Ask Paul: April 5 (Premium)

Happy Friday! Well, it’s another monster edition of Ask Paul, so buckle up: It’s time to kick off the weekend.

pAId?

will asks:

Right now AI is considered in many ways a paid add-on for most services from Microsoft and Google. However, there are some services that are starting to provide AI features at no additional cost such as Zoom (if you have the paid version). Do you think it is possible that AI functions will be “baked in” to products in the future and there would not be an up charge for AI if more services and products include a basic AI level? I mean if it is “good enough” will it be harder for Microsoft to justify a $20-$30 per month service in the long run?

I was just thinking along these lines because of a rumor that Google is considering charging for AI search. But this raises all kinds of questions. Here’s what I’m thinking now. This type of thing will evolve rapidly (along with everything else related to AI).

To your “good enough” comment, I recently (re)told the story about me download Slackware Linux floppy disk images, one by one, using a computer at the community college I then attended and worked at (they had a T1 connection, which was very fast for the day). But the real point of that story is that it introduced my brain to this notion of “good enough”: What if Linux with some free office productivity suite was good enough to supplant Windows and Office? I had the idea that if some organization could somehow deliver on just the 10 percent most-used/most useful parts of Office, it could be lights out.

That never happened, of course, and today, Linux use on the desktop remains tiny. But as I discovered recently, the free office suites that came out of that era, like LibreOffice, are, in fact, quite good. And there is a news story today that a state in Germany—it’s always Germany, they’ll never learn—has adopted Linux and LibreOffice and will reduce its reliance on proprietary software. This feels like a fight from another era, not to mention a broken record, but it’s fair to say that many people could use this combination quite successfully. And so that idea of almost 30 years ago remains. For many, good enough is, well, good enough.

Yesterday, I pulled the trigger on something I had expected to do by mid-February at the latest and signed up for ChatGPT Plus. Like Copilot Pro, it’s $20 per user per month, and the paid version supports custom GPTs, editing DALL-E images, and other features. I was surprised by how good Copilot Pro was, so I’ve kind of stuck with that. But for whatever reason I finally took a look.

And I have to say, the custom GPT stuff for image creation is pretty terrible. Universally terrible, really: I haven’t created an acceptable image with it yet, and while the things it purports to do are more advanced than what I’ve tried with Copilot/Designer—upload a photo and turn it into a Pixar-type cartoon, etc.—it was disappointing. And that’s a paid product: If that’s not good enough, those guys are screwed.

But there are absolutely free AI services that are good enough. Even for the stuff I really use Copilot Pro for, the free version would likely be good enough, but not ideal. (Which is the point of good enough, really.) It only create 4:3 images, for example. Differentiating free and paid versions of the same product, like creating different product SKUs (Windows, Office) can be more art than science.

I was also thinking about this as I wrote about Brave’s Leo AI coming to iOS yesterday: Leo is free, but there’s a paid version too. And having used it to summarize YouTube videos and articles, I feel like the free version is both good and good enough, and that the benefits of the paid version, which is a bit expensive, frankly, don’t seem to warrant the price. More to the point, it doesn’t matter what the price is: Leo (free) is good enough.

Microsoft is in a weird and unique position. It is betting the company on AI, and it needs to sell Copilot Pro and Microsoft 365 subscriptions. But it also has to give away AI, and all over the place. Microsoft Edge continues to be a distant third in web browsers overall (5 percent), and a distant number two on desktop (12.7 percent). And so it is pimping a lot of AI in the browser, and much of that needs to be free. This gives it a competitive position against other browsers, but it kind of screws them over in-house, where the paid services can be less compelling when you get much of it for free. It’s a tough problem.

Related, Microsoft (and others) gave away tons of functionality during the pandemic, in particular Teams- and Skype-based video calling features, that it would otherwise prefer to charge for. And when the pandemic ended, it scaled that back and introduced a Teams Premium subscription that included, among other things, some features that were previously free. This landed with a thud, it’s a big failure.

So Microsoft must have this recent experience in mind when it thinks about how to break down what’s free and what’s paid. It’s another wrench in an already complicated set of decisions it will need to make.

In some ways, this situation also mirrors what’s now a decade’s old situation that impacts individuals as well as companies of all sizes. Do you buy into a single technology stack, as many Apple fans do as consumers and many big businesses do with Microsoft and Microsoft 365, or do you pick and choose between cheaper solutions from a variety of vendors? The latter is more common with small businesses—I always use the example of Google Workspace + Slack + Notion to explain what this can look like. But this is also what my “right tool for the job” thing is about. And the related debate about whether you pay for something like cloud storage using a company like Dropbox, that pretty much only does that one thing and thus probably does it right, or do you use a service from a company like Microsoft that might not always have its eye on the ball in this area?

One could make a case for either approach. I have seen with myself lately that I am drifting more and more away from the monolithic single supplier scenario (Microsoft, in my case) towards individual solutions that best meet my needs. And just writing that makes it feel obvious: Of course, you should always use the best tools for your needs, whatever they are and wherever they come from. So for AI specifically, many will likely mix free and paid services. Many others will pay once and never shop around. Many will never pay for anything. I’m sure those choices are all going to stick with us to some degree.

I could kind of go on and on here, but to your “baked in” thing, yeah, that plays a role too. In the Microsoft space, specifically, the AI-based features that you get in Word, Excel, and the other desktop apps in Office to me feel like the types of updates that Microsoft would have just added in the past as part of your subscription. Now they are asking you to pay. But should they just be there for everyone? I believe so, yeah. But it’s their strategy to experiment with. I’m curious what this will look like in 6 months or one year or whatever. I bet it changes a lot over time. That’s the only guarantee with AI, really. Change.

A Tablet PC renaissance?

TheJoeFin asks:

It seems like the attitude around Windows on ARM has become positive with these latest Qualcomm chips. Assuming the new devices are good, what effect (if any) will these ARM devices have on Windows?

I just wrote up my hands-on experiences with Windows on Arm running on the Snapdragon X Elite, but I do feel good about the coming generation of PCs we’re going to see based on this chipset. I think the end result will be what Microsoft and its partners have been hoping for since the iPhone (and, to a lesser degree, the iPad) changed everything: A world in which Windows PCs can be thin, light, fanless, and noise-free, while still offering the performance and compatibility that PC users demand, combined with terrific battery life. This has been a holy grail of sorts for over 20 years, but it’s finally happening.

When I think about using a 10″ ARM Windows tablet most of my worries come from Windows and the apps/programs not being good enough to use it touch only. Can you see yourself using a Windows device as a pure tablet today or in the near future?

No, that ship has sailed.

PC makers will continue making 2-in-1s and convertibles because they’re lucrative, with higher margins. And though the Surface Pro continues to define the form category, few people would ever use that device as a pure tablet. There’s just no software for it, and with Microsoft walking away from the mobile apps platform that it hoped would make such a thing make sense, the PC in whatever form remains true to itself as a productivity device. That is, Surface Pro is successful specifically because it’s a good laptop. But like other convertibles, there’s a promise there of additional use cases that almost no one will ever take advantage of. This is probably just human nature: We tend to overly focus on the what-ifs and you-never-knows.

It’s far more likely that a mobile platform like Android or iPad could be made more PC-like, if optionally, and eventually replace Windows and the PC. Or that some combination of these things—More useful iPads, Chromebooks, whatever–just take a bigger share of that part of the market in time.

The song remains the same

JaviAl asks:

Last week I was writing to you about the new Outlook and Microsoft’s new web-based applications and I disagree with some of your responses.

For whatever it’s worth, I knew that would disagree with almost everything I wrote.

But we’re almost debating two different topics. This is, in a way, very similar to the conversation/debate that’s erupted in the wake of the story about the Microsoft engineer who miraculously accidentally discovered a backdoor that had been put into a compression utility in Linux: Security experts, including those involved in national security, are using this event to drive home a concern they’ve had for years about the security implications inherent in the lax oversight of much open source software. But open source software proponents believe that this episode shows the strength of open source, believing that this backdoor discovery might never have been possible with closed source software. So the two sides argue, the two sides in this case being people who are technical but have absolutely no stake in this fight whatsoever. And they’re arguing with each other … for some reason.

Similarly, the debate over Microsoft Outlook can be fought at a macro or micro level. It can involve high-level ideas (this kind of software is more efficient than that kind of software) or very specific features (the new Outlook doesn’t work offline, or whatever). But my point is only that none of that matters. Microsoft is doing what it’s doing for good reasons, and it is meeting the concerns that users have over a very protracted period of time.

The central reality here that new native app development on Windows is dead, and has been for decades. And that to address the world that we live in today, Microsoft is using web technologies for its new apps (Outlook, Teams, Loop, etc.) and the Windows App SDK as a way for developers of existing native apps (Win32 or UWP) to modernize legacy codebases that are never going to be rewritten from scratch. (It’s doing so itself, too. The latest File Explorer version, which is garbage, is a Win32 app that uses Windows App SDK for its front-end, and it just announced that it’s migrating the Photos app for Windows, which is a UWP app, to Windows App SDK as well.)

With Outlook specifically, Microsoft has this 25-year-old teetering mass of code that hundreds of millions of users, including its key business customers, use every day and rely on. As I noted last week, it evolved this app as much as it could using the existing code base over a very long time, and I pointed out the move to web technologies for extensibility, a change that impacted the other major Office apps as well. But it has clearly reached the point where it’s no longer viable to keep updating this huge, complex app. And it reached the “correct” conclusion, which is the same conclusion the rest of the industry reached long ago: Because there is no viable, modern Windows app platform and never will be, it is moving to web technologies in part so it can share code with the other versions of Outlook more efficiently.

Arguing over whether these apps are less efficient or can be as full-featured as native apps is beside the point. Because there is no such thing as new native apps anymore. Microsoft couldn’t find developers who could start a new Win32 project of this scope, let alone maintain it anyway. It’s just not possible. We’re decades past that making any sense.

That it tried to make this shift before, with the mobile/Metro versions of Office that came and went because that platform was so immature, is notable here, too. You may recall that Windows RT shipped with Arm versions of all the desktop apps except for Outlook initially. The reason is that Outlook—remember this was 12 years ago—was such a complex mess of legacy code that just running it impacted Surface RT battery life by up to 2 hours. Outlook couldn’t be modernized easily enough to make that make sense. (Microsoft eventually relented on Outlook for Windows RT after Surface RT failed.) It’s even less modern today. And to be clear, modern doesn’t mean mobile and pretty. It means secure and maintainable.

With all that in mind, to your concerns/disagreements:

Any web-based application compared to a native compiled OS appplication is “always” less efficient.

I’m not sure about that, but let’s take it as fact. So what? We have incredibly powerful PCs today.

An assembly language app would likewise be more efficient/performant than an app written in C/C++, or in C#, which is managed code and arguably not “native” other than the fact that its runtime comes with Windows. Or with WinForms or WPF, likewise not native. Or UWP/Windows App SDK, same thing. Debating “native” misses the point, it’s semantics now. Apps are apps.

There are a lot of new native apps for Windows (not from Microsoft, like the Affinity applications, the Skylum applicattions like Luminar Neo, etc.). Each platform, desktop, mobile and web has its own uses, advantages and disadvantages, but they are different things and uses.

My point is that there have been no major new native Windows-only apps developed in over 20 years. And for good reason: That’s not where the market is. The market for mainstream apps is heterogeneous, and mobile is much bigger than desktop. This is where the focus is, and where the talent is.

Without going through an app-by-app breakdown, none of the Affinity apps are strictly native: They are made with a combination of C++, C#, and JavaScript (and, on Mac and iPad, with Objective C, probably Swift today). But every developer is free to use whatever technologies make the most sense to them. Microsoft, as the creator of the OS and all its native apps platforms, has simply done what most of the industry is doing: Forego native apps, which would require a unique code base for every client version, and use web technologies for major new apps.

Put another way, pointing out an exception doesn’t prove the point. That it’s an exception is the point.

I don’t agree that web architecture is the right architecture for desktop applications.

Oh, I know.  🙂 You’re clear on that.

The important thing to understand is that Microsoft, which is creating this app, has decided it’s the right architecture.

‌A professional writer who writes books of hundreds or thousands of pages, will be more comfortable with a native desktop application and I do not think that a web application handles well so many pages with images, graphics, etc. (Word even costs a little and takes time to paginate so many pages).

This may not have been the best example to present me with.

As a professional and award-winning writer who has written over 30 books, several of them over 1,000 pages in length, I can tell you emphatically that this is not the case. The Windows 11 Field Guide, like all my recent books, was written entirely in Visual Studio Code, a cross-platform web app, and that book contains thousands of images. Related to this, I use Notion extensively, every day, and that’s a web app. And I use Clipchamp exclusively for video editing, another web app. It’s not 1997 anymore. The web has matured greatly as a platform.

‌Apart from the performance and fluidity of an application, web applications have some limitations: they are page-based instead of native applications that are window-based.

That’s a modern design pattern, not a limitation of web apps.

And modern Windows apps are also designed this way. Look at Notepad, which dates back to the earliest days of Windows. It’s a classic desktop app, but it’s been updated with a WinUI 3.0 front-end and there are no more dialogs or other windows, not even an About box: All those things open as a single Settings interface inside in the app window, and to exit Settings, you click a Back button, like in a browser. The default for multiple documents is tabs, too. None of this makes Notepad less usable. It just makes it more modern.

Most of your concerns come down to preference, which is rooted in familiarity and the expectations you have from decades of doing things a certain way. That’s completely normal and understandable. This isn’t really about one thing being better than the other. It’s that things are changing. Arguing against modernity won’t accomplish anything. We can adapt, or not. If you prefer these complex apps from the past that require documentation and work to master, you can keep using them. And with Outlook in particular, you have at least 5 more years before that thing is retired.

Proof of life

JustMe asks:

Given the number of Windows 10 users out there, what do you think will happen when W10 hits end of life? Do you think Microsoft will relax hardware requirements for W11 to get people to adopt it? Will we see a small surge in PC sales as people upgrade? Or will it be more business as usual with potentially unpatched machines out there until legacy machines fail?

I feel like Microsoft is playing a game of chicken with the Windows 10 end-of-life. Meaning, yeah, it just announced the pricing for the Extended Security Updates program ($61 per user the first year, with the price doubling in each of the next two years) which one might think indicates it intends to follow through on the October 25 end-of-support schedule. But this will really be dictated by its most important customers. And if enough of them want to stick to Windows 10, I could see it blinking and extending the time frame, as it did most famously with Windows XP.

That said, I personally expect Microsoft to simply stick to the schedule. That’s a full 10 years of support, which is reasonable. And while the Windows 11 hardware requirements feel arbitrary, the oldest mainstream chipset that’s not supported is from 2016, so the oldest computers based on that system will be 9 years old in October 2025. That’s also a reasonable support time frame.

To me, the big question is what the PC market looks like in October 2025. We’ve seen a pretty consistent two-thirds/one-third split between Windows 10 and Windows 11 usage for the past year, and Windows 11 has never really seen any dramatic usage spurts. And while there are probably many reasons for that, I suspect it’s tied mostly to disinterest in Windows and PCs in general: I’m fully invested in this stuff, but most people see the PC as a tool, not a personal device to be fawned over, personalized, and kept up-to-date. Certainly, that’s true in the corporate space.

Regardless, the issue here is obvious: Microsoft has never ended support for any version of Windows when it was the most-used version in the market, and doing so could be dangerous from a security standpoint.

For example, when Microsoft ended support for Windows 7, it had less than 25 percent usage share, and when that system’s Extended Security Update (ESU) program ended three years later, usage was down to less than 10 percent. (After that, it really dropped off a cliff.) We can only guess what Windows 10 usage will look like in October 2025. But it’s not unreasonable to believe it could be in the 40 to 50 percent range. Microsoft is gambling, or at least hoping, it won’t be. That it will be lower than that.

What will Microsoft do? For now, I suspect that it will wait and see what happens. It will watch what its enterprise customers do, and try to understand their migration schedules. Proactively extending the support timeline would send the wrong message, but so would cutting off support if too much of the market was still on Windows 10.

What’s a WhatsApp?

spacecamel asks:

Why isn’t WhatsApp more popular in the States? When I am out of the country, it is indispensable for communicating with people.

My guess is that the popularity of Apple’s iMessage service is to blame. The iPhone has long had an outsized share of mobile in the U.S. And this impacts those who choose Android, too, since iMessage is compatible with SMS/MMS. So it’s the default choice for most people.

I had, of course, heard of WhatsApp, but I never needed it or thought much about it until we went to Mexico. Now we use it all the time: It’s not the default messaging solution there, it’s the only messaging solution. And it’s used in ways that would confuse most Americans. Everyone gives you their WhatsApp number, including even your doctor. It’s just the way people communicate. The volume of messages I receive on WhatsApp today outstrips every other messaging app I used combined, and this is mostly from two chat groups tied to the apartment we’re in. It’s endless. (If anything, it’s tiring.)

Samsung apps suck

MichaelMDiv asks:

I recently moved from a Pixel 6 to a Samsung S23, despite the fact that I can’t stand duplicate apps. Samsung is the worst in this, as you have wrote about, but I thought that even a blind squirrel finds an acorn now and then, so there might be one or two Samsung apps that are worth using. Are there any Samsung apps you have liked when you tested Samsung devices?

No. Not even one.

If anything, Samsung continually amazes me at how horrible it is. For example, the other day, I got a notification that some apps needed to be updated in the Galaxy Store. I’ve never used this store, but I figured it was related to some of the in-box Samsung apps, so I tapped the notification to find out. It listed four apps, none of which I had installed through the Galaxy Store: Zoom, Firefox, Microsoft 365, and Audible. Though the latter was called “Audible for Samsung.” (Please imagine the exaggerated eye roll for that last one.). I tapped “Update all” and was told “Can’t update some apps because they weren’t downloaded from the Galaxy Store.” Sigh. That’s the Samsung apps experience in a nutshell.

And I just checked it again. Those apps are still sitting in there, waiting to be updated. Along with two more, Amazon Prime Video, which I downloaded from Google, and Samsung Visit In (whatever the hell that is).

I also just scanned through the all apps list just to make sure I hadn’t overlooked anything. But no, there’s nothing.

UWP is still dead

helix2301 asks:

Paul I’ve been messing with apps on windows is there still a UWP I’ve been looking into this and see much has changed recently any suggestions I wanted a windows n Xbox app.

Microsoft officially deprecated UWP in late 2021, promoting the Windows App SDK as the path forward for developers who want to modernize existing UWP (mobile) or desktop (Win32) apps. Since then, it’s been slowly retiring the in-box UWP apps, in some cases updating them using the Windows App SDK, most recently Photos (as noted above). There’s an Xbox app that ships with Windows, of course, but that’s a web app, not a UWP app. But if you mean an example of a UWP app that runs on both Windows and Xbox, I doubt there’s a major (or in-house Microsoft-made) example of such a thing. And now, of course, there won’t be.

In Windows 11, some of the remaining UWP apps include Movies & TV (which dates back to Windows 8.x and was never updated with a Windows 10 look and feel) and Media Player (which dates back to Groove Music and Xbox Music, and thus Windows 8 as well). I’m not sure of a definitive list and am guessing that many of the smaller apps—Calculator, Sound Recorder—are possibly still UWP or some way into a post-UWP transition. On Xbox, there are probably even fewer examples.

DSLR?

jrzoomer asks:

Paul do you use a DSLR camera at all for photos or videos, or is it now clear that with smartphones, there is no longer any need to carry one around anymore

I’ve not used a standalone camera of any kind since 2013: I switched to smartphones for photos when I got the Lumia 1020 and haven’t looked back since. YMMV.

Is that an Xbox in your hand or are you just happy to play against me?

hastin asks:

Any thoughts on the Xbox Handheld rumors?

This came up on First Ring Daily this morning, but I’m of two minds. On the one hand, an Xbox handheld gaming system is technically possible, and I think the Hyper-V-based architecture of the recent Xbox console generations points the way. On the other, is there any market for such a thing? We already have many ways to play games on the go, and it’s not clear we need yet another thing to carry around. Our phones, tablets, and laptops are great for games, including Xbox Cloud Gaming.

I feel like the Nintendo Switch is popular for all kinds of reasons, but part of it has to be that it’s the only way to play those games. That is, there isn’t a console version and a portable handheld version. There’s just the one thing that does both. This isn’t possible with Xbox, not if you want the handheld to be 100 percent compatible with the current-get consoles.

Personally, I think given the market and interest in the Steam Deck, Windows handhelds, and even the PS Portal – something that could run Xbox (or PC) games locally OR via the cloud seems like a great idea.

Yeah. Part of the conversation Brad and I had was about whether it would be possible to create a lightweight version of Windows using that Xbox/Hyper-V architecture, for a portable gaming PC device like Steam Deck that would be compatible with all Windows games. Of course, the question there, as with a standalone Xbox handheld, is whether the demand warrants making such a product. I’d like to see either version, personally. But I do wonder if either makes any sense.