I'm old and I'm looking back. I don't mean it like that. I'm not like into nostalgia for nostalgia's sake or whatever. And I also am clear-eyed about the things that were bad, right? It's not— not everything was great.

Paul Thurrott [01:56:54]:
It wasn't all strawberries and cream back in the 1980s. It was mostly cocaine, really, but whatever. Um, but I do feel like as this thing has matured, um, this thing being personal computing or personal technology, whatever you want to call it, you know, it's become It's not like for us anymore. This is what enshittification is in the big tech space. This is when Steven Sinofsky wrote his big book about his history at Microsoft. He talked about how early versions of Office were really aimed at individuals who were enthusiasts, individuals who were just consumers. Eventually it was for smaller businesses and organizations and then just businesses, and they left behind consumers. And then it became this enterprise play where it's just subscription service with multiple tiers, and it's just this giant hairball of terribleness.

Paul Thurrott [01:57:43]:
And I wish there was something like those home computers, because one of the things that was interesting— like the Atari that you mentioned, the Atari 800, which came out 3 or 4 years before the Commodore 64— was, I mean, a miracle of technology. It was better than computers of that day. That's something that Commodore did, or Amiga did with the Amiga as well.

Leo Laporte [01:58:05]:
You were an Amiga.

Paul Thurrott [01:58:06]:
That was your first personal computer. Well, it wasn't my first one, but it was— yeah, I mean, it was your first. So in the, uh, in the sense that people like us always have this conversation, what was your first personal, you know, your first computer? I win this. First love. I win this easily. Oh yeah? I don't care who I'm talking to, I will win this battle. My first computer was the ECS Entertainment Computer System add-on for the Mattel Intellivision. Um, was that a computer though.

Paul Thurrott [01:58:33]:
Really? Yeah, it had a BASIC language, and one of the cool things about it was that you could plug in a cartridge and access the sprites and other graphic assets on that cartridge to make your own games in BASIC.

Leo Laporte [01:58:44]:
Well, there you go.

Paul Thurrott [01:58:46]:
It was pretty cool. Yeah, my first real computer was a Commodore 64. Um, you know, at every step of the way, I, I, I always wanted something better but couldn't afford the better thing, so I had a Commodore Commodore 128. I had an Amiga 500, not an Amiga 2000. You know, it's just kind of like my life, you know, like the way it's always gone. But you know what? Those computers were incredible, and they were all better than other things of the day. So you— the best-selling, you know, computer in 1987, '88, whatever, was an IBM PC, or maybe a Compaq PC by that point, whatever it was, but it was PC. And those things were technically inferior.

Paul Thurrott [01:59:22]:
They just were. And that's the problem with the Raspberry Pi to me. If the Raspberry Pi was this this $100 to $200 computer that somehow outperformed in some way, um, you know, modern PCs, it would be like, wow, um, that would, that would be like the home computer era of the 1980s, right? Um, but it's not. And it does— I mean, it's, it's still inexpensive and great, and I love that it exists. And I like— and I'm not really into the hardware sensor maker thing, but, but I do love that that exists. Um, I wish there was something though. Like, I'm just trying to— I'm trying to understand. Like, you know, Commodore is back, sort of.

Paul Thurrott [01:59:57]:
Like, so they have these new Commodore 64s. I, I do not want one of those. I love that they make it, but I don't want to go back to a breadbox with a 40-column display. And, you know, it's ridiculous. Like, I, um, but I don't know. I— can you guys— can either of you think of anything that falls into that kind of enthusiast, you know, like the actual magic of what this world was like in the very late '70s, early '80s.

Leo Laporte [02:00:27]:
Like, I mean, we were talking yesterday about 3D printers, and there is that kind of same—

Paul Thurrott [02:00:33]:
yeah, okay, that's actually— that's a good one. Do-it-yourself. This notion that you could, as an individual, print a part, yeah, for a car that is no longer supported, or a device that— whatever it might be in the future, you'll do this for— well, just to repair things, right?

Leo Laporte [02:00:47]:
Like, this is how he— who is one of our regulars in the club, prints his own Crocs.

Paul Thurrott [02:00:57]:
Nice.

Leo Laporte [02:00:57]:
Wow, I'm sure they're very comfortable.

Paul Thurrott [02:01:00]:
Economical way to make Crocs. Well, but you might— you, you will have a unique design.

Leo Laporte [02:01:04]:
I mean, yeah, it's kind of interesting. Yeah.

Richard Campbell [02:01:08]:
What was your first computer? Uh, Richard, Model 1. Okay, 4K RAM.

Paul Thurrott [02:01:14]:
Yeah, yeah, 1977.

Richard Campbell [02:01:16]:
First, the other—

Paul Thurrott [02:01:17]:
yes, repairing those machines. This is the other aspect of this early computer market that fascinates me, the, the what could have been moments, like how Amiga could have taken over the world, or, um, or the Tandy. You know, the thing about the TRS, um, the Tandy model, the Tandy computers, was that they had this distribution network Yeah, right. They didn't believe RadioShacks were this wonder of, you know, bins of little bits and metal pieces and cassette tapes.

Richard Campbell [02:01:46]:
I was in there buying parts when they got the machine, and then it distracted me from the parts. Oh, interesting. Oh, electronic rocket countdown timer, because I'm the kind of lazy that can't even say 5, 4, 3, 2, 1, right?

Paul Thurrott [02:01:58]:
Like, I literally— but you're also the kind of person that would show up, like, hang out at a RadioShack Yeah, and meet like-minded people. Countdown counter for my rocket. Oh, you don't want that one. You want the Model 100.

Leo Laporte [02:02:10]:
You want to, you know, that was like— I missed that. There was also that user group mentality. There was that. Yeah.

Paul Thurrott [02:02:17]:
Yes, early days of makers. There were literally not literal user groups, right, which still struggle along today.

Richard Campbell [02:02:22]:
There's still user groups out in the world, of course, mostly with gray hairs like us. And that's the problem. Reminded me at a dinner the other day The first time we met was when I repaired his badly installed lowercase kit on his TRS-80 Model 1 in like 1979.

Leo Laporte [02:02:40]:
Wow, that's when we met.

Paul Thurrott [02:02:44]:
Yeah.

Leo Laporte [02:02:44]:
Wow, that's cool. What else is— what would you say, Richard? Is there, is there a culture similar to that? I mean, I think they're still there, they're just more subcultures. There's not—

Richard Campbell [02:02:53]:
yeah, it's all different little— and you're right about the 3D printing thing, but you know, they actually—

Leo Laporte [02:02:58]:
a lot of cool cultures, right?

Richard Campbell [02:03:02]:
I think that's what's happening. You know, this part of the world, the mountain biking groups, right? And just, you know, they organize on the internet and so forth, but they're going out to ride places where the bears are, you know.

Paul Thurrott [02:03:16]:
Right. But yeah, so the internet in many ways has kind of democratized access to information and so forth. I find the long tail— in the mid, uh, actually the early 1990s, I went to a like a, like a video game store, and they— the company had bought some of the assets of some company that had gone out of business in California, and they had all these brand new Atari and Intellivision and Coleco cartridges that in shrink wrap that had never been opened. It was in boxes. And when I first discovered this stuff, I was like, oh my God, like I still had an Intellivision at the time. So I was buying these things for $2 apiece, $2.50 maybe, But the internet started happening, so through AOL, I could go onto these Usenet newsgroups and say, "I have these cartridges I can sell you for $5 that I bought for $2 or whatever, or $10 or whatever they are," including some really rare ones. But that enabled me to reach this kind of bigger audience. It was great until they started charging me more because I was buying so many of them and they were like, "Wait, this stuff is actually valuable." But then, you know, the downside or the flip side of this is that the internet also enables every one of these little microgroups, right? So instead of like these big shared experiences that most people have or some large audiences have, you have these small experiences, you know? And so it's hard to find a single thing where you're like, we're all rallying around the Commodore 64 or whatever it might be, you know? You have, yeah.

Paul Thurrott [02:04:44]:
You have these little— there's a— instead of like 3 big groups, there are a million little groups, you know? Right. It's, it's, it's an interesting problem. I don't know. I mean, I'm not looking— I'm not calling for the end of evolution or anything, but— and I don't think that it was ever perfect at any point in time. We should go back to that. I don't mean it like that. But I do wish there was that sense of wonder to technology because some of it is really wonderful. Yeah.

Richard Campbell [02:05:09]:
But the companies that are providing this to us by and large are terrible, right? Well, the expectations of the average person, young person wanting to play with a machine— because they already have a phone, right? Like, how do you present them with a Raspberry Pi and a command line when they have an app?

Paul Thurrott [02:05:25]:
Well, maybe think about it this way. Maybe, um, when I saw a— I can't remember what it was, either a VIC-20 or Commodore 64— but I was in a Sears with my mother and I had these computers. I'd never seen anything like this, and I was like instantly I imagined what I was going to make with this thing, right? What I imagined was something I still could not make today, which was, uh, basically like a Star Wars type game where you're, you're flying in some ship down the side of a giant ship that's going by you and, you know, fighting things or whatever. And but that's where my brain went immediately. I was like, I want this because I want to build something. I want to— I'm going to learn how to program this thing. I'm going to do this. So if I were to give my, like, a phone to a kid on their birthday or Christmas or something, or do they have that? Are those people around? Are there people out there now who are like, they, they see this technology thing? I just wonder if this is just too much of it.

Richard Campbell [02:06:18]:
They don't think that is— my 3D printer gets that reaction out of certain people where it's like, I know what I want to make.

Leo Laporte [02:06:24]:
Yeah, creators on TikTok, you know, they're making videos or they're— so I think there's a huge explosion of creativity among young people, but I don't think it's, uh, yeah, it's not necessarily electronics as much as it was.

Richard Campbell [02:06:36]:
You know, we had Heathkit.

Leo Laporte [02:06:37]:
Robotics usually grabs them pretty hard too. Robotics, very cool. And FIRST is still a thing at many schools.

Paul Thurrott [02:06:43]:
Yeah, I want it to be something that's not nostalgia or fake nostalgia, which is how I describe these hipsters who have to have vinyl records and iPods now and whatever other nonsense. And it's like, guys, there's a reason we moved past all of these things. Like, the thing you don't remember because you weren't there and/or just don't remember is these things were all imperfect, you know. Um, and I— whatever, I'm not— I mean, look, I've been— like, all of us have been in this industry essentially our entire lives. Like, we're obviously—

Leo Laporte [02:07:16]:
we care about technology. A lot of energy got channeled into gamers. So the people who are gamers today probably would have been more like you, Paul? How can I write a game? Or more like you, Richard, how can I make music come out of my ham radio with this little dial? Well, maybe AI does it though, right? And I think AI is going to bring back some—

Paul Thurrott [02:07:35]:
make me a game hacking creativity. I have an idea for a game. Maybe. I think it might be exciting for people if they understood they could make things.

Leo Laporte [02:07:46]:
There was this whole maker movement, which I wish it'd gotten a little farther. Uh, you know, Dale Doherty is a good friend, and, and there are still makerspaces in many communities, but it's— and that's sort of bifurcated into the 3D printer crowd. But I don't think it's as—

Paul Thurrott [02:08:02]:
yeah, as big as it probably can't be because there's so much of it and, and because of this internet problem. The, you know, uh, if you had said to me 30, 35 years ago, uh, you can have news but it will be the news you care about. It won't be the stuff you don't care about. You don't have to sit through the sports if you don't care about the sports. You can, you know, it sounds wonderful. Yeah. But then you realize that all this has done is created these little silos where we all hate each other because you don't believe exactly what I believe. And it, you know, I don't think this is a solvable community.

Richard Campbell [02:08:32]:
Was like that back then too. Like, oh yeah, you're an Atari guy, 100%, not talking to a trash guy, and strongly too, vice versa.

Paul Thurrott [02:08:39]:
Like, yeah. But you know what, it was Terrible. I've experienced this with sports and I've definitely experienced it with technology. You grow up at— I grew up in Boston, big sports town, lots of good teams in the '80s, et cetera. So you have this kind of attitude about it. But then as you get older and you rewatch games from the past, maybe, or watch specials or whatever it is, you realize, man, that team from LA or from Chicago or from where else was excellent too. And you have a different level of respect for that stuff. And you're not this— well, hopefully you're not.

Paul Thurrott [02:09:08]:
My brother's still like this actually. An idiot, but like a nonpartisan, like, yeah, a total homer. And more of a, like, I just love the sport. And I, you know, and I think the way I got into it was when the Patriots had that huge run with Tom Brady, we would watch football all the time, obviously. And when the Patriots were done and it was another team, I was always like, oh, thank God, now I can just enjoy the game. I don't have to stress over who— I don't even care, you know. And At least it's like that with technology now. Yeah, technology is like that too.

Paul Thurrott [02:09:39]:
So I was a Commodore guy, you know, hated Atari, but I, I've watched so many videos and read so many books and I know so much more about this stuff. And there were these incredible technical achievements or just wonderful things that happened on all these platforms, and they were all pretty excellent in their own ways.

Leo Laporte [02:09:55]:
And back in the day, we did have general-purpose computing magazines like Byte, which— right, it is now very much all niche, uh, stuff.

Paul Thurrott [02:10:05]:
I mean, I just realized— so it's funny you mentioned Byte. So Byte was made, uh, published by Macmillan. Macmillan was a sponsor of that Computer Chronicles show, which is fantastic, right? I've watched a bunch of that in recent— yes, sadly. Um, it's a great show. Um, but I'm not going to get this exactly right, but in the beginning of one of the episodes, it was like, you know, Computer Chronicles is brought to you by Macmillan, the publishers of Byte magazine. Magazine. And it was like the magazine for people who love software, hardware, and programming. And I was like, oh my God, that's me today.

Paul Thurrott [02:10:40]:
I was like, that's an interesting way to— I mean, I might be off on how I describe it. It might not be that exact phrase, but it was those three things. And I was like, yeah, I was like, yeah, that's pretty close. It was 40 years ago, 30 years ago, whatever.

Leo Laporte [02:10:55]:
But yeah, that.

Paul Thurrott [02:10:56]:
That thing. That—

Leo Laporte [02:10:58]:
it's true. Anyway, oh well, a little nostalgia. I like going down memory lane, I really do.

Paul Thurrott [02:11:02]:
Um, I do too, but I, I mean, I feel like we're in this—

Richard Campbell [02:11:07]:
avoid—

Leo Laporte [02:11:08]:
I feel computing period ever now, to be honest.

Paul Thurrott [02:11:10]:
I think AI has really brought— yeah, well, okay. I mean, I will say, tied to that, when I— as I was growing up and as the late '80s occurred in the early '90s, there were all these books being written about Apple about Apple, because Steve Jobs had left, and about Microsoft in the early history and all this stuff. And I was going to get into this industry and I was going to learn to become a programmer. And I really felt this sense of loss that I had missed out on this. The genesis story never happens again, right? It's come and gone. It's never going to be that exciting again. But the thing that did happen was the internet and it got exciting again. And so I was lucky enough by the mid-'90s to be writing and publishing books and doing things and being part of it, a small part of it.

Paul Thurrott [02:11:50]:
I was able to be there as it happened, not read about it after it happened. And yeah, with AI now, if anything, I feel like it's accelerating.

Leo Laporte [02:12:03]:
It's rather insane. I mean, I'm imagining on Monday we're expecting, I think might be a very big announcement from NVIDIA.

Paul Thurrott [02:12:10]:
Right.

Leo Laporte [02:12:10]:
Oh yeah.

Paul Thurrott [02:12:11]:
Their new Gentic platform. I meant to say, I'm not happy with the GTC thing. Um, but I— maybe this will be a minor announcement for them in some ways, but they are about to come out with some kind of an ARM chipset thing within MediaTek. Um, their own, their own. Notice that Mobile World Congress came and went and there were no Snapdragon X2 announcements, there were no ARM, NVIDIA, anything announcements. And so they're holding on to it. It has to happen soon.

Richard Campbell [02:12:40]:
It's going to be the first half of the year. So that NVIDIA thing was Okay, so you've been doing well selling shovels, so you really— and now you decided you need to dig some gold?

Leo Laporte [02:12:49]:
Like, what?

Paul Thurrott [02:12:50]:
This will be called Nemo Claw. They need a, uh, a shovel adjacent, uh, something like, you know, Apple, uh, their app— they are to AI what Apple was to phones, right? So you have this one super explosive successful thing, awesome, it's going to slow down eventually. What do you do next? Like, you can't it's not going to last forever. Like, you have to do more. And, um, you know, we'll see if Nvidia can make that happen. But not that long ago, that company was just about nothing. Yeah.

Leo Laporte [02:13:20]:
And today they're just about everything. And, uh, flying cars are coming this summer.

Paul Thurrott [02:13:28]:
So finally. And then, and then after that, the paperless office. It's happening, folks.

Leo Laporte [02:13:34]:
We did it. We did it. We did it.

Paul Thurrott [02:13:39]:
Oh man.

Leo Laporte [02:13:39]:
All right. All right. So that's your tip of the week. Yeah.

Paul Thurrott [02:13:44]:
So think back to the good old days. Yeah. Now I'm going to look to the future. So I've been working on this series of kind of Notepad clones, having a lot of success, but also a lot of defeats, mostly lately because the Windows App SDK is so terrible. I've tried all kinds of AIs. I told you about the thing from last November where it kept making so many mistakes that I ran out of credits. And I'm like, what? Which is insane. But, and I had also said in the past, like, look, I would never pay for, I just said it today, I will not pay for AI.

Paul Thurrott [02:14:13]:
If I was a developer, I would absolutely pay for AI. And that's a different, it's just, that's stuff. Yeah, that's right. So there's different ways to go with this, but Stardock, the company that my friend Brad works at, which is run by Brad Wardell, another friend of a sort. A great guy. They make, you know, Start 11 window blinds, fences, etc. Came up with a, I guess I'll call it like an AI kind of a front end. It's not just for developers.

Paul Thurrott [02:14:41]:
I have only used it for developer stuff. But in the sense like OpenClaw or Is For Agents or, you know, Cursor was like a year ago, you can point it at a code base and you can it can help you with it, right? So you do, you do code review, which is incredible. To this, to literally to this point, and this has been a month and a half, maybe month, over a month anyway, it has never screwed me over, not one time. It has never created code that did not compile, which has been a huge problem with the GitHub Copilot stuff. It has fixed some embarrassing problems with my codebase, including a couple of async issues like I mentioned, which I can't I don't understand because you always use async for file operations and what the hell, Paul? Anyway, I can't explain that one. But there are features I had in old versions of the app that were on WPF that I could not implement in the Windows App SDK because it's so hard. And it did those. It added new features I've just wanted to add.

Paul Thurrott [02:15:44]:
And this has gone so well that I finally decided— I don't know why I kind of jumped the gun, but I was like, I want to build the thing I always wanted to build, which is that multi-app, sorry, multi-document, multi-tab version of this app. I meant to have it build it from scratch just to see if it could and to see what it would look like, but it actually used my code base to do it. And it did it. Goddamn it, it did it. And it's wonderful. And part of the reason it's wonderful is because I had actually solved the architectural issue of managing the state of multiple documents and tabs, but over a year ago. It's just that the Windows App SDK is so terrible that I— this is a hard thing to explain. I don't want to go into too much detail, but if you think about switching between tabs in an app and maintaining the state of whatever it contains, in WPF, there are events that fire before you switch.

Paul Thurrott [02:16:42]:
And in Windows App SDK, they only have events that fire after you switch. So it's actually very easy to lose state. Anyway, it took what I did, it fixed it, it spit out a version of this app. This, I swear to God, this kills me. I'm just going to mention one more thing about this because this kills me. There were things— I started working on this, some version of this app, starting with Visual Basic, by the way, several years ago. So I'm going to say 7-ish years ago. I don't remember.

Paul Thurrott [02:17:10]:
I've made multiple versions of it over time. So Windows Forms with the VB, Windows Forms with C#, Windows Forms with C#, and then at the time modern like.NET Core versions. I did a Universal Windows, what do you call it? UWP version. But the best version I ever did was WPF. It was fantastic. But one of the things dating back to literally V1, if you look at Notepad today, if you look at like, this is just an esoteric kind of, there's two things. In the bottom right of the corner, it says UTF-8, which is an encoding format. This is kind of the modern text encoding format, but it supports multiple formats.

Paul Thurrott [02:17:49]:
And it also says Windows CRLF, like carriage return line feed. And this is a line ending format. So Windows— well, Windows, yes, but Notepad supports multiple line format, line ending formats, I call them, I guess, and then encoding formats, right? Dating all the way back to literally ANSI. I'm pretty sure you could create a ANSI text file in this thing if you wanted to. I don't know why you would, but anyway, it's good because it can import those files, right? You could read everything. I was never going to figure that out ever, ever, ever. I looked into this. It's deep, deep archaic.

Paul Thurrott [02:18:23]:
It literally dates back to the early versions of Win32 and the Windows SDK. And I worked on this several years ago. And I was like, no, screw this. And I could justify it by saying, look, no one using this is ever going to do anything but UTF-8 and Windows line endings, right? Who cares? But I asked this thing to make an app that had all the features and goddammit, it freaking supports all the encoding formats and the line endings.

Richard Campbell [02:18:52]:
It's very well documented.

Paul Thurrott [02:18:52]:
The tool is going to be great at that. Goddammit. Like, what the fuck?

Richard Campbell [02:18:59]:
That's incredible. Like, that's incredible. You're disappointing the Uno guys. I had them on the show the other day and they were desperate to help you with your app.

Paul Thurrott [02:19:04]:
And now you've fixed it all. Oh no, no, I'm— one of my next steps will be making it cross-platform with Uno. I'm definitely—

Richard Campbell [02:19:12]:
I'm absolutely doing that.

Paul Thurrott [02:19:13]:
Those guys are great. You should definitely— they are great. They're the best. No, they're— they're— if anything, they're too great. They're, they're so nice. They were— they, uh, Jerome, or one of those guys, it was, it was like, um, he's like, just send us the code, we'll put an engineer on it. I'm like, what is wrong with you? No, this is, this This is like a hobby project. No, please don't do that.

Paul Thurrott [02:19:34]:
But this has enabled me to get over the, you know, get over this, right? So I've restarted it, the project. And so, you know, you can imagine like 3 instances of Visual Studio running, right? So the new app, my old app, and the app that it created, right? So I can go and I can see how it took the class, the model they would call it, that I created to manage state and extended it to support, by the way, the line ending and the formatting, the encoding rather, because those are part of that document that's in each tab, right? Because they could be all different. And so it actually manages that state as well. I've never seen anything like this. This has never not worked properly. One of the things that I didn't think worked that it did was it added AI-based writing help. Tools, right? And I was like, I can't get this to work. And it also said to me, I swear to God, it's like, here are 4 possible implementations.

Paul Thurrott [02:20:32]:
We recommend doing 2 of them. And in this case, it meant you can use Ollama, which is a locally downloadable SLM, and as a fallback if you don't have anything else, or you can have Gemini. If you have an API key, you can paste it into settings, you can use that. And I couldn't I was like, I installed Ollama. I'm like, this isn't working. I don't know, it's failing. And then it said, it was like, look, we're going to change the error message so we understand what's happening better. And is that okay? I'm like, yep.

Paul Thurrott [02:21:00]:
And then it did the thing and it came up and it said, yeah, you don't have the Ollama model installed. You have the Ollama app installed. And I was like, what the? I was like, and then I installed the model. Which it did for me.

Richard Campbell [02:21:17]:
Nice.

Paul Thurrott [02:21:17]:
And it worked. It's here to help. It's incredible. If I was a developer, I would pay— I would pay $200 a month for AI. There's no doubt about it. It would be the greatest gift.

Richard Campbell [02:21:31]:
It's so justifiable. Yeah, I think so.

Paul Thurrott [02:21:33]:
The productivity is astonishing. I cannot— anyway, so I, I recommend people look into this tool. Clairvoyance is magical. It also works for the co-work type things where you can make a game. In fact, it has all these sample projects you can do. It will generate on the fly. And so I was like, yeah, make a video. It makes this perfect Asteroids clone.

Paul Thurrott [02:21:55]:
And I'm like, oh, fucking— It's unbelievable. But you could also use it for productivity type things. Brad's daughter uses it to make flashcard tests for an exam she's going to have or whatever. It's going to enable people to do that kind of vibe coding based personalized app thing as well. Like I'm using it very specifically for software development, but it's rather incredible. The only problem— models they're using. So that's up to you. And so I'm right.

Leo Laporte [02:22:19]:
I should, I should have said that it's just a harness for—

Paul Thurrott [02:22:23]:
yeah, I should have. Yeah. So I use it with Anthropic Cloud. I have the Pro, the $20 a month thing. So you just plug it in and— but it can work with anything. You can use it with Gemini if you want. You can use it with Copilot if you're that mental. It's still on a waitlist, like it's preview.

Paul Thurrott [02:22:41]:
I think it was Leo mentioned earlier in the show, he was talking about how sometimes you almost just have to reboot or reset the thing because it kind of loses its place. And of course that passes through, right? So every once in a while it's like, yeah, you're not looking at the right thing and you just restart the app and it's like, oh, here we are again. And it does a really good job at that kind of thing. It does a really good job at everything. I've never used an AI like this that has— it's never caused a problem. Like, it's like, it's never— I can't say it's disturbing. It does. This is why people believe this stuff is magic or godlike, right? I mean, when you see it actually do this, it's like, oh my God.

Paul Thurrott [02:23:24]:
Like, it's— and especially because I've experienced so many failures. You know, that helps because, well, I mean, I didn't want to go through that, but I mean, having done that and going into this kind of gingerly and sort of also feeling like every 6 months or 9 months there's this new flavor of the month in AI, like, you know, Cursor had their moment in the sun like a year ago, whatever, and thinking like, well, you know, things are going to kind of, things are going to change again. And I don't know, but I cannot believe how well this works.

Leo Laporte [02:23:52]:
So you need to have an object desktop. Subscription?

Paul Thurrott [02:23:56]:
Well, that's one way to get into it right away. No, you don't need to have one. Oh, you don't? If you do have one, you can get it right now. If you don't have one, it's free, but you have to— it's a waitlist for now because they're still iterating through, or it's still early in some ways. But it makes these— they call them exhibits, but it's basically like a— it looks like a slide from a slide deck, which shows you all the options.— because one of the great things you can do is say, I want to do this thing. Like I said, with the AI writing tools, I did this for also recent files implementation, like a menu with recent files and how to store that. It came up in that case with 5— so these cards with the 5 versions, the pros and the cons of each, difficulty rating of each, how far do you want to go? And then a recommendation about the— and the one that it recommended which was the first one, was exactly how I had already started implementing this in a different version of the app. And I was like, yeah, now do this.

Paul Thurrott [02:24:55]:
And it's— it worked perfectly the first time.

Leo Laporte [02:24:59]:
It just worked. It looks like kind of their version of Cowork.

Paul Thurrott [02:25:05]:
Yeah, basically. Yeah. Yeah. I mean, I think we'll see where this goes, you know, but as of right now, I'm paying for Anthropic, and I'm not paying them anything, right?

Leo Laporte [02:25:18]:
And it is— did Brad—

Paul Thurrott [02:25:19]:
was Brad involved in this, or— yeah, yeah, he was. He's been talking to me about this for months. He's like, he's really excited. This, this could be a completely different direction for this company, you know?

Leo Laporte [02:25:29]:
Yeah, no kidding.

Paul Thurrott [02:25:30]:
This is very different from what they normally do.

Leo Laporte [02:25:34]:
Yeah, it's, it's cross-platform.

Paul Thurrott [02:25:37]:
Yeah.

Leo Laporte [02:25:37]:
Did they vibe code it?

Paul Thurrott [02:25:39]:
Um, I bet they did. I think parts of it may be, but, you know, Brad Wardell is an astonishingly good programmer, and they obviously have programmers. I mean, it's— there's a whole, I don't know, team of people or whatever. But yeah, I mean, I, I— at some point you use the thing to make the thing, you know, if that makes sense, right?

Leo Laporte [02:25:59]:
You know, when did she start— everybody's—

Paul Thurrott [02:26:01]:
even Claude and, you know, OpenAI are doing that now. Yeah, exactly right. So what Once a tool like this improves to the point where you can use it to help make that tool, right, I think you've crossed a very interesting milestone, right? Anyway, interesting. I just— you can't just— unless you do have an object desktop, you can't just go get it right now, but because they're in alpha, but get on the waitlist. And, and I, I think people are going to be blown away by— well, I, I know they are. It's incredible because he was very excited and he showed me like remote demos. And I was like, okay, let's go. And then I started doing it and I'm like, I'm gonna run— I've run into problems, I know exactly where the problem is going to occur.

Paul Thurrott [02:26:39]:
None of those problems occurred. None.

Leo Laporte [02:26:43]:
Like none, right? It's astonishing. AI workspace where agents work as persistent staff across your projects. Oh, that's the other thing.

Paul Thurrott [02:26:48]:
So this is a harness and again, yeah, multiple agents, right? I called my agent Brad, of course, because I like the idea of working on it. Um, you can also do things like where you It's an interactive chat experience, obviously. It's multithreaded in the sense that you can give new directions to go do other things while it's doing a thing. And you can see the agents kind of start over on the side.

Leo Laporte [02:27:16]:
And it's rather impressive. This is an explosive area.

Paul Thurrott [02:27:20]:
This is the year of the agents, basically. And this is, I think, when I talk about how people will have an aha moment with AI, This is— this could be it for a lot of people, this type of thing. It doesn't necessarily have to be clairvoyance. It is for me, but, um, this is the type of thing where you're gonna have a moment. It didn't just correct like the spelling of a word, it created a thing. And you're like, okay, so that was impossible, or that was something I was never going to do on my own. And in my case, it's happy— you could just go back and read the articles, but I think there were 20-ish instances of things where I'm like, I don't think I'm ever going to solve this problem. And it just did them all in like a day.

Paul Thurrott [02:27:59]:
I mean, well, did them all immediately, and then I did them, you know, I went through them and I was like, uh, how, how?

Richard Campbell [02:28:05]:
I keep waiting for it to screw up. Yeah, I have a friend with the 6 Anthropic Max accounts who's now configured his phone so he's just looking at the results of each of the agents and then telling them what they did wrong and what to do next. That's more than $1,000 a month he's spent. Yeah, but he's knocking out 6-week sprints in less than a week.

Paul Thurrott [02:28:26]:
Yeah, I mean, if this is your job, it's probably causing the tornadoes in the Midwest right now, but it's fine. He consumes enough tokens that Anthropic calls him on a regular basis and goes, uh, that's the other— well, I don't want to get too deep into this, but it's the way it consumes tokens on your behalf is fascinating because it actually tries to save You feel like it will do things in a way that will save tokens. Like, yeah, you know, things you can do tells you that, and it's like, yeah, like, it's, it's neat.

Leo Laporte [02:28:56]:
Okay, I'm sorry, I've been babbling, but it's really— the real, the real question for a lot of people is, uh, this is true of OpenClaw as well, is, is it Claude, is it Opus 4.6 that's really the impressive thing, or is it the harness?

Paul Thurrott [02:29:08]:
All right, so, right, so I'm sorry, I, I've written several articles about this, but, and I'm trying to condense this into like a 10 or 2,000-minute conversation, but, um, I did start using just Cloud Directly and it ran into the same kind of problems that I was having before. Yeah, um, it was better, but there's something about using this with that.

Leo Laporte [02:29:27]:
Yeah, you can improve on it for sure.

Paul Thurrott [02:29:29]:
Yeah, it just, it put it right over the— it's not even close, like it's a huge difference.

Leo Laporte [02:29:35]:
Yeah, I'll have to try it. Uh, we have a few more minutes, uh, to go on the show. Coming up in about 15 minutes, it is Guy Kawasaki's going to be our guest on Intelligent Machines.

Richard Campbell [02:29:45]:
Oh, really?

Paul Thurrott [02:29:45]:
Oh my God. A lot of fun. Speaking of early books from the computer industry in those areas, like The Macintosh Way. Yeah, that's right. A bunch of books about marketing and stuff I don't care about too much, but yeah, an amazing individual.

Leo Laporte [02:29:57]:
He just wrote a book on signal and we're going to talk about using signal to protect you.

Richard Campbell [02:30:01]:
I follow him on Substack, I think. Big signal fan, but I hang around with too many security people. It got me. I couldn't, I couldn't talk to him if I didn't use Signal.

Leo Laporte [02:30:09]:
Yeah, that's right, right. Uh, and actually, uh, Apple's 50th anniversary is, is April 1st, so the day before that, David Pogue will join us to talk about his new book, uh, 50 Years of Apple, which just came out yesterday. And that'll be on MacBreak Weekly. So if you're an Apple fan, it's a banner month. But if you're not, Run As Radio has something for you.

Richard Campbell [02:30:32]:
We don't hate Apple on RunAs Radio. It's ever so often we have to manage them. You know, it's just another piece of hardware to manage. Although I remember saying I did a show, uh, talking about Group Policy and Active Directory and the fact that it was becoming actually easier to manage non-Windows devices. Like, if you had a Windows device, things were harder because it was all this archaic management rules that Intune and the new approaches to Group Policy are much simpler on. Nice. But that was not this conversation. I just counted up.

Richard Campbell [02:31:07]:
This is the 17th appearance on RunAs Radio for Bob Ward. Admittedly, he's done a bunch of panels over the years, but I can't talk about more regular than Bob for just what do you want to know about SQL Server. He's been the advocate in that space forever. He runs the Microsoft Data Group, is an architect. Like he's the guy to talk to. He knows his stuff and he's a pleasure to talk to. Well admittedly we got distracted for a minute. They were talking about being granddads.

Richard Campbell [02:31:36]:
I got some nice comments on that. It's like congratulations. So nice. But what we really were talking about was SQL Server 2025, which came in full release in the Ignite timeframe last year. And then we finally got Bob on the show to go over that. And then a lot of conversation about some of the AI-related features related features, like there's now a, a RAG vector store as a type inside of SQL Server, along with a bunch of other constructs. And of course, the infinite number of Copilots that are related to SQL Server, both in SSMS and some other tools. And a great extension for Visual Studio Code if you need to work with SQL Server, make your life much easier with this new extension.

Richard Campbell [02:32:13]:
So just a rundown of all the good stuff, as is you would expect expect from any show I do with Bob Ward, we're going to talk about SQL Server and we're going to go deep. And as you might expect on any show we do with Richard Campbell, there will be whiskey. There will be.

Paul Thurrott [02:32:31]:
And I'm in Canada, so it's going to be Canadian whiskey.

Richard Campbell [02:32:36]:
Hooray! You know, I've talked with the Sons of Vancouver was such a hit. I was so happy about this. So I was looking for some other craft whiskey Well, this is not that crafty in the sense of it's not brand new. This is a distillery called Highwood, and it is actually in a town called High River in Alberta. So next province over, it's about 45 minutes drive south of Calgary. It's also about that same distance from the Rocky Mountains itself, but that would be due west. Today, that town's about 15,000 people, although there's been people living around there literally for centuries, possibly millennia. The Blackfoot First Nation's been in that area, and of course stretches down to the US and so forth.

Richard Campbell [02:33:18]:
European settlers don't show up in that part of the world till 1877, which is after Confederation, right, which is 1867. Uh, it's named of course for the Highwood River, and of course then the town is built at the ford at the river they call The Crossing. So back in the day, without a lot of technology, you're driving your cattle around, this is where you go through. So naturally a town conglomerates around there, although just a small one. Calgary is a multi-million person city or million and a half, and Highwoods just 15,000 people. This part of the world is considered what they call humid continental for the climatologist type. So this is mild and warm summers with very cool, with cool nights even in the summertime, and the winters are cold. Cold and snowy.

Richard Campbell [02:34:02]:
They average about 70 inches of snow most years. Southern Alberta is notorious for hitting negative 30s and negative 40s at certain points in the, in the wintertime, but then also having a warm wind they call a Chinook. So, uh, but different, different style of life here at higher altitude, you know, not like the coast life that I have. And they're gonna make whiskey there. So Highwood Distillers was founded in 1974.

Paul Thurrott [02:34:28]:
'74.

Richard Campbell [02:34:28]:
So that makes it an old distillery. Normally a distillery that old would have been grabbed up by a conglomerate by now, but they are fiercely independent even though they've gone through a few different owners. So in '74, uh, High River had about 3,000 people, a much smaller town, and then the original name was the Sunnyvale Distillery. So it was a group of investors out of Calgary. They hired a distillery— a distiller from Germany to build a distillery. It was the only one in Alberta at the time, although today Alberta is a massive producer of whiskey and other drinks. They rebranded as Highwood in '84, and this particular whiskey that I've got, which they call the Centennial, was actually launched in 1996. Uh, as they grew, they moved their sales marketing office back up to the city of Calgary.

Richard Campbell [02:35:10]:
So what's in High River is just a distillery facility and all of the storage and so forth. They acquired the only distillery in British Columbia in 2005. A distillery called Potter's Distillers, which was actually established in 1958, so even older than Highwood, by a guy named Harry Potter— no relation— um, who's, who's from Langley, which is where my daughter lives right now with my granddaughter. Uh, and then it was passed on to a guy named Harold John Cameron Terry, because he was Australian, so you get names like that, in 1962, who moved it, and then in 1990 to Kelowna, and it became a thing called Aschidia Brands, which still exists to this day. While they do a little bit of distilling, they mostly blended other whiskeys. It's a very common practice for Canadian whiskey. And when Highwood acquired them, they largely stopped their production. In fact, they only produce— they, they're actually on the grounds of a large brewery called Granville Island Brewery, and they only make whiskey there one week a year to maintain their distiller's license.

Richard Campbell [02:36:11]:
And other than that, they mostly do blends. Things like that. Uh, that acquisition scaled up Highwood for a fair bit, so they in 2006 built additional facilities in High River, their major employer there. And then just recently, this distillery— the Highwood was acquired by another distillery, Caldera Distilling, another Canadian one. This guy named Jarrett Stewart, who was originally from Alberta, worked in the energy industry, got disenfranchised of that at one point or another, and decided to get into whiskey to the point where he actually went to the UK to Rotherhithe and studied at the Institute of Brewers and Distillers, and then chose to build a distillery, build Caldera, in River John, Nova Scotia. And he was making product back in 2013, all very local to the island, Newfoundland Island. And then a friend of his from the energy industry let him know that Highwood was available for acquisition. And so in 2022, they decided that Caldera would buy that combined it together, and it's only been more successful than that.

Richard Campbell [02:37:11]:
In fact, this whiskey that I'm trying today, the Centennial, won the bronze medal in the Canadian Whiskey Awards in 2024. So we're talking about a 40-plus, almost 50-year-old distillery that's still not part of a big conglomerate. It's been around for decades. It's very rare. They also make a huge array of products. Whiskey is not their primary product. They mostly make vodka— well, premixes are their largest business, then vodka. Whiskey is only about 20% of the business.

Richard Campbell [02:37:38]:
What's a premix? Uh, this— so this is your like, uh, whiskey cocktails in a can. That comes in a can? Yeah.

Paul Thurrott [02:37:45]:
Okay.

Richard Campbell [02:37:45]:
So they make a bunch of vodka and they make a rum and a few other things as well. All that stuff is easy to produce, but premixes pay very well.

Leo Laporte [02:37:51]:
They're cheap to make and you make them properly.

Richard Campbell [02:37:53]:
It's like owning a bar. That's how you can— it's only about 30 employees. It's not a huge place, but they, they make money. Their distilling approach is very interesting. They do not do any distilling of corn. The most of their whiskeys are wheat and rye. They do occasionally use corn spirit, and I strongly suspect there's corn spirit in this whiskey, but they buy it from other producers because it's easy to come by. They also, because they're working with wheat and rye, no barley, they don't do any malting, but they do use a pressurized mash tun, which is very unusual.

Richard Campbell [02:38:29]:
So they have a mash tun that will actually cook under pressure. So you load about 2,500 kilos of grain, that's 5,500 pounds, with 12,000 liters of water, about 3,200 gallons. The water does not come from the river, it comes from a well. And then they cook it under pressure to break, to turn it into a kind of porridge, produce about 14,000 liters of mash. And after they cool it down, then they add the enzymes. So that's a biological process to convert the starches into sugar, and then they do a very long fermentation that results in about a 12% beer. That's high. Normally you look at about 8 or 9%.

Richard Campbell [02:39:09]:
Uh, in their documentation they say they distill in a beer still, but you would call it a column still and raise it to about somewhere between 70-75%, which is a very high pass, and then cut it down again and redistill it in a pot still. So they cut it down with water to 40%, then run it through a pot still back up into the 70s. Also, which is common in Canadian whiskeys, not common in other places, the wheat and the rye are distilled entirely separately and only meet in the mixing vats before they go to barrels. And it seems likely, and you'll see when we talk about the price for what they call the Centennial, they probably put some corn spirit as well, are basically neutral alcohol. They go into the barrel at 72%. That also is very high. And in the case of Centennial, this is ex-bourbon barrels. Normally you'd go in at about the, the Scots going at 63.5.

Richard Campbell [02:40:05]:
So 72 is pretty darn high. But you're talking about southern Alberta. So the humidity there is pretty low, which means they don't give up any angel share. They tend to lose more water than alcohol. It is common after 3 years to come out of the barrel at 75%, having gone in at 72. Now I want you to look at this bottle. Notice it says Centennial on it. Nowhere on it does it say 10 years old.

Richard Campbell [02:40:36]:
And there are rules, right, in Canada for an age statement. If you're going to put a 10 on it, it's because the youngest thing in that bottle is 10, same as the Scots. Same as the Americans too. It's not on there, although the website says 10 years old. It is called Centennial, which only makes me wonder, is it actually 10 years old? I don't know the answer to that for sure. People are pretty comfortable with it. The website itself, like the web, the URL that I'm providing here, highwooddistillers.com, Centennial 10-year-old Canadian rye whiskey. But the primary grain in here, I can't— I haven't been able to find the actual mash bill, mostly wheat.

Richard Campbell [02:41:13]:
The rye is kind of a flavor green. So no sharp nose, is only 40%, so not— no big buy on there. Boy, it drinks nice.

Paul Thurrott [02:41:26]:
It's not that—

Richard Campbell [02:41:27]:
it's a little sweet. I would bet there's corn in this, even though they don't say it anywhere in the documentation. But the reason I'm sure it's not just straight 10-year-old wheat and rye is this is a $25 bottle of whiskey. $25 Canadian. So that's maybe $20 in America if you could find it. That is really a very reasonable price for a whiskey that drinks so nice. Nice enough, I think it'd be a sin to make this into a cocktail, and it's a $25 bottle of whiskey.

Paul Thurrott [02:42:04]:
Wow.

Richard Campbell [02:42:04]:
And that seems to be the statement for most people who try this, is like, this is magic, really, because they've got a really curious distillation process. It's kind of strange, and yet they've done a great job. They make a bunch of other editions, but this one, people say great things about it, and I can see why. Like, it's, it's very drinkable. It wins awards for a reason, but it's almost so cheap you can't believe it's true. It's mostly— it's made in winter wheat. Soft Winter Wheat with rye added. But I think the only way you get down to that price is they probably got some corn spirit in it.

Richard Campbell [02:42:39]:
But I don't care, it's good. It's a bit— maybe a bit of a cheater, but boy oh boy, it drinks nice.

Leo Laporte [02:42:43]:
So I don't know why you'd be unhappy with this.

Paul Thurrott [02:42:49]:
Very nice.

Richard Campbell [02:42:49]:
Can you get it in the US? Wasn't able to find it at the BevMoza or any of those. Once again, he's teasing us. The production is is so low here, right? They only— they're not even—

Leo Laporte [02:43:01]:
they're not really craft distillery, they're not a little guy per se, but they just don't produce.

Richard Campbell [02:43:05]:
30 people is not very, very big. Yeah, yeah. They're not— you know, you want millions of liters, you go up the road to Alberta Distillers, one of the biggest in the world, right?

Paul Thurrott [02:43:15]:
Right.

Richard Campbell [02:43:15]:
These guys have stayed independent for a reason, in a little town for a reason.

Leo Laporte [02:43:19]:
They only make so much. Now, how is their Sweet Sippin' Maple Whiskey?

Richard Campbell [02:43:23]:
Because that sounds pretty good.

Paul Thurrott [02:43:24]:
Lots of people talk about it.

Leo Laporte [02:43:26]:
Pretty good.

Richard Campbell [02:43:27]:
Actually, it comes in a maple leaf bottle. They're not even spirits.

Paul Thurrott [02:43:30]:
They come in at 25 to 30%.

Leo Laporte [02:43:32]:
You could like dunk one of those maple candy things into it as you're drinking it.

Richard Campbell [02:43:37]:
It is 25%. That's nothing. Yeah, that means it's not a spirit per se. It's kind of an— you know what, that's probably not exportable because you don't know how to categorize it, right? It's too high to be a malted beverage, right? Too low to be a spirit. It's an orphan. It sounds good though. Yeah, no, and people say great things about it, and I love the bottle. All right, well, we're gonna— like, that's where you usually— that's like a maple syrup bottle, right?

Leo Laporte [02:44:02]:
When I want to buy maple syrup at the duty-free, that's what it's going to be in. It is. It looks like that. Well, now I know what I want to do when we come up and visit you in Vancouver.

Paul Thurrott [02:44:11]:
I want to try—

Richard Campbell [02:44:11]:
well, I think we're going to Sons of Vancouver first and foremost.

Leo Laporte [02:44:16]:
Oh, absolutely.

Richard Campbell [02:44:17]:
Thanks. Fantastic.

Leo Laporte [02:44:17]:
Thanks once again for the bottle of, uh, number 82 Amaretto. And once again, thank you for an excellent edition of Windows Weekly. You'll find Richard Campbell in British Columbia for a few days more.

Richard Campbell [02:44:32]:
Where are you going next? Um, home next week. Oh good, the MVP Summit, which, if you remember last year's MVP Summit, how did I get 8 bottles of whiskey? I don't know. Where is that held?

Leo Laporte [02:44:43]:
That's in Redmond, Washington. It's at the headquarters. People know where to find you though. That's the key. Not hard to find. Bring him whiskey. I promise. Uh, Richard's also found at— on the internet at runasradio.com, where you'll find Run As Radio and.NET Rocks!

Richard Campbell [02:45:01]:
And, uh, listen to the Geek Outs. Oh, you know, one of the things we're doing at the MVP Summit this year, we're recording episode 2 2000. Unbelievable.

Leo Laporte [02:45:13]:
That's impressive. Paul Theriot is at theriot.com, and of course, uh, you should become a premium member there. His books are at leanpub.com, including— now there's 3 of them— Field Guide to Windows 11 with Windows 10 built right in, uh, the Windows Everywhere book, and his newest, De-Insidify Windows. Cory Doctorow was on on Sunday, and I thank for the, the word, right? Yeah. Uh, you will find them both here every Wednesday around 11 AM Pacific, 2 PM Eastern, 1800 UTC. You can watch us live on YouTube, Twitch, x.com, Facebook, LinkedIn, uh, and Kick, or of course in the Club Twit Discord if you are a member. I hope you're a member. Please join us if you're not.

Leo Laporte [02:46:00]:
Please, I beg you, please. Uh, After-the-fact on-demand versions of the show available at twit.tv/ww. There's a YouTube channel dedicated to Windows Weekly, great place to share clips with friends. That's another way you can support us, of course, by, you know, spreading the word about the best darn Microsoft podcast in the world. And you can also subscribe in your favorite podcast client, you'll get it automatically that way. I guess that's all to say except See you back here next Wednesday for another gripping edition of Windows Weekly. Gripping like a stomach problem kind of gripping?

Paul Thurrott [02:46:41]:
Like grip.

Leo Laporte [02:46:41]:
Thank you, Paul. The classic French grip. Thank you, Richard, and thank you all for joining us. We'll see you next time.

Paul Thurrott [02:46:46]:
Bye.