Ask Paul: May 29 ⭐

Happy Friday! It’s the fifth Friday this month and the end of a short week that felt awfully long for some reason. But let’s kick it off early with some great reader questions.

🎶 Nashville

train_wreck asks:

You excited for the trip to (country) music city? (Nashville, Tennessee). It’s my hometown! Hope you have fun here.

Thanks, yes, though it’s mixed up in a lot of busyness between our return from Mexico and that week, plus the three industry shows (Computex, Build, WWDC) still to come. But yes.

Stephanie and I did visit Nashville once, over 20 years ago, because I was speaking in Huntsville, Alabama and we decided to make a long weekend out of it. Among other things, we visited the nearby Jack Daniel’s distillery, which was in a dry county at that time, so we couldn’t even taste anything during the tour. (You can easily imagine me in the back of a group muttering “Are you f#$king kidding me” when we found this out halfway through.) I guess that’s no longer the case.

But I do recall really liking the city, and the music scene, and it will be interesting to revisit it. I sometimes go into trips like this doing zero research, but I will try not to do that this time, and Steph and I have both passed around a few ideas for the kids to look at. There are still a few weeks to go before we leave.

One thing that came up in my “research,” so to speak, was that there’s an Amazon Prime series out now called Scarpetta, based on a series of books I’d never read, that’s filmed entirely in Nashville even though the show takes place in Virginia mostly. (This is the second recent example of this I’ve noticed; there’s a TV show version of Man on Fire, an excellent book and movie, that was filmed entirely in Mexico City but takes place, in the show, in Brazil for some reason. The movie does take place in Mexico City and was filmed there.) Anyway, hoping that this Scarpetta show would provide some good glimpses of Nashville, we’ve been watching it this week. And I cannot stress enough how terrible it is. We have two episodes to go as I write this and will endure the rest. But … terrible. Unless you enjoy 45 minutes of people just yelling at each other, I guess.

“New Oura Ring 5 unveiled with dramatically smaller design”

Now it only fits on your pinkie

🧑‍💻 Build it and you are dumb

Will asks:

Next week is Build, and I am curious if the entire focus will be on AI? I am assuming there will be some news on Copilot and maybe some Windows news in there as well, but what exactly will the focus of Build be and who is the audience today?

I am curious as well, but there were no pre-briefings and won’t be, per se, as press/bloggers won’t get materials until the night before the show, which I believe is a first. Mary Jo and I were chatting about this just yesterday, and I guess this year’s show will be much smaller than usual. I was invited, which apparently is unusual as many were not. But I won’t be attending in person, in part because I wasn’t sure about our Nashville schedule when it came up and in part because of the expense and uncertainties around the content.

If you care about the client side of things, as I do–Windows, mostly, but Office and related things–and of course whatever developer topics, Build has been a tough sell for many years. This is tied to Microsoft’s broader focus over the past 15 years-ish, of course, with its shift to cloud computing and then, more recently, AI. And to the fact that .NET is occurring separately on its own schedule, with a November launch each year.

Anyway, there’s no one audience at Build so you kind of have to cherry-pick what to pay attention to, and this can make the keynote(s) painful in particular. I don’t expect any new Windows frameworks or whatever. But I do expect Windows news from the show, which I can write publicly because I literally have no idea, whereas in years past I’d have had to be more careful as we are usually briefed on these shows before now. Framing what Microsoft is doing in terms of developer opportunity feels a bit forced these days, frankly. Thanks to AI, almost everyone will be a developer soon, of sorts. What does this leave for professional developers? Massive legacy codebases? I don’t know, but it’s vaguely depressing.

“LLMs believe false statements even after explicit warnings that they’re false”

So just like the people who made them

🐧 Linux musings

spacecamel asks:

I wanted to thank you for recommending Insync a few years ago. Your problems trying to connect your Google Drive reminded me about it again. It is the best solution for Google Drive that I have found for linux.

It’s curious to me that the sync/files on demand thing hasn’t been broadly solved on Linux, especially given how much this platform has improved in so many other areas for desktop users. I mentioned this in Switcher 2026: The Zen of Linux ⭐, but we joked about “the year of desktop Linux” for so long that we kind of stopped paying attention outside that community. The quality I see in Linux now would surprise a lot of people, I bet.

For whatever it’s worth, one of my (mostly unspoken) goals in experimenting recently with Linux was to just use the thing as it is, which is similar to what I did with macOS when I got the MacBook Air two years ago. But once I exhaust what’s possible there, I will inevitably do what I do with the Mac and figure out which workarounds I can use to get to what I think of as that “last mile” and eliminate, hopefully, any remaining missing features I want/need. And utilities like Insync will definitely be part of that.

One of the reasons why Gnome doesn’t connect anymore to Google Drive is that one library “libgdata” is not maintained and the whole thing was removed in the latest Gnome 50. Do you think we would see a day where AI could maintain some of these smaller but important projects that are the building blocks for many bigger projects?

Yep. This is something that I’ve been thinking about a lot lately, and it’s a natural progression from things like Rust in the Windows kernel, Mythos and other AI finding bugs in Firefox and other platform codebases, and the vibe coding stuff that’s starting to finally become real. For example, it seems like a gross oversimplification to make a statement like “Microsoft should just refactor the Windows/Azure/whatever kernel in Rust” … until it isn’t. I feel like that’s inevitable now, and not so far off.

More to your question, I also feel like we will go through a period of pain regarding AI triggering an exponential explosion in bug/security vulnerability findings until that finally settles down and we will use AI more proactively. And part of that, absolutely, will be assisting in the maintenance of open source codebases that are today managed by a single person, never thanked or compensated, or by no one at all. We as users will be able to tell whatever AI in whatever Linux distribution that we need to have full offline sync with files on demand capabilities for whatever cloud service and, if it’s not just built into the system, it will make that for us. Welcome to the matrix.

Tied to this, I’m partway through watching the welcome keynote for the Ubuntu Summit 26.04 event that’s occurring this week. After an opening bit with Mark Shuttleworth, Canonical vice president of engineering Jon Seager expands on the April news about how Ubuntu will embrace AI, it’s worth paying attention to. (It starts at about the 31:30 mark.) He makes a great case for the work Ubuntu is doing here, with the key argument being that ignoring AI won’t make it go away and that someone, or in this case, some company, in a position of authority needs to show some leadership to make sure it happens correctly.

“Why Ae We Still Driving?”

I ask drivers in Pennsylvania that question every single day

🗃️ Why, I do declare

thejoefin asks:

Microsoft just released a new experiment around UI Frameworks to make it possible to write WinUI more like Compose, React, Swift, etc. Any thoughts or feedback on this experiment?

Interesting. And on numerous levels. I hadn’t heard about this, so thanks for the heads-up.

For those unfamiliar, one of the key innovations of .NET from the early years was the introduction of XAML as an XML-based way to describe the user interface of an app. This was a then-modern replacement for the black box designers behind previous environments like Visual Basic, and it presented (and still does) an interesting separation of UI code (XAML) from business logic (typically C#). It’s been used by every Microsoft UI framework in the 25 subsequent years, from WPF to UWP, the Windows App Framework/WinUI 3, MAUI, and more (including some long gone, like Silverlight). And until fairly recently, Google used a similar XML-based Views approach for Android user interface development.

At a high level, XAML and things like XAML are what’s called declarative code, as opposed to imperative code, which is more explicit. With declarative code, you describe what you want a UI, in this case, to look like and the underlying framework/platform handles how to make that happen. With the previous imperative methods, you would explicitly describe how that UI would look, typically with pixel precision. Declarative UI is better for high-DPI and high resolution displays and the multitude of screen sizes that now exist, whereas imperative methods made sense when the whole user base was stuck on VGA or whatever displays.

XAML was obviously inspired by HTML and the web technologies of the era, but in more recent years, other ways of creating UI declaratively have emerged. I’m not sure of the exact origins of this, but we now have many UI frameworks that use the same language for UI code and logic code, with the two separated using modular design patterns. React is a JavaScript UI framework that seems to have kicked off or at least popularized this approach. And now we have things like Flutter (which uses the Dart language) and SwiftUI (which uses Swift) doing the same thing elsewhere. On Android, Google uses Jetpack Compose, a Kotlin-based declarative UI framework, and it’s pushing it on developers because it’s “reactive” and can accommodate all the screen sizes Android does and will run on, including automatically changing layouts if a user, say, switches from an outside screen to a bigger internal screen on a foldable.

During all this time, Microsoft has stuck with XAML, which now feels a bit old-fashioned. And though it’s familiar to those in the .NET/Microsoft space, the one thing we would all agree with is that XAML is incredibly verbose: You have to write an astonishing amount of XAML code to create most UIs, and it takes a lot of time and work. We were supposed to be spared from that by visual designers but that proved so difficult it never really happened. And so most .NET developers just write XAML code. And curse a lot, I bet.

That’s one reason why this new Microsoft experiment, called Microsoft UI Reactor, is so intriguing. If you had asked me yesterday or at any time in the past several years whether Microsoft would ever modernize its developer stack to include a C#-based declarative UI framework, I would have just laughed. There’s no way, right? Why even bother.

Why indeed.

Yes, this experimental project is just that, experimental, and it doesn’t necessarily mean we’re suddenly going to see a big new initiative to modernize .NET and C# to accommodate this style of UI development. But … maybe it does mean that. I never saw this coming.

When I was writing the Programming Windows series that became Windows Everywhere, I had to pause for many months in the middle when I got to the .NET era because that was when my first-hand experience with Microsoft’s developer technologies had faded: I had never formally learned C#, XAML, or WPF or any of those more recent frameworks. So I took the time to do that. Today, I feel like I know XAML well enough for my needs as a non-programmer, but now I find these newer declarative environments to be confusing. I’ve dabbled in them all, but I’ve never really had that ah-ha moment. And I’m not sure how much time or effort I want to spend on something so tangential to my life.

The style of coding used by Microsoft UI Reactor is not that, of course. Writing UI with C# today is not declarative. But this new thing brings the React style of declarative UI coding to C# and .NET (or, at least WinUI, which is technically not .NET). And that does fascinate me. I’m curious if this goes anywhere or nowhere at all. There is a case to be made that anyone writing or maintaining C# codebases today is, by definition, an older person well-schooled in these technologies and that moving to some new thing at this point would be counterproductive. But if you believe C#, .NET, and whatever else has a future for new apps/services and a new generation of developers, this makes all kinds of sense.

But there is this one thing lost in the whole C#/XAML/.NET story that those who learned this back in the day probably knew about and may have since forgotten or rarely use. There is nothing you can do in XAML that you can’t do in C#. That is, yes, you can create app UIs in XAML, and most do. But you can also create those UIs in C# code, and in the course of writing .NETpad and WinUIpad, I came across times when this made sense, for example, for dynamically creating the UI that might go in any number of tabs in a multi-tab version of the app. Charles Petzold was all over this, which you can see in his first .NET-based books from 20-25 years ago. I mean, of course he was. But it’s a lost art in some ways, I think.

I will take a closer look at this today if I have time or perhaps over the weekend. Thanks.

“Snapdragon C wants to be Windows’ answer to the MacBook Neo, targeting $300 laptops”

I doubt the chip is sentient and has wants

🎊 It’s a surprise no matter when I tell you

madpapist asks:

Hey Paul,I’ve not kept up with FRD in a while, but I do recall enjoying episode 1000 when Mrs. Thurrott and Mrs. Sams hosted. Any plans for them to celebrate # 2000 with another?

I did have the idea that our daughters should record episode 2000 as a sort of fun follow-up to that previous episode with the wives. But Brad’s daughter is young and he understandably isn’t interested in having her appear on camera. So we will have to think of something else. We will figure something out.

“Call of Duty is in need of a big win, and Modern Warfare 4’s multiplayer might be it”

Black Ops 7 was Call of Duty’s Vietnam